Improving the therapeutic potential of human granzyme B and … · Improving the therapeutic potential of human granzyme B and evaluation of granzyme M as novel effector molecules

Improving the therapeutic potential of

human granzyme B and evaluation of

granzyme M as novel effector molecules in

cytolytic fusion proteins for the treatment of

Serpin B9-positive cancer

Von der Fakultät für Mathematik, Informatik und Naturwissenschaften

der RWTH Aachen University zur Erlangung des akademischen Grades einer

Doktorin der Naturwissenschaften genehmigte Dissertation

vorgelegt von

Diplom-Bioingenieurin

Sonja Schiffer, geb. Hermes

aus Neuss

Berichter: Universitätsprofessor Dr. rer. nat. Rainer Fischer Universitätsprofessor Dr. rer. nat. Dr. rer. medic. Stefan Barth

Tag der mündlichen Prüfung: 29. November 2013

Diese Dissertation ist auf den Internetseiten der Hochschulbibliothek online verfügbar.

Content ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

i

ABBREVIATIONS ............................................................................................................................................... V

1 INTRODUCTION........................................................................................................................................ 1

1.1 CANCER .................................................................................................................................................... 1

1.2 TARGETED CANCER THERAPY ......................................................................................................................... 2

1.2.1 Immunotoxins .................................................................................................................................... 3

1.2.2 Humanization of immunotoxins ........................................................................................................ 4

1.3 GRANZYMES .............................................................................................................................................. 5

1.3.1 Granzyme B ....................................................................................................................................... 5

1.3.2 Granzyme M ...................................................................................................................................... 7

1.4 TUMOR ESCAPE MECHANISMS TO EVADE IMMUNOSURVEILLANCE ......................................................................... 8

1.4.1 The granzyme B inhibitor PI-9 ........................................................................................................... 9

1.5 HODGKIN LYMPHOMA AND THE CD30 RECEPTOR ........................................................................................... 11

1.6 CHRONIC AND ACUTE MYELOMONOCYTIC LEUKEMIA AND THE CD64 RECEPTOR .................................................... 14

1.7 MOLECULAR MODELING ............................................................................................................................ 16

1.8 AIMS AND OBJECTIVES ............................................................................................................................... 17

2 MATERIAL AND METHODS ..................................................................................................................... 19

2.1 CHEMICAL AND BIO-CHEMICAL MATERIAL ...................................................................................................... 19

2.1.1 Equipment ....................................................................................................................................... 19

2.1.2 Chemicals and consumables ............................................................................................................ 20

2.1.3 Kits, enzymes and standards ........................................................................................................... 20

2.1.4 Synthetic oligonucleotides and plasmids ......................................................................................... 21

2.1.5 Antibodies and antibody fragments ................................................................................................ 22

2.1.6 Bacterial strains ............................................................................................................................... 23

2.1.7 Mammalian cell lines....................................................................................................................... 24

2.1.8 Primary cells .................................................................................................................................... 25

2.1.9 Media and buffers ........................................................................................................................... 25

2.2 MOLECULAR-BIOLOGICAL METHODS ............................................................................................................. 27

2.2.1 Polymerase chain reaction .............................................................................................................. 27

2.2.2 Site-directed mutagenesis ............................................................................................................... 28

2.2.3 Agarose gel electrophoresis ............................................................................................................ 29

2.2.4 Zero blunt TOPO PCR cloning .......................................................................................................... 29

2.2.5 Endonuclease restriction of DNA ..................................................................................................... 29

2.2.6 Ligation of DNA fragments .............................................................................................................. 29

2.2.7 Transformation of plasmid DNA into competent E. coli .................................................................. 30

2.2.8 Preparation of plasmid DNA from E. coli ......................................................................................... 30

2.2.9 Sequencing of DNA .......................................................................................................................... 30

2.3 CELL CULTURE .......................................................................................................................................... 30

2.3.1 Cultivation of cell lines and primary cells ........................................................................................ 30

Content ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

ii

2.3.2 Transfection of cell lines .................................................................................................................. 31

2.3.3 Protein delivery into cells via streptolysin O .................................................................................... 31

2.4 PROTEIN-BIOCHEMICAL METHODS ................................................................................................................ 32

2.4.1 SDS-PAGE ......................................................................................................................................... 32

2.4.2 Western blot analysis ...................................................................................................................... 32

2.4.3 Bacterial expression and purification .............................................................................................. 33

2.4.4 Mammalian expression and purification ......................................................................................... 33

2.4.5 Protein quantification ...................................................................................................................... 34

2.4.6 Labeling of SNAP constructs ............................................................................................................ 35

2.4.7 Detection of endogenous PI-9 from cell lines or primary cells......................................................... 35

2.4.8 Serum stability assays ..................................................................................................................... 36

2.4.9 ELISA for detection of CD30 receptor in blood................................................................................. 36

2.5 FLOW CYTOMETRIC ANALYSIS ...................................................................................................................... 36

2.5.1 Binding analysis ............................................................................................................................... 36

2.5.2 Determination of Kd values .............................................................................................................. 37

2.5.3 Determination of specific cell death of primary cells ...................................................................... 37

2.6 CONFOCAL MICROSCOPY ............................................................................................................................ 37

2.7 FUNCTIONAL CHARACTERIZATION OF RECOMBINANT PROTEINS .......................................................................... 37

2.7.1 Protease assays ............................................................................................................................... 37

2.7.2 Apoptosis assays ............................................................................................................................. 39

2.7.3 Viability assay .................................................................................................................................. 39

2.8 IMMUNOHISTOCHEMISTRY ......................................................................................................................... 40

2.9 IN VIVO EXPERIMENTS ................................................................................................................................ 41

2.9.1 Mouse strains, housing and maintenance of animals ..................................................................... 41

2.9.2 Handling of mice and anesthesia .................................................................................................... 41

2.9.3 Establishment of xenograft subcutaneous tumor models ............................................................... 41

2.9.4 In vivo optical imaging by Cri-Maestro system ............................................................................... 42

2.9.5 Biological activity of Gb-Ki4(scFv) and GbR201K-Ki4(scFv) ............................................................. 42

2.10 STATISTICAL ANALYSIS ................................................................................................................................ 42

3 RESULTS ................................................................................................................................................. 43

3.1 PI-9 EXPRESSION IN CANCER CELL LINES ......................................................................................................... 43

3.2 IDENTIFICATION OF MUTATIONAL SITES BY MOLECULAR MODELING ..................................................................... 44

3.3 GENERATION AND PREPARATION OF GRANZYME B CONSTRUCTS AND ITS MUTANTS ................................................ 46

3.4 PREPARATION AND FUNCTIONAL CHARACTERIZATION OF RECOMBINANT PI-9 ....................................................... 48

3.5 ESTABLISHMENT OF A TEST SYSTEM FOR GRANZYME B MUTANTS ........................................................................ 49

3.6 IN VITRO FUNCTIONAL CHARACTERIZATION OF WILD-TYPE AND MUTANT GB-H22(SCFV) AND GB-KI4(SCFV) .............. 52

3.6.1 Proteolytic activity of granzyme B in absence and presence of PI-9 ............................................... 52

3.6.2 Target cell binding ........................................................................................................................... 54

Content ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

iii

3.6.3 Internalization activity ..................................................................................................................... 55

3.6.4 Cytotoxic activity against PI-9-negative cell lines............................................................................ 57

3.6.5 Apoptotic activity on PI-9-negative cell lines ................................................................................... 58

3.7 SELECTION OF THE MOST PROMISING GRANZYME B MUTANT ON PI-9-POSITIVE CELL LINES ..................................... 59

3.7.1 Cytotoxic activity on PI-9-positive cell lines ..................................................................................... 59

3.7.2 Apoptotic activity on PI-9-positive cell lines .................................................................................... 60

3.8 IN VIVO BIOLOGICAL ACTIVITY OF GB-KI4(SCFV) AND GBR201K-KI4(SCFV) ........................................................ 61

3.8.1 Stability of Gb-Ki4(scFv) constructs in mouse serum ....................................................................... 61

3.8.2 Establishment of Hodgkin lymphoma xenograft models ................................................................ 62

3.8.3 In vivo effects of GbR201K-Ki4(scFv) on PI-9-negative L540cy cells ................................................ 65

3.8.4 Comparative in vivo effects of Gb-Ki4(scFv) and GbR201K-Ki4(scFv) on PI-9-positive L428 cells .... 66

3.9 EX VIVO STUDIES WITH CMML AND AMML PATIENT-DERIVED MONOCYTES ........................................................ 68

3.9.1 PI-9 expression in patient-derived PBMCs ....................................................................................... 68

3.9.2 Receptor phenotyping of PBMCs ..................................................................................................... 69

3.9.3 Target cell binding and internalization............................................................................................ 70

3.9.4 Cytotoxic activity on primary cells ................................................................................................... 72

3.9.5 Specificity of target cell death ......................................................................................................... 74

3.9.6 Serum stability of Gb-H22(scFv) fusion proteins in patient-derived serum ..................................... 77

3.10 EVALUATION OF GRANZYME M AS NOVEL EFFECTOR MOLECULE IN CYTOLYTIC FUSION PROTEINS ............................... 77

3.10.1 Generation and preparation of granzyme M fusion proteins ..................................................... 77

3.10.2 Functional characterization of granzyme M fusion proteins ...................................................... 79

3.10.3 Target cell binding ...................................................................................................................... 80

3.10.4 Cytotoxic activity of granzyme M fusion proteins ....................................................................... 81

3.10.5 Ex vivo studies with Gm-H22(scFv) .............................................................................................. 84

4 DISCUSSION ........................................................................................................................................... 86

4.1 THE THERAPEUTIC POTENTIAL OF GRANZYME B-BASED CYTOLYTIC FUSION PROTEINS .............................................. 86

4.2 IMPACT OF ENDOGENOUS PI-9 EXPRESSION IN CANCER CELLS ON THE PRO-APOPTOTIC ACTIVITY OF GRANZYME B ........ 89

4.3 GENERATION AND IDENTIFICATION OF A PI-9 INDEPENDENT GRANZYME B VARIANT .............................................. 94

4.3.1 Evaluation of in silico findings and comparison with in vitro enzymatic activity ............................ 94

4.3.2 Selection of the optimal granzyme B mutant .................................................................................. 95

4.4 EFFICACY OF THE GBR201K MUTANT IN VIVO ................................................................................................ 98

4.5 EX VIVO DETERMINATIONS ON PRIMARY LEUKEMIC CELLS ................................................................................ 101

4.5.1 CMML and AMML primary cells as target for CD64-specific fusion proteins ................................ 101

4.5.2 Specific cytotoxic efficacy of CD64-specific fusion proteins on primary CMML and AMML cells .. 103

4.5.3 The role of PI-9 and the potential of the determined cytolytic fusion proteins for personalized

medicine ..................................................................................................................................................... 105

4.6 APPLICATION OF GRANZYME M AS AN ALTERNATIVE EFFECTOR MOLECULE IN CYTOLYTIC FUSION PROTEINS ............... 106

4.6.1 Efficacy of granzyme M based CFPs .............................................................................................. 107

Content ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

iv

4.6.2 Clinical potential of granzyme M as effector molecule in cytolytic fusion proteins ...................... 109

5 OUTLOOK ............................................................................................................................................. 111

6 SUMMARY ........................................................................................................................................... 113

7 LITERATURE ......................................................................................................................................... 115

8 APPENDIX ................................................................................................................................................. I

8.1 DNA AND AMINO ACID SEQUENCE OF GBR201K ............................................................................................... I

8.2 SYNTHETIC DNA SEQUENCE OF GM-H22(SCFV) IN PMS VECTOR ......................................................................... I

9 INDEX OF FIGURES AND TABLES .............................................................................................................. IV

10 PUBLICATIONS ........................................................................................................................................ VI

Abbreviations ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

v

Abbreviations

AA acrylamide

ADC antibody-drug-conjugate

AML acute myeloid leukemia

AMML acute myelomonocytic leukemia

AP alkaline phosphatase

APC allophycocyanin

BCIP 5-bromo-4-chloro-3-indolyl phosphate

BG O(6)-benzyl guanine

BID BH3-interacting domain death agonist

BSA bovines serum albumin

CFP cytolytic fusion protein

CML chronic myeloid leukemia

CMML chronic myelomonocytic leukemia

CTL cytotoxic T lymphocyte

dd double distilled

DMSO dimethylsulfoxide

DNA deoxyribonucleic acid

dNTPs deoxyribonucleotide triphosphates

DSMZ German Collection for Microorganisms and Cell Cultures

DT diphtheria toxin

DTT dithiothreitol

EBV Epstein-Barr virus

ECS enterokinase cleavage site

E. coli Escherichia coli

EDTA ethylenediamin tetraacetate

EGF epidermal growth factor

EGFP enhanced green fluorescent protein

ETA’ Pseudomonas exotoxin A, mutant with deleted binding domain

et al. et altera

FACS fluorescence-activated cell sorting

FC fragment crystallizable

FCS fetal calf serum

FITC fluorescein isothiocyanate

FPLC fast protein liquid chromatography

FSC forward scatter

g relative centrifuge force

GAM goat anti mouse

Gb granzyme B

GF gel filtration

Gm granzyme M

g gram

h hour(s)

HAMA human anti-mouse antibody

Hb hemoglobin

His-tag polyhistidine motif

HL Hodgkin lymphoma

HRPO horse radish peroxidase

HRS Hodgkin and Reed-Sternberg

IC50 half maximal inhibitory concentration

ICAD inhibitor of caspase-activated DNase

IFN γ interferon gamma

Ig immunglobulin

IL interleukin

IMAC immobilized metal ion affinity chromatography

IPTG isopropyl-β-D-thiogalactoside

IRES internal ribosomal entry site

IT immunotoxin

IVS synthetical intron

Kat2 Katushka2

kDa kilo Dalton

L liter

LB Luria Bertani

M molarity (mol/L)

mA milli ampere

min minute(s)

MFI mean fluorescence intensity

MTD maximum tolerable dose

Abbreviations ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

vi

n nano

NBT nitro blue tetrazolium chloride

n.d. not determined

NFκB nuclear factor 'kappa light chain enhancer' of activated B cells

NK natural killer

nm nanometer

ns not significant

p pico

PAGE polyacrylamide gel electrophoresis

PBMC peripheral blood mononuclear cell

PBS phosphate buffered saline

PBST PBS + 0.05 % (v/v) Tween-20

PCR polymerase chain reaction

PE phycoerythrin

pH negative log of the concentration of hydrogen ions in solution (pondus

hydrogenii)

PI propidium iodide

pI isoelectric point

PLT platelets

RCL reactive center loop

pNA para-nitroanilin

PO peroxidase

RNA ribonucleic acid

rpm rounds per minute

RT room temperature

scFv single chain variable fragment

SD standard deviation

SDS sodiumdodecylsulfate

sec seconds

siRNA small interfering RNA

SSC sideward scatter

TAE Tris-acetate-EDTA

Taq Thermus aquaticus

TB Terrific Broth

TNF tumor necrosis factor

Tris Tris(hydroxymethyl)aminomethane

Tween 20 polyoxyethylene (20) sorbitan monolaurate

U unit

UV ultraviolet

V volts

v/v volume per volume

WBC white blood cells

XIAP X-linked inhibitor of apoptosis protein

w/v weight per volume

XTT sodium3’-[1-(phenylaminocarbonyl)-3,4- tetrazolium]-bis(4-methoxy-6- nitro)benzene sulfonic acid hydrate

One and three letter code of amino acids

A Ala Alanine

C Cys Cysteine

D Asp Aspartic acid

E Glu Glutamic acid

F Phe Phenylalanine

G Gly Glycine

H His Histidine

I Ile Isoleucine

K Lys Lysine

M Met Methionine

N Asn Asparagine

P Pro Proline

A Ala Alanine

Q Gln Glutamine

R Arg Arginine

S Ser Serine

T Thr Threonine

V Val Valine

W Trp Tryptophan

Y Tyr Tyrosine

1 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

1 Introduction

1.1 CANCER

Cancer is one of the most severe health problems in the world. In the United States more than

1.6 million new cancer cases occurred in 2012 and more than half a million people died from this

disease (estimated by the American Cancer Society 2012). During the last five years, overall cancer

incidence rates remained almost stable whereby death rates decreased by 1.8 % per year in men and

by 1.6 % per year in women. Nevertheless, one in four deaths in the United States is still due to

cancer. The four leading types of cancer are lung, bronchus, colorectal and prostate for men or

breast cancer for women. The lifetime probability of developing an invasive cancer is 45 % for men

and 38 % for women. Cancer cells are characterized by genome instability, which generates the

genetic diversity that expedites carcinogenesis and tumor cell proliferation; and, inflammation, which

promotes the multiple hallmark functions of cancer. Tumors exhibit a high level of complexity having

many distinct cell types, which interact in a heterotypic way. The hallmark functions of cancer

comprise continuation of proliferative signaling, resistance to cell death, induction of angiogenesis,

evasion of growth suppressors, enabling of replicative immortality and activation of invasion and

metastasis [1]. During the last few years, further features have been identified, namely

reprogramming of energy metabolism and evasion of immune destruction. Most important for the

survival of tumors is their recruitment of normal cells in order to create a special ‘tumor

microenvironment’ [2].

Typical cancer treatment strategies including surgery, chemotherapy and radiotherapy are limited by

their corresponding serious and often life-threatening side effects, caused by the concurrent killing of

nonmalignant cells. Furthermore, many tumor cells become resistant to drugs, substantially reducing

any positive response to the treatment and resulting in a poor prognosis. In addition, surgical

resection of primary tumors results in high relapse rates due to incomplete removal of tumor cells as

well as the ineffective control of metastatic disease. Even more life-threatening are tumor stem cells

since they are the main initiators of tumor cell proliferation and often survive classical treatment

approaches. Tumor stem cells will often reactivate after therapy leading to a relapse. These facts

clearly illustrate the strong demand for both more potent and yet more selective, targeted treatment

options [3,4].

2 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

1.2 TARGETED CANCER THERAPY

Immunotherapeutic approaches have the potential to provide the needed selective, targeted

treatment options to specifically eliminate malignant cells. Immunotherapy generally involves the

induction of several mechanisms that interfere with both carcinogenesis and with tumor cell

proliferation. Active immunotherapy provides a more specialized approach, in which the immune

system itself is stimulated to remove malignant cells [5,6]. Targeted therapy involves either the direct

activation or inactivation of (soluble) enzymes, growth factors or surface receptors in signaling

pathways that are necessary for tumor maintenance and proliferation (e.g. growth factor receptors

such as EGFR (Epidermal Growth Factor Receptor), Her2, BRAF and Kit) or the indirect targeted

delivery of effector molecules to tumor cells, leading to growth arrest or remission [7,8]. This may

involve drugs that penetrate the tumor cell plasma membrane allowing the targeting of intracellular

tumor antigens, or the use of monoclonal antibodies (mAbs) that target cell surface molecules. Due

to these different mechanisms of action, the potential of targeted therapies is most promising for the

treatment of chemoresistant cells.

Monoclonal antibody therapy (passive immunotherapy) is a rapidly-expanding immunotherapeutic

treatment platform for cancer. At least 12 therapeutic mAbs for oncologic indications have already

been approved by the FDA [9] and several more are in clinical development [10]. However, the

activity of mAbs is limited by their relatively large size, resulting in poor tumor penetration.

Therefore high doses are needed to effectively compete with serum IgGs, which in turn lead to

severe side-effects. Furthermore unfavorable Fc receptor polymorphisms can reduce the impact of

Fc-mediated effector functions. For these reasons, unmodified antibodies are rarely suitable as

therapeutics [11,12].

Several strategies have been developed to enhance the anti-tumor activity of mAbs, including the

humanization of antibodies [13], the optimization of antibody effector functions [14] and the fusion

of antibodies to a potent cytotoxic drug. In the early 1900s, Paul Ehrlich considered the idea of a

‘magic bullet’ forming the basis of the antibody-drug conjugate (ADC); and, immunotoxin (IT)

concepts in which antibodies or antibody fragments specifically targeting tumor antigens deliver

cytotoxic payloads and trigger their internalization into abnormal cells [15].

ADCs usually comprise a full-size antibody chemically linked to a toxin, drug or radionuclide. The

therapeutic potential of ADCs has been demonstrated in numerous preclinical studies [11,16]. The

FDA has approved Brentuximab Vedotin, a CD30-directed ADC consisting of the anti-CD30 mAb

Brentuximab linked to the antimitotic agent monomethyl auristatin E for the treatment of Hodgkin

lymphoma and anaplastic large cell lymphoma [17]. After binding and internalization this agent binds

to tubulin which arrests the cell cycle leading to apoptosis of the targeted cells. In 2013, an ADC

3 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

called Ado-trastuzumab Emtansine (Kadcyla) received FDA approval for treatment of HER2-positive,

metastatic breast cancer in patients who have already undergone unsuccessful treatment with the

mAb Trastuzumab (Herceptin) and a taxane. Ado-trastuzumab Emtasine is a conjugate between the

Trastuzumab portion and a chemotherapeutic molecule called DM1, which attacks the cancer cells

after Her2 targeting [18]. The FDA also approved Gemtuzumab Ozogamicin (Mylotarg®), a humanized

anti-CD33 mAb conjugated to calicheamicin for the treatment of acute myeloid leukemia (AML) [19],

although this has now been withdrawn from the market due to adverse systemic effects.

1.2.1 Immunotoxins

ITs comprise an effector domain, which is a naturally-occurring toxin that inhibits protein synthesis or

induces apoptosis by modifying signal transduction pathways [20,21]. Toxins used in the design of

conventional highly effective recombinant ITs originate from plants, like Ricin A; or, bacteria, like

Pseudomonas exotoxin A (ETA) and Diphtheria Toxin (DT). The toxins inhibit protein biosynthesis by

ADP ribosylation of eukaryotic elongation factor 2, leading to cell death. Cell death-inducing toxins

use different mechanisms of action than chemotherapeutics. Therefore, ITs are promising for the

treatment of chemoresistant tumors or residual diseases to achieve for example a ‘chronic’ status of

the cancer. Furthermore ITs utilize a gentler approach than ADCs which indicates a high potential for

use in post-remission therapies. In addition, the use of ITs for patients who have endured lengthy

chemotherapy can result in improved quality of life and better clinical outcomes.

The first generation of ITs comprised full-size antibodies chemically conjugated to whole-protein

toxins. These ITs, however, have several drawbacks, including lack of specificity, poor stability, and

the tendency to be heterogeneous. Improvements were necessary. Investigation of protein structure

and function showed the presence of discrete cell binding domains within the toxins and led to the

design of the next generations of ITs. In second generation ITs, these cell binding domains were

removed, thereby diminishing the binding to normal cells. Third generation ITs were generated via

recombinant DNA techniques resulting in truncated toxins genetically linked to shorter antibody

fragments (scFv), leaving only the antigen binding variable domains of an antibody, or natural

ligands. These improvements yielded ITs having small size and diminished immunogenic response,

resulting in easier manufacturing, better tumor penetration and increased stability [22].

The selection of proper target molecules specific for tumor cells is one of the challenges in targeted

cancer therapy. But, advancements in phage display technology and protein engineering have

enabled and improved the selection and generation of high-affinity targeting residues [23].

Early generation ITs have shown promising results in in vitro studies and preclinical tests and have

also been tested in a small number of clinical trials [24]. The most prominent example is Denileukin

4 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Diftitox (Ontak®), a recombinant IT comprising DT and interleukin-2 (IL-2), which has been approved

by the FDA for the treatment of T-cell lymphoma [25].

However, the major limitation of clinical deployment of these early generation ITs is that they are

limited by their immunogenicity since they have non-human components. During long-term cancer

treatment requiring repeated doses, these components can induce neutralizing antibodies in patients

with a functional immune system. A murine origin of the binding moiety for example induces a

human anti-mouse antibody (HAMA) response in some patients. HAMA responses not only reduce

treatment efficiency, but also have been shown to cause allergic reactions, in extreme cases leading

to life-threatening complications such as renal failure [26].

These issues reduce the therapeutic potential of these early ITs. The complete humanization of ITs is

therefore an important criterion for the development of fourth-generation molecules called human

cytolytic fusion proteins (CFPs) [4].

1.2.2 Humanization of immunotoxins

There are several strategies that can be used to humanize the tumor-selective ligand. For example,

the effector moiety can be fused to human cytokines or growth factors that target tumor receptors,

e.g., IL-2, IL-4, EGF (epidermal growth factor), TGFα (tumor growth factor alpha) or the CD30 ligand

[24]. Recombinant DNA technology can also be used to engineer antibodies for reduced

immunogenicity. Chimeric antibodies comprising a murine Fab fragment and a human Fc region can

reduce HAMA responses, because the constant region makes the most significant contribution to

immunogenicity. Another strategy is the generation of humanized antibodies by grafting murine

complementarity determining regions (CDR) onto human V-region frameworks, an approach known

as CDR replacement [27]. Complete humanization is also possible by selecting the V-regions from

human antibody libraries or using transgenic knock-out mice or rats in which the gene loci for the

light and heavy antibody chains are replaced by their human counterparts [28]. ScFv antibodies used

for the production of CFPs can be derived from humanized or human monoclonal antibodies or

selected directly from a human scFv library [4].

There are a number of approaches that can be used to decrease the immunogenic responses of ITs

caused by the cell-death inducing domains found in bacterial or plant toxins. For example, the T or B

cell epitopes of ETA can be modified and ITs can be derivated by polyethylene glycol or

immunosuppressive agents can be co-applied [29,30]. A more favorable approach to achieve

complete humanization uses human cytolytic proteins, such as enzymes that induce cell death by a

variety of mechanisms. Pro-apoptotic human proteins that are potentially useful for targeted cancer

therapy include RNases such as angiogenin as well as death receptor ligands such as sTRAIL (tumor

5 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

necrosis factor (TNF) -related apoptosis-inducing ligand) and Fas ligand and pro-apoptotic members

of the Bcl-2 family [31]. In addition, enzymes with tumor suppressive functionality, such as DAPK2

(death-associated protein kinase 2), allow targeted restoration of tumor suppressor activity in cancer

cells [4]. Granzymes (serine proteases released by cytoplasmic granules within cytotoxic T

lymphocytes (CTLs) and natural killer (NK) cells) are another prominent group of human pro-

apoptotic molecules.

1.3 GRANZYMES

Cytotoxic granules contain a variety of granule-associated proteases. The secretion of cytotoxic

granules is the primary defense against virally infected and malignant cells. Among these granule-

associated proteases are granzymes, serine proteases, which play a significant role in this defense.

Five different human granzymes (A, B, H, K and M), each with unique species-dependent expression

patterns and substrate specificities have been discovered [32,33]. Granzymes promote apoptosis;

and regulate B-cell proliferation, the cleavage of extracellular matrix proteins, the induction of

cytokine secretion, the activation of cytokines and the control of viral infections [34]. The main focus

of this study was granzyme B and granzyme M.

1.3.1 Granzyme B

Recombinant granzyme B is the most important and well-studied member of the human granzyme

family. Its inherent pro-apoptotic activity has been demonstrated in vitro in several proof-of-concept

studies including Stahnke et al. (AML), Liu et al. (melanoma), Dalken et al. (breast carcinoma) and

Kanatani et al. (ovarian, breast, endometrial and prostate carcinoma) (Table 9). Granzyme B is also

the most powerful pro-apoptotic member of the serine proteases stored as a macro-molecular

complex with the proteoglycan serglycin and the pore forming protein perforin within CTLs and NK

cells [35]. During NK cell mediated immune surveillance, secretory granules are depleted into the

immunological synapse formed after specific recognition of abnormal cells via an elaborated

repertoire of cell surface receptors [36]. The precise mechanism by which granzyme B is taken up

into cells and released from the endosome into the cytosol is still controversial. Two possible models

for delivery of granzyme B into target cells, which have been reviewed recently, are an endocytosis

dependent mechanism of vesicle disruption or a direct pore-mediated delivery of the apoptosis

inducing cargo. Following the recognition of the target cell, granules release their cytolytic content

into the immunological synapse that directs granzyme B, most likely by the assistance of perforin, to

the target cell’s cytosol [37]. Once released, granzyme B recruits multiple signal transduction

pathways to fulfill its death inducing function.

6 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

The classical apoptosis induction by granzyme B is comparable to the Fas mediated pathway. It

involves the cleavage of several executive and downstream caspases and is mediated by both the

caspase-pathway and the BID (BH3-interacting domain death agonist) -pathway. Direct activation of

the effector caspase 3 is controlled by the inhibitor of apoptosis protein (IAP) that has to be

inactivated by e.g. HtrA2/OMI, a pro-apoptotic molecule released from disrupted mitochondria (see

below). Successful processing of pro-caspase 3 leads to the cleavage of several downstream death

substrates including the inhibitor of the caspase-activated DNase (CAD) ICAD, poly(ADP ribose)

polymerase (PARP), DNA-dependent protein kinase (DNA-PK), the nuclear mitotic apparatus protein

(NuMA) and the nuclear-envelope intermediate-filament protein lamin B. This results in the typical

morphological changes associated with classical apoptosis, i. e. DNA fragmentation, chromatin

condensation, membrane blebbing, cell shrinkage and the formation of apoptotic bodies. The

activation of the BID pathway, either directly by granzyme B or indirectly via the activation of

caspase-8, leads to the cleavage of mitochondrial BID involving BAX (BCL-2-associated X protein) and

BAK (BCL-2 antagonist). The translocation of tBID (truncated BID) to the mitochondria results in the

permeabilization of the outer mitochondrial membrane and thus mitochondrial depolarization and

the subsequent release of cytochrome c (cyt c) and other pro-apoptotic molecules like endonuclease

G and HtrA2/OMI. Cyt c presence in the cytosol is crucial for the apoptosome formation by assembly

to the apoptotic protease activation factor APAF-1 which in turn leads to autocatalytical activation of

pro-caspase-9 and subsequent processing of the effector caspase-3. In addition to the BID-mediated

loss of mitochondrial transmembrane potential, granzyme B can directly induce mitochondrial

depolarization without the assistance of cytosolic mediators, but the exact mechanisms remain

unclear [4,38].

Granzyme B is also known as ‘Asp-ase’ because it cleaves downstream of aspartic or glutamic acid

residues giving it substrate specificity similar to that of caspases. Consequently, granzyme B can

directly target and activate the downstream caspase substrates ICAD, PARP, DNA-PK, NuMA and

lamin B. Taken together, these modes of action show that granzyme B is not only a highly potent

inducer of apoptosis due to its status as the major mediator of cellular immunity, but also because of

its ability to induce apoptosis at multiple levels using distinct pathways. Thus granzyme B has the

potential to circumvent anti-apoptotic mechanisms of the tumor cells, as described below (1.4) [4].

Recombinant expression of granzyme B has been proven successful in different expression systems

of pro- and eukaryotic origin including Escherichia coli (E. coli), Pichia pastoris (P. pastoris), insect

cells and mammalian cells such as HEK293T, HeLa, Jurkat and COS [4,37]. To be active granzyme B

requires a free N-terminus. The native enzyme is expressed as an inactive zymogen containing an N-

terminal prepro-leader sequence. The signal peptide is cleaved by a signal peptidase in the

endoplasmic reticulum to produce an inactive proenzyme with an N-terminal pro-peptide, protecting

7 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

the cell from the induction of apoptosis, thus the pro-peptide must be removed to generate an active

protease [39]. In the Golgi apparatus, a mannose-6-phosphate tag is attached which directs

granzyme B to the cytotoxic granules. Here the pro-peptide is removed by the dipeptidyl peptidase I

(DPPI) (also known as cathepsin C) [40]. Several strategies have been developed to accomplish the

correct in vitro or in vivo processing and activation of recombinant granzyme B, including the here

applied insertion of an enterokinase cleavage site (ECS) upstream of the sequence of the mature

granzyme B polypeptide, allowing activation after purification by in vitro processing. Secretory

expression of the active protein in P. Pastoris can be achieved by the direct N-terminal fusion of the

mature polypeptide to the Saccharomyces cerevisiae (S. cerevisiae) mating factor α prepro-leader

sequence allowing activation by local Kex2-protease. Secretion and functional in vivo activation by a

host cell signal peptidase has been achieved in insect cells and COS cells using the native sequence of

the granzyme B precursor protein with the propeptide deleted, or the genetic fusion of a furin

recognition motif allowing in vivo processing by endogenous P. pastoris furin-like proteases. In

addition, in vitro and in vivo activation has been achieved by either co-expression of rat DPPI in COS

cells or subsequent incubation with bovine spleen DPPI [4].

The crystal structure of granzyme B has been elucidated before by two groups [41,42]. As for all

proteases, the catalytic activity of granzyme B depends on a serine residue at the active site, one of a

triad of residues corresponding to His57, Asp102 and Ser195 in chymotrypsin. The key residue for

contact with the P11 substrate residue is Arg226 [34]. The optimal tetrapeptide recognition motif (P4-

P1) is IEPD, which is similar to that of caspase 8 and caspase 9 and thereby reflects their common

role in the activation of downstream caspases such as caspase-3 [42].

1.3.2 Granzyme M

Granzyme M is classified as ‘Met-ase’ since it cleaves specifically after amino acids with long aliphatic

side chains such as methionine, leucine, or norleucine at the P1 position1 [43]. It is predominantly

expressed within NK cells, NKT cells, γd-T cells and CD8+ effector T cells [44,45]. Recently, it was

shown that granzyme M is involved in TLR4-driven inflammation and endotoxicosis, and as such it

contributes to the inflammatory response to bacterial lipopolysaccharide by inducing serum

cytokines such as interferon (IFN) γ, IL-1α, IL-1β and TNF α [46]. However, the exact mechanism

remains unclear.

Many groups have reported the cytotoxic efficacy of granzyme M towards tumor cells [47,48].

However, the precise molecular mechanism remains controversial in the literature. It has been

1 The active sites in enzymes can be divided in subsides (S) which consist of single amino acids. The corresponding positions (P) of the substrate have the same numbering as the subsides they occupy. The positions are counted from the point of cleavage. P1 is the primary determinant of specificity.

8 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

described that granzyme M activates a caspase-dependent pathway involving cleavage of ICAD,

poly(ADPribose) polymerase (PARP), heat shock protein 75 (HSP 75), survivin and ezrin as well as the

Fas-associated protein with death domain (FADD), but not Bid [49-52]. Thus, as described for

granzyme B, classical mechanisms of apoptosis are triggered such as phosphatidyl serine (PS)

exposure, mitochondrial swelling and loss of mitochondrial transmembrane potential, cyt c release,

and reactive oxygen species (ROS) production. Other groups did not report those effects, but rather a

caspase-independent form of cell death which is not blocked by Bcl-2 expression [53-55]. Other

substrates of granzyme M can be nucleophosmin [54] and α-tubulin leading to a disruption of the

microtubule network [53]. Granzyme M also specifically cleaves the granzyme B inhibitor PI-9 [56],

therefore the application of this enzyme increases the potential to kill even immune-resistant tumor

cells, particularly when applied in combination with granzyme B (see 1.4).

As with granzyme B, granzyme M requires a free N-terminus for enzymatic activity. Recombinant

expression of the active protein has so far only been reported to be successful in the eukaryotic

expression system P. pastoris featuring the S. cerevisiae mating factor α prepro-leader sequence as

described above for granzyme B (1.3.1).

1.4 TUMOR ESCAPE MECHANISMS TO EVADE IMMUNOSURVEILLANCE

There are many examples of the immune system operating as a significant barrier to tumor

formation and progression, especially with non-virus-induced cancer types [57]. This has been proven

by experiments where mice that are deficient in various components of the immune system (e.g.

missing functions of CTLs and NK cells) have increased tumor development [58]. This is further

illustrated in patients with colon and ovarian tumors where prognosis was better if the tumors were

heavily infiltrated with CTLs and NK cells [59].

However, there is also evidence that tumor cells are able to escape the body’s immune surveillance,

thus avoiding this barrier to tumorigenesis and tumor progression. Infiltrating immune cells can bind

to cancer cells and thereby take the place of the tumor’s disconnected epithelial surrounding cells,

allowing them to survive in ectopic microenvironments by suppression of cell death pathway

stimulation [2]. Furthermore, actively immunosuppressive tumors can recruit infiltrating immune

cells, which comprise regulatory T cells and myeloid-derived suppressor cells. Both types are

suppressive towards the activity of CTLs.

Another example of an escape mechanism of tumor cells is the secretion of immunosuppressive

factors such as TGF-β leading to paralysis of CTLs and NK cells [60]. In several studies on breast

cancer it was also shown that tumor-associated macrophages provide survival signals to cancer cells

9 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

such that tissue-protective and therapy-induced mechanisms no longer trigger cell death [61]. Tumor

cells’ strategies to overcome innate and adaptive immune responses also include impaired tumor

antigen presentation by the downregulation of MHC class I molecules, and the expression of factors

that either directly kill or inhibit effector cells or recruit regulatory T-cells and thereby induce

immunotolerance [62]. In addition, tumors can regulate apoptosis by the overexpression of different

classes of anti-apoptotic proteins. For example, FLICE-inhibitory proteins (FLIPs) interfere with signal

transduction by death receptor pathways, anti-apoptotic proteins of the Bcl-2 family regulate

apoptosis at the mitochondrial level and IAP (inhibitors of apoptosis proteins) family members inhibit

caspase activity [4]. Another escape mechanism from immune response, which is discussed in detail

in chapter 1.4.1, is displayed by the expression of serpin B9 (PI-9) in many tumor cells. PI-9 is the

main inhibitor of granzyme B (1.3.1).

Over the last decade the importance of cancer stem cells has also been discussed (1.1), noting that

tumors are maintained by their own stem cells. Central for this concept is the observation that not all

tumors cells are equal. Conventional therapeutic approaches like chemotherapy or radiation may kill

the majority of the tumor cells, but rare cancer stem cells may be able to survive and later multiply.

Additionally, quiescent tumor cells that have drifted to distant sites, will avoid removal by curative

surgical treatment of a primary tumor and this may be responsible for metastasis that can appear

many years after the surgical treatment [63].

The above described new findings, especially the complexity of the tumors, and the presence of

infiltrating immune cells and the maintenance of the tumors by cancer stem cells are connected with

unfavorable clinical outcome [2], greatly influence the development of new treatment approaches

against human cancer. Development of new treatment approaches requires consideration of the

many different aspects and influences within the tumor microenvironment. These aspects and

influences show that combinatorial therapies will be more and more important in eliminating as

many of these multiple characteristic malignant cells as possible, including the ‘heart’ of the tumor -

cancer stem cells. Therapeutics comprising multiple apoptosis induction mechanisms are highly

advantageous to overcome the resistance of tumors and to eliminate residual malignant cells.

1.4.1 The granzyme B inhibitor PI-9

As mentioned above, the expression of the granzyme B inhibitor PI-9 plays an important role in one

tumor cell escape mechanism from immune response (1.4). PI-9 is a member of the clade B serpin

superfamily, a group of structurally-related proteins that regulate proteases in the vertebrate

adaptive and innate immune systems [64,65]. Since granzymes are an integral component of the

human immune system they are tightly controlled at the posttranslational level by serpins. PI-9 is the

best-known granzyme-regulating serpin. This 42 kDa protein accumulates in the cytosol of

10 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

lymphocytes to avoid self-inflicted injury caused by granule leakage and the subsequently

misdirected activity of granzyme B. Cells at immune-privileged sites therefore express PI-9. These

sites include the lens capsula, ovaries, placenta, testes and embryonic stem cells [66,67]. The

inhibition of granzyme B by PI-9 involves a stoichiometric interaction, leading to the irreversible

inactivation of both the enzyme and the inhibitor [4]. Many studies have focused on the expression

profile of PI-9 within different cell types and its induction by different stimuli (see also 4.2). In

addition to effector T cells, PI-9 is expressed by inflammatory cells [68] and in human peripheral

blood mesenchymal stem cells [69]. It is also present in cells infected by cytomegalovirus (CMV),

Epstein-Barr virus (EBV) and the BK polyomavirus [70-72] as well as in endothelial, mesothelial cells

[73] and dendritic cells [66]. PI-9 can be induced in different cell lines by nuclear factor κB (NFκB),

activator protein-1 (AP-1), as well as pro-inflammatory cytokines such as TNFα, and

lipopolysaccharides [68,72,74]. It is also induced by phorbol myristate acetate and IL-2 in CD8+ T cells

[75,76], by viral dsRNA sensors in human renal tubular epithelial cells [70] and by estrogens and

estrogen receptor-α in breast cancer cells [77]. The expression of PI-9 in renal tubular cells is

upregulated following renal allograft rejection, suggesting that its induction could be used to confer

resistance against granzyme B-mediated cytotoxic effects leading to improved graft survival [78].

The widespread expression of PI-9 has to be taken into account when developing therapeutic

strategies based on granzyme B, particularly because malignant cells can also involve multiple

mechanisms to escape the pro-apoptotic signals induced by CTLs and NK cells [79]. Hence, many

studies have reported a direct correlation between the presence of PI-9 and a reduced pro-apoptotic

granzyme B activity [64,80]. Examples for tumor cells that have been found to express PI-9 to

circumvent immune response and therefore might be resistant to granzyme B based

immunotherapeutics are summarized in Table 10 and Table 11. In addition, high expression of PI-9

has been discussed in context of poor clinical prognosis (compare 4.2).

The crystal structure of PI-9 has not been elucidated yet, but serpins share a characteristic

metastable structure and efficient inhibitory activity [65]. The three-dimensional structure comprises

nine α-helices, three β-helices and a variable reactive center loop (RCL), which is the primary site of

interaction with target proteases and is therefore critical for serpin specificity. Members of this

protein family have high sequence homology in a region of about 350 amino acids and are presumed

to have structural similarity [81]. The amino acid residues in the RCL of PI-9 are necessary for its

activity and the residue at the P1 position is a glutamic acid [82,83], explaining its specificity for

granzyme B which favors acidic P1 amino acids (1.3.1). The RCL of serpins acts as a pseudo-substrate,

which is recognized by the protease so the two molecules dock and form a reversible ‘Michaelis

complex’ (initial enzyme-substrate complex) (Figure 1). Rapidly, the protease cleaves at the favored

P1 position resulting in major structural rearrangements within the serpin that involve alternative

Introduction _____________________________________________________________________________________________________________________________________________________________________________________________________________

conformations for the RCL, β-sheet A and the

stoichiometric complex, in which both proteins remain inactive

substrate. The interaction between

(stoichiometry of inhibition) ~ 1

intensively and during their studies they

binds less efficiently to PI-9 [86]. This mutant is used as

Figure 1: X-ray crystal structure of the non

(purple) and S195A trypsin (cyan) (PDB file 1K9O

To visualize the initial complex, mutated proteins were used. the surface of the protease. This complex can be used to superimpose the crystal structure of uncomplexed granzyme B (PDB file 1IAU) and the

1.5 HODGKIN LYMPHOMA

The above mentioned infiltrating immune cells

lymphoma (HL). 150 years ago Thomas Hodgkin already described several cases of a disea

was later associated with Hodgkin’s diseas

disease. Its hallmarks are the B

Reed-Sternberg cells (HRS cells), first published in 1900

1 % of cells in the tumor tissue, although some cases may show more than 10

majority of cells in Hodgkin lymphoma lesions are a mixed infiltrate of various types of cells of the

immune system including T cells, B cells, plasma cells, neutrophils,

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

sheet A and the attached strand 1 of β-sheet resulting in a covalent

stoichiometric complex, in which both proteins remain inactive [84,85]. Thus, PI

substrate. The interaction between granzyme B and PI-9 has a Kass value of 1.7 × 10

~ 1 [4,82]. Sun et al. investigated the importance of the RCL of PI

ensively and during their studies they revealed that a mutant of granzyme B

. This mutant is used as a reference in this study.

ray crystal structure of the non-covalent Michaelis complex of A353K Manduca sexta

(PDB file 1K9O [87]).

mutated proteins were used. The RCL of the serpin interacts This complex can be used to superimpose the crystal structure of uncomplexed

and the active serpin 1K (PDB file 1SEK) for further investigations of interactions

HODGKIN LYMPHOMA AND THE CD30 RECEPTOR

infiltrating immune cells (1.4) play a key role in the development of Hodgkin

Thomas Hodgkin already described several cases of a disea

was later associated with Hodgkin’s disease or better HL since it was realized that it is a lymphoid

B-cell derived mononucleated Hodgkin cells and the multinucleated

), first published in 1900 [89,90]. They usually account for only about

cells in the tumor tissue, although some cases may show more than 10


including T cells, B cells, plasma cells, neutrophils, eosinophils and mast cells

Protease

(S195A rat trypsin)

Serpin

(A353K Manduca sexta

RCL

11 ________________________________________________________________________________________________________

sheet resulting in a covalent

. Thus, PI-9 acts as a suicide

value of 1.7 × 106 M−1s−1; SI

investigated the importance of the RCL of PI-9

B (granzyme B K27A)

Manduca sexta serpin B1

interacts extensively with This complex can be used to superimpose the crystal structure of uncomplexed

for further investigations of interactions.

play a key role in the development of Hodgkin

Thomas Hodgkin already described several cases of a disease [88] that

since it was realized that it is a lymphoid

mononucleated Hodgkin cells and the multinucleated

usually account for only about

cells in the tumor tissue, although some cases may show more than 10 % HRS cells. The


eosinophils and mast cells [91].

Protease

(S195A rat trypsin)

Manduca sexta B1)

12 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Today HL is sub-divided according to differences in morphology and phenotype of the cells and the

composition of the cellular infiltrate into classical HL (cHL, 95 % of all cases) and nodular lymphocyte-

predominant HL (NLPHL, 5 % of all cases) [91]. CHL is further classified into nodular sclerosis, mixed

cellularity, lymphocyte depletion and lymphocyte-rich HL lymphoma. HRS cells of cHL probably

originate from germinal center B cells that have acquired disadvantageous immunoglobulin variable

chain gene mutations which usually results in cell death. Lymphocytic and histiocytic cells of NLPHL

apparently derive from antigen-selected germinal center B cells. Only a few cases of cHL originate

from T cells. The extent to which the cHL cells undergo reprogramming of gene expression is unique

among human lymphomas. They succeed in expression of multiple genes that are actually typical for

other types of cells within the immune system and lose the expression of most B cell-typical genes

[92]. NFkB, a key anti-apoptosis factor in HL, is one of the signaling pathways and transcription

factors that are de-regulated within HRS cells [93]. This transcription factor is activated by various

external stimuli of which many trigger inflammation as well. Its inactive form is present in the

cytoplasm of almost every cell type, bound to an inhibitory protein called IkB, of which IkBα is the

best studied. Upon stimulation of NFkB, IkB kinases are activated which lead to phosphorylation and

ubiquitin-mediated degradation of IkB, so active NFkB is released. NFkB activity comprises its

translocation into the nucleus where it induces the expression of dozens of different genes involved

in cellular processes as diverse as the immune system, in inflammation, cell proliferation, and

inhibition of apoptosis as well as in angiogenesis, invasion and metastasis [94,95]. Recent studies

showed that the constitutive activation of NFkB in chronic inflammation is a central molecular link

between inflammation and cancer development [96]. NFkB was found to be expressed in many HL

cell lines as well [93]. Discussions are ongoing if stimulation of CD30, a receptor highly over-

expressed on HL cells, activates NFkB, leading to increased resistance of the cells towards apoptosis

[97-100].

Other hypotheses on how HRS cells circumvent immune attack by CTLs or NK cells include the

formation of an inflammatory microenvironment resulting from the attraction of many nonmalignant

inflammatory cells into the lymphoma tissue. Thus, unlike most other human malignancies, the HRS

cells are vastly outnumbered by the surrounding nonmalignant inflammatory cells. 99 % of the tumor

mass consists of reactive inflammatory cells (lymphocytes, histiocytes, eosinophils, fibroblasts,

neutrophils, plasma cells) [96]. Thus, the histological appearance is reminiscent of an inflammatory

process as seen for viral infections, which supports the widely accepted finding of a causative role of

EBV infection in HL in epidemiological studies. In 40 % of cHL cases in the western world, HRS cells

contain EBV [101,102].

HL is one of the most curable cancer types. In recent years combinatorial treatment approaches

using chemotherapy and radiation therapy have led to promising cure rates. The estimated 5- and

13 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

10- year relative survival for patients with HL is 85 % and 80 %, respectively [103,104]. However,

patients suffer from high toxicity of current therapeutic regimens and in addition there are still

patients that relapse even after autologous hematopoietic stem-cell transplantation which is

reserved for patients with recurrent or progressive HL after failure of conventional therapy. Those

patients have bad prognosis and limited therapeutic options. Hence, efforts have been made to find

alternative therapies for those resistant cancer subtypes.

The CD30 receptor is an activation-induced antigen that is predominantly expressed on lymphoid

cells. CD30 has been shown to be over-expressed on malignant cells of HL, anaplastic large cell

lymphoma (ALCL) and adult T cell leukemia. Therefore it is a promising target for the treatment of

those HL patients which do not respond to standard therapy [105,106]. CD30 is a glycosylated type I

transmembrane protein and belongs to the TNF receptor superfamily [107]. It can be proteolytically

cleaved in close proximity to the cell membrane by the membrane-anchored metalloproteinase TNF-

α converting enzyme (TACE), leading to the release of the soluble ectodomain of CD30. This is found

at low levels in the serum of healthy donors but is detected at elevated levels in patients suffering

from CD30-positive neoplasms, viral infections, and autoimmune disorders. Hence, in most cases the

level of soluble CD30 provides a prognostic marker [108].

The most recent novel immunotherapeutic agent that targets CD30 is the above mentioned

Brentuximab Vedotin (1.2.1) for the treatment of HL patients that have not responded to prior

chemotherapy or autologous stem cell transplantation. Other immunotherapeutic agents targeting

this receptor have been described in context of in vitro, pre-clinical and clinical studies. Apart from a

CFP consisting of the human RNase angiogenin and the natural ligand of CD30 (CD30L) [109] there

are several other anti-CD30 human IL-2 fusion proteins that have Ber-H2(scFv), among others, as a

binding domain [110]. In addition the construct Ki4(scFv)-ETA’, which as with Ber-H2(scFv), does not

bind soluble CD30 receptor and even prevents shedding [111], has shown promising results in vitro

and in vivo [112]. However, it has the disadvantage of containing a toxic compound of non-human

origin that has the potential to provoke undesirable immune responses. Therefore, fusion to human

effector molecules would be advantageous.

14 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

1.6 CHRONIC AND ACUTE MYELOMONOCYTIC LEUKEMIA AND THE CD64

RECEPTOR

Chronic myelomonocytic leukemia (CMML) is one of the three rare myeloid neoplasms termed

myelodysplastic/ myeloproliferative diseases according to the World Health Organization (WHO).

This clonal hematopoietic stem cell disorder is characterized by persistent monocytosis in peripheral

blood, at least one dysplasia component within the bone marrow and a percentage lower than 20 %

of promonocytes and blasts in peripheral blood and bone marrow [113]. It can be further classified

depending on the amount of blasts into CMML 1 (< 5 % blasts in peripheral blood, < 10 % blasts in

bone marrow) and CMML 2 (< 20 % blasts in peripheral blood, 10 - 19 % blasts in bone marrow)

[114]. CMML mainly occurs in elderly people with median survival of 15 – 20 months and leukemic

transformation rates in 15 – 30 % (both depending on type 1 or 2) [115]. Other diagnostic criteria

include immunophenotyping, whereby CMML and monocytic AML show overlapping antigen

expression patterns. Only the expression of CD56 combined with underexpression of a myeloid

marker like CD15 or CD13 seems to be unique to CMML monocytes [116]. Furthermore an over-

expression of CD33 [117,118] and also of CD14 on marrow monocytes was found in CMML compared

to reactive monocytosis and normal marrow samples [116]. In addition, clonal cytogenetic

abnormalities are found in 20 – 40 % of patients with CMML, but none is specific [119-121]. The close

similarity among the other myelodysplastic/ myeloproliferative diseases such as atypical chronic

myeloid leukemia (CML) and juvenile myelomonocytic leukemia (JMML) means a constant challenge

for diagnosis as well as the appropriate therapy. New molecular markers have been described in

recent years such as mutations in TET2 (Tet Methylcytosine Dioxygenase 2) or CBL (Casitas B-lineage

Lymphoma) [122]. However, these aberrations are non specific for CMML and no targeted therapies

are available for these genetic changes today. Because of this and its rare occurrence, the

establishment of new therapeutic approaches specific to CMML is difficult.

Conventional treatment of CMML is very heterogeneous. It usually starts with the ‘watch and wait

strategy’. The only curative approach for CMML is allogeneic hematopoietic stem cell

transplantation, but this treatment option is often not applicable in patients affected by CMML due

to the age distribution of the disease. Actually, the gold standard treatment is to decrease the

amount of white blood cells in peripheral blood in the beginning with low dose cytarabine,

hydroxycarbamide or etoposide. Another option is supportive care with transfusion. Depending on

the variant of the disease, compounds which had been used to treat myelodysplastic syndrome were

explored for usage in CMML, especially the demethylating agents azacytidin und decitabin,

topoisomerase inhibitors such as valproic acid and farnesyltransferase inhibitors like lonafarnib.

However, information on the significance of these drugs is very limited due to the rare occurrence of

15 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

CMML [123]. In patients with CMML 2, a treatment with DNA methyltransferase-inhibitors has also

been shown to be effective by reducing the rate of transformation into AML and to reduce the

transfusion need. However approaches based on molecular or biological characteristics of the

disease are urgently needed [123]. Those new treatment strategies are unmet medical need for

these patients.

In some patients with a history of CMML, the disease converts into acute myelomonocytic leukemia

(AMML) which is also called AML-M4, the fourth sub-category of AML according to FAB (French-

American-British) classification. This disease is diagnosed in 15-20 % of all AML cases. It occurs in all

age groups, but is more common in older adults, and is initially treated with chemotherapy. AML-M4

cases have a population of < 20 % myeloblasts and monoblasts and < 80 % of total nucleated cells are

monoblasts. As indicated above, the expression patterns of AMML do not distinguish significantly

from those of CMML. In general, myeloblasts express CD13, CD33, CD34 and CD117 whereas

monocytoid cells including monoblasts and promonocytes express CD11b, CD11c, CD13, CD14, CD33

and CD64 [114,118]. Treatment of AMML includes usually intensive multidrug chemotherapy and in

some cases allogeneic hematopoietic stem cell transplantation.

Treatment approaches of other leukemic diseases such as AML comprise targeting of the 72 kDa

glycoprotein CD64, the high affinity receptor for human IgG (Fc γ receptor I, FcγRI), which is

expressed on cells of the myeloid lineage, i. e. monocytes, macrophages and dendritic cells, but is

also over-expressed on the surface of leukemic cells [124,125]. The corresponding cells can be

activated by several cytokines, such as IFNγ, which plays a crucial role in inflammatory responses. It

also leads to a 5- to 10-fold increase of CD64 expression as well as enhanced receptor activity [126],

whereby the amount of CD64 receptors is no prime requisite for sensitivity of the cells, but rather

their activation state. This means that an increased CD64 expression is not necessarily accompanied

by a greater sensitivity to cell-death inducing agents, but rather their corresponding activation via

e.g. IFNγ [127,128]. The advantages of CD64 for targeted delivery of cytotoxic effector molecules

include its highly efficient internalization rate and its limited expression, so normal, non-activated

CD64-expressing cells are not affected by the fusion proteins [127,129,130]. The specific killing of

target cells via CD64 can be accomplished by the fusion of cell-death inducing moieties to H22(scFv),

a humanized single chain antibody fragment [131,132]. The potential of fusions of H22(scFv) to both

Ricin A and ETA' as therapeutic agents against AML or inflammatory cells has been shown previously

[129,133] as well as the efficacy of Gb-H22(scFv) against AML cell lines and primary cells (Table 9).

16 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

1.7 MOLECULAR MODELING

Molecular modeling can be applied to model or mimic the behavior of molecules and for studying the

interaction between more than one molecule and large molecular systems. The underlying features

of the computational calculations use atoms as the smallest individual unit. Alanine scanning is a fast

computational approach for the prediction of energetically important amino acid residues in protein-

protein interfaces [134,135]. To determine the role of a specific side chain group of a specific residue,

systematic mutations to alanine are performed. Alanine has a small and inert functional group;

therefore it is possible to measure the effect of the deletion of potential crucial side chains. The

output is a list of ‘hot spots’ or amino acid side chains that are predicted to significantly destabilize

the interface when mutated to alanine [136]. Despite the large size of binding interfaces, single

residues can contribute to a large fraction of the binding free energy in an interface [134]. The

disadvantages of alanine scanning are that it does not allow the investigation of the impact of

mutations other than alanine; it does not include hydration at the protein-protein interface in spite

of the fact that water molecules may play a very important role for complex stability; and, it does not

fully take into account the complete flexibility properties of the complex. Hence, alanine scanning

provides only very approximate results. To overcome these limitations, these first predictions are

succeeded by molecular dynamics (MD) simulations in a vacuum (gas-phase simulation), in the

presence of solvent molecules (explicit solvent simulation) or in empirical mathematically expressed

solvents (implicit solvent simulation). Subsequently, those calculations allow predicting which of the

suggested mutation sites (from alanine scanning) destabilize the determined protein-protein

complex the most.

Today, molecular modeling methods are routinely used in pharmaceutical research to investigate

protein folding, enzyme catalysis, protein stability, conformational changes associated with bio-

molecular function, and molecular recognition of proteins, DNA, and membrane complexes to e.g.

design novel drug candidates.

17 Introduction

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

1.8 AIMS AND OBJECTIVES

Conventional cancer treatment usually involves chemotherapy and radiotherapy, which can cause

severe side effects. More targeted approaches are therefore preferred, such as immunotherapy

based on immunotoxins (ITs) comprising an antibody-based ligand that specifically binds to diseased

cells, and a pro-apoptotic protein. However, if the toxin component is derived from bacteria or plants

it can trigger a neutralizing immune response; therefore, human effector molecules are more

suitable. This results in the fourth generation of ITs, human CFPs. The most prominent example of

these fourth generation ITs, that has been successfully tested in cytotoxicity assays against different

cancer cells in vitro and ex vivo, comprises granzyme B as pro-apoptotic candidate. There are three

main reasons for the application of granzyme B as part of CFPs. First, it is a human effector molecule

so the risk of immunogenic responses is limited. Second, the use of granzyme B is not only beneficial

due to its high apoptotic efficiency as the major mediator of cellular immunity, but also because of its

capacity to induce apoptosis at multiple levels so the cancer cells’ lack of sensitivity due to

deregulated apoptotic pathways or overexpression of certain inhibitors may be overcome. Third, it is

relatively flexible concerning heterologous expression. On the downside, during the last decade, the

presence of the endogenous inhibitor PI-9 has been discovered in different types of cancer cells. This

diminishes the universal potential of granzyme B in targeted tumor therapy. Since the expression of

PI-9 has especially been described in the context of treatment-resistant cells and poor clinical

outcome, it becomes evident that alternative therapeutic approaches become necessary to kill these

cells. The aim of this work was the evaluation of human CFPs which induce cell-death independent of

the endogenous expression of PI-9. This should be achieved by highly cytotoxic granzyme B mutants

resistant to PI-9 inhibition.

The following steps will be pursued to achieve this goal (overview shown in Figure 2):

• Elucidation of the most important amino acids responsible for the interaction between the

serine protease and the RCL of its specific inhibitor by molecular homology modeling.

• Identification of suitable amino acids to replace the identified residues with the aim to

prevent an irreversible inactivation of the enzyme by covalent complex formation without

affecting proteolytic activity.

• Verification of the desired effects in silico (cooperation with German Research School for

Simulation Sciences in Juelich, Prof. Dr. rer. nat. Paolo Carloni).

• Generation of different fusion constructs comprising granzyme B and the CD30-targeting

Ki4(scFv) and CD64-targeting H22(scFv), mammalian secretory expression and purification of

the granzyme B mutants.

• Identification of the most promising mutants by in vitro PI-9 inhibition assay.

Introduction _____________________________________________________________________________________________________________________________________________________________________________________________________________

• Set up of a cell-based test system to identify the

• Establishment of xenograft tumor models

towards PI-9-positive and

• Ex vivo verification of the efficacy of the selected mutant on

primary monocytic cells from

Another potent granzyme, granzyme

of human effector molecules for CPFs

effective in killing tumor cells. I

PI-9, so its application increases the potential to kill even

the way for combinatorial treatment with granzyme

The following steps were taken

granzyme-based CFPs:

• Generation of different

Ki4(scFv) and CD64-targeting H22(scFv), mammalian secretory expression and purification

• Determination of the enz

• Verification of the cytotoxic functionality

lines in vitro and CMML and AMML patient

granzyme B-based constructs.

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

based test system to identify the optimal mutant.

Establishment of xenograft tumor models using HL cell lines to confirm anti

positive and -negative tumor cells in vivo.

verification of the efficacy of the selected mutant on PI-9-positive and

primary monocytic cells from CMML and AMML patients.

granzyme M, will also be investigated in this study to extent the portfolio

for CPFs. As with granzyme B, granzyme M has been

In addition, granzyme M specifically cleaves the granzyme

its application increases the potential to kill even immune-resistant tumor cells

the way for combinatorial treatment with granzyme B.

taken to verify the potential of this enzyme in context of

Generation of different fusion constructs comprising granzyme M and

targeting H22(scFv), mammalian secretory expression and purification

nzymatic activity of the novel constructs.

Verification of the cytotoxic functionality of PI-9-positive and PI-9-negative

CMML and AMML patient-derived cells ex vivo in comparison to

based constructs.

Figure 2: Workflow of this study.

18 ________________________________________________________________________________________________________

L cell lines to confirm anti-tumor activity

positive and -negative

to extent the portfolio

M has been proven highly

specifically cleaves the granzyme B inhibitor

resistant tumor cells or to pave

to verify the potential of this enzyme in context of novel human

nd the CD30-targeting

targeting H22(scFv), mammalian secretory expression and purification.

negative HL or AML cell

in comparison to

Material and methods ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

19

2 Material and methods

2.1 CHEMICAL AND BIO-CHEMICAL MATERIAL

2.1.1 Equipment

The equipment and devices used in this work are listed below.

Autoclaves: Varioklav (H+P) 75S (Thermo Electron Corporation)

Balances: Precision and analysis balances, Talent: TE6101, TE64, TE31025, TE12000 (Sartorius)

Cell counter: Casy1 (Schaerfe System)

Centrifuges: Table centrifuge 5415D and 5415R (Eppendorf), Multifuge 3 S-R / 1S (Heraeus Instruments), Avanti J-25I Centrifuge (Beckman Coulter), Rotina 420 R (‘Hettich Zentrifuge’)

Cryo sections: Cryostat CM 3050 (Leica)

DNA sequencing: ABI Prism 3730 Capillary-Sequencer (Applied Biosystems)

End-over-end shaker (CMV): Innova TM 4000 incubator shaker or Innova TM 4430 incubator shaker (New Brunswick Scientific)

Flow cytometry: FACScalibur (Becton Dickinson), FACS-Vantage with DiVa option (Becton Dickinson)

FPLC: Aekta FPLC (Amersham Pharmacia Biotech)

Gelelectrophoresis and Electroblotting:

Mini Protean III gel chamber (BioRad Laboratories, Inc.), Protein Gel-Apparatus and supplies Mini PROTEAN II™ (BioRad), Gel Air Dryer (BioRad), Mini Trans-Blot Cell (BioRad)

Ice machine: Icematic D201 (Castel Mac)

Incubators: Thermomixer compact (Eppendorf AG), Incubator Function Line Type UT12 (Heraeus Instruments); cell culture: CD210 (Binder)

Liquid handling: Pipetman P (Gilson), Multichannel Pipettes (Eppendorf)

Microplate reader: ELISAreader ELx808 (Bio-TEK Instruments) and Epoch (Bio-TEK Instruments)

Microscopes: TCS SP5 confocal laser microscope (Leica), DMIL inverted microscope with mercury fluorescence light source (Leica), DM2000 (Leica)

PCR-Cyclers: Primus 96 Plus (MWG-Biotech), Programmable Thermal Controller PTC-200™ (MJ Research Inc.)

pH-Meter: Basic Meter PB-11 (Sartorius)

Photometer: Biophotometer (Eppendorf),

Software: AIDA image analyzer (Raytest), GraphPad Prism 4.0, WinMDI 2.8

Sonication: Bandelin Sonopuls HD 2070 (Bandelin Electronic) Micro tip MS72/73

Transiluminator: Molecular Imager Gel Doc XR System (BioRad)

Waterbaths: TE31025, TE12000 (Sartorius)


20

2.1.2 Chemicals and consumables

If not otherwise indicated, the chemicals used were purchased from the following companies in the

highest available purity: BD Biosciences (Heidelberg), Bio-Rad (Munich), Qualilab (Olivet, France),

Merck (Darmstadt), Roche Diagnostics (Mannheim), Roth (Karlsruhe), Invitrogen (San Diego, USA),

Sigma (Munich), MWG-Biotech (Ebersberg), KMF Laborchemie Handels GmbH (Siegburg), Peprotech

GmbH (Hamburg).

Expendable materials were usually purchased from: BD Biosciences (Heidelberg), Biozym

(Oldendorf), Eppendorf (Hamburg), Greiner (Solingen), Invitrogen (San Diego, USA) , Labomedic

(Bonn), Millipore (Eschborn), Pierce Biotechnology (Rockford, USA), Qiagen (Hilden), Schott (Mainz),

Whatman (Bender & Hobheim, Bruchsal), Zeiss (Oberkochen).

2.1.3 Kits, enzymes and standards

If not otherwise indicated, the restriction enzymes, Quick ligase, T4-DNA-Ligase and Phusion High

Fidelity DNA-polymerase used were purchased from: Invitrogen (San Diego, USA), MWG-Biotech

(Ebersberg) and New Englands Biolab (NEB) (Frankfurt am Main).

The following kits from Macherey-Nagel (Dueren) were used for isolation and purification of DNA or

RNA according to the manufacturer’s instructions:

Plasmid isolation kit (Mini): NucleoSpin® Plasmid

Plasmid isolation kit (Maxi): NucleoBond® AX

DNA-purification kit: NucleoSpin® Extract II

RNA purification kit: NucleoSpin® RNA II

Protein standards:

Pre-stained broad range marker: 175 / 80 / 58 / 46 / 30 / 25 / 17 / 7 kDa; #P7708, NEB (Frankfurt am

Main)

Protein broad range marker: 212 / 158 / 116 / /97.2 / 66.4 / 55.6 / 42.7 / 34.6 / 27.0 / 20.0 / 14.3 /

6.5 kDa, #P7702, NEB (Frankfurt am Main)

Nucleic acid standards:

2-log ladder: 10.0 / 8.0 / 6.0 / 5.0 / 4.0 / 3.0 / 2.0 / 1.5 / 1.2 / 1.0 / 0.9 / 0.8 / 0.7 / 0.6 / 0.5 / 0.4 / 0.3

/ 0.2 / 0.1 kb, #N3200, NEB (Frankfurt am Main)


21

2.1.4 Synthetic oligonucleotides and plasmids

Synthetic oligonucleotides were purchased from MWG-Biotech (Ebersberg) in standard quality. The

DNA oligonucleotides used in this work are listed in Table 1.

Table 1: List of oligo nucleotides used for cloning and DNA sequencing.

Oligo name Oligo sequence

Sequencing

T7 promotor 5‘-TAATACGACTCACTATAGGG-3‘

Hinterhis rev 3‘-CCAACAGCTGGCCCTCGCAGACAG-5‘

3’Gb seq 3‘-CTTCACCTGGGCCTTGTTGC-5‘

5‘ pMS 5‘-GGGCGGTAGGCGTGTACGGTGGG-3‘

5‘ PI9 seq 5‘-AGTGGAATGAACCGTTTGAC-3‘

3‘ PI9 seq 3‘-CGAGCTTAAACGTGGCCTCC-5‘

Cloning

5‘ pMS 5‘-GGGCGGTAGGCGTGTACGGTGGG-3‘

3‘ Gb (BlpI) 3‘-CGCGCTCAGCGTAGCGTTTCATGGTTTTC-5‘

5‘ PI9 (XbaI) 5‘-GCGTCTAGAGAAACTCTTTCTAATGCAAGTGGTAC-3‘

5‘ PI9 (NheI) 5‘-GCCGCTAGCCACCATGGAAACTCTTTCTAATGC-3‘

3‘ PI9 (NotI) 3‘-CGGCGCGGCCGCTGGCGATGAGAACCTGCCACAG-5‘

3‘ PI9 (EcoRI) 3‘-CGCGAATTCTCATGGCGATGAGAACCTGCCACAG-5‘

SOE primer

5‘ Gb SOE K27A 5‘-CAGAAGTCTCTGGCGAGGTGCGGTGGCTTCC-3‘

5‘ Gb SOE R28A 5‘-CAGAAGTCTCTGAAGGCGTGCGGTGGCTTCC -3‘

5‘ Gb SOE R28E 5‘-CAGAAGTCTCTGAAGGAGTGCGGTGGCTTCC-3‘

5‘ Gb SOE R28K 5‘-CAGAAGTCTCTGAAGAAGTGCGGTGGCTTCC-3‘

5‘-EGb S195A 5‘-CTTTAAGGGGGACGCTGGAGGCCCTCTTGTGTG-3‘

3‘ Gb SOE K27A 3‘-GGAAGCCACCGCACCTCGCCAGAGACTTCTG-5‘

3‘ Gb SOE R28A 3‘-GGAAGCCACCGCACGCCTTCAGAGACTTCTG-5‘

3‘ Gb SOE R28E 3‘-GGAAGCCACCGCACTCCTTCAGAGACTTCTG-5‘

3‘ Gb SOE R28K 3‘-GGAAGCCACCGCACTTCTTCAGAGACTTCTG-5‘

3‘-EGb S195A 3‘-CACACAAGAGGGCCTCCAGCGTCCCCCTTAAAG-5‘


22

The underlying plasmids for all molecular cloning procedures were pMS for mammalian expression

[137] or pMT for bacterial expression [133]. All cloning work was adapted to the characteristics of the

chosen vectors, such as corresponding restriction sites, and performed according to standard

methods (2.2).

For expression of all recombinant granzyme B and granzyme M constructs the already published

plasmids pMS-L-EGb-H22-IV [138] and the earlier generated pMS-L-EGb-Ki4-IV and pMS-L-EGb-IV

were applied as template plasmids. The granzyme M sequence was prepared earlier during a

bachelor thesis by PCR using specific primers based on a full-length cDNA clone (IRATp970D0643D)

from the German Resource Centre for Genome Research (RZPD).

For expression of recombinant human PI-9, cDNA was derived from RNA preparations of target cell

lines by reverse transcription and specific primers (2.1.4). The derived sequence was either cloned

into the pMS vector, deleting the Igκ leader sequence for cytosolic mammalian expression

(pMS-PI9-IV) or into the pMT plasmid to express PI-9 in E. coli (pMT-L-PI9).

2.1.5 Antibodies and antibody fragments

The following antibodies and antibody conjugates were used in this work for western blot (2.4.2) or

flow cytometric (2.5) detection of recombinant or soluble proteins with the indicated dilutions:

- GAM-AP (1:5000) or GAM-PO (1:5000) or GAM-FITC (1:100): Goat-anti-mouse polyclonal

antibody directed against the Fc-part of murine IgGs, conjugated to Alkaline Phosphatase

(AP), horseradish peroxidase (PO) or Fluorescein-Isothiocyanat (FITC) (Sigma, Munich)

- anti-His (1:5000): monoclonal mouse-anti-penta-His antibody directed against C- or

N-terminal Histidin Tags (His-tags) of recombinant proteins (Sigma, Munich)

- anti-His Alexa488 (1:300): monoclonal mouse-anti-penta-His Alexa Fluor®488 antibody

directed against the C- or N-His-tags of recombinant proteins, conjugated to fluorescent dye

Alexa Fluor 488 (Qiagen, Hilden)

- anti-human Gb-PE (1:50): monoclonal mouse-anti-granzyme B-PE antibody directed against

human granzyme B, conjugated to R-Phycoerythrin (PE) (BD Bioscience, Heidelberg)

- anti-human Gm (1:500): monoclonal mouse-anti-granzyme M antibody directed against

human granzyme M, clone 4D11 (Abnova, Heidelberg)

- anti-human Gb (1:1000): monoclonal mouse-anti-granzyme B antibody directed against

human granzyme B, clone 2C5 (Santa Cruz, San Diego)

- anti-human PI-9 (1:300): monoclonal mouse-anti-PI-9 antibody directed against human PI-9,

clone 7D8 (Santa Cruz, San Diego)


23

The following antibodies were used for the flow cytometric (2.5) detection of surface antigens on

cancer cell lines or primary leukemic cells according to the manufacturer’s instructions. They are

monoclonal mouse-anti-human antibodies conjugated either to FITC, PE, Allophycocyanin (APC),

Alexa 647 or 488 and detect the antigens as indicated by their name.

- anti-CD64-FITC (AbD Serotec, Duesseldorf)

- anti-CD64-488 (Biolegend, San Diego, USA)

- anti-CD64-Alexa 647 (AbD Serotec, Duesseldorf)

- anti-CD14-APC (eBioscience, Frankfurt)

- anti-CD14-FITC (Dako, Hamburg)


- anti-CD56-PE (eBioscience, Frankfurt)


- anti-CD25-PE (eBioscience, Frankfurt)

In addition, the single chain antibody fragments (scFv) listed in Table 2 were used in this study.

Table 2: Origin and specificity of antibody fragments.

Antibody/ligand specificity origin/specification

H22(scFv) human CD64 humanized monoclonal antibody H22

Ki4(scFv) human CD30 mouse monoclonal antibody Ki4

425(scFv) human EGF receptor, extracellular domain III (no mouse cross reactivity)

mouse monoclonal antibody 425

2.1.6 Bacterial strains

The following E. coli strains were used for selection and amplification of plasmid DNA, and for

expression of recombinant protein respectively:

DH5α: F- φ80 lacZΔM15 Δ(lacZYA-argF) U169 recA1 endA1 hsdR17(rk-, mk+) phoA supE44thi-1 gyrA96

relA1 tonA (resistant to phage T1) (Invitrogen, Darmstadt)

BL21 StarTM (DE3): F- ompT hsdSB (rB-mB-) gal dcm rne131 (DE3) (Invitrogen, Darmstadt)


24

2.1.7 Mammalian cell lines

The list in Table 3 displays all mammalian cell lines used, including their background and indicating

the receptor that played a role in this work.

Table 3: Mammalian cell lines.

Name Background Receptor

used

Reference / Source

HEK293T highly transfectable derivative of the human primary embryonal kidney cell line 293

/ ATCC, CRL-11268

HL60 established from the peripheral blood of a 35-year-old woman with AML FAB M2 in 1976

have to be stimulated with IFNγ (50 U/mL) 24 h prior to cytotoxicity/ apoptosis assays [139]

CD64 DSMZ, ACC 3

L540 established from the bone marrow of a 20-year-old woman with HL

CD30 DSMZ, ACC 72

L540cy subline of L540: was recultured from a tumor grown in a nude mouse after injection of L540 and treatment of the mouse with cyclophosphamide

CD30 [140]

L428 established from the pleural effusion of a 37-year-old woman with HL 1978

for mouse experiments transfected with far red fluorescent protein Kat2 (pTag-Katushka2-N; Evrogen)

CD30 DSMZ, ACC 197

L1236 established from the peripheral blood of a 34-year-old man with HL in 1994

CD30 DSMZ, ACC 530

K562 established from the pleural effusion of a 53-year-old woman with CML in blast crisis in 1970

CD30 DSMZ, ACC 10

Jurkat established from the peripheral blood of a 14-year-old boy with acute lymphoblastic leukemia (ALL) at first relapse in 1976

/ DSMZ, ACC 282

A431 established from the solid tumor of an 85-year-old woman (epidermoid carcinoma)

EGFR DSMZ, ACC 91

Ramos established from the ascitic fluid of a 3-year-old boy with American-type Burkitt lymphoma in 1972

/ DSMZ, ACC 603

Karpas 299 established from the peripheral blood of a 25-year-old man with T cell non-HL in 1986

/ DSMZ, ACC 31

Kasumi-1 established from the peripheral blood of a 7-year-old Japanese boy with AML FAB M2 (in 2nd relapse after bone marrow transplantation) in 1989

/ DSMZ, ACC 220


25

2.1.8 Primary cells

Primary cells from patients were obtained through PD Dr. med. Edgar Jost (University Hospital of

Aachen) after informed consent and with the approval of the Clinical Research Ethics Board of the

RWTH Aachen University. The patients’ diagnosis was either AMML or CMML 1 (here abbreviated as

CMML). Diagnosis of CMML was made according to the commonly used WHO-criteria. No prior

treatment occurred for any patients except for CMML II (2000 mg/d hydroxyurea) and CMML III

(1500 mg/d hydroxyurea). Mononuclear cells were isolated from peripheral blood by density

gradient centrifugation using Biocoll separating solution (Biochrom AG) and subsequently cultured in

RPMI complex medium (2.3). Prior to cytotoxicity and apoptosis assays they were stimulated with

200 U/mL IFNγ for 1-2 h.

CMML I: male (83 years), 16 % monocytes, 31.5 G/L WBC, 20 G/L PLT, 95 g/L Hb

CMML II: male (66 years), 7 % monocytes, 48.7 G/L WBC, 1212 G/L PLT, 101 g/L Hb

CMML III: male (64 years), 29 % monocytes, 61 G/L WBC, 306 G/L PLT, 86 g/L Hb

CMML IV: male (74 years), 20 % monocytes, 117 G/L WBC, 471 G/L PLT, 11 1g/L Hb

AMML I: male (78 years), 30 % monocytes, 268 G/L WBC, 58 G/L PLT, 88 g/L Hb

AMML II: female (68 years), 70 % monocytes, 37 G/L WBC, 66 G/L PLT, 99 g/L Hb

AMML III: male (45 years), 37 % monocytes, 44 G/L WBC, 44 G/L PLT, 88 g/L Hb

2.1.9 Media and buffers

In Table 4, buffer, solutions and media used in this study are listed.

Table 4: Buffers and media.

Buffer Composition Concentration

10x PBS (pH 7.4) NaCl KCl Na2HPO4 x 12 H2O KH2PO4

1.37 27 81 15

M mM mM mM

1x PBST 1x PBS (pH 7.4) Tween 20

0.05

% (w/v)

5x TGS Tris-HCl (pH 8.3) Glycine SDS

125 960 0.5

mM mM % (w/v)

50x TAE Tris-HCl (pH 8.3) Glacial acetic acid EDTA

2 5.7 50

M % (v/v) mM

5x OrangeG loading buffer

OrangeG Glycerol in 1x TAE

0.01 30

% (w/v) % (v/v)

Coomassie staining solution

Coomassie bb. G-250 Methanol Glacial acetic acid

0.25 50 9

% (w/v) % (v/v) % (v/v)


26

Table 4: Cont.

Buffer Composition Concentration

Coomassie destaining solution

Methanol Glacial acetic acid

10 10

% (v/v) % (v/v)

5x (reducing) protein loading buffer

Tris-HCl (pH 6.8) Glycerol SDS Bromphenolblue (ß-Mercaptoethanol)

62.5 30 4 0.05 10

mM % (v/v) % (w/v) % (w/v) % (v/v)

SDS PAGE gel (separating gel)

Acrylamide/Bisacrylamide (30/1) Tris-HCl (pH 8.8) SDS (10 % (w/v)) TEMED APS (20 % (w/v))

x 375 0.1 0.1 0.1

mL mM % (w/v) % (v/v) % (v/v)

SDS PAGE gel (stacking gel)

Acrylamide/Bisacrylamide (30/1) Tris-HCl (pH 6.8) SDS TEMED APS

5 150 0.1 0.1 0.1

% (w/v) mM % (w/v) % (v/v) % (v/v)

AP buffer Tris-HCl (pH 9.6) NaCl MgCl2

100 100 5

mM mM mM

Transfer buffer for western blotting

Tris-HCl (pH 8.3) Glycine SDS Methanol

25 192 0.05 20

mM mM % (w/v) % (v/v)

Stripping buffer

Tris-HCl (pH 6.8) SDS β-mercaptoethanol

62.5 2 100

mM % (w/v) mM

For FPLC protein purification: 4x binding buffer Tris-HCl (pH 7.4)

NaCl Imidazol

200 1.2 40

mM M mM

Wash buffer Tris-HCl (pH 7.4) NaCl Imidazol

50 300 40

mM mM mM

Elution buffer Tris-HCl (pH 7.4) NaCl Imidazol

50 300 500

mM mM mM

Media

LB (pH 7.5) NaCl Peptone Yeast extract (Agar

1 0.5 1.5 1.6

% (w/v) % (w/v) % (w/v) % (w/v))

SOC Peptone Yeast extract NaCl KCl MgCl2 MgSO4 Glucose

2 % 0,5 % 10 25 10 10 20

(w/v) (w/v) mM mM mM mM mM


27

Their application will be described in later chapters. All media were autoclaved for

20 min/121 °C/1 bar. If necessary, buffers were sterile filtrated with a 0.2 µm PES membrane. The

following antibiotics were applied in media for bacterial cultivation, if necessary, with the indicated

concentrations: Ampicillin (100 µg/µL), Kanamycin (50 µg/µL) and Chloramphenicol (25 µg/µL).

2.2 MOLECULAR-BIOLOGICAL METHODS

The described methods followed standard techniques according to Sambrook and Russell [141].

2.2.1 Polymerase chain reaction

The polymerase chain reaction (PCR) for the amplification of DNA segments was performed using

Phusion High-Fidelity DNA Polymerase (NEB), a proofreading polymerase with 3´→5´ exonuclease

activity resulting in higher accuracy during amplification. The standard volume of a PCR procedure

was 50 μL and composed of 1x HF buffer including MgCl2, 10 pmol 5’ primer, 10 pmol 3’ primer,

0.2 mM dNTPs. After the addition of 50 ng of the template and double distilled water (ddH2O),

0.5 % (v/v) polymerase was added.

The following standard protocol with 30 cycles was used:

30 sec 98 °C Loop30 cycles start 10 sec 98 °C

20 sec 55 °C 30 sec 72 °C

Loop30 cycles end 5 min 72 °C

To verify the successful transformation of a specific insert into bacterial cells, a colony PCR reaction

was run using home-made Taq DNA polymerase. Therefore colonies were picked and heated at 95 °C

for 15 min in 20 µl ddH2O in order to destroy the bacterial cells. After spinning off the cell debris

10 µl of the mixture was added to the PCR mixture containing 1x PCR buffer, 1 mM MgCl2, 10 pmol

5’ primer, 10 pmol 3’ primer, 0.2 mM dNTPs, 1 U Taq DNA polymerase in 25 µL total volume.

The following protocol with 30 cycles was applied:

5 min 96 °C Loop30 cycles start 1 min 96 °C

30 sec 55 °C 45 sec 72 °C

Loop30 cycles end 5 min 72 °C


28

The purification of the PCR products was done according to the manufacturer’s instructions

(Macherey-Nagel Kits, 2.1.3).

2.2.2 Site-directed mutagenesis

Site-directed mutagenesis was conducted by overlap extension PCR using specific primers during an

SOE (splicing by overlapping extension) PCR. The primer pairs for PCR 1 containing the corresponding

mutations are listed in Table 1. To obtain the two fragments, the primer 5’ pMS was used to generate

the 5’ to 3’ fragment and the primer 3’ Gb (BlpI) was used to generate the 3’ to 5’ fragment. A

construct containing the desired single chain sequence as ligand, including the wild-type sequence of

granzyme B, was used as a template. (2.1.4). The PCR mixtures and programs were as follows.

PCR 1:

1x HF buffer

10 ng template DNA

10 pmol 5’ primer

10 pmol 3’ primer

0.2 mM dNTPs

0.02 U/µL Phusion DNA polymerase

Ad 50 µL ddH2O

Program 1:

30 sec 98 °C

Loop30 cycles start 10 sec 98 °C

20 sec 60 °C

30 sec 72 °C

Loop end

5 min 72 °C

Annealing:

2 µL fragment 1

2 µL fragment 2

1x HF buffer

0.2 mM dNTPs


Ad 50 µL ddH2O

Program 2:

30 sec 98 °C


1 min 50 °C

30 sec 72 °C

Loop end

5 min 72 °C

Amplification:

10 µL annealing product

1x HF buffer

10 pmol 5’ primer (5’ pMS)

10 pmol 3’ primer (3’ Gb (BlpI))

0.2 mM dNTPs


Ad 50 µL ddH2O

Program 3:

30 sec 98 °C


20 sec 60 °C

30 sec 72 °C

Loop end

5 min 72 °C


29

2.2.3 Agarose gel electrophoresis

To analyze the size and concentration of the PCR products (2.2.1), isolated plasmids (2.2.8) and

restriction fragments (2.2.5), analytical agarose gel electrophoresis was carried out in 1.2 % (w/v)

agarose at field strength of 120 V after addition of 5x OrangeG loading buffer (Table 4).

Electrophoresis was run in 1x TAE buffer with 0.3 μg/mL ethidium bromide. 0.33-0.6 μg 2-Log DNA

Ladder (NEB) was used as the standard for determining the size and concentration of the fragments

(2.1.3). DNA bands were visualized on a UV transilluminator at a wavelength of 302 nm. The bands

obtained were documented with a video documentation system using the software ‘Quantity One

Document Software’ version 4.6 (BioRad, Munich).

2.2.4 Zero blunt TOPO PCR cloning

Since the digestion of PCR products (2.2.1) was often inefficient, the PCR products were cloned into a

TOPO vector according to the manufacturer’s instructions of Zero Blunt® TOPO® PCR Cloning Kit

(Invitrogen, Darmstadt). After growing of the corresponding bacteria, the purified plasmids could be

digested as described below (2.2.5).

2.2.5 Endonuclease restriction of DNA

The double digestion with endonucleases and buffers from NEB (Frankfurt am Main) was performed

according to the manufacturer’s instructions in a total volume of 50 µL with 1 to 2 µg DNA. The

isolation of restricted DNA occurred via preparative agarose gel electrophoresis (2.2.3). After

separation of DNA fragments, desired DNA bands were excised using a clean scalpel on a UV

transilluminator. The isolation of the DNA from the gel was performed according to the

manufacturer's protocol (2.1.3). The concentration of the purified DNA was determined using

analytical agarose gel electrophoresis (2.2.3) or by measuring the absorbance at 260 nm. The

extracted and purified DNA was then used for ligation (2.2.6). For control digestions, 5 times less

than the recommended volume was used and the specific bands were visualized as described above

(2.2.3).

2.2.6 Ligation of DNA fragments

Two purified restriction digested fragments (2.2.5) were ligated with T4 DNA ligase and Quick ligase

respectively, using the buffer system specified by the manufacturer with a total volume of 20 μL. The

ligation mix contained the insert in triplicate molar excess of vector quantity, and the mixture was

incubated overnight at 16 °C and for 5 min at room temperature (RT) respectively. Afterwards, the

whole ligation mixture was transformed into heat shock competent bacteria (2.2.7), as explained


30

below.

2.2.7 Transformation of plasmid DNA into competent E. coli

Heat shock competent cells (2.1.6) were prepared with Rhubidium-chloride according to Hanahan

[142]. The aliquoted cells were thawed on ice before the DNA was added (using either 20 µL ligation

mix (2.2.6) or 30 ng purified plasmid DNA (2.2.8) for re-transformation). After incubation for 30 min

on ice, the cells were incubated at 42 °C for 30 sec (heat shock). The transformed bacteria were

stored on ice for 2 min and then mixed with 800 μL of sterile SOC medium. For regeneration, the

bacteria were incubated for 1 h at 37 °C before they were plated on LB agar plates (2.1.9) containing

the corresponding antibiotic (ampicillin for pMS-plasmids, kanamycin for pMT-plasmids). The plates

were incubated overnight at 37 °C and colonies were analyzed using colony PCR (2.2.1) the next day.

Replica plates were prepared for at least three clones.

For the final clones glycerol stocks were prepared in 25 % (v/v) glycerol solution and stored at -80 °C.

2.2.8 Preparation of plasmid DNA from E. coli

For small scale DNA preparation, 3 mL of LB medium supplemented with the appropriate antibiotic

(2.1.9) was inoculated and then the recombinant plasmid DNA (2.2.8) was isolated using a

NucleoSpin plasmid kit according to the manufacturer’s instructions (2.1.3). The concentration was

determined at 260 nm.

2.2.9 Sequencing of DNA

To verify the desired sequence of recombinant clones (2.2.7, 2.2.8), sequencing was carried out at

the Fraunhofer IME sequencing facility by the chain termination method using the ‘Applied

Biosystems 3700 DNA Analyzer’ and the BigDye™ cycle sequencing kit. 160 ng/kb plasmid DNA was

added to ddH2O in a total volume of 28 μL with 20 pmol of sequencing primer (2.1.4). The resulting

sequences were evaluated with the ‘CLC Main Workbench 6.0’ and ‘Clone Manager 9.0’ software.

2.3 CELL CULTURE

2.3.1 Cultivation of cell lines and primary cells

All mammalian cell lines (Table 3) as well as the primary monocytic cells (2.1.8) were cultured in T25

or T75 cell culture flasks (Greiner) in RPMI1640 (Gibco) supplemented with 10 % (v/v) heat-

inactivated fetal calf serum (FCS) (Gibco), 50 μg/mL penicillin and 50 μg/mL streptomycin. The cells


31

were maintained at 37 °C in a humidified atmosphere of 5 % CO2. After transfection of HEK293T cells,

100 µg/mL Zeocin was added for selection purposes.

When large amounts of secretory expressed protein were intended, triple flasks (VWR, Darmstadt) or

cell factories (Thermo Fisher Scientific, Waltham, USA) were inoculated, so supernatant volumes

above 1 L could be collected easily for purification.

For viability and apoptosis assays 200 or 50 U/mL Interferon (IFN) γ was added for stimulation of

CD64 expression in the primary cells or HL60 cell line, respectively.

2.3.2 Transfection of cell lines

The physical or chemical transfection of different mammalian cell lines was performed according to

the manufacturer’s instructions (Table 5) and using their protocols for optimization. As

recommended, the cells were kept in serum-free RPMI medium or RPMI containing 10 % (v/v) FCS

without antibiotics. The efficiency of transfection was monitored via flow cytometric analysis.

Table 5: Reagents for physical or chemical transfection of mammalian cell lines.

Name Company

Physical:

Amaxa Nucleofection Kit V Lonza (Cologne)

Neon Transfection systems Life technologies (Darmstadt)

Chemical:

Nanofectin Kit PAA (Pasching, Austria)

Attractene Transfection reagent

Qiagen (Hilden)

GeneJuice Transfection Reagent

Novagen (Darmstadt)

TransIT-Jurkat MirusBio (Madison, USA)

Ingenio Electroporation Kit MIR 50109

MirusBio (Madison, USA)

Roti-Fect Reagent Roth (Karlsruhe)

2.3.3 Protein delivery into cells via streptolysin O

Reversible permeabilization of cells can be achieved via streptolysin O (Sigma, Munich). For protein

delivery 5*106 cells/mL were washed once with HEPES buffer (20 mM HEPES, pH 7.4, 150 mM NaCl)

and re-suspended in 100 µL HEPES buffer containing 0.5 % (w/v) BSA, 5 mM DTT and 10 mM glucose.


32

The cells were incubated with 8 µg/mL streptolysin O for 15 min at 37 °C in a 96-well plate.

Afterwards the cell-death inducing protein was added and cells were further incubated for 4 h. The

permeabilization step was conducted in the absence of Ca2+, whereby resealing of the cells was

induced by addition of 1.25 mM Ca2+ 1 h after addition of the cell-death inducing protein. Apoptotic

effects were monitored via Annexin V/ PI assay as described in 2.7.2.1.

2.4 PROTEIN-BIOCHEMICAL METHODS

2.4.1 SDS-PAGE

To determine the amount and identity of proteins (2.4.3, 2.4.4) according to their molecular weight,

analytical protein separation was performed via discontinuous SDS- poly-acryl-amide (SDS-PAA) gel

electrophoresis (SDS-PAGE) based on the method of Laemmli [143]. The denaturing SDS-PAA gel was

prepared with the above mentioned buffers (Table 4). Before the gel was loaded, the samples were

mixed with Laemmli sample buffer and heated for 5 min at 95 °C to destroy the structure of the

proteins. The gel was run with 160 mV in ‘Mini-Protean III’-protein gel electrophoresis chamber

(BioRad) until the samples reached the bottom of the gel. The gel was dyed in a coomassie brilliant

blue solution for 10-30 min and destained with the described buffer (Table 4). With the help of the

loaded standard (2.1.3) the proteins could be identified.

2.4.2 Western blot analysis

Western blot analysis was performed using the ‘Mini Trans-Blot® Electrophoretic Transfer Cell’ (BIO-

RAD) according to the manufacturer's instructions with transfer buffer (Table 4). After the transfer

(250 mA for 90 min or 40 V overnight at 4 °C) of proteins, separated by SDS-PAGE, on a nitrocellulose

membrane (Protran, Whatman), blocking was performed using 2.5 % (w/v) milk powder resolved in

PBST either for 1 h at RT or overnight at 4 °C. After washing three times with PBST, the membrane

was incubated with the primary antibody in PBST for 1 h at RT. The same procedure was followed for

incubation with the secondary antibody. After final washing in PBST and water, the membrane was

developed in the dark depending on the conjugation of the second antibody (2.1.5) if not otherwise

indicated using NBT/BCIP [3.33 % (w/v) NBT, 1.67 % (w/v) BCIP in dimethylformamide] added to the

100x volume of AP buffer (Table 4) for AP-labeled antibodies or with an enhanced

chemiluminescence (ECL) system (Thermo Scientific) and LAS-3000 reader (Fujifilm) for PO-

conjugated ones.


33

2.4.3 Bacterial expression and purification

For the expression of recombinant PI-9, the E. coli expression strain BL21 (DE3) was used (2.1.6). To

evaluate successful protein expression, optimal conditions and sufficient purification steps, small

scale experiments were performed in a volume of 300 mL according to the published stress-

expression protocols [144] or the protocols from Qiagen (‘The QiaExpressionistTM’). In order to obtain

enough protein for further determinations, fermentation with 4 L volume (BioFlo 110 fermenter,

400-1000 rpm) was run in synthetic minimal medium (16.6 g/L KH2PO4, 4.6 g/L NH4(H2PO4), 70 mg/L

CaCl2*2H2O, 140 g/L FeSO4*7H2O, 1.5 g/L MgSO4*7H2O, 2.1 g/L citric acid, 0.2 g/L L-methionine,

0.2 g/L L-arginine-hydrochloride, 1 mL/L PTM1-trace metals, 25 mg/L Kanamycin, 0.5 mL/L Antifoam)

supplemented with 20 g/L glucose. Inoculation was prepared with 400 mL bacteria culture grown at

37°C overnight in LB medium. The bacteria were harvested 24 h after induction with 1 mM Isopropyl

Thiogalactoside (IPTG).

After harvesting and centrifugation (4000 g, 15 minutes, 4°C) the bacterial pellet was re-suspended in

lysis buffer (50 mM NaH2PO4, pH 8.0, 500 mM NaCl, 0.5 mM DTT) and sonicated six times for 60 s,

70 %, 9 cycles (2.1.1). Due to its enzymatic activity towards proteases, the addition of protease

inhibitor was disclaimed. Cell debris was removed by centrifugation at 30.000 g, 4 °C for 30 min. The

supernatant was supplemented with 10 mM imidazol and purified via Immobilized Metal-ion Affinity

Chromatography (IMAC) and Fast Performance Liquid Chromatography (FPLC) as described below

(2.4.4.2). After a second IMAC step, the protein was re-buffered into 20 mM Tris, pH 7.4, 1 mM DTT

and further purified via anion exchange chromatography (Q-Sepharose XL, 1 mL, GE Healthcare) with

a salt gradient of 0.05–1 M NaCl. To achieve satisfying purity, gel filtration chromatography followed

using a Superdex 75 (GE Healthcare) column in 20 mM Tris, pH 7.5, 50 mM NaCl and 1 mM DTT.

2.4.4 Mammalian expression and purification

All granzyme B- and granzyme M-based constructs were produced secretory in mammalian HEK293T

cells. In addition, PI-9 was expressed cytosolically within these cells.

2.4.4.1 Expression of recombinant protein in HEK293T cells

HEK293T cells were used as the expression cell line (2.1.6). The cells were transfected with 1 µg DNA

according to the manufacturer’s instructions using RotiFect (Roth) (2.3.2). The pMS plasmid used

(2.1.4) encodes for the bicistronic EGFP reporter, so expression of the target protein could be verified

by the green fluorescence via fluorescence microscopy. To obtain high producing cultures

fluorescence activated cell sorting was performed to select high reporter EGFP producing cells.


34

The secretory expression was enabled by the signal peptide sequence of immunoglobulin kappa light

chain (Igκ) encoded by the pMS plasmid (2.1.4). Protein containing supernatants were kept sterile

and were stored at 4 °C until sufficient volumes were collected for purification (2.4.4.2). In case of

cytosolically expressed protein, the cells were harvested, washed with PBS and stored at -20 °C until

further processing (2.4.4.3).

2.4.4.2 Purification of recombinant protein from supernatant

The secreted protein could be purified from the cell culture supernatant via IMAC and FPLC. The

cleared supernatant was supplemented with 10 mM imidazole before loading onto an XK16/20

column (Amersham/ GE Healthcare) containing 8 mL Sepharose 6 Fast Flow resin (GE Healthcare).

The buffers used are listed in Table 4. Afterwards the elution buffer was exchanged with adequate

buffer to enable enterokinase digestion (20 mM Tris-HCl, 200 mM NaCl) via a HiPrep 26/10 desalting

column (GE Healthcare). To obtain higher purity of proteins especially for in vivo experiments, the

protein solution was additionally loaded onto a gel filtration chromatography column (Superdex 75,

GE Healthcare) using the enterokinase buffer.

In order to concentrate the protein solutions, Vivaspin 6 ultrafiltration spin columns (10 kDa MWCO;

Sartorius) were centrifuged at 4000 g and 4 °C. Aliquoted proteins were stored at -80 °C. To activate

granzyme-based constructs enterokinase was added to the protein (0.02 U/µg) with 2 mM CaCl2 for

16 h incubation at 23 °C prior to use. Activated protein was stored after the addition of 1 mM DTT at

-80 °C. The purified proteins or cell lysates were analyzed via SDS-PAGE under reducing conditions

and coomassie staining (2.4.1) or western blots (2.4.2).

2.4.4.3 Purification of recombinant protein from cell lysate

HEK293T cells transfected with pMS-PI9-IV (2.1.4) were harvested and lysed within a native buffer

(50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 % (v/v) Triton X100). Incubation of the cells in buffer took

place for 30 min on ice. Cell debris was then removed by centrifugation at 13.000 g for 15 min at 4 °C

and the supernatant containing the desired protein was purified as described in 2.4.4.2.

2.4.5 Protein quantification

Three different methods were applied for the determination of protein concentration. For purified

protein the concentration was determined after SDS-PAGE and coomassie staining using ‘AIDA Image

Analyzer’ Software (Raytest Isotopenmessgerate). The concentration could be correlated to the

additionally loaded bovine serum albumin (BSA) standard (100 ng, 250 ng, 500 ng, 1000 ng). The

values were confirmed by Bradford assay in a 96-well plate using 1x Roti-Nanoquant (Roth). All

samples were pipetted in triplicates. A stock solution with 1 μg/μL purified BSA (0, 0.5, 1, 2, 4, 6, 8,


35

10 μg/well) was used as a calibration standard. In order to determine the total protein concentration

of cell extracts Bradford or BCA assay (Thermo Scientific Pierce) was performed according to each

manufacturer’s instructions. Absorption was measured at 595 nm for BSA and at 562 nm for BCA in a

microplate reader. Data were analyzed with MS Office Excel software, and the total protein

concentration of the samples was determined according to the BSA standard curve.

2.4.6 Labeling of SNAP constructs

The establishment of the SNAP (engineered version of the human DNA-repair enzyme O(6)-

alkylguanine DNA alkyltransferase) -tag technology and the preparation of the corresponding

constructs and their covalent conjugation to substrates containing O(6)-benzylguanine (BG) via a

stable thioether bond in a rapid and highly specific self-labeling reaction has been described before

[145]. Here, in HEK293T cells, secretory expressed and subsequently purified SNAP-tag fusion

proteins (Ki4(scFv)-SNAP and H22(scFv)-SNAP) were conjugated with the BG-modified dyes BG Alexa

Fluor 647 (BG 647), BG Alexa Fluor 747 (BG 747) or BG Vista Green (Covalys Biosciences AG,

Witterswil) by incubation in the dark with 1.5-3-fold molar excess of the dye overnight at 4 °C.

Residual dye was removed using PD 10 columns (GE Healthcare). Stained proteins were visualized

after separation by SDS-PAGE with either a UV transilluminator (BioRad Gel Doc XR) or with a CRi

Maestro imaging system using blue, red and near-infrared filter sets.

2.4.7 Detection of endogenous PI-9 from cell lines or primary cells

For the detection of cytosolic expressed PI-9 within tumor cell lines or primary cells, 106 cells were

lysed within 50 µL lysis buffer (PBS supplemented with 1 % (v/v) Triton X-100) for 30 min on ice. The

cell extract was cleared via centrifugation and the protein concentration was determined with

Bradford reagent (2.4.5). 40 µg total protein amount were loaded on a SDS-gel for western blot

analysis. PI-9 was detected by incubation for 1 h with anti-human PI-9 (2.1.5) in PBST. After washing

with PBST, GAM-PO was added and signals could be detected as described in 2.4.2.

In parallel, detection was performed via flow cytometry (2.5.1) which allows more quantitative

analysis. Therefore 106 cells were washed with PBS and resuspended in 500 µl Cytofix/ Cytoperm (BD

Bioscience, Heidelberg). After incubation for 20 min on ice and washing of the cells, a blocking step

was included with 5 % (w/v) BSA in 200 µl PBS for 20 min on ice. The cells were washed with PBS

containing 0.2 % (v/v) Tween 20 and incubated with anti-human PI-9 (2.1.5) for 30 min on ice. The

second antibody GAM-FITC (2.1.5) was added to the washed cells and incubated as described above.

The final wash step and resuspension of the cells was performed in PBS comprising 0.2 % (v/v)

Tween 20. The measured shift during flow cytometric analysis due to a specific binding of the


36

antibody to the fixed and permeabilized cells was quantified as MFI (Median Fluorescence Intensity)

in WinMDI 2.8. MFIs due to low unspecific binding of the secondary antibody were subtracted.

2.4.8 Serum stability assays

70 ng/µL purified protein was incubated with 50 % murine serum (AbD Serotec, Duesseldorf) or

human serum (obtained by centrifugation of 500 to 1000 µL patient-derived blood samples (2.1.8)

for 10 min, 300x g at RT) at 37 °C for different time periods between 0 and 48 h. The amount of

functional remaining protein binding was measured via flow cytometric analysis (2.5.1), here with

400-fold dilution of 7.5 µL of the mixture.

2.4.9 ELISA for detection of CD30 receptor in blood

Mouse serum of L540cy bearing mice was obtained by centrifugation of 10-30 µL mouse-derived

blood samples for 10 min and 300x g at RT and stored at -20 °C until use. The concentration of

soluble CD30 receptor within blood serum was evaluated by an ELISA set-up as previously published

[146]. High binding ELISA plates were coated with 50 µl Ki2 monoclonal antibody (240 µg/mL) in PBS

overnight at 4 °C. After washing with PBST and blocking in PBS containing 5 % (w/v) milk powder for

1 h at RT, 50 µl CD30-containing mouse serum (1:5 dilution in PBS) was added and incubated for 2 h

at RT. Plates were washed three times and 50 µl Biotin coupled Ki3 monoclonal antibody (270 ng/µL,

1:100 dilution in PBS) was added for a 1 h incubation at RT. For detection, plates were first incubated

with 50 µl Streptavidin-PO (1:200 in PBS, Abcam, Cambridge, UK) for 1 h; then washed in PBST and

then incubated for 1 h with 50 µl ABTS solution (Roche Diagnostics, Mannheim). Plates were read out

at 405 nm in a microplate reader.

2.5 FLOW CYTOMETRIC ANALYSIS

2.5.1 Binding analysis

The binding of fusion proteins to the target cells (Table 3) or the endogenous protein thereof (2.4.7)

was evaluated via flow cytometric analysis (CellQuest Version 3.3 (Becton Dickinson, Heidelberg))

and WinMDI 2.8. For standard binding analysis 4*105 cells were washed with PBS and incubated with

50 nM purified protein in 100-300 µL PBS for 30 min on ice. After 2 wash cycles the cells were

incubated with anti-His Alexa 488 (2.1.5) for 30 min on ice, in the dark. Negative controls with cell

lines not expressing the target receptor were run in parallel. Monocytic cells of primary PBMCs were

gated in forward and sideward scatter (FSC/ SSC) dotplot prior to evaluation.


37

2.5.2 Determination of Kd values

For the determination of the affinity constant (Kd) different concentrations of the fusion proteins

(0-16.7 nM, 1:2 dilution) were incubated with the cells and evaluated by flow cytometric analysis

(2.5.1) and non-linear-regression determinations in GraphPad Prism 4.0.

2.5.3 Determination of specific cell death of primary cells

The specific reduction of the target cell population was determined by staining primary cells with

anti-CD64 Alexa647 (AbD Serotec, Duesseldorf) to localize the population within the FSC/SSC dotplot

during flow cytometry. These cells were gated with WinMDI 2.8 so their reduction after treatment

could be evaluated accordingly and compared to untreated control.

To further specify reduction of the target cell population specific antibodies displayed in 2.1.5 were

incubated with the treated cells and reduction of the stained target cell population could be

compared to the untreated control by flow cytometric analysis.

2.6 CONFOCAL MICROSCOPY

All images were taken with a LEICA confocal microscope (TCS SP5) equipped with the following

lasers: 458 nm, 488 nm, 561 nm and 633 nm. A 40x optical magnification was used. Cells were

stained with SNAP-based constructs (2.4.6) as indicated and prepared as described above for flow

cytometry (2.5), resuspended in 50 µL PBS and pipetted onto a microscope slide.

2.7 FUNCTIONAL CHARACTERIZATION OF RECOMBINANT PROTEINS

2.7.1 Protease assays

2.7.1.1 Colorimetric substrate assays

The activity of 100 nM granzyme B-based constructs after enterokinase digestion was detected by

cleavage of 200 µM of the synthetic substrate Ac-IETD-pNA (Calbiochem/Merck, Darmstadt) which

mimics the cleavage site of pro-caspase 3. The reaction was monitored in 96-well plates in a

microplate reader at 405 nm and 37 °C at 1 or 2 min intervals for 1 h. Afterwards, time was plotted

against the absorbance so differences in activity could be evaluated. After complex formation

(2.7.1.4) and measurement of the remaining activity the velocity of the reaction was calculated from

the linear slope of the reaction within the first 10-12 min and converted to pmol/min with the help of


38

the corresponding conversion factor (µM/A405nm) according to the manufacturer’s protocol of the

Caspase-8 Assay kit (Calbiochem).

Granzyme M activity, after enterokinase digestion, was evaluated using the colorimetric substrate

Suc-Ala-Ala-Pro-Leu-pNA (Bachem, Bubendorf, Switzerland) according to a published protocol [56].

The change in absorbance at 405 nm was measured as above with the difference that here only the

absolute change was recorded, not the kinetics.

2.7.1.2 Determination of Michaelis-Menten Kinetics

To analyze the activity of the wild-type protein and its most promising mutant, Km values were

evaluated according to the manufacturer’s protocol of the Caspase-8 Assay kit (Calbiochem). The

enzymatic reaction was prepared and measured as described above (2.7.1.1). For evaluation, change

in absorbance at 405 nm during the first 12 min of the reaction was determined. Km values were

derived by non-linear curve fitting to the Michaelis–Menten equation using GraphPad Prism 4.0.

2.7.1.3 Cleavage of PI-9 by granzyme M

The activity of the granzyme M-based constructs was verified by their ability to cleave recombinant

PI-9 (1.3.2). The fusion protein was incubated at a 1:3 molar ratio with PI-9 at 25 °C in 100 mM HEPES

(pH 7.4), 200 mM NaCl, 0.01 % (v/v) Tween 20 and 1 mM DTT. After 24 h the reaction mixture was

analyzed by SDS-PAGE and western blotting using anti-human PI-9 and GAM-PO as described in 2.4.2.

2.7.1.4 Formation of the granzyme B and PI-9 Complex

The complex of recombinant PI-9 and Gb-H22(scFv) and its mutants was formed using a 5:1 molar

ratio (PI-9:Gb-H22(scFv)) under reducing conditions in 20 mM Tris-HCl (pH 7.4), 50 mM NaCl and

1 mM DTT. 600 ng of wild-type or mutant granzyme B construct was incubated with or without PI-9

for 1 h at 37 °C and the retained enzymatic activity was evaluated as described above (2.7.1.1). For

western blot analysis (2.4.2) adequate amounts containing at least 100 ng of granzyme B-based

protein was prepared for SDS-PAGE in the same manner. For detection of specific bands, anti-human

PI-9, anti-human Gb or anti-His were used followed by the second antibody GAM-PO (2.1.5).

Endogenous complex formation was determined after incubation using Gb-Ki4(scFv) as described for

apoptosis assay (2.7.2.1). Analysis was done as described above using anti-human Gb and GAM-PO.


39

2.7.2 Apoptosis assays

Apoptosis was documented via Annexin V/ propidium iodide (PI) staining (2.7.2.1) or via

determination of caspase 3/7 activation (2.7.2.2) within the target cells.

2.7.2.1 Annexin V/ PI Assay

Annexin V binds to phosphatidylserine which is exposed on the surface of apoptotic cells. Thus, dead

cells can be visualized using labeled Annexin V protein. In addition, PI enters through the permeable

membrane of dead cells, so necrotic and apoptotic cells can be distinguished. 2*105 cells/mL were

incubated at 37 °C, 5 % CO2, with different concentrations of fusion protein (between 11 and

100 nM), depending on target cells and construct, for 48 h in 12-well plates containing 1 mL

supplemented RPMI medium (2.3.1). Afterwards, the cells were washed in PBS and the pellet was re-

suspended either in 100 µL 1x Annexin V binding buffer (10 mM HEPES (pH 7.5), 140 mM NaCl,

0.5 mM CaCl2) supplemented with 2.5 µL Annexin V-FITC (eBioscience, Frankfurt) or in 450 µL cell-

free culture supernatant from HEK293T cells expressing Annexin V-labeled green fluorescent protein

(EGFP, [147]) supplemented with 10x Annexin V binding buffer and 5 µg/mL PI. The incubation with

labeled Annexin V protein took place in the dark, either at RT for 15 min or on ice for 20 min

respectively. Results were analyzed via flow cytometry (2.5).

2.7.2.2 Caspase 3/7 assay

To determine the caspase 3/7 activity within apoptotic, granzyme B-treated cells (2.7.2.1), a pre-

luminescent caspase 3/7-DEVD-aminoluciferine substrate was applied (Caspase-Glo™ 3/7 Assay,

Promega). Caspase 3/7 is a direct substrate of granzyme B. If it is cleaved after granzyme B delivery

into the target cells and thereby activated, it can cleave the pre-luminescent caspase 3/7-DEVD-

aminoluciferine substrate so free unbound aminoluciferine is released. This is subsequently used by

luciferase whereby a luminescence signal is produced. For the evaluation, cells were transferred into

96-well plates and the read out was done in a microplate reader. The amount of luminescence was

directly proportional to the activity of caspase 3/7.

2.7.3 Viability assay

The cytotoxic effect of the granzyme B-based fusion proteins was monitored using the ability of

metabolic active cells to reduce the tetrazolium salt XTT ((sodium 2,3,-bis(2-methoxy-4-nitro-5-

sulfophenyl)-5-[(phenylamino)-carbonyl]-2H-tetrazolium) inner salt, Life Technologies (Invitrogen)) to

orange colored compounds of formazan. The dye intensity was measured by a microplate reader and

is directly proportional to the number of living cells.


40

For Gb-Ki4(scFv) and its mutants 2*105 cells were plated in 1 mL in 12-well plates with 1:3 dilutions

and incubated for 48 or 72 h at 37 °C and 5 % CO2. For Gb-H22(scFv) and its variants and

Gm-H22(scFv), cells were incubated in 96-well plates with 5.000 IFNγ stimulated HL60 cells (or

control cell lines) or 20.000 primary cells per 100 µL with 1:3 or 1:5 dilutions of protein for 72 h at

37 °C. In a different approach, XTT was used for read out of viable cells after incubation with single

doses of fusion protein (concentrations as indicated, depending on target cells and construct) in 12-

well plates and evaluated after transfer of 100 µL of each approach into 96-well plates as described

above. 50 µL XTT was added to a 100 µL cell suspension and extinction was measured after a 4 h

incubation at 450 nm at a reference wavelength of 630 nm.

Viability curves and IC50 values were evaluated with GraphPad Prism4.0. Therefore dilution effects by

the protein volume used were normed using respective buffer controls. Zeocin was used as the

negative control.

Competition assays were carried out with H22(scFv) by adding 10 nM Gb-H22(scFv) or Gm-H22(scFv)

to IFNγ stimulated HL60 cells and incubating them for 48 h in the presence or absence of 10 or

100 nM H22(scFv). The buffer control was assigned a viability of 100 % and the converted XTT signal

was read out as above.

2.8 IMMUNOHISTOCHEMISTRY

Resected cryo-conserved tumors were cut into 8 µm sections on a freezing microtome (Cryostat CM

3050 (Leica) and mounted on coated slides. After drying for 48-72 h, sections were fixed for 10 min

with dry acetone and air-dried. In order to detect endothelial cells, slides were incubated with

naphthol AS-BI phosphate (sodium salt, 50 mg/100 mL; Sigma, Munich) as substrate and new fuchsin

(10 mg/100 mL; Merck, Darmstadt) as chromogen dissolved in 0.1 M Tris-HCl (pH 8.5) resulting in

pink/red staining. By omitting the addition of levamisole to the reaction mixture, endothelial cells

became visible. Slides were counterstained with hematoxylin.

In order to visualize remaining CD30 receptor expression in resected tumors, 12.5 nM

Ki4(scFv)-SNAP-BG-Vista-Green diluted in PBS was added to the slides and incubated for 30 min in

the dark at RT. After washing, images were collected via a Leica confocal microscope (2.6).


41

2.9 IN VIVO EXPERIMENTS

The experiments were officially approved by the local Animal Care and Use Review Committee (AZ:

87-51.04.2010.A278). All animals received human care in accordance with the requirements of the

‘German Tierschutzgesetz’, §8 Abs. 1 and in accordance with the Guide for the Care and Use of

Laboratory Animals published by the National Institute of Health.

2.9.1 Mouse strains, housing and maintenance of animals

For animal experiments 6 to 8 week-old female BALBc nu/nu mice (Charles River GmbH) were used.

Mice were housed in the ‘Institut fuer Versuchstierkunde’ of the university hospital, Aachen, during

tumor take rate studies, and in the animal facility of the Fraunhofer IME during the treatment

experiments.

2.9.2 Handling of mice and anesthesia

Anesthesia of the mice was performed either by intramuscular administration of 75 mg/kg Ketamin –

10 mg/kg Xylazin for periods up to 30 min or by isoflurane for shorter times up to 2 min.

Proteins were administered intravenously (i. v.) with volumes up to 100 μL. Samples were brought up

to RT before injection. Before and after image acquisition, anesthetized mice were kept warm on a

hot-water bag. As in previous imaging experiments, the mice injected with far red fluorescent protein

Kat2 (pTag-Katushka2-N; Evrogen) transfected cells were fed a purified, chlorophyll-free diet

(AIN93G, SSNIFF GmbH) 7 days before the imaging experiments were started whereas mice injected

with L540cy were fed a normal diet (SSNIFF GmbH).

Blood samples (10-30 µL) were obtained through the tail vein of the mice.

2.9.3 Establishment of xenograft subcutaneous tumor models

Before treatment of the mice with fusion proteins could be started, the tumor take rates of the

potential target cell lines had to be determined. For injection of tumor cells, L428 cells transfected

with Kat2 or L540cy cells (not transfected) were used after washing with PBS and resuspension in

50 % BD Matrigel™ Basement Membrane Matrix High Concentration, Growth Factor Reduced (BD

Bioscience, Heidelberg). 5*106 cells in a volume of 30 µL were injected subcutaneously into the right

hind limb of 5 mice for L428 and 5 mice for L540cy. Tumor growth was monitored by molecular

imaging with the CRi-Maestro Imaging System (2.9.4) in the case of Kat2 transfected L428; for L540cy

triplet caliper measurements were executed. Based on the evaluated tumor take rates and tumor

sizes, adequate treatment schedules could be set up.


42

2.9.4 In vivo optical imaging by Cri-Maestro system

In vivo optical imaging was performed with the CRi-Maestro Imaging System (Cri Inc., Woburn, MA,

USA) and analyzed with the Maestro Spectral Imaging Software. The Maestro imager has a Xenon

light source and allows measurements through the complete visible and near infrared light spectrum

by applying different filter sets. Whole mouse images were taken with constant stage height and

illumination settings. The Kat2 signal of the transfected L428 cells was visualized with the yellow filter

set (630–850 nm). To identify and discriminate the individual spectra within the images, spectral

libraries were defined. Therefore the different dye spectra were unmixed from the cube images. The

component images, which display the targeted signal, were used to define the tumor to background

ratios (TBRs) by drawing a region of interest (ROI) around the tumor and a larger area on the back

representing average background signal. The thereby calculated values for the tumor size in mm² for

the region of interest were used to calculate the tumor regression.

2.9.5 Biological activity of Gb-Ki4(scFv) and GbR201K-Ki4(scFv)

To determine the in vivo biological activity of Gb-Ki4(scFv) and GbR201K-Ki4(scFv), the cell lines L428

(Kat2 transfected) and L540cy were injected into the mice respectively as described in 2.9.3. The

mice were randomized equally into groups and received 50 µg GbR201K-H22(scFv) (as non-binding

control), Gb-Ki4(scFv) or GbR201K-Ki4(scFv) i. v. per treatment day. Tumor size was measured as

described in 2.9.3 and 2.9.4. Relative tumor size was calculated by setting the tumor size on first day

of treatment to 100 % and adapting the following values accordingly.

In vivo targeting of CD30-positive L540cy cells was accomplished by injection of 0.75 nmol

Ki4(scFv)-SNAP-BG-747 into L540cy bearing mice. 6 h later, the mice were imaged with the CRi-

Maestro system using the deep red filter set (730-950 nm) (2.9.4).

2.10 STATISTICAL ANALYSIS

Data were analyzed using a two-tailed t-test in GraphPad Prism 4.0, with p < 0.05 considered as

statistically significant.

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

43

3 Results

3.1 PI-9 EXPRESSION IN CANCER CELL LINES

The high potential of granzyme B-based CFPs has already been reported by several groups in vitro

and in vivo; however, one of the major challenges for its therapeutic application is the expression of

the specific granzyme B inhibitor PI-9 in many tumor cells (see 1.4.1). Based on these findings and the

observed limitations of cytotoxic effects induced by different granzyme B-based CFPs within this

group, the PI-9 expression profiles of the most commonly used cancer cell lines were evaluated.

Therefore total soluble protein of native cell extracts (2.4.7) of different cancer cell lines (2.1.6) was

separated by SDS-PAGE and transferred to a nitrocellulose membrane. The expression of

endogenous PI-9 could be confirmed by specific anti-human PI-9 staining of a 42 kDa protein band

correlating to the calculated molecular weight. An overview of the results for all tested cell lines is

displayed in Table 6. Representative results are shown for CD30-positive cell lines in Figure 3A.

Table 6: Overview of the PI-9 expression profile of different cancer cell lines.

(++: strong, +: moderate, (+): low, -: no expression)

Cell line Cell type Target receptor PI-9 expression

(protein level)

L428 HL CD30 ++ L1236 HL CD30 ++

L540/ L540cy HL CD30 - K562 Chronic myeloid leukemia in blast crisis CD30 +

Karpas 299 T cell lymphoma CD30 - L3.6pl Pancreatic carcinoma EGFR ++ PT45 Pancreatic adenocarcinoma EGFR - A431 Epidermoid carcinoma EGFR - HL60 AML CD64 - U937 Histiocytic lymphoma CD64 - Jurkat T cell leukemia / -

Kasumi 1 AML / (+)

PI-9-positive as well as PI-9–negative cell lines expressing the tumor marker CD30 were identified.

Corresponding qualitative results from western blotting were confirmed by flow cytometry providing

more quantitative data. For this purpose triplicate flow cytometric analysis of fixed and

permeabilized cells (2.4.7) was performed (Figure 3B).

Jurkat cells were PI-9-negative according to Godal et al. [148] so they could be used as a negative

control. The CML cell line K562, also previously published to express PI-9 at moderate level [148],

showed the lowest expression level compared to the Hodgkin lymphoma (HL) cell lines L428 and

L1236 which express similar amounts of PI-9. L428 retained their PI-9 expression after transfection

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

with far red fluorescent protein Kat2 (

inhibitor. The latter was recultured from a tumor grown in a nude mouse after injection of L540 so it

was advantageous for potential

their apparent cellular characteristics or sensi

work was continued with L540cy only.

No dependency between cultivation time and

cell lines (data not shown).

(A)

Figure 3: Expression of endogenous

(A) Total soluble (40 µg/lane) protein from cell lysates (transferred to a nitrocellulose membraneanti-human PI-9 and GAM-PO (2.1.5fixed, permeabilized (2.4.7) and stained with antidetected by flow cytometry (2.5). mean fluorescence intensities (MFIs) ± SD

The data confirmed the presence of endogenous PI

group and showed that the

expression. These findings are

granzyme B variant (3.5).

3.2 IDENTIFICATION OF MU

At the molecular level a reversible Michaelis complex is formed

inhibitors, followed by a 1:1 covalent complex whereby a covalent bond is

chain oxygen atom of GbS195

mature granzyme B) and the carbonyl carbon of the PI

conformational changes occur in both proteins after covalent binding, it was important to refer to a

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

with far red fluorescent protein Kat2 (Table 3). L540 and L540cy do not express the granzyme

The latter was recultured from a tumor grown in a nude mouse after injection of L540 so it

for potential in vivo studies. During in vitro studies no differences according to

their apparent cellular characteristics or sensitivity to the fusion proteins used was observed, so the

work was continued with L540cy only.

cultivation time and PI-9 expression level was observed in the investigated

(B)

L540cy L428 L12360

5

10

15

20

25

30

35

PI-

9 [M

FI]

endogenous PI-9 in HL cell lines L428, L1236, L540cy and CML cell line K562.

protein from cell lysates (2.4.7) was separated by transferred to a nitrocellulose membrane (2.4.2). Endogenous PI-9 was detected in a western b

2.1.5). M: Pre-stained broad range marker in kDa (NEB). and stained with anti-human PI-9 and GAM-FITC (2.1.5).

Jurkat cells were used as PI-9-negative control. The bar graph shows the fluorescence intensities (MFIs) ± SD of three independent experiments.

The data confirmed the presence of endogenous PI-9 in many of the cancer cell lines used in this

showed that the CD30-positive cell lines are heterogeneous regarding their PI

important for further evaluations to select the most promising

IDENTIFICATION OF MUTATIONAL SITES BY MOLECULAR MODELING

reversible Michaelis complex is formed between serine proteases and their

, followed by a 1:1 covalent complex whereby a covalent bond is formed

(corresponds to S184 according to the numbering of residues in

and the carbonyl carbon of the PI-9 P1 residue (1.4.1


_______________________________________________________________________________________________

44

L540cy do not express the granzyme B

The latter was recultured from a tumor grown in a nude mouse after injection of L540 so it

studies no differences according to

was observed, so the

ion level was observed in the investigated

L1236 K562 Jurkat

9 in HL cell lines L428, L1236, L540cy and CML cell line K562.

SDS-PAGE (2.4.1) and 9 was detected in a western blot (2.4.2) using

stained broad range marker in kDa (NEB). (B) 1*106 cells were . Endogenous PI-9 was

The bar graph shows the

9 in many of the cancer cell lines used in this

heterogeneous regarding their PI-9

important for further evaluations to select the most promising

LECULAR MODELING

between serine proteases and their

formed between the side

(corresponds to S184 according to the numbering of residues in

1.4.1). Because severe


Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

45

structure describing the initial Michaelis complex, in which the protease was inactivated by mutating

S195 to alanine; allowing the interacting amino acids responsible for complex formation to be

identified. The published rat trypsin - M. sexta serpin B1 complex [87] was found to be suitable for

fulfilling the prerequisites. In addition, the crystal structure of human granzyme B bound to a five-

residue peptide, solved at 2 Å resolution [42], is known. Human granzyme B and rat trypsin have a

sequence identity of 36 % according to results from homology modeling (MODELLERv7 software). In

addition, it is known that these proteins have similar length and are closely functionally and

structurally related. Thus, the backbone of human granzyme B structure could be fitted with that of

rat trypsin. Because the crystal structure of human PI-9 has not yet been solved, its structure had to

be built using molecular homology modeling based on the X-ray structure of the rat trypsin - M. sexta

serpin B1 complex. The sequence identity between M. sexta serpin B1 and human PI-9 was 27 %.

In cooperation with the German Research School for Simulation Sciences in Juelich (Prof. Dr. rer. nat.

Paolo Carloni), the modeled complex was inserted into a water box and key interactions at the

interface of the two proteins were screened by Baker’s alanine scanning procedure using ROBETTA

server (AMBER ff99SB, TIP3P and Aaqvist force fields were used for the protein, water, and counter

ions respectively, evaluation of long-range electrostatic interactions via particle mesh Ewald

method). The residues R28 and R201, according to the numbering of residues in mature granzyme B,

were found to have the largest destabilization in terms of loss of contacts. According to the

respective charge of the amino acids (neutral, opposite and same charge as the original amino acid,

respectively) arginine was substituted with alanine, glutamic acid and lysine by site-directed

mutagenesis (2.2.2). To investigate a potential supportive effect of both sites at once, the double

mutant R28A/R201A was considered as well, so in total, seven putative PI-9 resistant variants were

generated in silico: R28A, R28K, R28E, R201A, R201K, R201E and R28A-R201A. Their influence on

complex formation was analyzed during 50 ns molecular dynamic simulations in aqueous solution

(NAMD 2.7 package). In all of these mutated complexes the structure of the active sites was

maintained according to molecular modeling results. The three mutants R28K, R201A and R201K

were predicted to be the most destabilized, whereas the others were almost as stable as those of the

wild-type complex. Further details about the molecular simulation studies are published in Losasso et

al. [149].

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

3.3 GENERATION AND PREPARAT

MUTANTS

The effect of the amino acid exchanges described above on the

sensitivity towards PI-9 was evaluate

both generated prior to this work

malignancies has been demonstrated before

cytotoxic effects on the HL target cell line L540.

cytotoxicity and apoptosis could be shown on L540 as well as L540cy

the protein granzyme B lacking a cell

structures are displayed in Figure

(A)

(B) I

Figure 4: Overview of initial granzyme B

and after Enterokinase cleavage (B)

(A) Schematic structure of the binary eukaryotic expression vectorH22(scFv) and EGb. Igκ: signal peptide sequencecleavage site, Gb: granzyme B encoding sequence, scFv: single chain variable fragment, tag, IVS/IRES: synthetic intron and internal ribosome entry site, Mutated granzyme B sequences could be inserted via expressed in HEK293T cells and purified from supernatant via chromatography (2.4.4). Activation of the protein was (2.4.4) resulting in higher running band in presenceseparated by SDS-PAGE and stained with coomassie(NEB); II: (E)Gb-Ki4(scFv), M: Pre-stained broad range marker in kDa (NEBmarker in kDa (NEB).

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

PREPARATION OF GRANZYME B CONSTRUCTS AND

amino acid exchanges described above on their functionality as well as their

9 was evaluated using the initial constructs Gb-H22(scFv) and Gb

both generated prior to this work. The cytotoxic potential of Gb-H22(scFv) against CD64

malignancies has been demonstrated before [138]. Gb-Ki4(scFv) was cloned earlier

cytotoxic effects on the HL target cell line L540. In this study, the protocol was optimized

apoptosis could be shown on L540 as well as L540cy cells (Figure

B lacking a cell-binding ligand was prepared (Gb). The corresponding

Figure 4A.

II III

granzyme B constructs (A) and SDS-PAGE of the purified fusion proteins before

(B).

Schematic structure of the binary eukaryotic expression vector pMS encoding for EGbIgκ: signal peptide sequence (L) of immunoglobulin kappa light chain, ECS: enterokinase

B encoding sequence, scFv: single chain variable fragment, ron and internal ribosome entry site, EGFP: enhanced green fluorescent protein.

B sequences could be inserted via XbaI/ BlpI cloning site. (B) expressed in HEK293T cells and purified from supernatant via IMAC and subsequent gel filtration

. Activation of the protein was achieved via incubation with 0.02resulting in higher running band in presence (a) and lower bands in absence (b)

PAGE and stained with coomassie. I: (E)Gb-H22(scFv), M: Protein broad range marker in kDa stained broad range marker in kDa (NEB), III: (E)Gb, M: Protein broad range

_______________________________________________________________________________________________

46

CONSTRUCTS AND ITS

functionality as well as their

H22(scFv) and Gb-Ki4(scFv),

H22(scFv) against CD64-positive

earlier but did not show

the protocol was optimized (2.7.3) so

Figure 12A). In addition

corresponding schematic

PAGE of the purified fusion proteins before

pMS encoding for EGb-Ki4(scFv), EGb-of immunoglobulin kappa light chain, ECS: enterokinase

B encoding sequence, scFv: single chain variable fragment, H: 6x histidine (His6)-GFP: enhanced green fluorescent protein.

Fusion proteins were and subsequent gel filtration

0.02 U/µg enterokinase of ECS (E). Protein was

in broad range marker in kDa ), III: (E)Gb, M: Protein broad range

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

The secretory expression, enabled by the Igκ leader (

expressed in an inactive form by insertion of an enterokinase cleavage site at its N

prevent killing of the expressing cells, was performed

fusion proteins into the supernatant should allow correct post

granzyme B has two glycosylation sites. The protein was purified by affinity chromatography

selecting for the His6-tag and gel filtration (

PAGE (2.4.5). The identity of the protein was confirmed by western blot analysis using an anti

Gb antibody (2.4.2). To obtain a free N

granzyme B, the protein was incubated with enterokinase (0.02

Gb-H22(scFv) and Gb-Ki4(scFv) are shown in

whereby successful cleavage was verified by

the proteins was quantified by AIDA analysis (

The putative PI-9 resistant mutants described in

splice-overlap extension (SOE)-PCR

cloned into the corresponding pMS vector encoding for the wild

addition, the enzymatically inactive variant S184A was generated as control protein

granzyme B mutant K27A which has been described

overview of all constructs generat

Figure 5: List of generated point-mutated variants (

modeling (3.2).

All of the constructs needed,

determinations, have successfully been generated and produced in HEK293T cells, followed by

cleavage by enterokinase.

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

The secretory expression, enabled by the Igκ leader (2.4.4), of the granzyme

expressed in an inactive form by insertion of an enterokinase cleavage site at its N

prevent killing of the expressing cells, was performed in HEK293T cells (2.4.4.1). The secretion of the

fusion proteins into the supernatant should allow correct post-translational modification because

wo glycosylation sites. The protein was purified by affinity chromatography

tag and gel filtration (2.4.4.2), with a yield of 1-2 mg/L, as de

). The identity of the protein was confirmed by western blot analysis using an anti

). To obtain a free N-terminus, which is important for the proteolytic activity of

B, the protein was incubated with enterokinase (0.02 U/µg). The purified fusion proteins

are shown in Figure 4B before and after digestion with enterokinase,

whereby successful cleavage was verified by the reduction of molecular weight. The concentration of

the proteins was quantified by AIDA analysis (2.4.5).

mutants described in 3.2 were generated via site-directed mutagenesis

PCR (2.2.2) with specific primers (2.1.4), and the new sequences were

cloned into the corresponding pMS vector encoding for the wild-type-based con

the enzymatically inactive variant S184A was generated as control protein

which has been described before to be resistant to PI

generated is summarized in Figure 5.

mutated variants (2.2.2) arisen from literature (#: [87]

needed, including the respective mutations suggested from

have successfully been generated and produced in HEK293T cells, followed by

_______________________________________________________________________________________________

47

B proteins, which is

expressed in an inactive form by insertion of an enterokinase cleavage site at its N-terminus to

). The secretion of the

translational modification because

wo glycosylation sites. The protein was purified by affinity chromatography

mg/L, as determined by SDS-

). The identity of the protein was confirmed by western blot analysis using an anti-human

terminus, which is important for the proteolytic activity of

U/µg). The purified fusion proteins

B before and after digestion with enterokinase,

the reduction of molecular weight. The concentration of

directed mutagenesis by

and the new sequences were

based constructs (2.1.4). In

the enzymatically inactive variant S184A was generated as control protein and a

to be resistant to PI-9 inhibition [86]. An

];

##: [86]) or molecular

including the respective mutations suggested from in silico

have successfully been generated and produced in HEK293T cells, followed by

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

48

3.4 PREPARATION AND FUNCTIONAL CHARACTERIZATION OF RECOMBINANT PI-9

Recombinant PI-9 had to be produced in order to evaluate the enzymatic activity of the new

granzyme B mutants in the presence of the inhibitor in vitro during a substrate assay to pre-select the

most promising mutants. As during the initially planned mammalian expression in HEK293T cells, only

low amounts of PI-9 could be produced (2.1.4, data not shown); therefore another alternative for the

production of PI-9 was pursued, namely expression in E. coli. The PI-9-encoding sequence was cloned

into the E. coli expression vector pMT (2.1.4) by HindIII/NotI restriction sites (Figure 6A). The

periplasmic expression under osmotic stress (2.4.3) has been proven to be successful in small scale

by SDS-PAGE and western blotting (data not shown); so a 4 L batch fermentation was carried out. As

shown in Figure 6B, the four subsequent purification steps included two IMAC purifications, one

purification step via anion exchange chromatography and a final run using a gel filtration column

(2.4.3). The purified protein was sterile filtered and stored at -80 °C in the presence of 1 mM DTT.

Functionality of the produced PI-9 was confirmed by its ability to form covalent, SDS-stable

complexes with recombinant wild-type granzyme B-based proteins. Therefore PI-9 and wild-type

Gb-H22(scFv) were incubated in a 5:1 molar ratio for 1 h at 37 °C under reducing conditions in

presence of 1 mM DTT (2.7.1.4). Successful complex formation and thereby functionality of the

produced PI-9 was verified via western blot analysis (Figure 6C).

Complexes were formed between PI-9 and free Gb which was present as a degradation product

within the Gb-H22(scFv) sample running at approximately 77 kDa and between full length

Gb-H22(scFv) and PI-9 resulting in a complex above 100 kDa (correlates with 175 kDa band of marker

which is inaccurate at high molecular sizes). The complex could not be verified using anti-His, maybe

due to sterical hindrance of the antigen binding site; therefore the complex was verified using anti-

human Gb and anti-human PI-9. During analysis with anti-His it turned out that the band running

below the more intense band was not functional, since it remained within the sample after

incubation with granzyme B in contrast to the higher running band (Figure 6C, III, lane 2).

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

(A)

(C)

I

II

Figure 6: Preparation of recombinant PI

type Gb-H22(scFv).

(A) Schematic structure of the prokaryotic expression vector pMT encoding for PIsequence for periplasmic localizatioNotI cloning site. (B) Recombinant PIsubsequently via anion exchange chromatography (MonoQ column) and finally via gel filtrationchromatography (GF) (2.4.4). the band of active PIfurther experiments. The band below the 42assays (C,III). (C) To demonstrate the wild-type Gb-H22(scFv) for 1 h at 37°C transferred to a nitrocellulose membrane (western blot using anti-human Gb ((2.1.5); 1: Gb-H22(scFv), 2: Gb-H22(scFv)complex are indicated in grey, full-length protein and complex (red rectangle) thereof in blackbroad range marker in kDa (NEB).

Functional PI-9 could be produced

proteolytic activity of the new granzyme

3.5 ESTABLISHMENT OF A T

For selection of the most promising

suitable in vitro test system based on PI

Within this thesis different strategies were pu

cell lines expressing cytosolic PI-

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

(B)

II

III

Preparation of recombinant PI-9 and functionality determination via complex formation with

Schematic structure of the prokaryotic expression vector pMT encoding for PI-9. pelBlocalization, H: 6x histidine tag, PI-9: PI-9 encoding sequence, inserted via

Recombinant PI-9 was expressed in E. coli (2.4.3) and purified subsequently via anion exchange chromatography (MonoQ column) and finally via gel filtration

the band of active PI-9 protein is indicated by an arrow. Fraction 6 was used for and below the 42 kDa band was inactive protein, as prove

the functionality of the produced recombinant PI-9 at 37°C (2.7.1.4) before the protein mixture was separated by SDS

transferred to a nitrocellulose membrane (2.4.2). Complexed and uncomplexed protein was detected in a human Gb (I), anti-human PI-9 (II) or anti-His monoclonal antibodies

(scFv) and PI-9, 3: PI-9. Gb cleaved during purification and the corresponding length protein and complex (red rectangle) thereof in black

9 could be produced by bacterial expression and is essential to determine

proteolytic activity of the new granzyme B mutants in presence of PI-9 (3.6.1).

ESTABLISHMENT OF A TEST SYSTEM FOR GRANZYME B MUTANTS

promising mutant, able to induce apoptosis within PI

based on PI-9-positive and –negative cell lines had to be

sis different strategies were pursued: First, the generation of transiently transfected

-9 or the delivery of recombinant granzyme B protein

_______________________________________________________________________________________________

49

omplex formation with wild-

9. pelB: signal peptide 9 encoding sequence, inserted via HindIII/

and purified twice via IMAC, subsequently via anion exchange chromatography (MonoQ column) and finally via gel filtration

Fraction 6 was used for proven during functionality

it was incubated with protein mixture was separated by SDS-PAGE and

and uncomplexed protein was detected in a monoclonal antibodies (III) and GAM-PO

purification and the corresponding length protein and complex (red rectangle) thereof in black. M: Pre-stained

essential to determine the retained

RANZYME B MUTANTS

o induce apoptosis within PI-9-expressing cells, a

had to be established.

the generation of transiently transfected

protein into inherently

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

50

PI-9-positive cells. Second, the identification of PI-9-positive as well as -negative cell lines

overexpressing identical target receptors. The different approaches which were evaluated in parallel

are described in detail in the following.

1) The direct delivery of the protein into potential target cells is independent of receptor

expression and specific binders, and thereby parameters based on affinity, internalization and

endosomal release. This approach was followed by generating a granzyme B construct lacking

the binding moiety (Figure 4A). Protein delivery into the cytosol of PI-9-positive L428 and L1236

cells as well as PI-9-negative Jurkat cells, among others, (3.1) was analyzed with the pore-

forming reagent streptolysin O (2.3.3), an oxygen-labile streptococcal hemolytic exotoxin. The

delivery of granzyme B variants was only successful for Jurkat cells as shown by western blot

analysis of total soluble protein of cell lysates and resulting cytotoxic effects. The efficacy was

determined via apoptosis assays done 4 h after Gb delivery into the cells (2.7.2.1). In contrast,

translocation efficiencies for L428 and L1236 cells were not sufficient to compare the effect of

different granzyme B variants to inherently PI-9-positive or -negative cells. Since Jurkat cells

could not be transfected with PI-9 even though all common transfection reagents and methods

were tested (Table 5), this approach was not further pursued.

2) Since the functionality of Gb-H22(scFv) was established before with AML cell line U937 and

primary CD64-positive leukemic cells, this construct could be used to test the new granzyme B

mutants targeting the CD64-positive AML cell line HL60 (U937 lost its sensitivity to Gb-H22(scFv)

for unknown reasons and was therefore no longer applicable for this purpose). HL60 cells were

tested to be PI-9-negative (Table 6), so PI-9 transfection of this cell line would allow comparable

assays to confirm resistance of granzyme B mutants against PI-9. For this purpose different

transfection methods were evaluated using varying conditions as listed in Table 5. However, as

for Jurkat cells (3.5, 1)) all transfection efficiencies were too low to establish a stable transfected

cell line.

3) The other available granzyme B-based construct was Gb-Ki4(scFv) whose cytotoxic potential

could be evaluated in this study using the CD30-positive cell lines L540 and L540cy (2.1.6). Both

target cell lines were tested to be PI-9-negative (Figure 3). As was seen with HL60 cells,

transfection with PI-9 was not successful for L540cy cells (3.5, 2). Alternatively, the in vitro anti-

tumor efficacy in cells expressing both endogenous PI-9 and the CD30 receptor was tested.

Here, depending on the ability of the target cell lines to express PI-9 or not, no induction of

apoptosis was determined by an Annexin V/ PI assay after incubation with wild-type

Gb-Ki4(scFv) for 48 h. Results are shown in Figure 7A. Apoptotic cells were only observed in the

PI-9-negative cell line L540cy. No cell death was induced in PI-9-positive L428 cells. To verify that

complex formation with the inhibitor was the cause for the cytotoxic limitation, the cells were

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

lysed after treatment with Gb

expected height of Gb-Ki4(scFv)

demonstrated via western blot analysis

L540cy cells.

(A)

Figure 7: Influence of endogenous PI

(A) 33 nM Gb-Ki4(scFv) was incubated with Apoptotic effects were measured via Annexincells; bottom right: early apoptosis; top right: late apoptosis/dead cells. Numbers in the quadrants represthe percentage of cells of each category.treated (+) or untreated (-) cells was separated by SDSmembrane (2.4.2). PI-9-Gb-Ki4(scFv) complex and uncomplexed Gb(2.4.2) using anti-human Gb and GAMin L428 cells. M: Pre-stained broad range marker in kDa (NEB).

In conclusion, the CD30-positive cells

the cytotoxic potential of wild-type and mutant granzyme

The results were confirmed during

the UK Aachen Dr. Edgar Jost) using CD64

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

lysed after treatment with Gb-Ki4(scFv). Despite the non-specifically detected band at the same

Ki4(scFv) the inactivated SDS-stable granzyme B

rn blot analysis (Figure 7B) in contrast to the findings in PI

(B)

PI-9 on apoptotic activity of Gb-Ki4(scFv).

was incubated with PI-9-negative L540cy and PI-9-positive L428 cellsApoptotic effects were measured via Annexin V/ PI assay (2.7.2.1). Bottom left: viable cells; top left: necrotic cells; bottom right: early apoptosis; top right: late apoptosis/dead cells. Numbers in the quadrants represthe percentage of cells of each category. (B) Total soluble protein from cell lysates (40 µg/lane) of

) cells was separated by SDS-PAGE (2.4.1) and transferred to a nitrocellulose Ki4(scFv) complex and uncomplexed Gb-Ki4(scFv) was detected in a western blot

human Gb and GAM-PO (2.1.5). Unspecific band was detected with equal size stained broad range marker in kDa (NEB).

positive cells used turned out to be suitable for the intended

type and mutant granzyme B on PI-9-positive and

The results were confirmed during ex vivo studies on patient-derived leukemic cells (obtained from

the UK Aachen Dr. Edgar Jost) using CD64-targeting Gb-H22(scFv) and its mutants

_______________________________________________________________________________________________

51

specifically detected band at the same

B-PI-9 complex was

in contrast to the findings in PI-9-negative

positive L428 cells (both HL) for 48 h. Bottom left: viable cells; top left: necrotic

cells; bottom right: early apoptosis; top right: late apoptosis/dead cells. Numbers in the quadrants represent (B) Total soluble protein from cell lysates (40 µg/lane) of Gb-Ki4(scFv)

) and transferred to a nitrocellulose Ki4(scFv) was detected in a western blot

equal size of Gb-Ki4(scFv)

for the intended evaluation of

positive and –negative cells (3.7).

derived leukemic cells (obtained from

and its mutants (3.9).

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

52

3.6 IN VITRO FUNCTIONAL CHARACTERIZATION OF WILD-TYPE AND MUTANT

GB-H22(SCFV) AND GB-KI4(SCFV)

The main challenge during generation of granzyme B mutants was to avoid a negative effect of the

mutations on the active site of the enzyme and its tertiary structure. To exclude hampering of

functionality, the wild-type and mutated granzyme B-based fusion proteins were tested in

comparative assays regarding their enzymatic activity, their binding affinity, and their specific cell

cytotoxicity.

3.6.1 Proteolytic activity of granzyme B in absence and presence of PI-9

The influence of the single point mutations on the granzyme B proteolytic activity was determined

via a colorimetric substrate assay based on the synthetic granzyme B substrate Ac-IETD-pNA, which

mimics the cleavage and activation site of human procaspase-3 (2.7.1.1). Results for Gb-Ki4(scFv) and

GbR201K-Ki4(scFv) are shown in Figure 8.

0 10 20 30 40 50 600.0

0.1

0.2

0.3

0.4

0.5

0.6

GbR201K-Ki4Gb-Ki4

Time [min]

E40

5 n

m

Figure 8: Enzymatic activity of Gb-Ki4(scFv) and GbR201K-Ki4(scFv).

The proteolytic activity of activated protein was measured via a colorimetric assay based on the synthetic granzyme B substrate Ac-IETD-pNA (2.7.1.1). Reaction was documented for 1 h with a 2 min interval at 37 °C and 405 nm in an Elisa plate reader. Graphs were plotted using GraphPad Prism 4.0.

No differences were observed regarding the measured enzymatic kinetics so it could be concluded

that the active site was not affected by these mutations.

Enzymatic activity of Gb-H22(scFv) and GbR201K-H22(scFv) was additionally quantified via Michaelis-

Menten kinetics. KM values were comparable for both constructs with 104 ± 32 µM for the wild-type

and 105 ± 13 µM for the mutant.

A comparative overview of the activity of all generated mutants, including GbK27A, a mutant

published before and supposed to bind less efficiently to PI-9 (1.4.1), is depicted as grey bars in

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

53

Figure 9. Except for the double mutant R28A/R201A and the single mutant R201E, no significant

differences were measured in comparison to the wild-type protein.

Gb-H22

GbK27

A-H22

GbR28

A-H22

GbR28

E-H22

GbR28

K-H22

GbR28

1A/R

201A-H

22

GbR20

1A-H

22

GbR201E

-H22

GbR20

1K-H

22

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035without PI-9with PI-9

E40

5nm

/min

Figure 9: Enzymatic activity of Gb-H22(scFv) wild-type and its mutants in the presence and absence of PI-9.

Gb-H22(scFv) was incubated either with our without recombinant PI-9 according to 2.7.1.4. The proteolytic activity was measured via a colorimetric assay based on the synthetic granzyme B substrate Ac-IETD-pNA (2.7.1.1, 2.7.1.4). The graph shows means ± SD of 3 to 6 independent experiments [149].

To confirm the earlier obtained in silico findings (3.23.2) in vitro, the proteolytic activity of wild-type

and mutant granzyme B was analyzed after challenge with recombinant PI-9 in a comparative

enzymatic assay. In the presence of PI-9, GbR28K retained 76 % of its original activity, and GbR201A

retained 46 %. The enzymatic activity of GbR201K was almost the same in the presence or absence of

PI-9 (6 % reduction after PI-9 challenge), so this variant was the least inhibited variant according to

this assay. The catalytic activity of the other mutated granzyme B variants decreased relative to wild-

type granzyme B, or was even lower in the case of the double mutant. Their remaining activity after

complexing ranged from 0.5 % for the double mutant (R28A-R201A) to 25 % for R28E. An exception

was R28A with a remaining activity of 54 % as opposed to the computed predictions [149]. Even

though the result of the latter was promising, this variant was not further tested since it did not show

good results during preliminary cytotoxicity studies on PI-9—positive and –negative cells (data not

shown). The PI-9 insensitivity of the earlier described granzyme B variant K27A (see above) could not

be reproduced either during the in silico or during the in vitro experiments.

The most promising constructs with retained proteolytic activity after incubation with recombinant

PI-9, GbR28K, GbR201A and GbR201K, were further tested, always in comparison to GbK27A and Gb

wild-type.

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

3.6.2 Target cell binding

To exclude an influence of the point

activity of the granzyme B variants, flow

Gb-Ki4(scFv) or Gb-H22(scFv) variants

human Gb-PE or anti-His Alexa 488

Representative results for Gb-Ki4(scFv) and GbR201K

(A)

(B)

Figure 10: Binding analysis of granzyme B

50 nM Gb-based constructs were incubated with binding was detected with anti-His Alexa488 (GbR201K-Ki4(scFv) on CD30-positive L428 and CD30GbR201K-H22(scFv) on CD64-positive HL60 and

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

To exclude an influence of the point-mutations on the tertiary structure and possibly the binding

variants, flow cytometry-based binding assays were performed with all

variants. The specific detection of cellular binding was done by anti

His Alexa 488 (2.1.5).

Ki4(scFv) and GbR201K-Ki4(scFv) are shown in Figure

granzyme B-based CFPs.

based constructs were incubated with 4*105 target and control cell lines His Alexa488 (2.1.5) and flow cytometry (2.5.1). (A) Binding of Gb

ve L428 and CD30-negative HL60 cell line. (B) Binding of Gbpositive HL60 and CD64-negative Ramos cell line.

_______________________________________________________________________________________________

54

mutations on the tertiary structure and possibly the binding

were performed with all

of cellular binding was done by anti-

Figure 10A.

target and control cell lines for 30 min on ice and (A) Binding of Gb-Ki4(scFv) and

. (B) Binding of Gb-H22(scFv) and

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

55

No difference in binding activity was found for any of the new constructs compared to the wild-type,

on CD30-positive HL cell lines L428, L1236 and L540cy, or on the CML cell line K562. No unspecific

binding to the CD30-negative AML cell line HL60 was detected.

The corresponding Kd values for Gb-Ki4(scFv) and GbR201K-Ki4(scFv) on all of the cell lines used are

summarized in Table 7 and are all in the same nM range, independent of both the mutation and the

level of the CD30 receptor expression. The latter was evaluated since significant differences were

observed during previous flow cytometric analysis. CD30 expression level was set to 100 % for L540cy

since for this cell line the highest shift in fluorescence was measured. K562 was determined to have

the lowest CD30 expression.

Table 7: Affinity constants and relative receptor expression level of CD30-positive cell lines.

(n. d.: not determined)

Cell line KD of Gb-Ki4(scFv)

[nM]

KD of GbR201K-Ki4(scFv)

[nM]

CD30 expression

[% MFI]

L540cy 3.6 5.7 100

L428 1.5 1.9 32

L1236 n. d. n. d. 35

K562 5.2 4.8 7

Specific binding of Gb-H22(scFv) and its mutants was determined using CD64-positive HL60 cells and

CD64-negative Burkitt lymphoma cell line Ramos as a control. Data are shown representatively for

Gb-H22(scFv) and GbR201K-H22(scFv) in Figure 10B. The affinity constant (Kd) for each of these

constructs was measured to be 2.1 nM; which shows specific and high affinity to the receptor.

The described data indicated that the point-mutations within granzyme B did not affect the high

affinity binding of the determined CFPs to their respective target cells.

3.6.3 Internalization activity

Since marked dissimilarities in CD30 receptor expression were observed (Table 7) and to assure that

cytotoxic differences were not due to subsequent diminished internalization, target cell

internalization was analyzed by confocal microscopy (2.6). For that purpose cells were incubated

with Ki4(scFv)-SNAP-BG-647, a red fluorescent protein based on the SNAP-tag technology (2.4.6), at

37 °C or 4 °C. Incubation times were adapted to CD30 expression levels so K562 was incubated for 4 h

with the fluorescent fusion protein whereas the other cell lines were investigated after 2 h. This

longer incubation time for K562 was chosen based on the low CD30 expression to allow for additional

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

56

accumulation of fluorescent protein to bind and internalize allowing it to be visualized via confocal

microscopy. Internalization and binding was also shown for CD64-positive HL60 cells after 2 h

incubation with H22(scFv)-SNAP-BG-647 (Figure 11). Results are depicted in Figure 11.

Construct Cell line 37 °C 4 °C 37 °C (-)

Ki4

(scF

v)-

SN

AP

-BG

-64

7

L540cy

L428

K562

H2

2(s

cFv

)-S

NA

P-

BG

-64

7

HL60

Figure 11: Internalization analysis of Ki4(scFv)-SNAP-BG-647 and H22(scFv)-SNAP-BG-647 into corresponding

target cell lines.

50 nM fluorescent fusion protein or buffer only (-) were incubated with 4*105 target cell lines at 37 °C or 4 °C and internalization was detected with confocal microscopy after 2 h for L540cy, L428, L1236 and HL60 and after 4 h for K562 (2.6). The (scFv)-SNAP-BG-647 signal (upper left), a transmitted light picture (upper right) and the overlay of the two (lower left) are shown.

External binding only was observed at 4 °C; expression level signals differed in intensity depending on

the receptor used. Incubation at 37 °C led to receptor-mediated endocytosis in all cell lines tested

(Figure 11). The more intense fluorescent signal observed for H22(scFv)-SNAP-BG-647 was probably

due to a better coupling efficiency of the SNAP-based protein.

10 µm 10 µm 10 µm

10 µm 10 µm 10 µm

10 µm 10 µm 20 µm

10 µm 10 µm 10 µm

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

57

Taken together, on a qualitative basis successful protein delivery into the CD30-positive cells could

be verified despite of their differences in receptor expression level. In addition, successful

internalization could be demonstrated for CD64-positive HL60 cells.

3.6.4 Cytotoxic activity against PI-9-negative cell lines

The next step was to verify that the three most promising mutants, and the mutant described in the

literature (3.6.1), were still able to induce cell death of PI-9-negative target cells in the same manner

as the wild-type. For this purpose comparative cell proliferation assays (2.7.3), using the

corresponding wild-type granzyme B constructs as standards, were conducted. As described above,

the sensitivity of CD64-positive HL60 cells depends on their activation with adequate concentrations

of IFNγ (1.6, 2.1.6). The optimal concentration resulting in sensitive and at the same time viable HL60

cells after 24 h was evaluated to be 50 U/mL. CD30-positive cells did not have to be treated with

cytokines prior to cytotoxicity assays.

L540cy

100 101 102 103 1040

25

50

75

100

GbR28K-Ki4GbR201A-Ki4GbR201K-Ki4

GbK27A-Ki4Gb-Ki4

CFP [ng/ml]

Via

bilit

y [%

]

HL60

100 101 102 103 104 105 1060

25

50

75

100Gb-H22GbK27A-H22GbR28K-H22GbR201A-H22GbR201K-H22

CFP [ng/ml]

Via

bilit

y [%

]

CFP IC50

Gb-Ki4 GbK27A-Ki4

GbR28K-Ki4 GbR201A-Ki4 GbR201K-Ki4

2.5 nM 2.2 nM 5.1 nM 3.4 nM 1.7 nM

CFP IC50

Gb-H22 GbK27A-H22

GbR28K-H22 GbR201A-H22 GbR201K-H22

1.2 nM 5.3 nM 1.8 nM 3.9 nM 2.4 nM

Figure 12: Cytotoxic effects of Gb-Ki4(scFv) and Gb-H22(scFv) wild-type and their mutants respectively on

PI-9-negative target cell lines.

L540cy and stimulated (50 U/mL IFNγ) HL60 cells were incubated for 72 h at 37 °C with decreasing protein concentrations (serial dilution 1:3) of Gb-Ki4(scFv) and its mutants or Gb-H22(scFv) and its mutants, respectively. Cell viability was evaluated with a colorimetric XTT-based assay (2.7.3). Viability of untreated cells was set to 100 %. The graphs show means ± SD of triplicates per dilution. IC50 values were evaluated with GraphPad Prism 4.0 and are listed for all constructs.

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

58

Due to the absence of PI-9 in L540cy and HL60 cells, no differences of functionality should occur

during the performed assays. This could be confirmed by the dose-depending curves whose values

were determined after incubation of the target cells with the fusion proteins Gb-Ki4(scFv) and

Gb-H22(scFv) and their mutants. The IC50 values (between 1.7 and 5.1 nM or 1.2 and 5.3 nM,

respectively, summarized in Figure 12) were all in the same range not differing significantly from the

wild-type. No cytotoxic effects were determined on target receptor negative cell lines HL60 for

Gb-Ki4(scFv) constructs and Ramos for Gb-H22(scFv) constructs (data not shown).

The specificity of cell death induced by GbR201K-H22(scFv) was additionally confirmed by a

competitive assay (2.7.3). Therefore cells were incubated with 10 nM of the fusion protein in the

absence and presence of 10 or 100 nM H22(scFv). 48 h incubation of HL60 cells with

GbR201K-H22(scFv) alone resulted in a remaining cell viability of 44 % compared to 100 % of

untreated cells determined via XTT assay. Cytotoxic effects were significantly reduced by 14 % or

25 % in presence of 10 nM H22(scFv) or 100 nM H22(scFv), respectively. Incubation with H22(scFv)

alone did not reduce cell viability significantly (data not shown).

3.6.5 Apoptotic activity on PI-9-negative cell lines

To verify apoptosis induction as underlying mechanism of the observed cell death, an Annexin V/ PI

assay was performed as described in 2.7.2.1. Results are depicted in Figure 13.

(A)

Buffer

Gb-Ki4

GbK27

A-Ki4

GbR28

K-Ki4

GbR20

1A-K

i4

GbR201K

-Ki4

0

25

50

75

100 L540cy

Via

bilit

y [%

]

(B)

Buffer

Gb-H22

GbR201K

-H22

0

25

50

75

100 HL60K562

Via

bilit

y [%

]

Figure 13: Apoptotic effects of Gb-Ki4(scFv) or Gb-H22(scFv) wild-type and their mutants respectively on PI-9-

negative target cell lines.

21 nM CFPs were incubated with target cell lines for 48 h at 37°C. Apoptotic effects were measured via Annexin V/ PI assay (2.7.2.1), normalized to buffer control and converted to viability. The bar graphs show means ± SD of three independent experiments. (A) Result after incubation of Gb-Ki4(scFv) and its mutants with CD30+ L540cy cell line. (B) Result after incubation of Gb-H22(scFv) and its mutants with stimulated (50 U/mL IFNγ) CD64+ HL60 cell line. CD64 K562 cell line was used as negative control.

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

59

Here, up to 60 % of the targeted L540cy cells were apoptotic after incubation with Gb-Ki4(scFv) or its

mutants (Figure 13A). Receptor-negative HL60 cells were not affected by the fusion proteins (data

not shown) confirming specificity of apoptosis induction.

In case of Gb-H22(scFv) and its most promising mutant GbR201K-H22(scFv), apoptosis induction after

incubation with these CFPs was observed in more than 50 % of the stimulated HL60 cells, whereas no

unspecific effect could be detected on the CD64-negative CML cell line K562 (Figure 13B).

Apoptosis induction was additionally confirmed via a colorimetric caspase 3/7 assay (2.7.2.2) which is

a good indicator for apoptosis after treatment with granzyme B since caspase 3 is a direct substrate

of the enzyme. After incubation of HL60 cells with 33 nM Gb-H22(scFv) or GbR201K-H22(scFv) for

48 h the specific signal for caspase activity was in both cases increased 2.5-fold compared to the

buffer control. The same increase was determined for the positive control treated with Zeocin (data

not shown).

Since the above data showed no differences in cytotoxic potential for the tested granzyme B variants

on PI-9-negative cell lines, they were all further tested on PI-9-positive cells to select the most

promising one.

3.7 SELECTION OF THE MOST PROMISING GRANZYME B MUTANT ON PI-9-

POSITIVE CELL LINES

As described in 3.5, the in vitro test system used for this purpose was based on CD30-positive cell

lines which can be targeted with the construct Gb-Ki4(scFv). Decreased apoptotic effects were

observed in PI-9-expressing cells after incubation with Gb-Ki4(scFv) in contrast to PI-9-negative cells

(Figure 7A). Based on the assumption that the observed resistance to cell death is related to the

expression of the endogenous granzyme B inhibitor (Figure 7B) and subsequently depending on the

remaining sensitivity of the novel constructs towards the inhibitor, differences in their cytotoxic

potential were expected here.

3.7.1 Cytotoxic activity on PI-9-positive cell lines

To select the optimal granzyme B variant, the in vitro anti-tumor efficacy of the granzyme B mutants

GbR28K-Ki4(scFv), GbR201A-Ki4(scFv) and GbR201K-Ki4(scFv) was tested on the PI-9-positive HL cell

lines L428 and L1236 as well as the PI-9-positive CML cell line K562 and compared with the effects

triggered by the wild-type Gb-Ki4(scFv) or GbK27A-Ki4(scFv), respectively (2.7.3).

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

Figure 14: Cytotoxic effect of Gb-Ki4(scFv) and its

11 nM CFPs were incubated with PICell viability was evaluated with a colorimetric XTT100 %. The bar graphs show means ± SD of three to five determined via two-tailed t-test, (*)

As depicted in Figure 14, GbK27A

GbR201A-Ki4(scFv) showed significantly

GbR201K-Ki4(scFv), reduced the viability to 62

(CML) compared to the untreated control (p

3.7.2 Apoptotic activity on PI

Induction of apoptosis within all three cell lines was

mutant GbR201K-Ki4(scFv) by activating

Here, a statistically significant increase of the relative caspase

effects induced by the wild-type protein was detected.

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Ki4(scFv) and its mutants on PI-9-positive target cell lines.

PI-9+ HL cell lines L1236 and L428 or PI-9+ CML cell lineiability was evaluated with a colorimetric XTT-based assay (2.7.3). Viability of untreate

s ± SD of three to five independent experiments. Statistic(*): p<0.05, (***): p<0.001 (2.10).

, GbK27A-Ki4(scFv) performed as poorly as the wild

showed significantly reduced cell viability. The most effective

Ki4(scFv), reduced the viability to 62 % in L1236 (HL), 70 % in L428 (HL) and 58

(CML) compared to the untreated control (p < 0.001).

PI-9-positive cell lines

Induction of apoptosis within all three cell lines was additionally verified for

activating endogenous caspase 3/7 (2.7.2.2), as shown in

Here, a statistically significant increase of the relative caspase 3/7 activity by 25

type protein was detected.

_______________________________________________________________________________________________

60

positive target cell lines.

K562 for 48 h at 37 °C. Viability of untreated cells was set to

tatistical significance was

as the wild-type whereas

reduced cell viability. The most effective variant,

% in L428 (HL) and 58 % in K562

y verified for the most promising

, as shown in Figure 15.

3/7 activity by 25 - 35 % compared to

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

Figure 15: Apoptotic effect of Gb-Ki4(scFv) and

11 nM CFPs were incubated with PIApoptosis was evaluated by caspaseof Gb-Ki4(scFv). The bar graphs show means ± SD of three to five independent experiments. significance was determined via two

In summary, as predicted from the

promising mutant in all performed

3.8 IN VIVO BIOLOGICAL ACTIVITY O

3.8.1 Stability of Gb-Ki4(scFv) constructs in mouse serum

The stability of Gb-Ki4(scFv) and Gb

prior to conducting in vivo

determination of residual binding activity.

fluorescence-labeled antibodies specific for both the N

anti-His Alexa 488) were used. Exemplary results are shown in

For both constructs, around 90

37 °C whereas the binding activity of the granzyme

50 % after 48 h. However, within the first 10

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Ki4(scFv) and GbR201K-Ki4(scFv) on PI-9 positive target cell lines.

PI-9+ HL cell lines L1236 and L428 or PI-9+ CML cell line K562 for 48Apoptosis was evaluated by caspase 3/7 assay (2.7.2.2). Data are shown as normalized to

The bar graphs show means ± SD of three to five independent experiments. significance was determined via two-tailed t-test, (*): p<0.05, (***): p<0.001 (2.10).

as predicted from the in silico calculations, GbR201K-Ki4(scFv) was found to be the most

promising mutant in all performed in vitro studies, so it was further tested in vivo

IOLOGICAL ACTIVITY OF GB-KI4(SCFV) AND GBR201K

Ki4(scFv) constructs in mouse serum

Ki4(scFv) and GbR201K-Ki4(scFv) in mouse serum compared to PBS

studies. This was accomplished indirectly by

determination of residual binding activity. To verify stability of the full-length fusion protein

labeled antibodies specific for both the N- as well as the C-terminus (anti

Exemplary results are shown in Figure 16.

% binding activity was detected after incubation for 48

°C whereas the binding activity of the granzyme B constructs incubated in serum decreased to

h. However, within the first 10 h of incubation, a reduction of only 10

_______________________________________________________________________________________________

61

9 positive target cell lines.

CML cell line K562 for 48 h at 37 °C. Data are shown as normalized to the background signal

The bar graphs show means ± SD of three to five independent experiments. Statistical

Ki4(scFv) was found to be the most

in vivo.

K-KI4(SCFV)

compared to PBS was examined

indirectly by flow cytometric

length fusion protein,

terminus (anti-Gb PE and

% binding activity was detected after incubation for 48 h in PBS at

B constructs incubated in serum decreased to

a reduction of only 10 % was measured.

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

(A)

(C)

Serum stability [%]

Time [h] Gb-Ki4(scFv)0

0.5 2 4

10 24

32.5 48

100 100 96 99 90 65 63 49

Figure 16: Serum stability of Gb-Ki4(scFv) and GbR201K

70 ng/µL of each fusion protein wasafter different time points (2.4.8). Ranti-His Alexa 488 (2.5.1). Data exemplary of percentage serum and PBS stability

Serum stability was comparable for both the wild

point mutation.

3.8.2 Establishment of Hodgkin lymphoma

Before testing Gb-Ki4(scFv) and GbR201K

based on PI-9-negative cell line L540cy

To do this, L540cy cells were resuspended in

injected into 5 mice. Each mouse was injected in

(1*107 cells and 5*106 cells in 30

concentrations. Tumors, 2-5 mm in diameter,

visibility and accessibility, the tu

sizes were above 10 mm, so the

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

(B)

Serum stability [%] PBS stability [%]

(scFv) GbR201K-Ki4(scFv) Gb-Ki4(scFv) GbR201K100 107 97 99 90 69 62 53

100 - -

99 -

91 -

84

Ki4(scFv) and GbR201K-Ki4(scFv).

as incubated in 50 % mouse serum (A) or PBS (B) and samples were taken ). Residual binding activity to L540cy was measured via flow

exemplary shown for GbR201K-Ki4(scFv) and selected time pointsstability based on MFIs and related to 0 h respectively.

Serum stability was comparable for both the wild-type and the mutant indicating no influence of the

Hodgkin lymphoma xenograft models

Ki4(scFv) and GbR201K-Ki4(scFv) in vivo, subcutaneous xenograft

negative cell line L540cy, and on the PI-9-positive cell line L428, had to be established.

cells were resuspended in 50 % Matrigel™ to support tumor growth

. Each mouse was injected in both hind legs each with a different cell densit

cells in 30 µL) (2.9.3). A 100 % tumor take rate was obtained

mm in diameter, became clearly visible after 12 days.

tumors could be measured with a caliper. After three weeks

so the mice were sacrificed.

_______________________________________________________________________________________________

62

PBS stability [%]

GbR201K-Ki4(scFv) 100

- -

99 -

91 -

89

% mouse serum (A) or PBS (B) and samples were taken measured via flow cytometry using

and selected time points. (C) Overview

type and the mutant indicating no influence of the

subcutaneous xenograft tumor models

had to be established.

% Matrigel™ to support tumor growth and were then

different cell density

% tumor take rate was obtained from both cell

became clearly visible after 12 days. Due to their clear

After three weeks tumor

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

63

As the tumors seemed to be encapsulated, the resected tumors were also examined for the presence

of blood vessels. Tumor-derived tissue sections were stained with alkaline phosphatase without

levamisole to visualize any endothelial cells (2.8). Low vascularization was determined as

representatively shown in Figure 17A, therefore it had to be determined if intravenously applied

constructs could be delivered to the tumors. For that purpose, the reported CD30 ectodomain

shedding (1.5) was turned to account, since its level in mouse blood serum could provide information

on a connection of tumors with the bloodstream. The CD30 concentration in blood serum was

quantified with an earlier established ELISA set-up (2.4.9). It was found that it clearly correlated to

tumor size, such that the higher the CD30 level in blood serum the bigger the tumors (Figure 17B). In

addition, specific in vivo targeting of tumors was performed using Ki4(scFv)-SNAP-BG-747 and

molecular imaging (2.9.4). After injection into L540cy tumor-bearing mice localization of the signal in

the tumor was observed (Figure 17C) confirming continuous in vivo CD30 receptor expression and

successful delivery of the Ki4(scFv)-derived construct.

No obvious tumor growth was detected with L428 cells after tumor cell injection as described above,

so caliper measurements were impossible. Therefore, based on earlier obtained findings in this group

indicating the successful correlation between caliper determinations and the signals obtained by

optical imaging [150], Kat2 transfected L428 cells (2.1.6) were used that enable the visualization of

even slow and transient tumor growth via molecular imaging (Figure 17D). 5*106 cells were injected

into the right hind limb of 5 mice. Tumor growth was monitored by optical imaging using a Cri-

Maestro system with the yellow filter set (630-850 nm) and an exposure time of 500 ms every other

day. Tumor sizes remained constant for at least one week. Afterwards spontaneous tumor remission

was observed. A tumor take rate of 90 % was determined.

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

64

(A)

(B)

M1 M2 M30.0

0.1

0.2

0.3

0.4

0.5

0100200300400500600700800900

CD30 level Tumor size

CD

30 le

vel [

µg/µ

l]

Tumor size [m

m²]

(C) I II

(D)

I II

Figure 17: Evaluation of the established L540cy and L428 tumor models.

(A) L540cy-tumor was resected from sacrificed animals and 8 µm cryo-sections were prepared in a freezing microtome. Endothelial cells were stained pink with alkaline phosphatase without levamisole to determine vascularization of the tumor (2.8). (B) Blood serum was isolated from tumor-bearing mice and CD30 concentration was quantified via an ELISA assay (2.4.9). The results were correlated to the corresponding tumor sizes. M1, M2 and M3: three different mice determined. (C) 0.75 nM Ki4(scFv)-SNAP-BG-747 was injected into an L540cy-tumor bearing mouse. Measurement was performed with deep red filter set (730–950 nm), exposure time 500 ms. I) background signal, II) image with fluorescence signal (yellow, Ki4(scFv)-SNAP-BG-747). Receptor depending targeting of the tumor (left side) was detected after 6 h via in vivo molecular imaging with the CRi-Maestro system (2.9.4). Unspecific accumulation of the 747 signal in the kidney was observed (right side) as well as retained signal at the injection site (right eye). (D) 5*106 Katushka transfected L428 cells resuspended in 50 % Matrigel™ were injected subcutaneously into the right hind limb (2.9.3). Tumor area was detected and calculated via in vivo molecular imaging with CRi-Maestro system (2.9.4). Measurement was performed with yellow filter set (630–850 nm), exposure time 500 ms. I) fluorescence image with red circled tumor area of one representative mouse, II) same picture as I, edited to visualize the mouse.

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

65

Based on all described pre-testing, the L540cy tumor model can be regarded as well established so a

treatment schedule starting approximately after two weeks when tumor size was 2 to 5 mm was set-

up. Also, an appropriate L428-based tumor model could successfully be established with the help of

in vivo optical imaging technique so a treatment schedule starting one day after cell injection could

be implemented.

3.8.3 In vivo effects of GbR201K-Ki4(scFv) on PI-9-negative L540cy cells

No difference in cytotoxic activity between Gb-Ki4(scFv) and its mutant GbR201K-Ki4(scFv) was

detected on PI-9-negative cells in vitro, so only the mutant was used to confirm its effectiveness on

PI-9-negative cells in vivo. As a control the non-binding fusion protein GbR201K-H22(scFv) was used.

The treatment schedule followed is shown in Figure 18A. Mice were injected with cells as described

above (3.8.2), and then randomized in two groups (N = 3 for control group, N = 6 for treatment

group) as soon as the tumor size was 2 – 5 mm according to caliper measurements (after 16 days).

The mice were treated with 50 µg protein for 5 consecutive days. Treatment was then continued

every other day until day 13. CFPs were administered i. v. by retro-orbital injection. The size of the

tumors on day 1 of the treatment was set to 100 %. Results are shown in Figure 18B.

In addition, L540cy tumor derived tissue sections were stained after excision by

Ki4(scFv)-SNAP-BG-Vista-Green to confirm continued expression of CD30 after treatment with the

Ki4(scFv)-based construct (Figure 18C).

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

(A)

(B)

Figure 18: Evaluation of the in vivo

(A) Treatment schedule indicating intravenous (i.v.) injection of into the right hind limb (2.9.3), protein injection and days of measurement. (B) GbR201K-Ki4(scFv) (N = 6) or with nonwith caliper and related to initial tumor test, (***): p < 0.001 (2.10). (C) L540cand cut into 8 µm sections in a freezing microtome. CD30 receptor Ki4(scFv)-SNAP-BG-Vista-Green (2.8

The data obtained clearly indicate

successfully triggered an arrest of tumor growth, whereas tumors in mice receiving non

GbR201K-H22(scFv) more than doubled in size

3.8.4 Comparative in vivo

positive L428 cells

The established model based on Kat2 transfected L428 cells (

used to compare the therapeutic efficacy

wild-type Gb-Ki4(scFv) in vivo against PI

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

(C)

anti-tumor activity against L540cy-induced tumors.

(A) Treatment schedule indicating intravenous (i.v.) injection of 5*106 L540cy cells resuspended in Matrigel, protein injection and days of measurement. (B) Mice were treated either with

6) or with non-binding control GbR201K-H22(scFv) (N = 3). Tumor sizecaliper and related to initial tumor size (100 %). Statistical significance was determined via two

L540cy-tumor was resected from sacrificed GbR201K-Ki4(scFv) a freezing microtome. CD30 receptor expression was visualized by

2.8) and confocal microscopy (2.6).

clearly indicates that treatment with GbR201K-Ki4(scFv) was CD30

triggered an arrest of tumor growth, whereas tumors in mice receiving non

doubled in size (mm²) (p < 0.001).

effects of Gb-Ki4(scFv) and GbR201K-Ki4(scFv)

he established model based on Kat2 transfected L428 cells (3.8.2) and molecular imaging (

o compare the therapeutic efficacy of the most promising mutant GbR201K

against PI-9-positive tumors. 21 mice were divided

_______________________________________________________________________________________________

66

induced tumors.

resuspended in Matrigel™ Mice were treated either with

umor sizes were measured Statistical significance was determined via two-tailed t-

Ki4(scFv) treated mouse visualized by staining with

was CD30-specific and

triggered an arrest of tumor growth, whereas tumors in mice receiving non-binding

Ki4(scFv) on PI-9-

) and molecular imaging (2.9.4) was

of the most promising mutant GbR201K-Ki4(scFv) and its

divided randomly into

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

3 groups and injected with cells as described above (

GbR201K-Ki4(scFv), administered

Figure 19A. GbR201K-H22(scFv) was used as

Only the mutant GbR201K-Ki4(scFv)

reduction was statistically significant compared to both wild

whereas the difference between wild

statistically significant (Figure 19

(A)

(B)

Figure 19: Evaluation of the in vivo

(A) Treatment schedule indicating intravenous (i.v.) injection of in Matrigel™ into right hind limb (2.9.3either with GbR201K-Ki4(scFv) (N =(N = 7). Tumor size was measured viarelated to initial tumor size (100 %). (**): p < 0.005, ns: not significant (2.10and cut into 8 µm sections on a freezing microtome. Ki4(scFv)-SNAP-BG-Vista-Green (2.8Kat2 signal from transfected cells, III)

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

injected with cells as described above (3.8.2). Treatment with 50

administered i.v., was started one day after injection of cells

H22(scFv) was used as the non-binding control.

Ki4(scFv) killed the PI-9-positive cells in vivo. Here, the tumor size

reduction was statistically significant compared to both wild-type (p < 0.05) and control (p

whereas the difference between wild-type induced tumor size decrease and control group was not

19B).

(C)

I

anti-tumor activity against L428-induced tumors.

(A) Treatment schedule indicating intravenous (i.v.) injection of 5*106 Kat2-transfected L428 cells2.9.3), protein injection and days of measurement. (B) = 7), Gb-Ki4(scFv) (N = 7) or with non-binding control GbR201K

size was measured via in vivo molecular imaging with the CRi-Maestro System (%). Statistical significance was determined via two-tailed t

2.10). (C) L428-tumor was resected from sacrificed controlµm sections on a freezing microtome. CD30 receptor expression was visualized by staining with

2.8) and confocal microscopy (2.6): I) Ki4(scFv)-SNAP-BGKat2 signal from transfected cells, III) overlay of an enlarged image.

_______________________________________________________________________________________________

67

. Treatment with 50 µg Gb-Ki4(scFv) or

jection of cells, as depicted in

Here, the tumor size

0.05) and control (p < 0.005),

type induced tumor size decrease and control group was not

II

L428 cells resuspended , protein injection and days of measurement. (B) Mice were treated

binding control GbR201K-H22(scFv) Maestro System (2.9.4) and

tailed t-test, (*): p < 0.05, tumor was resected from sacrificed control-treated animals

CD30 receptor expression was visualized by staining with BG-Vista-Green signal, II)

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

68

The continuous presence of CD30 receptor after treatment with Ki4(scFv)-specific CFPs in L428-

derived tumors was verified by immunohistochemical analysis of tissue sections (2.8). Successful

receptor staining with Ki4(scFv)-SNAP-BG-Vista-Green as well as the sustained fluorescence of the

cytosolically expressed Kat2 protein is shown in Figure 19C.

During the animal experiments, all mice involved were active and did not show any symptoms of

disease caused by treatment.

These results confirmed the earlier in vitro data. GbR201K-Ki4(scFv) was able to kill PI-9-positive as

well as PI-9-negative cells in contrast to the wild-type which only killed the latter. To further verify

these findings, ex vivo studies were performed.

3.9 EX VIVO STUDIES WITH CMML AND AMML PATIENT-DERIVED MONOCYTES

Effective apoptosis induction of Gb-H22(scFv) in AML patient-derived primary cells which are known

to overexpress CD64 has been shown before [138]. Therefore they are a valuable target for this CFP

(1.6). In this study the rare diseases CMML as well as AMML were put into focus to determine if

these cells are also suitable for an immunotherapeutic granzyme B-based approach. In addition,

during these ex vivo studies, the potential of the newly generated granzyme B variants could be

further evaluated.

Peripheral blood samples of leukemic patients, mostly prior to any clinical treatment, were obtained

from the university hospital Aachen (PD Dr. med. Edgar Jost) after pre-testing for over-expression of

CD64. PBMCs were isolated via Ficoll gradient, cultivated in supplemented RPMI medium (2.1.8) and

used for the following investigations.

3.9.1 PI-9 expression in patient-derived PBMCs

As described for the investigation of cancer cell lines previously (3.1), primary cells were tested for

their ability to express PI-9 via western blot analysis (2.4.7). To evaluate if the cells up-regulate

endogenous PI-9 during cultivation, or as a result of external stimuli, primary cells were harvested

and lysed immediately after isolation from peripheral blood and again after different time points

either in the presence or absence of IFNγ. The addition of IFNγ prior to cytotoxicity assays was shown

to be essential in this study to retain the activation status of the isolated cells and thereby sensitivity

to the used CFPs (1.6).

The results, representatively displayed in Figure 20A for CMML IV, indicated that PI-9 expression was

independent of the addition of IFNγ, but dependent on cultivation time. Except for CMML II, all

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

CMML samples exhibited expression of PI

CMML I and CMML III, PI-9 was detected at time point 0 as well, albeit significantly lower than after

14 or 24 h respectively (data not show

positive. Results for all determined cells are summarized in

PI-9-negative did not even express PI

(A)

Figure 20: Expression of endogenous PI

CMML or AMML patients (2.1.8).

(A) Exemplary data for patient CMMLfor different time intervals, before harvesting. separated by SDS-PAGE (2.4.1) and transferred to a nitrocellulose membrane. Endogenousa western blot (2.4.2) using anti-human PI(NEB); Pos.: recombinant PI-9 as positive control. patient cells after at least 14 h cultivation duration.

In summary, three out of the four investigated CMML samples and one out of the three determined

AMML samples exhibited PI-9 expression after 5 days in culture.

3.9.2 Receptor phenotyping of PBMCs

So far it is not clear if the receptor CD64 is

overexpression on the isolated cells and to determine its co

markers for AMML and CMML (CD56, CD33 and CD14, compare

accordingly by flow cytometry and confocal microscopy

shown in Figure 21A.

All patient samples showed a high CD64 expression ranging from 56 to 98

staining revealed that in most patients co

occurred. Co-expression of CD64 and CD14 also became evident during the investigations.

and C confirm the double staining of cells during confocal micro

CMML I clearly showed double staining via anti

for CMML III double staining with anti

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

CMML samples exhibited expression of PI-9 after at least 14 h of incubation in R10 medium. In

9 was detected at time point 0 as well, albeit significantly lower than after

h respectively (data not shown). Only one AMML probe (AMML I) turned out to be PI

ll determined cells are summarized in Figure 20B, whereby the cells indicated as

negative did not even express PI-9 after 5 days in supplemented RPMI medium

(B)

Patient

CMML I CMML II CMML III CMML IV AMML I AMML II AMML III

Expression of endogenous PI-9 (2.4.7) in primary leukemic cells derived from peripheral blood of

) Exemplary data for patient CMML IV. Cells were incubated in R10 in presence or absence of 200before harvesting. Total soluble protein from cell lysates

) and transferred to a nitrocellulose membrane. Endogenoushuman PI-9 and GAM-PO (2.1.5). M: Pre-stained broad range marker in kDa

9 as positive control. (B) Overview of PI-9 expression for all determined primary ivation duration.


9 expression after 5 days in culture.

henotyping of PBMCs

if the receptor CD64 is a valuable target for AMML and CMML. To investigate its

expression on the isolated cells and to determine its co-expression with the known tumor

markers for AMML and CMML (CD56, CD33 and CD14, compare 1.6), the cells were phenotyped

by flow cytometry and confocal microscopy. An overview of the flow cytometric r

All patient samples showed a high CD64 expression ranging from 56 to 98 % on all cells. Double

staining revealed that in most patients co-expression of CD64 and CD33 as well as CD64 and CD

expression of CD64 and CD14 also became evident during the investigations.

and C confirm the double staining of cells during confocal microscopy (2.6). Here, cells from patient

I clearly showed double staining via anti-CD56-APC and anti-CD64-FITC (

III double staining with anti-CD56-APC + anti-CD64-FITC, anti-CD33-APC + anti

_______________________________________________________________________________________________

69

h of incubation in R10 medium. In

9 was detected at time point 0 as well, albeit significantly lower than after

I) turned out to be PI-9-

whereby the cells indicated as

9 after 5 days in supplemented RPMI medium (2.3.1).

PI-9 expression

(protein level)

+ - + + + - -

leukemic cells derived from peripheral blood of

in presence or absence of 200 U/mL IFNγ Total soluble protein from cell lysates (40 µg/lane) was

) and transferred to a nitrocellulose membrane. Endogenous PI-9 was detected in stained broad range marker in kDa

9 expression for all determined primary


d CMML. To investigate its

expression with the known tumor

), the cells were phenotyped

An overview of the flow cytometric results is

% on all cells. Double

expression of CD64 and CD33 as well as CD64 and CD56

expression of CD64 and CD14 also became evident during the investigations. Figure 21B

). Here, cells from patient

FITC (Figure 21B) whereas

APC + anti-CD64-FITC

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

70

and anti-CD14-APC + anti-CD64-FITC showed heterogeneity of the cell population, as indicated by

flow cytometry before (Figure 21C).

(A)

Tumor marker Receptor expression [%]

CMML I CMML II CMML III CMML IV AMML I AMML II AMML III

CD64 79 59 56 98 (7.4) 82 59 (18) 82 (31) CD56 68 3 57 65 n.d. 7 30

CD64+CD56 60 2 36 64 n.d. n.d. 28 CD14 50 39 59 97 n.d. 31 30

CD64+CD14 45 n.d. 44 97 n.d. n.d. 30 CD33 n.d. n.d. 75 98 n.d. 63 31

CD64+CD33 n.d. n.d. 50 97 n.d. 59 30 CD56+CD14 46 3 32 97 n.d. 2 n.d. CD33+CD14 n.d. n.d. 46 51 n.d. 23 20

(B) CD56-APC/

CD64-FITC

(C) CD56-APC/ CD33-APC/ CD14-APC/

CD64-FITC CD64-FITC CD64-FITC

Figure 21: Phenotyping of primary leukemic cells.

PBMCs were isolated from peripheral blood of leukemic patients (CMML or AMML, 2.1.8) and incubated with specific antibodies (2.1.5) according to manufacturer’s instructions (2.5.3). Binding was detected by flow cytometry (2.5.3) or confocal microscopy (2.6) (A) Overview of expression profile of patient-derived primary leukemic cells determined via flow cytometry; %: portion of receptor positive cells within the whole isolated cell population; n.d.: not determined; numbers in brackets: strong CD64 expression detected. (B-C) Confocal microscopy of cells from patient CMML I (B) and CMML III (C) stained with antibodies (2.1.5) as indicated.

As a result of the described observations, targeting AMML and CMML patient-derived cells and

potentially killing them with the CD64-specific H22(scFv)-based CFPs was pursued.

3.9.3 Target cell binding and internalization

After detection of CD64 expression, specific binding of the H22(scFv)-derived fusion proteins was

likewise determined. Exemplary results are shown in Figure 22A for cells from patient CMML IV after

incubation with Gb-H22(scFv), GbR201K-H22(scFv) or H22(scFv)-ETA’, respectively. All constructs

bound specifically to the cells.

10 µm 30 µM 10 µM 30 µM

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

Specific CD64-based internalization was demonstrate

H22(scFv)-SNAP-BG-647 or H22(scFv)

CMML III and CMML IV in Figure

indicating that no unspecific binding occurred to CD64

(A)

I

(B) I

Figure 22: Binding and internalization a

(A) 50 nM H22(scFv)-fused constructs were incubated with peripheral blood of patient CMML IV for 30(2.1.5) and flow cytometry (2.5.1). incubated with 4*105 CD64-positive primary cells from patient 10 min at 37 °C and internalization was detected with c

Since specific binding and internalization

performed to evaluate the cell death inducing

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

based internalization was demonstrated after incubation of primary cells with

or H22(scFv)-SNAP-BG-Vista-Green as representatively depicted for patient

Figure 22B. Some cells were not stained by the CD64

indicating that no unspecific binding occurred to CD64-negative cells.

II

II

inding and internalization analysis with primary leukemic cells (2.1.8).

fused constructs were incubated with 4*105 CD64-positive primary cells CMML IV for 30 min on ice. Specific binding was detected with anti

). (B) 50 nM H22(scFv)-SNAP-BG-647 or H22(scFv)-SNAPpositive primary cells from patient CMML III (I) or CMML IV (II) respectively

and internalization was detected with confocal microscopy (2.6).

specific binding and internalization to the primary cells was hereby verified

cell death inducing efficacy of the corresponding CFPs.

10 µM

_______________________________________________________________________________________________

71

d after incubation of primary cells with

y depicted for patient

B. Some cells were not stained by the CD64-specific construct,

positive primary cells derived from ected with anti-His Alexa488

SNAP-BG-Vista-Green was (I) or CMML IV (II) respectively for

verified, studies could be

corresponding CFPs.

10 µM

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

3.9.4 Cytotoxic activity on primary cells

Primary cells were incubated with the

GbR201K-H22(scFv) to evaluate differences in their efficacy depending on

targeted cells. The PI-9 insensitive

viability was measured via XTT

Prior to cytotoxic studies, cells were stimulated with IFNγ to retain activation of the cells, which was

confirmed via flow cytometry using CD64

did not induce apoptosis within the targeted cells (data not shown).

CMML

(A)

(B)

Figure 23: Cytotoxic and apoptotic

primary leukemic cells.

21 nM fusion protein was incubated with stimulated (200peripheral blood (2.1.8) of CMML (leftevaluated by XTT metabolization (A) ((2.7.2.1) (B), normalized to buffer control and converted to viabilityindependent experiments. For CMMLH22(scFv)-ETA’. Statistical significance was determined via two

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Cytotoxic activity on primary cells

Primary cells were incubated with the CFPs Gb-H22(scFv) and the PI-9

H22(scFv) to evaluate differences in their efficacy depending on PI-

sensitive IT H22(scFv)-ETA’ was used as positive control

XTT (Figure 23A) and Annexin V/ PI staining (Figure

cells were stimulated with IFNγ to retain activation of the cells, which was

confirmed via flow cytometry using CD64-specific antibodies. Without the addition of IFNγ, the CFPs

did not induce apoptosis within the targeted cells (data not shown).

AMML

and apoptotic effects of Gb-H22(scFv), GbR201K-H22(scFv) and H22(scFv)

nM fusion protein was incubated with stimulated (200 U/mL IFNγ) primary mononuclear cells derived from ) of CMML (left side) or AMML patients (right side) for 48 h at 37°C. Viability was

by XTT metabolization (A) (2.7.3) and apoptotic effects were measured via Annexin, normalized to buffer control and converted to viability. The bar graphs show means ± SD of three

For CMML I, AMML I and AMML II no experiments were performed with Statistical significance was determined via two-tailed t-test, (*): p<0.05, (***)

_______________________________________________________________________________________________

72

9 insensitive variant

-9 expression of the

positive control. Apoptosis and

Figure 23B), respectively.

cells were stimulated with IFNγ to retain activation of the cells, which was

specific antibodies. Without the addition of IFNγ, the CFPs

H22(scFv) and H22(scFv)-ETA’ on

IFNγ) primary mononuclear cells derived from h at 37°C. Viability was

) and apoptotic effects were measured via Annexin V/ PI assay The bar graphs show means ± SD of three

II no experiments were performed with p<0.05, (***): p<0.001 (2.10).

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

73

A statistically significant increased reduction in cell viability was observed after XTT evaluation for

GbR201K-H22(scFv) compared to the wild-type protein for all PI-9-positive cells CMML I, CMML III

and CMML IV as well as AMML I. Hence, more than 50 % of cells died in the presence of

GbR201K-H22(scFv) in sample CMML I whereas only 30 % died after incubation with the wild-type

protein. For CMML II almost 30 % of the cells were eliminated. The results for PI-9-positive CMML III

after incubation with GbR201K-H22(scFv) were in the same range, whereby the effect of the wild-

type fusion was 10 % lower. In CMML IV the difference between the mutant and Gb-H22(scFv) is

even more evident. Here more than 40 % of the monocytes died in the presence of

GbR201K-H22(scFv) but only 18 % were killed by the unmodified version. The largest difference and

thereby the most significant evidence for the superior performance of the mutant compared to the

wild-type was observed in PI-9-positive cells from patient AMML I. Here 75 % of the cells could be

killed by GbR201K-H22(scFv), but only 15 % by Gb-H22(scFv). In contrast the cytotoxic efficacy of

both constructs was the same on PI-9-negative CMML II, AMML II and AMML III cells (Figure 23A).

The cytotoxic efficacy H22(scFv)-ETA’ was as high as for the mutant granzyme B construct in case of

patient CMML III. In contrast, for all other determined cells, the performance of H22(scFv)-ETA’ was

significantly lower than for Gb-H22(scFv) or GbR201K-H22(scFv).

To confirm that cell death occurred through apoptosis, Annexin V/ PI assays were additionally

performed. The viability detected here were in general higher than the ones measured via XTT;

however the same tendencies were seen (Figure 23B).

Incubation of the primary cells with the inactive variant GbS184A-H22(scFv) or the non-binding

GbR201K-Ki4(scFv) did not reduce cell viability (data not shown).

In addition, the concentration-dependent cytotoxic effect of the granzyme B-based constructs was

compared using XTT. As indicated by the data above, the most evident results for the better

performance of the mutant were obtained for AMML I and are shown in Figure 24.

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

74

10 -1 100 101 102 103 104 1050

25

50

75

100

125 Gb-H22GbR201K-H22

CFP [ng/ml]

Via

bilit

y [%

]

Figure 24: Cytotoxic effects of Gb-H22(scFv) and GbR201K-H22(scFv) on primary leukemic cells of AMML I.

Stimulated (200 U/mL IFNγ) target cells isolated from peripheral blood (2.1.8) were incubated for 72 h at 37 °C with decreasing protein concentrations (serial dilution 1:3) of Gb-H22(scFv) (grey curve) and GbR201K H22(scFv) (black curve). Cell viability was evaluated with a colorimetric XTT-based assay (2.7.3). The graphs show means ± SD of triplicates per dilution.

The determined sigmoidal viability curve confirmed specific and dose-depending cell death for

AMML I with superior efficacy of the mutant. The IC50 value for GbR201K-H22(scFv) was 8.4 nM.

3.9.5 Specificity of target cell death

Since the viability and apoptosis assays were performed with the entire cell population isolated from

blood after Ficoll gradient, it had to be confirmed that cell death was exclusive for malignant cells.

Different approaches were followed.

The challenge here is that positive selection for CD64 expression (either prior to treatment for pre-

selection of only the target cells or after treatment to identify specific reduction of CD64-positive

cells) was not possible, because limited binding of the CD64-specific antibody was observed on cells

after incubation with non-toxic H22(scFv)-binding constructs (data not shown).

Therefore indirect approaches based on flow cytometric analysis had to be evaluated to confirm

successful elimination of the target cells. The first approach was to identify the location of the target

cell population (R1) within the FCS/ SSC dotplot via a CD64-specific antibody. After treatment with

Gb-H22(scFv) or GbR201K-H22(scFv), this particular population decreased by half which is shown in

Figure 25A for the investigated CMML patients. Representative data for CMML I are shown in Figure

25B where a higher decrease for the mutant than for the wild-type protein was detected.

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

(A)

(B)

Buffer Gb

Figure 25: Specific reduction of target

H22(scFv)-ETA’ detected in FSC/SSC dotplot after flow cytometry

21 nM fusion protein was incubated with stimulated (200peripheral blood (2.1.8) of all determined with CD64-specific antibody the location ofFSC/SSC dotplot (R1) and their reduction evaluated with WinMDISD of three independent experiments for CMML samples. t-test, (**): p < 0.005, (***): p < 0.001 (gated CD64-positive R1 region; portion

The data obtained correspond

Annexin V/ PI measurements (

activity.

Other methods to verify specificity of cell death within the mixed cell population

staining of treated cells with Annexin

significant co-expression with CD64. The latter was chosen accord

however, only gave reliable results for cells from patient CMML

explanation see 4.5.2). Here CD56

the amount of apoptotic cells within the CD56

treatment with Gb-H22(scFv) and even

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Buffer Gb-H22(scFv) GbR201K

target cancer cell population by Gb-H22(scFv), GbR201K

ETA’ detected in FSC/SSC dotplot after flow cytometry.

nM fusion protein was incubated with stimulated (200 U/mL IFNγ) primary mononuclear cells derived from all determined CMML patients for 48 h at 37 °C. By labeling of mononuclear cells

ic antibody the location of the target cell population was identified via flow cytometry and their reduction evaluated with WinMDI 2.8 (2.5.3). (A) The bar graphs show means ±

SD of three independent experiments for CMML samples. Statistical significance was determined via two0.001 (2.10). (B) Exemplary data shown for treated cells of patient

portion of target cells shown in percentage.

corresponds to the results of the previously obtained

V/ PI measurements (3.9.4). This confirms the specificity of the determined cytotoxic

to verify specificity of cell death within the mixed cell population

Annexin V/ PI and a cell-specific APC-labeled antibody

expression with CD64. The latter was chosen according to the results from

however, only gave reliable results for cells from patient CMML I as shown in

CD56-positive cells were gated so it could clearly be

the amount of apoptotic cells within the CD56-positive target cell population increased after

H22(scFv) and even more significantly after treatment with GbR201K

‚

_______________________________________________________________________________________________

75

GbR201K-H22(scFv)

GbR201K-H22(scFv) and

primary mononuclear cells derived from By labeling of mononuclear cells

via flow cytometry within (A) The bar graphs show means ±

Statistical significance was determined via two-tailed treated cells of patient CMML I with

obtained data from XTT and

specificity of the determined cytotoxic

to verify specificity of cell death within the mixed cell population included the parallel

labeled antibody which showed

ing to the results from 3.9.2. This,

I as shown in Figure 26A (for

clearly be demonstrated that

positive target cell population increased after

GbR201K-H22(scFv).

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

(A)

Buffer

(B)

Buffer

Figure 26: Reduction of target cancer cell population

specific antibodies (2.1.5).

21 nM fusion protein was incubated with stimulated (200peripheral blood (2.1.8) of CMML patients for 48were labeled in parallel with CD56Dotplots show Annexin V/ PI stained cells after specific CD56cells; bottom right: early apoptosis; top right: late apoptosis/dead cells. Numbers in the quadrants represent the percentage of cells of each category.CD14-specific antibody (anti-CD14-APC,

The third approach to verify specificity

treated cells with certain antibodies

monitoring of the decrease of

phenotyping experiments (Figure

remaining CD14-positive cells after treatment

cells. After incubation with the

Results are displayed as histograms in

to the buffer control was detected

With the help of all described approaches, specificity of cell death within the heterogeneous primary

cell population induced by the CD64

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Gb-H22(scFv) GbR201K-H22(scFv)

Gb-H22(scFv) GbR201K-

eduction of target cancer cell population by Gb-H22(scFv) and GbR201K-H22(scFv)

nM fusion protein was incubated with stimulated (200 U/mL IFNγ) primary mononuclear cells derived from ) of CMML patients for 48 h at 37°C. (A) Treated and untreated cells

were labeled in parallel with CD56-specific antibody (anti-CD56-APC, 2.1.5) and AnnexinV/ PI stained cells after specific CD56- gating. Bottom left: viable cells; top left: necrotic

cells; bottom right: early apoptosis; top right: late apoptosis/dead cells. Numbers in the quadrants represent the percentage of cells of each category. (B) Treated and untreated cells of patient CMML II were

APC, 2.1.5). M1 represents cells with strong CD14 expression.

to verify specificity of CD64-selected cell death was accomplished by staining the

certain antibodies which were co-expressed with CD64 (Figure

the decrease of specific receptor expressing cells. For example, for

Figure 21) indicated a co-expression of CD14 and CD64 so

after treatment could be correlated to the targeted

fter incubation with the CFPs, the cells were harvested and labeled with anti

Results are displayed as histograms in Figure 26B. A clear decrease for the treated samples compared

to the buffer control was detected so it could be concluded that cell death was specific.

approaches, specificity of cell death within the heterogeneous primary

cell population induced by the CD64-targeting constructs could be confirmed.

_______________________________________________________________________________________________

76

(scFv)

-H22(scFv)

H22(scFv) detected with

IFNγ) primary mononuclear cells derived from Treated and untreated cells of patient CMML I

) and Annexin V/ PI (2.7.2.1). Bottom left: viable cells; top left: necrotic

cells; bottom right: early apoptosis; top right: late apoptosis/dead cells. Numbers in the quadrants represent Treated and untreated cells of patient CMML II were labeled with

M1 represents cells with strong CD14 expression.

was accomplished by staining the

Figure 21). This allowed

example, for CMML II,

CD64 so the amount of

the targeted CD64-expressing

the cells were harvested and labeled with anti-CD14-APC.

B. A clear decrease for the treated samples compared

so it could be concluded that cell death was specific.

approaches, specificity of cell death within the heterogeneous primary

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

77

3.9.6 Serum stability of Gb-H22(scFv) fusion proteins in patient-derived serum

Serum stability is an important pre-requisite for later systemic, intravenous application of the

immunotherapeutic. Hence, Gb-H22(scFv) and GbR201K-H22(scFv) were incubated in either 50 %

human patient serum derived from patient CMML III or PBS at 37 °C for 1, 6 or 24 h (2.4.8). Stability

was determined by measuring residual binding activity via flow cytometry as described for the

Ki4(scFv)-based constructs (3.8.1). Results are summarized in Table 8.

Table 8: Serum stability of Gb-H22(scFv) and GbR201K-H22(scFv) in human patient serum (CMML III) or PBS.

Stability in serum [%] Stability in PBS [%]

Time [h] Gb-H22(scFv) GbR201K-H22(scFv) Gb-H22(scFv) GbR201K-H22(scFv) 0 100 100 100 100 1 98 96 96 96 6 80 74 92 93

24 66 62 92 92

Serum stability was comparable for both variants and could be confirmed for at least 6 h. After 24 h

serum stability was clearly reduced whereas retained binding activity after incubation in PBS was

significantly higher (reduction of 8 %). Hence, the serum- and time-dependent stability for both

constructs was determined.

3.10 EVALUATION OF GRANZYME M AS NOVEL EFFECTOR MOLECULE IN CYTOLYTIC

FUSION PROTEINS

Granzyme M has been described to possess anti-tumor activity (1.3.2), so in this study CFPs

containing this serine protease as an alternative to granzyme B were generated. To confirm the

potential of granzyme M, it was fused to different tumor-specific single chains and functionality

assays were performed. To evaluate the results properly, experiments were prepared in comparison

to the corresponding granzyme B-based CFP.

3.10.1 Generation and preparation of granzyme M fusion proteins

The fusion constructs comprised of granzyme M and the CD64-specific H22(scFv) or CD30-specific

Ki4(scFv) (2.1.5) are shown in Figure 27A. They were generated as described in 2.1.4 by replacing the

granzyme B sequence with that of granzyme M within the corresponding constructs.

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

(A)

(B)

Figure 27: Overview of granzyme

Gm-H22(scFv) (B).

(A) Schematic structure of the binary eukaryotic expression vector pMS EGm-Ki4(scFv). Ig-κ: signal peptide sequencecleavage site, H: 6x histidine tag, IVS/IRES: synthetic intron and internal ribosome entry site, EGFP: enhanced green fluorescent protein. Enterokinase cleavage site indicated with in HEK293T cells and purified from supernatant via 0.02 U/µg Enterokinase (2.4.4). Protein w(2.4.1) or transferred to a nitrocellulose membrane (2.4.2). Specific protein bands were detected at 61.5kDa (NEB).

As in the case of the granzyme

N-terminus to facilitate expression in eukaryotic cells.

supernatant should allow correct post

glycosylation sites. The protein was purified by affinity chromatography selecting for the His

with a yield of 1 mg/L, as determined by SDS

The identity of the protein was confirmed by western blot analysis using an anti

(2.4.2). Under denaturing conditions, SDS

150 kDa, a band with the expected molecular weight of 61.5

Hence, the intended granzyme

expressed in HEK293T cells.

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

I II

Overview of granzyme M-based constructs (A) and SDS-PAGE and western blot analysis

of the binary eukaryotic expression vector pMS encoding forκ: signal peptide sequence (L) of immunoglobulin kappa light chain, ECS: enterokinase

cleavage site, H: 6x histidine tag, IVS/IRES: synthetic intron and internal ribosome entry site, EGFP: enhanced Enterokinase cleavage site indicated with arrow. (B) Fusion proteins were expressed

in HEK293T cells and purified from supernatant via IMAC (2.4.4). Activation of the protein was done with Protein was separated by SDS-PAGE and either stained with coomassie (I)

) or transferred to a nitrocellulose membrane for western blotting with anti-human Specific protein bands were detected at 61.5 kDa and ~150 kDa. M: Pre-stained broad range marker in

granzyme B-based CFPs, an enterokinase site precedes the granzyme

terminus to facilitate expression in eukaryotic cells. The secretion of the fusion proteins into the

allow correct post-translational modification because granzyme

ation sites. The protein was purified by affinity chromatography selecting for the His

mg/L, as determined by SDS-PAGE (2.4.5) and activated with enterokinase (

The identity of the protein was confirmed by western blot analysis using an anti-

). Under denaturing conditions, SDS-PAGE revealed, apart from a band at approximately

a band with the expected molecular weight of 61.5 kDa (Figure 27B).

Hence, the intended granzyme M constructs could successfully be generated and

_______________________________________________________________________________________________

78

and western blot analysis of

encoding for EGm-H22(scFv) and of immunoglobulin kappa light chain, ECS: enterokinase

cleavage site, H: 6x histidine tag, IVS/IRES: synthetic intron and internal ribosome entry site, EGFP: enhanced Fusion proteins were expressed

). Activation of the protein was done with stained with coomassie (I)

human Gm and GAM-AP stained broad range marker in

an enterokinase site precedes the granzyme M

secretion of the fusion proteins into the

translational modification because granzyme M has three

ation sites. The protein was purified by affinity chromatography selecting for the His6-tag,

d with enterokinase (2.4.4.2).

-human Gm antibody

a band at approximately

generated and secretory

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

3.10.2 Functional characterization of granzyme M

Successful activation after enterokinase cleavage

ability of granzyme M to cleave PI

synthetic granzyme M substrate, Suc

Western blot analysis with anti-human PI

Gm-H22(scFv) in a 3:1 molar ratio. Recombinant PI

unknown reasons (3.4) but cleavage by granzyme

lower band (Figure 28A). The same results were obtained for Gm

The enzymatic functionality of the fusion protein was also verified by the applied colorimetric assay.

No activity was detected for the uncleaved variant EGm

dependent increase in extinction was observed for

(A)

Figure 28: Enzymatic activity of Gm

(A) Recombinant PI-9 (3.4) and Gm-was separated by SDS-PAGE and transferred to a nitrocellulose membrane (was detected in a western blot using antiin kDa (NEB). (B) The proteolytic activity of activated protein was measured via a colorimetric assay based on the synthetic granzyme M substrate Suc405 nm for different concentrations as indicated and compared to uncleaved variant EGmgraphs show means ± SD of three independent experiments.

These results confirm that enzymatically

expression in HEK293T cells and subsequent enterokinase digestion

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Functional characterization of granzyme M fusion proteins

after enterokinase cleavage was tested by two methods, one based on the

to cleave PI-9 (2.7.1.3), and the other one using a colorimetric assay with a

, Suc-Ala-Ala-Pro-Leu-pNA (2.7.1.1).

human PI-9 confirmed the cleavage of PI-9 after 24

H22(scFv) in a 3:1 molar ratio. Recombinant PI-9 was partially degraded during purification for

eavage by granzyme M was nevertheless clearly visible, intensifying the

ame results were obtained for Gm-Ki4(scFv) (data not shown).


tected for the uncleaved variant EGm-H22(scFv), whereas a concentration

dependent increase in extinction was observed for active Gm-H22(scFv) (Figure 28

(B)

Buffer

EGm-H

22 1µ

M

Gm-H22

200

nM

Gm-H22

1 µM

Gm-H

22 2

µM0.00

0.25

0.50

0.75

1.00

Abs

orba

nce

405n

m

Gm-H22(scFv) tested by cleavage of PI-9 or a colorimetric substrate

-H22(scFv) were incubated for 24 h at 37 °C (2.7.1.3) before protein mixture PAGE and transferred to a nitrocellulose membrane (2.4.2). Active and inactivated PI

was detected in a western blot using anti-human PI-9 and GAM-PO (2.1.5) M: Pre-stained broad range marker oteolytic activity of activated protein was measured via a colorimetric assay based on

substrate Suc-Ala-Ala-Pro-Leu-pNA (2.7.1.1). Reaction was documentednm for different concentrations as indicated and compared to uncleaved variant EGm

graphs show means ± SD of three independent experiments.

enzymatically active granzyme M fusion proteins could be produced via

K293T cells and subsequent enterokinase digestion.

_______________________________________________________________________________________________

79

ed by two methods, one based on the

ng a colorimetric assay with a

9 after 24 h incubation with

9 was partially degraded during purification for

M was nevertheless clearly visible, intensifying the

Ki4(scFv) (data not shown).


H22(scFv), whereas a concentration-

28B).

Gm-H22

1 µM

Gm-H

22 2

µM

9 or a colorimetric substrate.

) before protein mixture ). Active and inactivated PI-9 stained broad range marker

oteolytic activity of activated protein was measured via a colorimetric assay based on ). Reaction was documented at 37 °C and

nm for different concentrations as indicated and compared to uncleaved variant EGm-H22(scFv). The bar

could be produced via

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

3.10.3 Target cell binding

The antigen-specific binding of all

corresponding target cell lines.

CD64-positive AML cell line HL60 (

(Figure 29B). CD64-negative L428

(A) (B

(C)

Figure 29: Binding analysis of Gm-H22(scFv

50 nM Gm-H22(scFv) was incubated with 4*10blood of patient CMML IV (B) for 30 min on icecytometry (2.5.1). CD64- L428 cell line was

An obvious fluorescence shift was detected on CD64

attested for the granzyme M-based CFP. No binding was observed on the control cell line L428

confirming that the binding activity was s

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

specific binding of all granzyme M-based CFPs was tested by flow cytometry using the

target cell lines. Exemplary results are shown for Gm-H22(scFv) targeting human

positive AML cell line HL60 (Figure 29A) and CD64-positive primary cells from a CMML patient

negative L428 cells were used as the control cell line (Figure 29

(A) (B)

H22(scFv).

H22(scFv) was incubated with 4*105 HL60 cells (A) or primary leukemic cells derived from peripheral for 30 min on ice. Binding was detected with anti-His Alexa488 (

L428 cell line was used as negative control.

fluorescence shift was detected on CD64-positive cells, so specific binding could be

based CFP. No binding was observed on the control cell line L428

confirming that the binding activity was selective for the targeted receptor.

_______________________________________________________________________________________________

80

ed by flow cytometry using the

H22(scFv) targeting human

positive primary cells from a CMML patient

29C).

(A) or primary leukemic cells derived from peripheral His Alexa488 (2.1.5) and flow

so specific binding could be

based CFP. No binding was observed on the control cell line L428

The same exclusive

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

81

target-specific binding was confirmed for Gm-Ki4(scFv) using CD30-positive cell lines (data not

shown).

3.10.4 Cytotoxic activity of granzyme M fusion proteins

In accordance with the viability and apoptosis assays performed and described for Gb-H22(scFv)

(3.6.4 and 3.6.5), the CD64-positive AML cell line HL60 was used for the evaluation of Gm-H22(scFv).

The concentration-dependent cytotoxicity of Gm-H22(scFv) was evaluated using an XTT-based

colorimetric cell proliferation assay (2.7.3) with IFNγ stimulated HL60 cells. Unstimulated HL60 cells

and CD64-negative L428 cells were used as negative control.

As shown in Figure 30A, the viability of the stimulated HL60 cells declined significantly, whereas CD64

receptor negative L428 cells and unstimulated HL60 cells remained unaffected. The IC50 values were

between 1.2 and 6.4 nM which is in the same range as shown before in Figure 12B for Gb-H22(scFv).

A non-specific granzyme M fusion protein showed no evidence of cytotoxicity (data not shown).

The antigen-dependency of Gm-H22(scFv)-induced cytotoxicity was confirmed by carrying out a

competitive assay which showed a statistically significant dose-dependent reduction in cytotoxicity in

the presence of free H22(scFv) (Figure 30B). There was no significant difference between the viability

of cells incubated with free H22(scFv) alone or with the buffer alone.

The mechanism of Gm-H22(scFv) toxicity in HL60 cells was further characterized by carrying out an

Annexin V/ PI apoptosis assay (Figure 30C). When stimulated HL60 cells were incubated with 66 nM

Gm-H22(scFv) for 48 h, the number of apoptotic cells stained with Annexin V/ PI was 65 % greater

than in control experiments lacking the protein. This was in accordance with the results obtained

after treatment with Gb-H22(scFv) and demonstrated efficient programmed cell death. No difference

between experimental cultures and controls was detected for the CD64-negative CML cell line K562.

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

(A)

100 101 102 1030

25

50

75

100

CFP [ng/ml]

Via

bilit

y [%

]

(C)

Figure 30: Cytotoxic and apoptotic effects of Gm

(A) Target cells were incubated for 72 Gm-H22(scFv) on stimulated (50 U/mbased assay (2.7.3). Viability of untreated cells was set to 100dilution. CD64- cell line L428 and nonHL60 cells were incubated for 72 h free H22(scFv). Viability was evaluated viathree independent experiments. Statistical significance was determined via two(2.10); ns: not significant. (C) 66 nM G48 h at 37°C. CD64 K562 cell line was used as PI assay (2.7.2.1). Bottom left: viable cells; top left: necrotic cells; bottom right: early apoptosis; top right: late apoptosis/dead cells. Numbers in the quadrants represent the percentage of cells of each category.

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

104 105

L428

HL60HL60 unst.

(B)

totic effects of Gm-H22(scFv).

Target cells were incubated for 72 h at 37°C with decreasing protein concentrations (U/mL IFNγ) HL60 cells. Cell viability was evaluated with a colorimetric XTT

Viability of untreated cells was set to 100 %. The graphs show means ± SD of triplicates per non-stimulated HL60 cells were used as control. (B) Stimulated (50

at 37°C with 10 nM Gm-H22(scFv) in competition with different dilutions of Fv). Viability was evaluated via XTT metabolization (2.7.3). The bar graphs show means

Statistical significance was determined via two-tailed tnM Gm-H22(scFv) was incubated with stimulated (50 U/m

K562 cell line was used as negative control. Apoptotic effects were measured via AnnexinBottom left: viable cells; top left: necrotic cells; bottom right: early apoptosis; top right: late

cells. Numbers in the quadrants represent the percentage of cells of each category.

*

_______________________________________________________________________________________________

82

protein concentrations (serial dilution 1:3) of iability was evaluated with a colorimetric XTT-

%. The graphs show means ± SD of triplicates per Stimulated (50 U/mL IFNγ)

H22(scFv) in competition with different dilutions of The bar graphs show means ± SD of

tailed t-test, (*): p < 0.05 U/mL IFNγ) HL60 cells for

were measured via Annexin V/ Bottom left: viable cells; top left: necrotic cells; bottom right: early apoptosis; top right: late

cells. Numbers in the quadrants represent the percentage of cells of each category.

*

ns

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

To evaluate the potential of granzyme

granzyme B Gm-Ki4(scFv) was tested in comparison to Gb

L428, L1236 and L540cy. The former

resistant against apoptosis induction by wild

PI assay are shown in Figure 31.

Figure 31: Apoptotic effect of Gm-Ki4(scFv) compared to

11 nM fusion protein was incubated with HL cell lines L428 (PIat 37 °C. Apoptotic effects were measured via Annexinnecrotic cells; bottom right: early apoptosis; top right: late apoptosis/dead cells. Numbers in the quadrants represent the percentage of cells of each category

Gm-Ki4(scFv) is able to kill PI-9

L540cy. As expected, Gb-Ki4(scFv) only kills

effectively than Gm-Ki4(scFv). During a dose

for Gm-Ki4(scFv) on L540cy was 272

(Figure 12A). No unspecific toxic effects were determined on CD3

shown).

To confirm the potential of granzyme

Gm-425(scFv) and Gb-425(scFv)

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

To evaluate the potential of granzyme M with PI-9-positive cells, in comparison to wild

Ki4(scFv) was tested in comparison to Gb-Ki4(scFv) on the CD30

he former of these two are PI-9-positive and have been shown to be

apoptosis induction by wild-type Gb-Ki4(scFv) (Figure 7). Results for an Annexin

Ki4(scFv) compared to Gb-Ki4(scFv).

nM fusion protein was incubated with HL cell lines L428 (PI-9+), L1236 (PI-9+) and L540cy (PI°C. Apoptotic effects were measured via Annexin V/ PI assay (2.7.2.1). Bottom left

necrotic cells; bottom right: early apoptosis; top right: late apoptosis/dead cells. Numbers in the quadrants represent the percentage of cells of each category.

9-positive cells L428 and L1236 and also the PI

Ki4(scFv) only kills the PI-9-negative cell line L540cy but

Ki4(scFv). During a dose-depending cytotoxicity assay the measured IC

Ki4(scFv) on L540cy was 272 nM which is 100x lower than was determined for Gb

A). No unspecific toxic effects were determined on CD30-negative cell lines (data not

To confirm the potential of granzyme M-based CFPs for the treatment of further tumor entities

425(scFv), which target the epidermal growth factor receptor (EGFR)

_______________________________________________________________________________________________

83

in comparison to wild-type

CD30-positive cell lines

and have been shown to be

Results for an Annexin V/

) and L540cy (PI-9-) (3.1) for 48 h Bottom left: viable cells; top left:

necrotic cells; bottom right: early apoptosis; top right: late apoptosis/dead cells. Numbers in the quadrants

PI-9-negative cell line

0cy but does so more

depending cytotoxicity assay the measured IC50 value

determined for Gb-Ki4(scFv)

negative cell lines (data not

based CFPs for the treatment of further tumor entities,

epidermal growth factor receptor (EGFR)

Results ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

84

overexpressed in a variety of tumor cells [151], were analyzed in comparative assays. Both granzyme

constructs showed a similar specific cytotoxic as well as apoptotic effects against the EGFR-positive

epidermoid cancer cell line A431, but did not affect the EGFR-negative cell line L540cy (data not

shown).

Concluding, granzyme M was fused to three different cell-binding moieties and showed specific anti-

tumor activity towards the corresponding target cells while leaving receptor negative cells

unaffected. It is also able to kill cells in presence of the granzyme B inhibitor PI-9 in contrast to the

wild-type granzyme B construct.

3.10.5 Ex vivo studies with Gm-H22(scFv)

Apart from in vivo studies, ex vivo experiments with primary cells are important assays to determine

the efficacy of novel immunotherapeutics. The characteristics of these patient-derived cells predict

the later effects in humans much better than artificial cell lines (see also 4.5). Hence, the potential of

the novel human effector molecule granzyme M fused to H22(scFv) was tested comparative to

Gb-H22(scFv) using the same primary cells as used in the experiments described in 3.9. In addition,

by parallel experiments with GbR201K-H22(scFv) the prospective of granzyme M on PI-9-expressing

cells in comparison to this newly established PI-9 insensitive variant could also be assessed.

Cytotoxic effects were monitored using the XTT assay (Figure 32A) or by Annexin V/ PI staining

(Figure 32B) as described in 3.9.4. For Gm-H22(scFv) both assays indicated that cell viability was

reduced by about 40 % in all samples. The same was observed for GbR201K-H22(scFv). Significant

differences of these both constructs compared to the wild-type Gb-H22(scFv) were evident for cells

from patient CMML IV, tested PI-9-positive before (Figure 20A). Here only 15 % of the cells could be

killed by the wild-type granzyme B-based construct.

The specificity of the cytotoxic effect and thereby the selective reduction of the CD64-positive target

cell population was demonstrated here in the same manner as described earlier in 3.9.5 (data not

shown).

To further evaluate the cytotoxic effect of Gm-H22(scFv) on primary cells, dose-dependent cell

viability assays were prepared for cells from patient AMML II (Figure 32C) and AMML III (Figure 32D).

The sigmoidal curves evidently demonstrated the dose-dependency of cell death triggered by

Gm-H22(scFv) ex vivo, even though general cell viability remained rather high.

Results _____________________________________________________________________________________________________________________________________________________________________________________________________________

(A)

(C)

100 101 102 1030

25

50

75

100

concentration [log ng/ml]

Via

bilit

y [%

]

Figure 32: Cytotoxic and apoptotic effects of Gm

(A,B) 21 nM fusion protein was incubated with stimulated (200from peripheral blood (2.1.8) of AMML or CMML patients (as indicated) for 48been tested according to 3.9.1. Viability was evaluated by XTT metabolization (A) (were measured via Annexin V/ PI assay (Data for Gm-H22(scFv) were compared to Gbshow means ± SD of three independent experiments.test, (*): p < 0.05 (2.10); n.d.: not determined.peripheral blood (2.1.8) of patient AMML II (PIwith decreasing protein concentrations (serial colorimetric XTT-based assay (2.7.3).

These data clearly indicated that Gm

contrast to the wild-type Gb-H22(scFv),

in PI-9-expressing target cells .

________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

(B)

104 105


(D)

100 101 1020

25

50

75

100


Via

bilit

y [%

]

Cytotoxic and apoptotic effects of Gm-H22(scFv) on primary leukemic cells.

nM fusion protein was incubated with stimulated (200 U/mL IFNγ) primary mononuclear cells derived ) of AMML or CMML patients (as indicated) for 48 h at 37°C. PI

. Viability was evaluated by XTT metabolization (A) (2.7.3PI assay (2.7.2.1), normalized to buffer control and converted to viability (B).

H22(scFv) were compared to Gb-H22(scFv) and GbR201K-H22(scFv) from Figure show means ± SD of three independent experiments. Statistical significance was determined via two

); n.d.: not determined. (C,D) Stimulated (200 U/mL IFNγ) target cells ) of patient AMML II (PI-9-) (A) and AMML III (PI-9-) (B) were incubated for 72

with decreasing protein concentrations (serial dilution 1:3) of Gm-H22(scFv). Cell viability was evaluated with a ). The graphs show means ± SD of triplicates per dilution.

These data clearly indicated that Gm-H22(scFv) could specifically kill leukemic cells

H22(scFv), it performed as well as the new mutant GbR201K

_______________________________________________________________________________________________

85

103 104 105


IFNγ) primary mononuclear cells derived at 37°C. PI-9 expression has 2.7.3) and apoptotic effects

), normalized to buffer control and converted to viability (B). Figure 23. The bar graphs

Statistical significance was determined via two-tailed t-U/mL IFNγ) target cells isolated from

were incubated for 72 h at 37 °C iability was evaluated with a

The graphs show means ± SD of triplicates per dilution.

leukemic cells ex vivo. In

as well as the new mutant GbR201K-H22(scFv)

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

86

4 Discussion

The use of immunotoxins (ITs) highly increases the efficacy of cancer therapy leaving healthy cells

mostly unaffected. However, the clinical deployment of early generation ITs was limited due to their

immunogenicity, resulting in the generation of neutralizing antibodies that exclude repeated

applications. Therefore complete humanization of ITs is an important prerequisite for the successful

development of more effective ITs [24,31], called cytolytic fusion proteins (CFPs). This concept is based

on the use of endogenous human enzymes that induce cell death, fused to humanized or even fully

human antibody fragments. Candidate human enzymes include proteases, RNases and kinases

[4,31,109,152,153]. In this thesis, the focus was granzymes, which are human immunoproteases

involved in the granule-mediated apoptosis of virus-infected or transformed tumor cells. The

efficiency of granzyme B was elucidated in relation to the expression of its specific inhibitor PI-9.

Since limited effects were observed for the wild-type protein on PI-9-positive cells, the intention was

to generate novel variants via molecular modeling and site-directed mutagenesis, which are

enzymatically active even in presence of PI-9. When fused to the CD30-targeting Ki4(scFv) or CD64-

targeting H22(scFv), the most promising variant was eventually selected and its efficacy was proven

in vitro and in vivo on CD30-positive Hodgkin lymphoma and CML cell lines, as well as ex vivo on

CD64-positive primary leukemic cells. In addition, another member of the granzyme family,

granzyme M, was introduced in this study bearing the potential to become a novel effector molecule

in context of CFPs. Combined with different cancer cell-specific ligands, its cytotoxic efficacy could

successfully be shown in vitro as well as ex vivo, with the advantage to be active despite the presence

of the inhibitor PI-9. In the following, the hallmarks of granzyme B and its PI-9 insensitive mutants as

well as of granzyme M will be discussed in detail focusing on their therapeutic potential and

applicability for later clinical use as immunotherapeutics.

4.1 THE THERAPEUTIC POTENTIAL OF GRANZYME B-BASED CYTOLYTIC FUSION

PROTEINS

Granzyme B has been evaluated as a CFP targeting CD64-positive cells during a previous PhD thesis in

this work-group and was published before by Stahnke et al. [138]. In this study Gb-H22(scFv) was

successfully used to selectively kill AML-derived HL60 cells and primary leukemic cells (Figure 12 and

Figure 23). In addition the in vitro protocols for another granzyme B-based CFP, Gb-Ki4(scFv), have

been optimized to allow IC50 value determination. The IC50 values were comparable with the ones

determined for Gb-H22(scFv).

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

87

Other groups have also shown the great anti-tumor potential of granzyme B-based CFPs as

summarized in Table 9. Apart from its low immunogenicity, granzyme B has several inherent

advantages as an immunotherapeutic component like high cytotoxic efficacy reflecting its status as

an effector molecule of the cellular immune system, broad range of apoptosis inducing mechanisms

with the potential to circumvent anti-apoptotic proteins (1.3.1). Furthermore, compared to other cell

death inducing proteins like microtubule-associated proteins which exclusively kill proliferating cells

[153], it is independent of cell proliferation, which is advantageous for non- or low-proliferating

malignant diseases like CMML. Also, the production of granzyme B (and CFPs derived therefrom) in

various heterologous expression systems (E. coli, Pichia pastoris, insect cells and mammalian cells

such as HEK293T, HeLa, Jurkat and COS [4]) is relatively straightforward, which reflects a significant

prerequisite before biopharmaceuticals can enter clinical studies.

The involvement of granzyme B in natural cellular defense, however, is a mixed blessing in targeted

cancer therapy since the efficacy of ITs/ CFPs depends on the sensitivity of tumor cells to pro-

apoptotic signals. One of the major challenges in both targeted cancer therapy and standard

chemotherapy approaches, that induce programmed cell death, is the tendency for tumor cells to

develop either escape mechanisms or resistance to pro-apoptotic signals allowing host

immunosurveillance. This limits the clinical application of granzyme B insofar as that it is specifically

inhibited by PI-9, which usually protects cells from misdirected granzyme B (1.4.1), but is also

upregulated in certain tumors (4.2). Other limiting factors are, that granzyme B has a high isoelectric

point (pI ~ 10), resulting in a positive surface charge contributing to non-specific binding to negatively

charged cell surface proteoglycans via electrostatic interactions. In addition, it is, as all CFPs in

general, released slowly after uptake by the receptor-mediated endocytosis which reduces its

cytotoxic impact [4]. In this study, the PI-9-related challenge is focused and in detail discussed in the

following sections, whereas solutions to the latter factors will be introduced in chapter 5.

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

88

Table 9: Overview of different granzyme B-based cytolytic fusion proteins (CFPs) published so far.

(blue background: data generated in this study and internal publication of this group)

CFP (as

published)

(expression

system)

Indication Target

receptor /

antigen

Target cells IC50 values PI-9 expression

reported from

others

Activation /

generation

of free N-

terminus

Ref.

Gb-H22(scFv)

(HEK293T)

AML, CD64+

malignancies

CD64 U937

AML cells (ex

vivo)

HL60

CMML; AMML

(ex vivo)

1.7–17 nM

--

~ 1.2 nM

--

No*, yes [155]

yes [155]

No*, yes [156]

Partially (see

Figure 20))

In vitro

activation

via EK

[138]

This

work

Gb-Ki4(scFv)

(HEK293T)

CD30+

malignancies,

e.g. Hodgkin

lymphoma

CD30 L540cy,

L428, L1236,

K562

~ 2.5 nM No*

Yes*

In vitro

activation

via EK

[157]

This

work

GrB-scFvMEL

(E.coli)

Melanoma gp240 A375-M

~ 20 nM

n.d. In vitro

activation

via EK

[158]

GrB-scFvMEL

(E.coli)

Melanoma gp240 A375-M (in

vivo)

MEL-526

TXM-18L

~ 30 nM

~ 50 nM

~ 150 nM

n.d.

n.d.

n.d.

In vitro

activation

via EK

[159]

GrB-TGFα

GrB-scFv

(FRP5)

(P.pastoris)

Breast

carcinoma

EGFR,

ErbB2

(Her2)

MDA-MB468 0.25 nM

0.29 nM

(+chloro-

quin for

both)

n.d. In vivo

activation

by Kex2

[160]

B3-GzmB,

GzmB-CD8

(E.coli)

Breast

carcinoma

Lewis Y

antigen

SK-BR3

MCF-7

(MCF-7casp3)

98 nM,

1595 nM

140 nM

(35 nM),

198 nM

(394 nM)

Yes [80]

Yes, after

induction by

estrogens [77]

Polyionic

adapters

between

granzyme B

and dsFv-B3

[161]

GrB-VEGF121

(E.coli)

Tumor

vascular

endothelial

cells

VEGF PAE/FLK-1

PAE/FLT-1

~ 10 nM

--

n.d.

n.d.

In vitro

activation

via EK

[162]

ImmunoGrB-

PEAII

(HeLa, Jurkat)

Breast and

ovarian

carcinoma

Her2 SK-BR3 (in

vivo)

SKOV-3

Not

applicable

since

supernatant

was used

Yes [80]

n.d.

PEAII trans-

location

motif to

allow

endogenous

furin

activation:

e23sFv-

PEAII-GrB

[163]

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

89

Table 9: Cont.

CFP (as

published)

(expression

system)

Indication Target

receptor /

antigen

Target cells IC50 values PI-9 expression

reported from

others

Structural

hallmarks

Ref.

ImmunoGrB-

Fpe/Fdt/R9 #

(HeLa, Jurkat)

Breast, gastric

and

hepatocellular

carcinoma

Her2 SK-BR3

SGC-7901 (in

vivo)

AGS

Hep G2

Not

applicable

since

supernatant

was used

Yes [80]

n.d.

n.d.

n.d.

Introduction

of novel

furin

cleavable

sites

between

e23Fv and

granzyme B

(compare

above)

[164]

GrB-YCG

(insect Sf9

cells)

Ovarian,

breast,

endometrial

and prostate

carcinoma

hLHR MA-10

(murine)

MCF-7

PC-3

0.16 µM

--

--

No PI-9

expected in

murine cell line

In vitro

activation

via EK prior

to use

[165]

GrB/4D5/26

(HEK293T)

Breast,

gastric,

lung,

cervical

carcinoma

Her2/neu BT474 M1

Calu3

eB-1

MDA MB435

NCI-N87

MDA MB453

Me180

29.3 nM

40.5 nM

93.1 nM

< 500 nM

90.4 nM

56.8 nM

< 500 nM

Yes**

Yes (strong)**

No**

No**

Yes**

Yes**

Yes**

Insertion of

a pH-

sensitive

fusogenic

peptide for

improved

delivery of

granzyme B

into the

cytosol

[166]

Abbreviations: GrB, Gb, GzmB: granzyme B; hLHR: human luteinizing hormone receptor; Her2: human epidermal growth factor receptor 2; Kex2: killer expression defective 2; EGFR: epidermal growth factor receptor; VEGF: vascular endothelial growth factor; dsFv: disulfide stabilized Fv fragment; n.d.: not determined in other publications; EK: Enterokinase; PEAII: second domain of Pseudomonas Exotoxin A (residues 253-364); (in vivo): cell line was successful killed in mice; * compare 3.1; # Fpe: furin site sequence from PEA (amino acids 273-282), Fdt: furin site sequence from diphtheria toxin (amino acids 187-196), R9: poly-arginine tract; **as determined in the same study.

4.2 IMPACT OF ENDOGENOUS PI-9 EXPRESSION IN CANCER CELLS ON THE PRO-

APOPTOTIC ACTIVITY OF GRANZYME B

A direct correlation between PI-9 expression and resistance to apoptosis induction by wild-type

granzyme B was verified in this study (Figure 7). Several groups have already determined the

expression of the granzyme B inhibitor PI-9 in primary cancer cells. An overview of the results is

shown in Table 10 and Table 11. The impact of resistance on granzyme B-based immunotherapeutics

is controversial, since some reports claim that resistance to perforin-dependent pathways is not a

relevant escape mechanism in lymphoma [148] and that the induction of apoptosis by granzyme B

can be even achieved in cells expressing or upregulating PI-9 (compare Table 9). For example, Cao et

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

90

al. have not observed any relationship between PI-9 levels and cell sensitivity in their granzyme B

fusion protein targeting Her2/neu-positive cell lines [166]. Another example is a CFP comprising

granzyme B and dsFv-B3, targeting the Lewis Y carbohydrate antigen overexpressed e. g. on

squamous cell lung carcinoma or lung adenocarcinoma, which has been shown to kill the human

breast cancer lines SK-BR3 and MCF-7 in an antigen-dependent manner [161]. The reported IC50

values are rather low compared to the ETA-based reference, which might reflect the expression of

PI-9 in SK-BR3 cells and the ability of MCF-7 cells to upregulate PI-9 in the presence of estrogen [77].

However, it is difficult to confirm in retrospect whether these cells actually expressed PI-9, because

cell lines can differ depending on their source, cultivation conditions, and cell cycle phases [167].

Indeed, several studies have revealed a direct correlation between PI-9 expression and the loss of

granzyme B pro-apoptotic activity in vitro and in xenograft mouse models [64,80,168-170].

As a consequence of these published data, and since a limited effect induced by some granzyme B-

based CFPs has been observed in our group, a closer look at the PI-9 protein expression profiles in

potential target cell lines was carried out in this work (3.1). Hereby, the influence of cultivation time

on the PI-9 level could be excluded via investigations with a Hoechst sorting reagent indicating that

all cells in one cultivation flask were in the same maturation status. Jurkat cells were found to be

PI-9-negative as suggested by published data [148], so they could be used as negative control. The

PI-9 expression shown within K562 (CML cell line) is consistent with earlier findings whereas the

absence of PI-9 in the AML cell lines HL-60 and U937 is controversial [82,148,155,156]. The latter

might be related to differences in detection methods, cultivation conditions and the fact that PI-9 can

be triggered by external stimuli (compare Table 10 and Table 14).

It was found that the PI-9 expression level differed among the EGFR-positive and CD30-positive target

cell lines (Table 6). The latter were chosen for further analysis (3.5) to compare cell lines which were

initially derived from patients with the same cancer type (compare details in 4.3.2). Flow cytometric

analysis indicated a heterogeneous expression of PI-9 among these cells (Figure 3B). This is in

accordance to an earlier study (compare Table 11) showing that 6 out of 57 tested HL patients

expressed the inhibitor [171]. In addition, PI-9 expression was also upregulated upon EBV infection,

which is a main cause of B-type HL [72] and was recently even directly linked to EBV-positive cases of

HL [172]. Other data, shown in Table 10 and Table 11, also indicate that PI-9 is expressed in a

heterogeneous manner within the tumor cell population, leading to the survival of a small number of

malignant cells surrounded by innate immune cells. PI-9 was expressed primarily in tumors

containing many granzyme B-positive infiltrating CTLs and the high incidence of PI-9-positive tumor

cells often predicted an unfavorable clinical outcome [173]. This suggest that PI-9-positive tumor

cells are positively selected as they evade immunosurveillance, which is clinically relevant for the

treatment of hematological disorders or solid tumors in which PI-9 expression is upregulated during

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

91

immunosurveillance or tumor therapy. Therefore the development of drugs that are effective against

such resistant cells is highly advantageous [4].

Table 10: Overview of PI-9 expression in different carcinomas / melanomas.

Cancer type Detection Hallmarks Ref.

Lung cancer In vitro (cell lines)

In vivo (primary cancer

cells)

PI-9 highly expressed in lung cancer cell lines

PI-9 expression was increased in primary lung cancer

cells and significantly correlated with cancer stage

(dependent on granzyme B expression of CTLs)

[170]

Prostate cancer In vitro (cell lines →

qPCR and flow

cytometry)

In vivo (prostate tumor

tissue → qPCR and

immunohistochemistry)

PI-9 expression in cell lines PC3 and DU-145

PI-9 expression is up-regulated in pre-cancerous

states, which is dysregulated in later stages whereas

it remains in some tumors (pilot study)

[169]

Non-small cell

lung carcinoma

cells (NSCLCs)

and tissues

In vitro (cell lines → RT-

PCR and western blot

analysis)

In vivo (lung tissue

samples from biopsies →

RT-PCR)

Strong PI-9 expression in 6 and low in 4 of 10 cell

lines on mRNA level correlating with detection on

protein level

PI-9 mRNA and protein expression in all of 150

patients at variable levels in NSCLC cells and tumors,

the less differentiated lung adenocarcinomas

showed significantly higher expression of PI-9 mRNA

as compared to the well-differentiated tumors

[174]

Breast cancer In vitro (cell line MCF-7

→ western blot and qRT-

PCR)

In vivo (MCF-7 xenograft

mouse model)

PI-9 expression is induced by estrogens and depends

on an interplay between estrogens, estrogen

receptor and EGF/EGFR

Induction of PI-9 by the estrogen ‘genistein’

[77]

[175]

Stage III and IV

melanoma

In vitro (cell lines →

western blot of cell

lysates)

In vivo (paraffin

embedded tissue of

patients →


PI-9 expression in 6 of 14 melanoma cell lines

PI-9 expression in 21 of 26 cases of primary

melanoma and in 22 of 28 metastases

After categorizing in respect to percentage of PI-9

expressing tumor cells (+ if > 50 % of cells expressed

PI-9, −if < 50 % of cells expressed PI-9) 15 of 26

primary tumors and 12 of 28 metastases were +

PI-9 expression in melanoma metastases correlates

with poor clinical outcome following active specific

immunotherapy (ASI) of stage III and IV melanoma

patients

[176]

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

92

Table 10: Cont.


Nasopharyngeal

carcinoma

In vivo (formalin-fixed,

paraffin-embedded

tumor biopsies →


PI-9 expression in 3 of 43 cases

Presence of many tumor infiltrating activate CTLs

within patient biopsies is related to bad clinical

outcome

[177]

Melanoma,

breast,

cervical and

colon carcinoma

In vitro (cell lines → PCR)

In vivo (primary colon

carcinoma cells → PCR)

PI-9 expression in a subset of determined tumor

lines (e.g. MCF-7, SK-BR3)

PI-9 expression in 2 of 4 primary surgical specimens

[80]

Strategies to optimize the efficacy of granzyme B-based CFPs on primarily PI-9-positive cells include

either the downregulation of PI-9 expression/ activity in target cells or the generation of granzyme B

variants that are insensitive to PI-9 (3.5). The former may be achieved by RNA interference (RNAi) or

antisense technology to knock down PI-9 gene expression. The successful use of siRNA to inhibit the

expression of PI-9 has been described in stem cells [69] and in breast cancer [77]. However, interest

in development of siRNA-based therapies has declined following the Phase III failure of Bevasiranib,

which targeted the expression of vascular endothelial growth factor (VEGF) for the treatment of age-

related wet macular degeneration [178]. In alternative approaches, the NFκB inhibitor pyrrolidin

dithiocarbamate (PTDC) has been used to suppress PI-9 gene expression in PBMCs cultivated in vitro

[72], whereas the serine protease granzyme M (found in the cytotoxic granules of NK cells) can

inhibit PI-9 at the protein level [56]. Hence, the genetic fusion of granzyme M to a specific binding

domain facilitating cellular uptake could lead to the inactivation of endogenous PI-9 so the co-

administration of granzymes M and B could potentially clear the way for granzyme B-based therapy

(further discussed in 4.6.2) [4].

Those strategies, however, are indirect approaches which require the administration of granzyme B

in combination with a separate therapeutic modality. In contrast, as primarily intended in this study,

granzyme B can be modified in such a way that it retains its enzymatic activity but becomes

insensitive to PI-9. In an analogous approach, a human tissue-type plasminogen activator was

mutated at a few critical contact residues to achieve resistance against a complex mixture of

endogenous serpins [37,179].

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

93

Table 11: Overview of PI-9 expression in different lymphomas.


Classical

Hodgkin

lymphoma

In vivo (lymph node

sections → gene

expression via

microarrays)

PI-9 expression determined in 6 EBV+ and 10 EBV- cases:

EBV+ are PI-9+, EBV- are PI-9-

[172]

Leukemia In vitro (cell lines →

western blot analysis)

Ex vivo (monocytes

from peripheral blood)

PI-9 expression in 4 of 6 cell lines

PI-9 expression in 0 of 2 ALL cases, 3 of 4 AML cases, 2

of 3 CLL cases

[180]

NK/ T cell

lymphoma

In vivo (deparaffinized

tissue sections of

patient biopsies →

immunohisto-

chemistry)

PI-9 expression in 26 of 39 cases

PI-9 expression level was heterogeneous from case to

case with clusters of negative cells suggesting the

emergence of PI-9 down-regulated subclones associated

with aggressiveness and invasive potential

High levels of PI-9 associated with favorable clinical

outcome

[181]

Lymphomas In vitro (cell lines →

flow cytometry)

Ex vivo (isolation of

monocytes from

peripheral blood →

flow cytometry)

PI-9 expression in 10 of 18 lymphoma cell lines (e.g.

K562)

PI-9 expression in 9 of 14 primary lymphomas

Using highly activated CL in vitro no inhibition of the

perforin-dependent cytotoxic pathway has been

observed.

[148]

ALCL

(anaplastic

large cell

lymphoma)

In vivo (biopsies of

systemic ALCL patients

→ immunohisto-

chemistry)

PI-9 expression in 6 of 45 cases (percentages of PI-9+

cells ranged from 5 % to 75 %), primarily found in

tumors harboring many Gb+ infiltrating CTLs

High numbers of PI-9+ tumor cells predict resistance to

chemotherapy-induced apoptosis and unfavorable

outcome

[173]

Different non-

Hodgkin and

Hodgkin

lymphomas

In vivo (formalin-fixed,

paraffin-embedded

tumor biopsies →

immunohisto-

chemistry)

Sub-division into 5 categories depending on percentage

of PI-9+ cells

PI-9 expression in neoplastic cells in 36 of 92 cases of T-

cell lymphoma, 20 of 75 of B-cell lymphoma and 6 of 57

of Hodgkin lymphoma (expression varied between 5 and

100 %)

For further differentiation of PI-9 expression and

lymphoma type see reference.

[171]

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

94

4.3 GENERATION AND IDENTIFICATION OF A PI-9 INDEPENDENT GRANZYME B

VARIANT

4.3.1 Evaluation of in silico findings and comparison with in vitro enzymatic activity

As described in the previous chapter, a granzyme B variant that can kill both PI-9-positive and PI-9-

negative tumor cells will provide a significant therapeutic advantage, particularly for the treatment of

relapsing tumors. Since the underlying reaction is typical for the interaction between proteases and

serpins the necessary structural hallmarks could be predicted in silico using a structurally related

Michaelis complex formed between a rat trypsin and M. sexta serpin B1 (3.2).

Using state-of-art molecular modeling techniques key residues at the interface of the two proteins

which might decrease their affinity could be identified (3.2). Among wild-type granzyme B residues

involved in complexing, only residues not located in its active site were considered so the catalytic

activity was not expected to be affected. First results were generated using computational alanine

scanning mutagenesis, a convenient and fast strategy to investigate the determinants of protein

complex interactions [182]. However, hydration at the protein/ protein interface is not included in

these calculations even though water molecules may play a very important role for complex stability

[183]. Another disadvantage is that only alanine mutations can be simulated and the flexibility

properties of the complex are omitted. Therefore, molecular dynamics simulations of wild-type and

mutated complexes were performed in water solution which on the one hand demands a greater

computational cost but on the other hand allows the qualitative assessment of the effect of all

proposed mutations. Those mutations were decided considering the charge of the original amino

acid, arginine, so in the end, including one double mutant, seven new granzyme B mutants were

investigated (R28A, R28E, R28K, R28A/R201A, R201A, R201E, R201K) [149].

In vitro the corresponding mutations were introduced into the wild-type sequence of granzyme B via

site-directed mutagenesis. The protein was produced via the established secretory expression in

HEK293T cells [137,138,184] allowing natural glycosylation patterns within granzyme B which

harbors two N-glycosylation sites [185]. These sites were not affected by the mutations. No

differences in production rates were observed compared to the wild-type protein and all variants

could successfully be cleaved by enterokinase (3.3).

To test the remaining in vitro proteolytic activity after site-directed mutagenesis and challenge with

PI-9, the inhibitor had to be recombinantly produced since pure protein was not commercially

available. Bacterial expression turned out to be the most efficient method to obtain sufficient

amounts of protein (3.4). The identity of the protein was confirmed by mass spectrometric analysis

(data not shown). The successful complex formation of the produced protein could be verified after

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

95

incubation with recombinant granzyme B-based CFPs by detection of a SDS-stable complex on

western blots by different antibodies (Figure 6). Because only the upper band disappeared after

incubation with granzyme B (3.4), it could be concluded that the lower band might be the earlier

described inactive variant of PI-9 lacking the RCL which resulted in the small decrease in molecular

weight compared with intact PI-9 [56]. The same samples were used in a colorimetric granzyme B

substrate assay and the inhibition of enzymatic activity by complex formation between wild-type

granzyme B and the produced recombinant PI-9 could be directly confirmed. This assay was

consequently used as the read-out system to identify the most promising PI-9-insensitive granzyme B

mutants. Here, it was found that the in silico calculations were highly consistent with the in vitro

results: GbR201A-, GbR201K- and GbR28K-based fusion proteins (to H22(scFv)) in the absence of PI-9

feature catalytic activity similar to that of wild-type granzyme B indicating that the active site was not

affected. In the presence of PI-9, GbR201K-H22(scFv) retains almost its entire catalytic activity

whereas GbR28K-H22(scFv) and GbR201A-H22(scFv) kept three quarter or half of their enzymatic

potential, respectively. The measured Km values (~ 104 µM) for GbR201K-H22(scFv) was almost

identical to the one determined for the wild-type fusion protein (3.6.1) and consistent with published

data for recombinant granzyme B [83,86].

The PI-9 insensitivity of the recently described granzyme B variant K27A by Sun et al. [86] could not

be reproduced neither during in silico nor in vitro experiments. The reason might be that this earlier

structural in silico prediction was based on a complex with a modeled truncated PI-9 version

consisting of only 12 residues (334–345) of the RCL. The latter is a solvent exposed stretch of amino

acids representing the primary site of interaction with human granzyme B. Subsequently, different H-

bonds and salt bridges have been considered by Sun et al. resulting in different predictions.

Inconsistencies seen in the in vitro assays could be explained by a different expression system for

human granzyme B and PI-9 (P. pastoris with Sun et al.) leading to differences in glycosylation

patterns [186], which can influence complex formation. Differences in protein quality used in this

thesis and in the previous study could also have an effect.

With the combination of the applied in silico and in vitro techniques it was possible to pre-select

three out of eight generated granzyme B mutants, which retained their enzymatic activity even after

incubation with recombinant PI-9.

4.3.2 Selection of the optimal granzyme B mutant

The above described results give a first insight into the functionality of the newly generated mutants.

The three most promising variants (R28K, R201A, R201K) were further tested in vitro according to

their binding activity and affinity to target cells, stability and cytotoxic as well as apoptotic efficacy.

K27A was used as a comparative control. A sufficient evaluation of their PI-9 sensitivity required PI-9-

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

96

negative as well as PI-9-positive cells. CD30-positive cell lines were used based on reasons described

in detail earlier (4.2). To target those cells Ki4(scFv)-based fusion proteins were generated whose

potential had been confirmed previously in combination with ETA’ whereby IC50 values of 43 - 85 pM

were determined [112]. CD30-targeting murine Ki4(scFv) is an outstanding anti-CD30 single chain as

it does not bind soluble CD30 receptor [187]. This is important since it has been reported that CD30

ectodomain shedding occurs at high incidences [108], which could lead to an off-target capturing of

the CFP. The H22(scFv)-based fusion proteins which were used during the above discussed in vitro

characterization (4.3.1) were also employed during the evaluation of the mutants on PI-9-negative

cells as a reference in comparative assays.

It was found that both wild-type and mutants specifically bound to the target cells and did not exhibit

unspecific binding to CD30- or CD64-negative control cell lines, respectively (Figure 10). Kd for the

wild-type and the R201K mutated variant were also in the same nM range on the respective cell lines

(3.6.2) and correlated to the affinity of 3*10-9 M published for Ki4(scFv) alone [187] and 10-8 to

10-10 M reported for H22(scFv) [188]. This excluded a steric hindrance of antibody fragment binding

by the mutated granzyme B fusions. Also a negative influence of the mutations on apoptosis inducing

potential (Figure 13) as well as the cytotoxic activity could be excluded, since the observed IC50 values

were comparable to the respective values determined with the wild-type protein (Figure 12).

Compared to other published granzyme B-based CFPs (Table 9), the determined IC50 values exceed

the average.

The next step was to compare the ability of the generated mutants to kill PI-9 expressing cells and

eventually select the most promising one. For this goal, different strategies were considered: either

direct delivery of granzyme B and its variants into the cells or indirect delivery based on specific

targeting of tumor antigens by genetic fusion to tumor specific ligands. Since the former approach

was not successful (compare (3.5, 1)), strategies involving indirect, receptor-mediated delivery of the

protein into the cells were pursued instead. Here, the modification of originally PI-9-negative cell

lines to express PI-9 was first intended to exclude any cell-based differences (other than the

expression of PI-9) which occur when different types of naturally PI-9-positive or -negative cell lines

are used. The resistance of the cell lines used towards transfection with pMS-PI9-IV, however, limited

this approach. Since HEK293T cells successfully produced cytosolic PI-9 after transfection with

pMS-PI9-IV, it could be excluded that the reasons for the transfection difficulties were based on the

chosen plasmid DNA, but rather were due to the cells themselves. It has been reported that certain

cell lines are intrinsically easier to transfect than others, although the exact reason for these

differences is currently unknown. Additionally, even clonal variability in DNA uptake has been

reported in certain mouse cells [189].

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

97

An alternative solution was to use cell lines equal in antigen-expression, but different in PI-9

expression. As described in 4.2, CD30-positive cell lines belong to this category which also showed

differences in sensitivity towards Gb-Ki4(scFv), whereby a limited apoptotic effect was observed on

PI-9-positive L428 cell line whereas successful cell death induction was seen on PI-9-negative L540cy

cells (Figure 7B). Limited internalization due to differences in receptor expression (Table 7) could be

excluded as a reason for variations in cytotoxicity since complete internalization could be verified

with the help of Ki4(scFv)-SNAP-BG-647 (Figure 11) and similar Kd values were measured for all cell

lines used (Table 7).

Based on these findings, an adequate cell-based test system could be set up for the further

evaluation of the generated mutants and selection of the most promising variant. As suggested by

the previous substrate assay (4.3.1), the wild-type and the already published mutant GbK27A were

only effective in PI-9-negative cells (Figure 14). GbR201K-Ki4(scFv) was the most effective as it was

able to significantly reduce the viability of both PI-9-positive (L428, L1236, K562) and PI-9-negative

(L540cy) cell lines. These results were verified in an assay based on caspase 3/7 (Figure 15), which is a

strong indicator for apoptosis induction, especially in case of granzyme B induced cell death (1.3.1).

The cytotoxic effects triggered by GbR201K-Ki4(scFv) in L428 and L1236 were rather low compared to

those in L540cy. The addition of endosomolytic agents like chloroquine led to improved effects in

previous studies [160], but did not increase cytotoxicity here; neither did the increase in protein

concentrations. However, the observed differences can be explained by the published, lower

sensitivity of L428 and L1236 to apoptosis induction [97]. As described in 1.5, one of the signaling

pathways characteristic for HL involves NFκB. In these previous studies L428 and L1236 were much

more resistant to apoptosis induction by the anti-CD30 antibody 5F11 than L540 due to the

expression of this anti-apoptotic protein leading to induced resistance to CD30-mediated cell death

in HL cells. These effects can be explained by the defective NFκB inhibitor IκBα in L428 or the instable

inhibitor in L1236 cells causing constitutively active NFκB, and thereby apoptosis resistance [190-

192]. In one study it was shown that NFκB could be down-regulated through the proteasome

inhibitor bortezomib after CD30-mediated NFκB activation with the human CD30 antibody 5F11 [97].

The same protocol was adapted in this thesis as well. However, here, the addition of bortezomib did

not improve cytotoxicity. Nevertheless, even the currently observed low cytotoxic effects can have

high medical impact, especially if they lead to the elimination of treatment resistant cells responsible

for the minimal residual disease which are potentially related to PI-9 expression (4.2). On the other

hand the CD30-based system applied here has to be regarded as a model system to select the most

promising granzyme B mutant and this was successfully accomplished.

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

98

In summary, based on a computational model of the complex between granzyme B and its inhibitor

PI-9, seven new granzyme B variants were generated and their in silico predicted potential was

successfully confirmed during comparative in vitro studies with the wild-type protein and an earlier

suggested mutant. Finally, the variant GbR201K was selected as the most promising and was used for

further evaluation as fusion to CD30-targeting Ki4(scFv) in vivo (4.4) as well as CD64-targeting

H22(scFv) ex vivo on patient derived leukemic cells (4.5).

4.4 EFFICACY OF THE GBR201K MUTANT IN VIVO

To confirm the differences between the in vitro performance of wild-type and mutant granzyme B

R201K, comparative studies on the biological activity were done employing the above used CD30-

positive cell lines L540cy (PI-9-negative) and L428 (PI-9-positive).

First, the serum stability of GbR201K-based CFPs was compared to that of the wild-type, and no

influence of the point-mutations on stability was observed (Figure 16 and Table 8). Murine serum

itself was found to already be very cytotoxic towards the target cells. Therefore a proper

discrimination between cell death induced by the fusion proteins or the serum was unfeasible. The

isolation of the protein from the serum via IMAC was not done as then the protein concentration

would be depending on His6-tag binding efficacy and not just serum stability. Alternatively, residual

binding activity of the constructs to the target cells was used as measure for serum stability, whereby

the detection of full length CFP was ensured by the employment of both His6-tag and granzyme B-

specific antibodies. Via this method, stability for at least 10 h was observed for both constructs. The

half life of biopharmaceuticals strongly depends on their composition and the intended application

and varies highly ranging from minutes to days [193]. The stability measured here is lower than the

one published for the ETA’-based construct Ki4(scFv)-ETA’ which retained 85% to 90% of initial

binding activity for at least 24 h [112]. However, according to Kampmeier et al. a 10 h stability should

suffice for our application. In their study the tumor-specific SNAP-based signal was strongest after

this time point in a subcutaneous mouse tumor model [194].

Two different xenograft mouse models were established in this study, one based on the PI-9-positive

L428 cells and one on PI-9-negative L540cy cells. So far in vivo studies for HL have been strongly

restricted to a single cell line, L540cy, that achieves a sufficient tumor take rate in subcutaneous and

disseminated mouse models [140]. However, tumor take rate related problems for L540cy have been

experienced before in this group (Dr. rer. nat. Theo Thepen, personal communication). Therefore, it

was decided to focus on a subcutaneous tumor model that is easier to monitor and faster to evaluate

than a disseminated one. Via resuspension of the cells in Matrigel™, which leads to a fixed

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

99

accumulation of the cells and has been described to be advantageous to establish difficult tumor

models [195], a high tumor take rate could be achieved (3.8.2). Because of the observed

encapsulation and limited vascularization (Figure 17B) the connection of the tumor to the

bloodstream was tested by quantitative analysis of soluble tumor-derived CD30 receptor in blood

samples of tumor-bearing mice (3.8.2). The direct correlation observed between tumor size and

CD30 concentration in blood suggested that intravenous application of the CFP could be successful.

In addition, the localization of the fluorescent signal after injection of Ki4(scFv)-SNAP-BG-747 into

L540cy tumor-bearing mice (Figure 17A) verified the specific delivery and the continuous CD30

receptor expression on the tumors themselves in vivo. The latter was also confirmed by staining of

tissue sections prepared from both L540cy and L428 tumors after treatment with the Ki4(scFv)-

derived constructs (Figure 18C and Figure 19C). Following, the usability of the established tumor

model was completely demonstrated. The translocation of the construct to the low vascularized

tumors was possibly supported by extravascular transport playing an important role within tumor

tissue [196]. Several studies in the past have changed the understanding of the interstitial-lymphatic

continuum of tumors from a passive to a more active one, so the delivery of tumor therapeutics is

possible [197]. To optimize transport conditions in future animal experiments, the intravascular

transport could be improved by the addition of dopamine, which supports the distribution and

normalization of dysfunctional blood vessels in tumors and has significantly increased the

concentration of anticancer drugs in tumor tissues in previous studies [198].

As mentioned, no tumor mouse model was available for PI-9-positive L428 cells. Such a model was,

however, necessary for the evaluation of the novel granzyme B mutant. Via Kat2 expression by these

slowly and transiently growing cells in vivo the monitoring of tumor growth became possible via

optical imaging (Figure 17C). The direct correlation between caliper measurements and tumor size

determined by the imaging technique was proven recently confirming the reliability of the imaging

measurements [150]. Hence, a novel HL tumor model based on the cell line L428 was established in

this work.

Since no difference in cytotoxic activity between Gb-Ki4(scFv) and its mutant GbR201K-Ki4(scFv) was

detected on PI-9-negative cells in vitro, only the mutant was used to confirm its effectiveness on PI-9-

negative cells in vivo. Up-regulation of PI-9 in L540cy as an immune escape mechanism could be

excluded by cell lysis and subsequent western blots of L540cy tumor-derived cells using PI-9-specific

antibodies (data not shown). To treat L428 tumors both constructs were used, to confirm differential

influence of endogenous PI-9 on the biological activity of wild-type and mutant granzyme B. To

exclude unspecific effects by granzyme B itself, instead of using a buffer control, the non-binding

control GbR201K-H22(scFv) was chosen. The obtained results clearly demonstrated that only the

mutant was able to kill the PI-9-negative tumor cells as well as the PI-9-positive ones in vivo so this

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

100

variant has the potential to replace the wild-type in the future, since this only kills PI-9-negative cells

(Figure 18B and Figure 19B). With this, the results of the previously obtained in vitro analyses could

be confirmed after the establishment of appropriate tumor mouse models.

In this study 50 µg of CFP per treatment day and injection was applied which was well tolerated by

the mice. In a previous study by Gehrmann et al. much higher injected concentrations of 20 µg

human granzyme B per gram body weight have been reported without inducing adverse side-effects,

meaning that after injection of active granzyme B the investigated organs exhibited no

pathophysiological changes and individual organs exhibited normal tissue architecture, with no

evidence of infarcts or internal bleeding [199]. Therefore it seems that active human granzyme B at

doses to efficiently reduce tumor size, does not induce detectable histological or morphological

changes in normal organs. However, those conclusions have to be regarded with care since in our

case immune suppressed mice were used and also species-dependent substrate specificities have to

be considered which can distort the results. Those species-dependent substrate specificities leading

to lacking cell death induction of granzyme B in foreign species have been described by others in

vitro [200]. In addition the same could be shown and published in the time course of this work as

well ex vivo (data not shown) when the cytotoxic efficacy of CD64-specific human CFPs and ETA’-

based IT was tested on murine and human derived macrophages [201]. In that study, Gb-H22(scFv)

only killed human derived macrophages, leaving the murine derived cells unaffected. These data

suggest that human granzyme B is not able to cause any side-effects in mice since it cannot even kill

target cells after specific delivery. Therefore, the estimation of maximum tolerated doses (MTDs) of

human granzyme B in humans becomes a challenge because the results from other species might be

misleading. Possible side-effects in humans could be expected from unspecific binding to cells or

tissue, due to the high positive surface charge of granzyme B [202], which can mediate electrostatic

interactions with the negatively-charged heparin sulfate proteoglycans found on the surface of

almost all cells. Also its potentially excessive extracellular activity, which is important for tissue

remodeling during an immune response, can cause pathophysiological tissue destruction [203]. It

seems that the only choice for a predictive pre-clinical model system will be primates due to their

close relationship to humans. Corresponding studies have not yet been reported and neither has the

efficacy of human granzyme B been tested in any primate cells. Another important fact regarding the

in vivo experiments described here is that we use subcutaneous tumor models. These models are

artificial because it does not reflect the biology of HL in humans (1.5). However, the focus here was

on the differences in cell death induction by the wild-type versus the mutant and not on the disease

itself. Therefore this model was sufficient for our studies. In summary, the exclusive biological

activity of the novel granzyme B mutant R201K on PI-9-positive as well as PI-9-negative cells in

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

101

contrast to the wild-type protein could successfully be demonstrated in vivo based on the newly

established xenograft tumor models.

4.5 EX VIVO DETERMINATIONS ON PRIMARY LEUKEMIC CELLS

Before biopharmaceuticals are approved for clinical studies, their in vitro and in vivo efficacy has to

be tested extensively. These experimental data are based either on animal models or on cell-based

assays. While animal models are commonly used for toxicity and metabolism studies, their

extrapolation to humans is restricted. Furthermore, they are both expensive and have low-

throughput which usually limits their use to the late stages of preclinical testing. Cell-based assays

are continuously being improved to help reverse the increasing trend of costly late-stage drug

failures. These improvements are of interest to the pharmaceutical industry. Use of standard well

established cell lines has the benefit of phenotypic stability and homogeneity, unlimited life span and

ease of culture. However, they cannot mimic the complexity and heterogeneity of e. g. tumor cells

and for some diseases appropriate cell lines are not even available. As a consequence, ex vivo

models, comprising primary cells, are coming more and more into focus because they are thought to

simulate more closely the actual cells. With the help of ex vivo models, the molecular mechanisms

underlying tumor behavior can be elucidated and the potential of new anticancer treatment

candidates evaluated. In addition to the great value for the characterization of new drugs in general,

another advantage of ex vivo determinations is that individual patients can be studied enabling

predictions on effectiveness of new drug candidates prior to clinical application [204]. These aspects

are especially valuable in context of personalized medicine. However, primary cells are variable and

often can only be cultured for a few passages before they reach senescence or undergo undesirable

phenotypic changes. In addition, the availability of patient material is often very limited and the

heterogeneity due to differences in patients’ treatment history restricts the comparison between

different experiments. As a consequence, established cell lines have their value in providing first

indication into the functionality of new drug candidates yielding reproducible data with standardized

methods. Those results, however, should always be confirmed using appropriate primary cell

material to approach the complex clinical situation and the real diverse hallmarks of the disease.

4.5.1 CMML and AMML primary cells as target for CD64-specific fusion proteins

Based on the above rationale, the potential of wild-type granzyme B and its most promising mutant

was evaluated ex vivo in addition to in vitro and in vivo studies. For this purpose the CD64-specific

constructs Gb-H22(scFv) and its mutant GbR201K-H22(scFv), whose in vitro efficacy has extensively

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

102

been tested during the time-course of this work (3.6.4 and 3.6.5), were applied and their efficacy was

compared to the established IT H22(scFv)-ETA’. As target, the leukemic rare diseases CMML as well

as AMML (1.6) have been chosen, since specific therapeutics against CMML and AMML represent

unmet medical need for the patients. Research of corresponding targeted therapeutics is restricted

by the lack of frequent and specific genetic aberrations known until now in this disease and the lack

of mouse and in vitro models of the chronic disease. Thus, studies with primary cells, as performed

within this thesis, are the only available system to test new immunotherapeutics.

The hematological malignancy leukemia is especially suitable for immunotherapeutic treatment with

CFPs because the circulating cancer cells are easily accessible and incomplete tumor penetration due

to the absence of vascularization is not a limiting factor [205]. Numerous studies have been

described in literature dealing with immunotherapeutics against primary leukemic cells, e.g. via IL-3

[206], IL-4 [207] or CD33 [208,209]. Examples have been introduced in 1.2.1 and 1.6.

Before cytotoxic assays on those primary cells were performed it was first evaluated, if CD64 is a

valuable target to treat CMML and AMML patients, since this tumor marker has not been described

for these diseases so far (1.6). Primary monocytes were isolated from fresh peripheral blood of

largely newly diagnosed patients (university hospital, Aachen) (2.1.8). The high affinity receptor CD64

is generally a promising candidate target in leukemia therapy. It is superior to most of the other

investigated targets due to its characteristics of efficiently internalizing fused ligands as well as its

restricted expression pattern, leaving normal, non-activated CD64-expressing cells unaffected (see

below and 1.6). The potential of CD64-targeting H22(scFv)-ETA’ as well as the ability of Gb-H22(scFv)

to specifically kill primary AML cells has been shown before [129,138]. The advantage of the latter is

that its cytotoxic efficacy is based on a human effector molecule preventing expected immunogenic

responses provoked by the non-human cell-death inducing agents and thereby allowing multiple

treatment cycles.

During phenotyping of the primary cells for CD64, CD56, CD14 and CD33 receptor expression (3.9.2),

a clear over-expression of CD64 cells was found on the CMML cells, enabling specific binding of

H22(scFv)-derived constructs (Figure 22A). The expression of CD64 has been discussed before in

context of CMML, where a decrease of this marker has been described [115,210]. However,

published data are based on immunohistochemical staining of bone marrow which is different to our

results based on flow cytometric analyses of monocytic cells. The detected overexpression of CD64

on the primary AMML cells was in accordance to published data [114]. Through double staining of

patient derived cells, significant co-expression of CD64 and CD56 as well as co-expression of the

malignant cell markers CD33 and CD14 with CD64, respectively, was observed via flow cytometry and

confocal microscopy (Figure 21). The finding that the same cells express CD64 in parallel to the

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

103

already known tumor markers confirmed the assumption that CD64 is a promising target to eliminate

malignant cells in CMML and AMML patients. In addition, this knowledge is valuable for the

verification of the specific cell death described in the following section.

To evaluate if Gb-H22(scFv) and GbR201K-H22(scFv) are actually stable in CMML and AMML patient-

derived serum, which is important for their potential clinical application, their serum stability was

determined (Table 8). The stability of the IT H22(scFv)-ETA’ has been evaluated before, whereby after

24 h 77 % of the original binding ability was retained after incubation in mouse serum [211]. This is

10 % more than determined for the CFPs after 24 h in patient serum and corresponds to the

determined stability of the CFPs after 6 h. Since H22(scFv) internalizes very fast and its affinity is very

high (Kd = 2.1 nM (3.6.2)) [132], this was sufficient (see also 1.6 and 4.4).

4.5.2 Specific cytotoxic efficacy of CD64-specific fusion proteins on primary CMML and

AMML cells

As indicated above, the characterization of CFPs using primary cells has some drawbacks (4.5). Firstly,

patients have a different clinical history and samples of patients whose treatment has not started yet

are obtainable just once. Secondly, in the present case it has to be taken into account that CMML

and AMML are rare diseases limiting the number of available patients and that the cell number

within their blood samples is irregular. These facts excluded a reproduction of the obtained results

using the cells from the individual patient or direct comparison of the investigations. For these

reasons, as many evaluation techniques as possible were used to obtain as much information as

possible and to increase reliability and elucidate specificity.

The standard XTT and Annexin V measurements after treatment with the H22(scFv)-based constructs

confirmed cell death of up to 50 % even though the results differ slightly; with the XTT-based data

showing higher toxic effects than observed during flow cytometry (Figure 23). This difference might

be due to the fact that classical cell proliferation assays do not discriminate between apoptotic and

necrotic cells as is the case for apoptosis assays. Hence, necrotic cells and very small cell fragments

could skew flow cytometric measurements. Additionally, dose-depending viability assays showed a

typical sigmoidal course despite the fact that still high overall viability remained (around 50 %). The

measured cytotoxic impact and especially the high overall remaining viability after incubation of the

primary cells with the CFPs were in contrast to the results obtained using homogeneous cell lines.

This is probably due to the heterogeneous nature of the primary cell population since it contains a

variable proportion of CD64-positive target cells together with many others; and the applied

measurement techniques are influenced by those non-target cells. This phenomenon was also

described previously in cytotoxicity studies conducted on primary AML cells [208,209,212]. The

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

104

enrichment of those cells, e. g. by magnetic cell sorting using CD64-specific antibodies, would have

occupied the receptors and interfered with the binding and internalization of the fusion protein.

It was observed that the percentage of dead cells did not correlate with the initially detected

proportion of CD64 cells within the cell population. This discrepancy is due to the fact that the

amount of the CD64 receptors is not related to the sensitivity of the cells. It is rather their activation

state, influenced by mediators like IFNγ as used in the current tests (1.6). Thus, if the CD64-

expressing cells are not activated, they are not affected by the immunotherapeutic. This is also

important for later clinical studies with CD64-specific CFPs since thereby it can be avoided that

healthy CD64-positive cells are also killed [213]. Controls using the antibody alone and also the

inactive granzyme B variant GbS184A-H22(scFv) [214] confirmed that cell death was induced by the

effector domain and not just by receptor binding and internalization. Therefore, the proof-of-concept

of the immunotherapeutics tested here could successfully be demonstrated.

Determinations regarding the specificity of primary cell death were performed by four different

approaches. First, using a competitive viability assay the antigen-specificity of cell death could be

confirmed (3.6.4). Hence, it could be concluded that even in a mixture of different cells only the

target cells would be specifically targeted. Second, efficient CD64-depending internalization into

primary cells was shown by confocal microscopy (Figure 22B). Third, non-specific effects by

granzyme B were excluded by incubation of the cells with the non-binding control Gb-Ki4(scFv).

Fourth, based on co-staining experiments of the heterogeneous cell population (4.5.1), the selective

reduction of the target cells was determined via flow cytometry using specific antibodies. The parallel

staining with Annexin V and a target cell-specific antibody to specifically determine apoptosis of the

targeted cells was only possible in one case (Figure 26A), likely due to the loss of receptor expression

by apoptotic cells. Thus, many apoptotic cells would have been omitted during evaluation of the

data. The same was seen during preliminary in vitro assays as well, where co-staining with Annexin V

and CD56 of the CD56/CD64-positive cells and subsequent gating of CD56-positive cells prior to

Annexin V-based evaluation significantly differed from the results obtained without prior gating (data

not shown). As an alternative the position of the CD64-positive target cell population in FSC/ SSC

dotplot has been identified so that the decrease of this population could be estimated after

treatment of the cells (3.9.5). Because the obtained results correlated reproducibly with the results

from viability and apoptosis assays (Figure 23 and Figure 25), this method was verified to be a

reliable measure for specific cell death as was proposed earlier [212]. Another measure for specific

cell death was the reduction of the target population identified and stained with an adequate

antibody (shown for CD14-staining in Figure 26B) which has been chosen according to phenotyping

(Figure 21). Since the incidence of blocking by the H22(scFv)-based constructs was observed on HL60

in pre-tests, the use of CD64-specific antibodies was denied. In summary, the results obtained here

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

105

provide clear evidence that CMML and AMML cells can both be selectively killed by granzyme B-

based CFPs.

4.5.3 The role of PI-9 and the potential of the determined cytolytic fusion proteins for

personalized medicine

To correlate the cytotoxic effects of the new PI-9 resistant granzyme B variant GbR201K-H22(scFv) on

primary cells with the expression of the inhibitor, endogenous PI-9 levels were investigated

immediately after isolation from peripheral blood and also after cultivation for different time

intervals by western blotting. The reason for the observed upregulation of PI-9 in cell culture after

day one is not known (Figure 20), but has been observed before in the AML cell line U937 [180].

Since the expression of PI-9 in lymphocytes and monocytes of PBMCs derived from normal blood has

been described before [72] and was determined here as well (data not shown), it could not be

explicitly concluded if the detected PI-9 expression is an escape mechanism in CMML and AMML to

protect against CTL attack, or the usual PI-9 level in PBMCs. The previously found mild up-regulation

of PI-9 by IFNγ in primary monocytes after in vitro cultivation could not be corroborated here (Figure

20B), however this might be due to the applied different detection method based on flow cytometry

[72]. Nevertheless, the detection of PI-9 within the target cells potentially diminishes the efficacy of

wild-type granzyme B based immunotherapy. Therefore these cells also represented a potential ex

vivo test system to evaluate the new granzyme B mutant R201K, especially in comparison with the

granzyme M-based construct Gm-H22(scFv) which will further be elucidated in 4.6.1. Indeed a

correlation between PI-9 expression and cytotoxic efficacy of the mutant was observed. In

comparison to wild-type Gb-H22(scFv) a significantly higher cytotoxic activity of GbR201K-H22(scFv)

was observed against cells with a distinct PI-9 expression. As observed during the in vitro studies on

cell lines, no difference between the effect of the mutant and its wild-type occurred in cells not

expressing the granzyme B inhibitor. However, to state that PI-9 expression is a hallmark of CMML or

AMML in vivo, further investigations are required involving a higher number of patients and a more

sensitive and in particular more quantitative detection method such as qPCR. The PI-9 expression

level might then be an important marker for patient stratification in terms of personalized therapies

(see below).

Unexpectedly and in contrast to results obtained on cell lines, lower cytotoxic effects by

H22(scFv)-ETA’ were observed, which acts independently of PI-9, compared to Gb-H22(scFv) and

GbR201K-H22(scFv) ex vivo, even though the bacterial toxin-based IT has been proven to be highly

effective before in preclinical in vitro and in vivo studies [211]. The reason might be that, as described

previously for the similarly operating DT [215], cells were rather resistant to ETA’ (1.2.1) due to the

cells’ lack in proliferation ex vivo. This is especially characteristic for CMML, where even in vivo cells

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

106

exhibit low cell cycle activity. In addition, this finding and also the fact that compared to wild-type

Gb-H22(scFv) the other two CFPs GbR201K-H22(scFv) and Gm-H22(scFv) were partly more effective,

might indicate that ex vivo studies can be used as a predictive screening method to determine

optimal sensitivity of malignant cells from specific patients. This is one of the long-term aims in

personalized medicine: providing the correct drug to the right patient with the appropriate dose.

Over the past decade, the proof of this concept has been reported for imatinib, a competitive

tyrosine-kinase inhibitor, which has dramatically improved the prognosis of CML patients with BCR–

ABL translocation (‘breakpoint cluster region’-‘Abelson’, chromosomal defect in Philadelphia

chromosome) [216].

In summary, three major conclusions can be drawn from the here presented data: First, due to the

co-expression with several CMML specific markers like CD56, CD33 and CD14, CD64 is a potential

specific target for the treatment of CMML. Second, the applied CD64-targeting CFPs kill the target

cells in a selective manner whereby the PI-9 independent granzyme B mutant R201K is as active as its

wild-type on PI-9-negative cells but is superior if the target cells have up-regulated the inhibitor.

Third, the applied CFPs are superior in cell death-induction compared to the bacterial based IT

H22(scFv)-ETA’.

4.6 APPLICATION OF GRANZYME M AS AN ALTERNATIVE EFFECTOR MOLECULE IN

CYTOLYTIC FUSION PROTEINS

As described earlier (1.2.2), research on targeted immunotherapy focuses more and more on human

effector molecules fulfilling the cell-death inducing function instead of bacterial or plant toxins.

Proteins of the granzyme family are most promising candidates for immunotherapeutic approaches

and the successful application of granzyme B fusion proteins in immunotherapeutic approaches has

been discussed and investigated extensively in this study. Depending on treatment history and

developed, specific anti-apoptotic responses (1.4), it will be advantageous for individualized medicine

to extend the portfolio of highly efficient human effector molecules for immunotherapeutic

approaches. Hence, an alternative serine proteinase, granzyme M, was evaluated in this study for its

use as effector molecule in human CFPs. The cytotoxic potential of granzyme M has already been

confirmed independently by several groups (1.3.2) [49-55]. Its ability to kill tumor cells has previously

been described in vivo, whereby it was demonstrated that adoptively transferred NK cells from

granzyme M deficient mice were unable to inhibit tumor growth compared to wild-type NK cells [48].

In addition, the prospective of granzyme M is especially supported by the fact that it induces

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

107

apoptosis via distinct pathways compared to granzyme B and that it is not inhibited by PI-9, but

instead is able to inactivate the inhibitor (1.3.2).

Here, granzyme M was fused to two different well-established cancer cell-targeting antibody

fragments, H22(scFv) and Ki4(scFv). All constructs were successfully expressed secretory in

mammalian HEK293T cells to ensure correct post-translational modification, in contrast to other

groups which used yeast as an expression system. After purification higher band at 150 kDa was

observed during SDS-PAGE and western blot analysis (Figure 27B). Such a band was detected for the

granzyme B constructs as well, but since it was much weaker, it was not further investigated. In this

case, mass spectrometric analysis (performed in context of a master thesis) identified an additional

protein apart from granzyme M within this band, called α-2-macroglobulin precursor [217]. This is a

plasma protein, which is synthesized mainly in the liver, but also locally by macrophages, fibroblasts,

and adrenocortical cells; it is also a component of the medium supplement FCS and of human serum.

After secretory expression in HEK293T cells cultivated in synthetic, complete serum-free medium

(Panserin293S, PAN Biotech GmbH) this band disappeared in the purified sample (data not shown),

which confirmed the assumption that granzyme M complexes with α-2-macroglobulin from FCS. The

major function of α-2-macroglobulin is the inhibition of proteinases. It does not inactivate the

proteinase, but hinders the access of large substrates to its active site. This is accomplished via

cleavage of a so called ‘bait’ region in the structure of α-2-macroglobulin, located near the middle of

the polypeptide chain, which triggers a conformational change and consequent entrapment of the

proteinase [218]. An inhibition, however, was not observed in this study, as the constructs were still

highly enzymatically active as well as cytotoxic towards the target cells even if, according to the SDS-

PAGE, the complete fusion protein was trapped (only higher band visible and only weak band at

expected height). It could be the fact that the complex, whose half life has been described to be only

a few minutes [219], disintegrates before granzyme M is internalized into the targeted cells. This

suggests an advantage for later use in humans, as by complexing the granzyme M-based

immunotherapeutic with the human serum component α-2-macroglobulin, its serum stability could

potentially be increased as the plasma protein acts as a carrier after i. v. injection of the

immunotherapeutic. This, however, has to be further investigated to exclude unspecific binding,

internalization, and killing of non-target cells as the α-2-macroglobulin receptor is expressed in many

tissues and cells, especially in fibroblasts and hepatocytes [219].

4.6.1 Efficacy of granzyme M based CFPs

All generated and purified constructs were enzymatically active after purification as proved by

cleavage of recombinant PI-9 and a synthetic substrate (Figure 28) and bound specifically to their

respective target cells (Figure 29). Subsequently a cell-specific cytotoxic activity could be attested for

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

108

all constructs, whereas control cell lines remained unaffected (Figure 30 and Figure 31). To evaluate

the potential of granzyme M-based CFPs their cytotoxic effects were compared with the equivalent

constructs based on granzyme B. Thereby, it could be demonstrated that Gm-H22(scFv) triggers

comparable cytotoxic and apoptotic effects as Gb-H22(scFv) (Figure 12 and Figure 30A,C) in a specific

manner (Figure 30B). The same was true when granzyme M was fused to the epidermal growth

factor receptor (EGFR)-specific antibody 425(scFv), which has previously been fused to ETA' and the

efficacy of this construct has been demonstrated in vitro and in vivo [150,220]. Again Gm-425(scFv)

was as cytotoxic active as Gb-425(scFv) during viability as well as in apoptosis assays using the EGFR-

positive, PI-9-negative cell line A431 (data not shown). These results indicated that granzyme M has

indeed high potential as a novel effector molecule in CFPs.

Differences were observed for CD30-specific constructs. Here, 100x lower cytotoxic efficacy, was

observed for Gm-Ki4(scFv), which was in correlation with repeatedly observed lower apoptotic

effects during Annexin V/ PI assays (3.10.4). During later studies in context of a Master thesis,

however, the performance of both constructs was comparable, so the differences detected here

might be due to the protein preparation used. The observations on L540cy were independent of PI-9.

In presence of PI-9, Gm-Ki4(scFv) was advantageous regarding cell death induction compared to wild-

type Gb-Ki4(scFv) (Figure 31). Comparisons with GbR201K-Ki4(scFv) were not performed here due to

time restrictions. Even though the detected effects were low (compare explanations in 4.3.2), it still

suggested that granzyme M could be active despite of the ability of the target cells to express PI-9.

To further evaluate these findings, the cytotoxic potential of the CD64-targeting CFPs based on

granzyme M, granzyme B as well as granzyme B R201K, was comparatively tested on PI-9-positive

primary leukemic cells ex vivo (4.5.2). Hereby, due to a limited availability of patient material and the

delayed production of Gm-H22(scFv), only the last three patients could be considered, of which only

one was PI-9-positive. The obstacles which have to be overcome when working with primary cells

have already been discussed earlier (4.5). As shown for granzyme B constructs, Gm-H22(scFv) was

also able to bind CD64-positive AMML and CMML primary cells and selectively kill them (3.10.5).

When comparing the effects induced by Gm-H22(scFv), Gb-H22(scFv) and GbR201K-H22(scFv) in the

presence of PI-9, it became obvious that the two new constructs are both advantageous compared to

the wild-type Gb-H22(scFv) (Figure 32). Hence, as indicated on CD30-positive cell lines, the

insensitivity of granzyme M-based CFPs towards PI-9 could be confirmed.

In summary, apart from the verified ability to specifically kill cancerous cells in vitro and ex vivo,

granzyme M is obviously able to induce apoptosis in the presence of PI-9 to the same extent as the

new granzyme B mutant, whereas no difference in cytotoxicity between the granzyme B-based and

granzyme M-based constructs was seen in cells lacking PI-9. Consequently, granzyme M has the

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

109

potential to serve either as an alternative to granzyme B or to improve the therapeutic efficiency of

wild-type granzyme B by combinatorial immunotherapeutic treatment of PI-9 positive tumor entities.

These aspects are further discussed in the following section.

4.6.2 Clinical potential of granzyme M as effector molecule in cytolytic fusion proteins

Based on the above described findings, granzyme M can extent the palette of available human pro-

apoptotic compounds for CFPs, which can be of great importance in the context of immunotherapy

and especially for personalized medicine. Hereby, most effective cytolytic moieties and also best

binding ligands need to be selected during pre-tests to achieve optimal therapeutic outcome for the

patient. Preliminary indications for this have already been reported here during ex vivo studies by the

application of three different CFPs on primary patient material (Figure 32A,B). In addition,

granzyme M cannot be just valuable as such but also in combination with other immunotherapeutics,

especially granzyme B-based constructs. For the latter there are two rationales:

First, combination therapy with both granzymes has significant potential as these enzymes trigger

different apoptotic pathways. Details on granzyme B and granzyme M substrates have been

introduced in 1.3. The differences reflect the sequence specificity of each enzyme, i.e. methionine or

leucine for granzyme M [43] and aspartate or glutamate for granzyme B [34]. Hereby it is evident

that granzyme M is unique among the granzyme family, due to its specific expression by NK cells,

which may have phylogenetically evolved early to serve a specialized function in innate immunity

[45]. Furthermore it has been suggested that there are still unidentified substrates, which, in the

presence of the protease, transiently induce a functional active-site conformation and leads to even

more pathways for granzyme M to induce cell death [221]. Taken together, a combination of

granzyme B and granzyme M may increase the cytotoxicity by synergistic effects and reduce the

likelihood that cells can evolve resistance, e. g. by overexpression of particular inhibitors. Hereby, it

would be more effective to use different target receptors for each enzyme, to prevent competition

or steric hindrance of cell surface binding sites. Such combinations have been tested during a Master

thesis related to this work in vitro on L540cy cells, which express both HL-specific receptors CD30 and

CD25. These can be targeted by the already evaluated ligands Ki4(scFv) and RFT5(scFv) [222]

respectively. Results after combination treatment with Gm-Ki4(scFv) and GbR201K-RFT5n(scFv) or

GbR201K-Ki4(scFv) and GbR201K-RFT5n(scFv) indeed suggest synergistic effects of granzyme B and

granzyme M.

Second, the described ability of granzyme M to cleave PI-9 [56] holds potential for a combination

approach, since thereby it could prevent inhibition of granzyme B. The ability of the granzyme M

constructs to directly cleave PI-9 was only demonstrated in vitro using recombinant PI-9 during this

work (Figure 28A). Corresponding determinations should be done in the future by either incubating

Discussion ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

110

native cell lysates of PI-9-positive cells with a granzyme M based CFP or by cell lysis and subsequent

western blot analysis of those cells after treatment with a corresponding construct. In fact, the high

potential of this approach is supported by the finding, that the interaction between PI-9 and

granzyme M has a high stoichiometry of inhibition (SI) value of 63, which means that PI-9 is a much

better substrate than inhibitor of granzyme B [56]. By cleaving PI-9, granzyme M could therefore

clear the way for apoptosis induced by granzyme B. Hence, granzyme M could be useful, not only in a

therapeutic context, especially against treatment-resistant cells on its own, but also in combination

with granzyme B.

In contrast to granzyme B, granzyme M has not been investigated as extensively, so in the future

both promising, as well as adverse effects may be detected on its prospective role as

immunotherapeutic. Known so far is e.g. that on the down side, as described for granzyme B, a

specific endogenous inhibitor, called serpin B4, has recently been suggested to form a typical SDS-

stable serpin–protease complex with granzyme M in vitro and thereby inhibited NK cell-mediated cell

death in Hela cells [223]. However, the affinity between the two proteins is of a second-order rate

constant of 1.3 x 104 M-1 s-1 which is two orders of magnitude below that of granzyme B and PI-9 and,

in general, below the range of biologically-relevant interactions. This rather weak inhibition might be

overcome by CFP-related excessive delivery of the effector molecule to the cells. Furthermore, the in

vivo expression of this inhibitor has not been reported and any potential escape mechanisms in

tumors remain to be investigated. Other serpins present in plasma, like α1-antichymotrypsin and α1-

proteinase inhibitor, could also become a limiting factor for granzyme M activity in blood as

suggested earlier [56]. Their inhibitory relevance has to be evaluated in the future in serum-stability

assays, and respective modifications, as shown in this study for granzyme B, should be considered.

Other modifications, also in accordance to granzyme B, might involve the cationic sites within

granzyme M that may lead to off-target effects [221]. Hence, current literature on this topic should

be monitored carefully, as in vivo studies should be pursued to further evaluate the potential of

granzyme M in immunotherapeutic approaches.

Concluding, with the novel granzyme B mutant, identified during in silico homology modeling, it was

demonstrated on in vitro, in vivo as well as ex vivo level that residual therapy-resistant cancer cells

can be killed, provided the resistance is due to PI-9 expression as an escape mechanism. Additionally,

the cytotoxic potential and the PI-9 insensitivity of granzyme M-based fusion proteins was verified in

vitro and ex vivo clearing the way for a potential combination treatment with both granzymes.

Outlook ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

111

5 Outlook

The presented data emphasize the great efficacy of granzyme B-based cytolytic fusion proteins (CFPs)

in vitro, in vivo as well as ex vivo and especially verify the necessity and potential of directed

modifications, here to pave the way for inhibitor-insensitive treatment. On this basis further

investigations and optimizations are now ongoing in our group concerning the structure of the

corresponding CFPs as well as their extended application.

Structural modifications include the introduction of an additional modification to reduce the high

positive surface charge of granzyme B (pI ~ 10) which is mainly caused by two major heparan-binding

motifs and might result in off-target effects by intrinsic binding to heparin sulfate proteoglycans. In a

previous study, the basic amino acid residues involved in heparan sulfate binding were substituted

for alanine which did not interfere with ligand binding or enzyme activity and reduced the negative

impact of extracellular matrix effects [224]. Hence, specific cytotoxicity of granzyme B could be

increased by using this granzyme B variant. Furthermore, the improvement of endosomal release

after receptor mediated endocytosis within the target cells by certain adapter sequences is now

being investigated in-house (confidential) which suggest that the engineering of granzyme B-based

CFPs by the insertion of functional adapter elements will improve their therapeutic potential. As an

alternative, a pH-sensitive fusogenic peptide has been introduced recently which increased specific

cytotoxicity compared with the construct lacking the peptide, and demonstrated enhanced cellular

uptake and improved delivery of granzyme B to the cytosol of target cells [166]. In addition, another

promising mutational site is presently evaluated in our group which was also suggested from in silico

determinations to be promising to render granzyme B resistant to PI-9 inhibition. The potential of

structure related modifications of the human molecules, however, introduce novel sequences with

potential antigenicity in vivo which have to be evaluated carefully.

Further structural related improvements which are required for translation into clinical application to

reduce potential immunogenicity include either removal of the His6-tag after IMAC purification or

applying alternative purification strategies like ion exchange chromatography independent of any

tags. In addition, the applied murine Ki4(scFv) should either be humanized or replaced by the human

CD30L, which has been shown promising results in combination with human RNase angiogenin

before [187] to prevent HAMA response.

Regarding the extended application of the CFPs, the economic and up-scaled production of active

granzyme B and granzyme M has to be considered. Therefore the yeast Hansenula polymorpha is

now under investigation as a potential alternative expression system. It harbors the additional

advantage to become independent of expensive enterokinase processing of the granzymes by

insertion of the mating factor α prepro-leader sequence allowing straightforward purification of

Outlook ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

112

active granzyme from the culture supernatant based on local Kex2-protease activity in the trans-

Golgi compartment. Furthermore, expression of E. coli is presently explored to increase protein yield

with lower expenses.

The described ex vivo findings need to be verified on additional patient material and evaluated in

terms of personalized medicine. Those investigations will be most valuable in combination with

information on the previous treatment and anamnesis. To further evaluate CD64 as valuable

potential therapeutic target also for CMML and AMML in the clinical it will now be investigated if

CD64 is indeed over-expressed on bone marrow samples of AMML and CMML patients. A

corresponding set-up for immunohistochemical determinations with H22 antibody fragments is

already established.

In addition, further in vivo experiments are planned with the described constructs which include the

establishment of a disseminated tumor model using the AML cell line HL60. The same might be

indicated for Hodgkin lymphoma since for this disease a disseminated model is less artificial than the

subcutaneous tumor model applied in this study. Furthermore, the novel variant GbR201K is now

used in combination with several other binders to kill target cells despite the presence of

endogenous PI-9.

The CFP-based treatment of tumor cells, even in presence of PI-9, was shown to be feasible using the

new effector molecule granzyme M. Further combination experiments are planned to test the

potential of granzyme M to clear the way for granzyme B activity by inactivation of PI-9 within the

target cells. These studies will be performed as soon as an adequate PI-9-positive cell line with both

receptors co-expressed becomes available. In addition, the potential of granzyme M-based CFPs has

to be shown in vivo as well using the xenograft tumor models already established for the

granzyme B-based constructs. For both constructs the species-dependent substrate-specificity has to

be taken into account when these enzymes undergo further studies using other than xenograft

tumor models or the in our groups well established mouse inflammation model [213], especially

when proceeding towards clinical application. These specificities can significantly obscure e.g.

immunogenic or unspecific effects, so special care has to be taken when planning pre-clinical studies.

In accordance to granzyme B, novel variants for granzyme M independent of the plasma serpins α1-

antichymotrypsin, α1-proteinase inhibitor and the endogenous inhibitor serpin B4, if detected in vivo

as well, can also be established. Additionally, a granzyme M variant, which does not bind to α-2-

macroglobulin anymore, is planned to prevent its entrapment in human serum.

The here developed novel constructs may eventually be used in the clinic in combination with

conventional treatment approaches helping to reduce the necessary dose of non-specific drugs,

avoiding adverse effects and promoting long-term remission.

Summary ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

113

6 Summary

Current cancer treatments often lack specificity and therefore also kill healthy cells, resulting in

severe side effects. To improve the specificity of cancer treatment, recombinant chimeric proteins

such as immunotoxins have been developed, consisting of a tumor-selective binding domain and a

cytotoxic protein primarily of bacterial or plant origin. To circumvent immunogenic responses, recent

studies have focused on the development of fully human cytolytic fusion proteins (CFPs), containing

pro-apoptotic components of human origin. Most promising are enzymes from granule-associated

immune defense. The best studied is the serine protease granzyme B. However, its anti-tumor

efficacy is limited by irreversible inhibition by serpin B9 (PI-9) which is naturally co-expressed in the

cytosol of cytotoxic immune cells to avoid auto-toxicity by misdirected granzyme B. Several tumors

have been described to overexpress PI-9 as a mechanism for immune escape. As a very high number

of PI-9-positive tumor cells predicts unfavorable clinical outcome, the specific killing of these often

treatment-resistant cells may strongly increase therapeutic success and life span.

In this study different cell lines were found to express PI-9, including CD30-positive HL cell lines or

primary leukemic cells, which decreased the cytotoxic efficacy of granzyme B-based CFPs during

former studies. The aim of this work was therefore to improve the therapeutic potential of

granzyme B by the generation of PI-9 resistant mutants. Therefore the interacting amino acids

between granzyme B and PI-9 were elucidated by in silico simulations based on the known crystal

structure of granzyme B as well as a complex between rat trypsin and Manduca sexta serpin B1. Two

corresponding amino acids, R28 and R201, were identified and exchanged to Ala, Lys or Glu

respectively, by site-directed mutagenesis. The resulting eight mutants, including a double mutant

(R28A-R201A) and a published mutant (K27A), were further tested in vitro with regard to an

increased resistance against PI-9 inhibition. The new mutants were fused to H22(scFv) and Ki4(scFv)

which bind CD64 (high-affinity receptor for IgG, over-expressed e. g. on AML cells) and CD30 (cell

membrane protein of the tumor necrosis factor receptor family, over-expressed e. g. on HL cells)

respectively. Proteins were expressed in HEK293T cells and purified via affinity chromatography from

the supernatant. Three mutants were identified by enzymatic activity assays to be as active as the

wild-type in absence of PI-9 and especially retained most of their proteolytic activity after challenge

with recombinant PI-9. These mutants were comparable to the corresponding wild-type constructs in

respect to binding affinity (between 1.5 and 5.7 nM), cytotoxic and apoptotic activity (IC50 between

1.2 and 5.3 nM) as well as serum stability (6 - 10 h).

The mutant R201K demonstrated the best cytotoxic activity against both CD30-expressing PI-9-

positive and -negative tumor cells in vitro. To evaluate the influence of the mutation on the biological

activity appropriate xenograft tumor models were set up. For HL, only PI-9-negative L540cy cells are

Summary ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

114

currently established. In addition, a novel tumor model with the PI-9-expressing HL cell line L428 was

established here based on optical in vivo imaging. In these proof-of-principle models, it could be

verified in vivo that only the mutant is able to kill cells both in the presence and absence of the

inhibitor, this in contrast to the wild-type which only killed PI-9-negative L540cy tumor cells in vivo.

This is analogous to ex vivo studies with primary cells from patients with the rare diseases CMML and

AMML. These cells were targeted via the CD64 receptor. The data additionally demonstrated that the

new constructs perform even better than H22(scFv)-ETA’. The differences in apoptotic efficacy

between the applied constructs suggest ex vivo studies to be a predictive screening method for

personalized medicine to enable a patient-specific treatment selection.

In addition, a novel human granzyme-based CFP with a pro-apoptotic moiety based on granzyme M is

firstly described here. This serine protease is known to be able to specifically cleave and inactivate

PI-9. It may be a valuable addition to the palette of human effector molecules, in particular for the

development of personalized medicine in which the optimal cytolytic protein and targeting molecule

combination can be selected on a per-patient basis. Different fusion constructs, comprising

granzyme M targeting CD64- and CD30-positive cancer cell lines were generated, expressed and

purified analogous to the granzyme B-based constructs. During in vitro and ex vivo studies

granzyme M was proved to be a potent human anti-tumor agent (IC50 between 1.2 and 6.4 nM for

Gm-H22(scFv)), even in presence of PI-9. CFPs combining granzyme B and granzyme M offer potential

for improved therapeutic approaches by simultaneously triggering different apoptotic pathways to

circumvent possible overexpression of inhibitors within the target cell or by potentiating granzyme B

activity by inactivating PI-9.

In summary, during this thesis two new effector molecules, granzyme M and a PI-9 resistant variant

of granzyme B, granzyme B R201K, have successfully been generated, produced and evaluated in

vitro, ex vivo and in parts also in vivo in regard to their cytotoxic potential and for use in

immunotherapy to treat leukemic diseases or HL. These effector molecules, either individually or in

combination, also bear clinical relevance for the treatment of solid tumors or other hematological

disorders that potentially express PI-9 or selected PI-9-positive cells under immune surveillance or

tumor therapy.

Literature ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

115

7 Literature

1. HANAHAN, D. and WEINBERG, R.A., The hallmarks of cancer. Cell, 2000. 100(1): p. 57-70. 2. HANAHAN, D. and COUSSENS, L.M., Accessories to the crime: functions of cells recruited to the tumor

microenvironment. Cancer Cell, 2012. 21(3): p. 309-22. 3. HILLEN, T., ENDERLING, H., and HAHNFELDT, P., The tumor growth paradox and immune system-mediated

selection for cancer stem cells. Bull Math Biol, 2013. 75(1): p. 161-84. 4. HEHMANN-TITT, G., SCHIFFER, S., BERGES, N., MELMER, G., and BARTH, S., Improving the Therapeutic Potential of

Human Granzyme B for Targeted Cancer Therapy. Antibodies, 2013. 2(1): p. 19-49. 5. KLEBANOFF, C.A., ACQUAVELLA, N., YU, Z.Y., and RESTIFO, N.P., Therapeutic cancer vaccines: are we there yet?

Immunological Reviews, 2011. 239: p. 27-44. 6. KREITMAN, R.J., Immunotoxins for targeted cancer therapy. AAPS J, 2006. 8(3): p. 532-51. 7. VANNEMAN, M. and DRANOFF, G., Combining immunotherapy and targeted therapies in cancer treatment. Nat

Rev Cancer, 2012. 12(4): p. 237-51. 8. WALDMANN, T.A., Immunotherapy: past, present and future. Nat Med, 2003. 9(3): p. 269-77. 9. SCOTT, A.M., ALLISON, J.P., and WOLCHOK, J.D., Monoclonal antibodies in cancer therapy. Cancer Immun, 2012.

12: p. 14. 10. REICHERT, J.M., Antibody-based therapeutics to watch in 2011. MAbs, 2011. 3(1): p. 76-99. 11. SCHRAMA, D., REISFELD, R.A., and BECKER, J.C., Antibody targeted drugs as cancer therapeutics. Nat Rev Drug

Discov, 2006. 5(2): p. 147-59. 12. CARTER, P., Improving the efficacy of antibody-based cancer therapies. Nat Rev Cancer, 2001. 1(2): p. 118-29. 13. RYBAK, S.M., HOOGENBOOM, H.R., MEADE, H.M., RAUS, J.C., SCHWARTZ, D., and YOULE, R.J., Humanization of

immunotoxins. Proc Natl Acad Sci U S A, 1992. 89(8): p. 3165-9. 14. NATSUME, A., NIWA, R., and SATOH, M., Improving effector functions of antibodies for cancer treatment:

Enhancing ADCC and CDC. Drug Des Devel Ther, 2009. 3: p. 7-16. 15. CARTER, P.J. and SENTER, P.D., Antibody-drug conjugates for cancer therapy. Cancer Journal, 2008. 14(3): p. 154-

69. 16. LAMBERT, J.M., Drug-conjugated monoclonal antibodies for the treatment of cancer. Curr Opin Pharmacol, 2005.

5(5): p. 543-9. 17. YOUNES, A., YASOTHAN, U., and KIRKPATRICK, P., Brentuximab vedotin. Nat Rev Drug Discov, 2012. 11(1): p. 19-

20. 18. TRAYNOR, K., Ado-trastuzumab emtansine approved for advanced breast cancer. Am J Health Syst Pharm, 2013.

70(7): p. 562. 19. BROSS, P.F., BEITZ, J., CHEN, G., CHEN, X.H., DUFFY, E., KIEFFER, L., et al., Approval summary: gemtuzumab

ozogamicin in relapsed acute myeloid leukemia. Clin Cancer Res, 2001. 7(6): p. 1490-6. 20. KREITMAN, R.J., Recombinant toxins for the treatment of cancer. Curr Opin Mol Ther, 2003. 5(1): p. 44-51. 21. ALLEN, T.M., Ligand-targeted therapeutics in anticancer therapy. Nat Rev Cancer, 2002. 2(10): p. 750-63. 22. PASTAN, I., HASSAN, R., FITZGERALD, D.J., and KREITMAN, R.J., Immunotoxin treatment of cancer. Annu Rev Med,

2007. 58: p. 221-37. 23. WINTER, G., GRIFFITHS, A.D., HAWKINS, R.E., and HOOGENBOOM, H.R., Making antibodies by phage display

technology. Annu Rev Immunol, 1994. 12: p. 433-55. 24. MADHUMATHI, J. and VERMA, R.S., Therapeutic targets and recent advances in protein immunotoxins. Curr Opin

Microbiol, 2012. 15(3): p. 300-9. 25. DANG, N.H., PRO, B., HAGEMEISTER, F.B., SAMANIEGO, F., JONES, D., SAMUELS, B.I., et al., Phase II trial of

denileukin diftitox for relapsed/refractory T-cell non-Hodgkin lymphoma. Br J Haematol, 2007. 136(3): p. 439-47. 26. HWANG, W.Y. and FOOTE, J., Immunogenicity of engineered antibodies. Methods, 2005. 36(1): p. 3-10. 27. JONES, P.T., DEAR, P.H., FOOTE, J., NEUBERGER, M.S., and WINTER, G., Replacing the complementarity-

determining regions in a human antibody with those from a mouse. Nature, 1986. 321(6069): p. 522-5. 28. MENORET, S., ISCACHE, A.L., TESSON, L., REMY, S., USAL, C., OSBORN, M.J., et al., Characterization of

immunoglobulin heavy chain knockout rats. Eur J Immunol, 2010. 40(10): p. 2932-41. 29. SIEGALL, C.B., HAGGERTY, H.G., WARNER, G.L., CHACE, D., MIXAN, B., LINSLEY, P.S., et al., Prevention of

immunotoxin-induced immunogenicity by coadministration with CTLA4Ig enhances antitumor efficacy. J Immunol, 1997. 159(10): p. 5168-73.

30. TSUTSUMI, Y., ONDA, M., NAGATA, S., LEE, B., KREITMAN, R.J., and PASTAN, I., Site-specific chemical modification

with polyethylene glycol of recombinant immunotoxin anti-Tac(Fv)-PE38 (LMB-2) improves antitumor activity and

reduces animal toxicity and immunogenicity. Proc Natl Acad Sci U S A, 2000. 97(15): p. 8548-53. 31. MATHEW, M. and VERMA, R.S., Humanized immunotoxins: a new generation of immunotoxins for targeted cancer

therapy. Cancer Sci, 2009. 100(8): p. 1359-65. 32. BOTS, M. and MEDEMA, J.P., Granzymes at a glance. J Cell Sci, 2006. 119(Pt 24): p. 5011-4. 33. DE POOT, S.A.H., WESTGEEST, M., HOSTETTER, D.R., VAN DAMME, P., PLASMAN, K., DEMEYER, K., et al., Human

and mouse granzyme M display divergent and species-specific substrate specificities. Biochem J 2011. 437: p. 431-42.

Literature ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

116

34. TRAPANI, J.A., Granzymes: a family of lymphocyte granule serine proteases. Genome Biol, 2001. 2(12): p. REVIEWS3014.

35. TRAPANI, J.A. and SUTTON, V.R., Granzyme B: pro-apoptotic, antiviral and antitumor functions. Curr Opin Immunol, 2003. 15(5): p. 533-43.

36. LANIER, L.L., NK cell recognition. Annu Rev Immunol, 2005. 23: p. 225-74. 37. KURSCHUS, F.C. and JENNE, D.E., Delivery and therapeutic potential of human granzyme B. Immunol Rev, 2010.

235(1): p. 159-71. 38. CHOWDHURY, D. and LIEBERMAN, J., Death by a thousand cuts: granzyme pathways of programmed cell death.

Annu Rev Immunol, 2008. 26: p. 389-420. 39. KAM, C.M., HUDIG, D., and POWERS, J.C., Granzymes (lymphocyte serine proteases): characterization with natural

and synthetic substrates and inhibitors. Biochim Biophys Acta, 2000. 1477(1-2): p. 307-23. 40. EDWARDS, K.M., DAVIS, J.E., BROWNE, K.A., SUTTON, V.R., and TRAPANI, J.A., Anti-viral strategies of cytotoxic T

lymphocytes are manifested through a variety of granule-bound pathways of apoptosis induction. Immunol Cell Biol, 1999. 77(1): p. 76-89.

41. ESTEBANEZ-PERPINA, E., FUENTES-PRIOR, P., BELORGEY, D., BRAUN, M., KIEFERSAUER, R., MASKOS, K., et al., Crystal structure of the caspase activator human granzyme B, a proteinase highly specific for an Asp-P1 residue. Biol Chem, 2000. 381(12): p. 1203-14.

42. ROTONDA, J., GARCIA-CALVO, M., BULL, H.G., GEISSLER, W.M., MCKEEVER, B.M., WILLOUGHBY, C.A., et al., The

three-dimensional structure of human granzyme B compared to caspase-3, key mediators of cell death with

cleavage specificity for aspartic acid in P1. Chem Biol, 2001. 8(4): p. 357-68. 43. SMYTH, M.J., O'CONNOR, M.D., TRAPANI, J.A., KERSHAW, M.H., and BRINKWORTH, R.I., A novel substrate-binding

pocket interaction restricts the specificity of the human NK cell-specific serine protease, Met-ase-1. J Immunol, 1996. 156(11): p. 4174-81.

44. BADE, B., BOETTCHER, H.E., LOHRMANN, J., HINK-SCHAUER, C., BRATKE, K., JENNE, D.E., et al., Differential

expression of the granzymes A, K and M and perforin in human peripheral blood lymphocytes. International Immunology, 2005. 17(11): p. 1419-28.

45. SAYERS, T.J., BROOKS, A.D., WARD, J.M., HOSHINO, T., BERE, W.E., WIEGAND, G.W., et al., The restricted

expression of granzyme M in human lymphocytes. J Immunol, 2001. 166(2): p. 765-71. 46. ANTHONY, D.A., ANDREWS, D.M., CHOW, M., WATT, S.V., HOUSE, C., AKIRA, S., et al., A role for granzyme M in

TLR4-driven inflammation and endotoxicosis. J Immunol, 2010. 185(3): p. 1794-803. 47. SUSANTO, O., TRAPANI, J.A., and BRASACCHIO, D., Controversies in granzyme biology. Tissue Antigens, 2012.

80(6): p. 477-87. 48. PEGRAM, H.J., HAYNES, N.M., SMYTH, M.J., KERSHAW, M.H., and DARCY, P.K., Characterizing the anti-tumor

function of adoptively transferred NK cells in vivo. Cancer Immunol Immunother, 2010. 59(8): p. 1235-46. 49. HU, D., LIU, S., SHI, L., LI, C., WU, L., and FAN, Z., Cleavage of survivin by Granzyme M triggers degradation of the

survivin-X-linked inhibitor of apoptosis protein (XIAP) complex to free caspase activity leading to cytolysis of target

tumor cells. J Biol Chem, 2010. 285(24): p. 18326-35. 50. HUA, G., ZHANG, Q., and FAN, Z., Heat shock protein 75 (TRAP1) antagonizes reactive oxygen species generation

and protects cells from granzyme M-mediated apoptosis. J Biol Chem, 2007. 282(28): p. 20553-60. 51. WANG, S., XIA, P., SHI, L., and FAN, Z., FADD cleavage by NK cell granzyme M enhances its self-association to

facilitate procaspase-8 recruitment for auto-processing leading to caspase cascade. Cell Death Differ, 2012. 19(4): p. 605-15.

52. LU, H., HOU, Q., ZHAO, T., ZHANG, H., ZHANG, Q., WU, L., et al., Granzyme M directly cleaves inhibitor of caspase-

activated DNase (CAD) to unleash CAD leading to DNA fragmentation. J Immunol, 2006. 177(2): p. 1171-8. 53. BOVENSCHEN, N., DE KONING, P.J., QUADIR, R., BROEKHUIZEN, R., DAMEN, J.M., FROELICH, C.J., et al., NK cell

protease granzyme M targets alpha-tubulin and disorganizes the microtubule network. J Immunol, 2008. 180(12): p. 8184-91.

54. CULLEN, S.P., AFONINA, I.S., DONADINI, R., LUTHI, A.U., MEDEMA, J.P., BIRD, P.I., et al., Nucleophosmin is cleaved

and inactivated by the cytotoxic granule protease granzyme M during natural killer cell-mediated killing. J Biol Chem, 2009. 284(8): p. 5137-47.

55. KELLY, J.M., WATERHOUSE, N.J., CRETNEY, E., BROWNE, K.A., ELLIS, S., TRAPANI, J.A., et al., Granzyme M mediates

a novel form of perforin-dependent cell death. J Biol Chem, 2004. 279(21): p. 22236-42. 56. MAHRUS, S., KISIEL, W., and CRAIK, C.S., Granzyme M is a regulatory protease that inactivates proteinase inhibitor

9, an endogenous inhibitor of granzyme B. J Biol Chem, 2004. 279(52): p. 54275-82. 57. TENG, M.W., SWANN, J.B., KOEBEL, C.M., SCHREIBER, R.D., and SMYTH, M.J., Immune-mediated dormancy: an

equilibrium with cancer. J Leukoc Biol, 2008. 84(4): p. 988-93. 58. KIM, R., EMI, M., and TANABE, K., Cancer immunoediting from immune surveillance to immune escape.

Immunology, 2007. 121(1): p. 1-14. 59. NELSON, B.H., The impact of T-cell immunity on ovarian cancer outcomes. Immunol Rev, 2008. 222: p. 101-16. 60. YANG, L., PANG, Y., and MOSES, H.L., TGF-beta and immune cells: an important regulatory axis in the tumor

microenvironment and progression. Trends Immunol, 2010. 31(6): p. 220-7. 61. SHREE, T., OLSON, O.C., ELIE, B.T., KESTER, J.C., GARFALL, A.L., SIMPSON, K., et al., Macrophages and cathepsin

proteases blunt chemotherapeutic response in breast cancer. Genes Dev, 2011. 25(23): p. 2465-79. 62. VESELY, M.D., KERSHAW, M.H., SCHREIBER, R.D., and SMYTH, M.J., Natural innate and adaptive immunity to

cancer. Annu Rev Immunol, 2011. 29: p. 235-71.

Literature ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

117

63. CLEVERS, H., The cancer stem cell: premises, promises and challenges. Nat Med, 2011. 17(3): p. 313-9. 64. BOTS, M., L, V.A.N.B., RADEMAKER, M.T., OFFRINGA, R., and MEDEMA, J.P., Serpins prevent granzyme-induced

death in a species-specific manner. Immunol Cell Biol, 2006. 84(1): p. 79-86. 65. HUNTINGTON, J.A., READ, R.J., and CARRELL, R.W., Structure of a serpin-protease complex shows inhibition by

deformation. Nature, 2000. 407(6806): p. 923-6. 66. BLADERGROEN, B.A., STRIK, M.C., BOVENSCHEN, N., VAN BERKUM, O., SCHEFFER, G.L., MEIJER, C.J., et al., The

granzyme B inhibitor, protease inhibitor 9, is mainly expressed by dendritic cells and at immune-privileged sites. J Immunol, 2001. 166(5): p. 3218-25.

67. BUZZA, M.S., HOSKING, P., and BIRD, P.I., The granzyme B inhibitor, PI-9, is differentially expressed during

placental development and up-regulated in hydatidiform moles. Placenta, 2006. 27(1): p. 62-9. 68. KANNAN-THULASIRAMAN, P. and SHAPIRO, D.J., Modulators of inflammation use nuclear factor-kappa B and

activator protein-1 sites to induce the caspase-1 and granzyme B inhibitor, proteinase inhibitor 9. J Biol Chem, 2002. 277(43): p. 41230-9.

69. EL HADDAD, N., MOORE, R., HEATHCOTE, D., MOUNAYAR, M., AZZI, J., MFARREJ, B., et al., The novel role of

SERPINB9 in cytotoxic protection of human mesenchymal stem cells. J Immunol, 2011. 187(5): p. 2252-60. 70. HEUTINCK, K.M., KASSIES, J., FLORQUIN, S., TEN BERGE, I.J., HAMANN, J., and ROWSHANI, A.T., SerpinB9

expression in human renal tubular epithelial cells is induced by triggering of the viral dsRNA sensors TLR3, MDA5

and RIG-I. Nephrol Dial Transplant, 2012. 27(7): p. 2746-54. 71. ROWSHANI, A.T., STRIK, M.C., MOLENAAR, R., YONG, S.L., WOLBINK, A.M., BEMELMAN, F.J., et al., The granzyme

B inhibitor SERPINB9 (protease inhibitor 9) circulates in blood and increases on primary cytomegalovirus infection

after renal transplantation. J Infect Dis, 2005. 192(11): p. 1908-11. 72. CLASSEN, C.F., BIRD, P.I., and DEBATIN, K.M., Modulation of the granzyme B inhibitor proteinase inhibitor 9 (PI-9)

by activation of lymphocytes and monocytes in vitro and by Epstein-Barr virus and bacterial infection. Clin Exp Immunol, 2006. 143(3): p. 534-42.

73. BUZZA, M.S., HIRST, C.E., BIRD, C.H., HOSKING, P., MCKENDRICK, J., and BIRD, P.I., The granzyme B inhibitor, PI-9,

is present in endothelial and mesothelial cells, suggesting that it protects bystander cells during immune

responses. Cell Immunol, 2001. 210(1): p. 21-9. 74. BARRIE, M.B., STOUT, H.W., ABOUGERGI, M.S., MILLER, B.C., and THIELE, D.L., Antiviral cytokines induce hepatic

expression of the granzyme B inhibitors, proteinase inhibitor 9 and serine proteinase inhibitor 6. J Immunol, 2004. 172(10): p. 6453-9.

75. HORIE, O., SAIGO, K., MURAYAMA, T., and RYO, R., Differential expression of proteinase inhibitor-9 and granzyme

B mRNAs in activated immunocompetent cells. Tohoku J Exp Med, 2005. 205(2): p. 103-13. 76. HIRST, C.E., BUZZA, M.S., BIRD, C.H., WARREN, H.S., CAMERON, P.U., ZHANG, M., et al., The intracellular granzyme

B inhibitor, proteinase inhibitor 9, is up-regulated during accessory cell maturation and effector cell degranulation,

and its overexpression enhances CTL potency. J Immunol, 2003. 170(2): p. 805-15. 77. JIANG, X., ELLISON, S.J., ALARID, E.T., and SHAPIRO, D.J., Interplay between the levels of estrogen and estrogen

receptor controls the level of the granzyme inhibitor, proteinase inhibitor 9 and susceptibility to immune

surveillance by natural killer cells. Oncogene, 2007. 26(28): p. 4106-14. 78. ROWSHANI, A.T., FLORQUIN, S., BEMELMAN, F., KUMMER, J.A., HACK, C.E., and TEN BERGE, I.J., Hyperexpression

of the granzyme B inhibitor PI-9 in human renal allografts: a potential mechanism for stable renal function in

patients with subclinical rejection. Kidney Int, 2004. 66(4): p. 1417-22. 79. WALKER, P.R., SAAS, P., and DIETRICH, P.Y., Role of Fas ligand (CD95L) in immune escape: the tumor cell strikes

back. J Immunol, 1997. 158(10): p. 4521-4. 80. MEDEMA, J.P., DE JONG, J., PELTENBURG, L.T., VERDEGAAL, E.M., GORTER, A., BRES, S.A., et al., Blockade of the

granzyme B/perforin pathway through overexpression of the serine protease inhibitor PI-9/SPI-6 constitutes a

mechanism for immune escape by tumors. Proc Natl Acad Sci U S A, 2001. 98(20): p. 11515-20. 81. GETTINS, P., PATSTON, P.A., and SCHAPIRA, M., The role of conformational change in serpin structure and

function. Bioessays, 1993. 15(7): p. 461-7. 82. SUN, J., BIRD, C.H., SUTTON, V., MCDONALD, L., COUGHLIN, P.B., DE JONG, T.A., et al., A cytosolic granzyme B

inhibitor related to the viral apoptotic regulator cytokine response modifier A is present in cytotoxic lymphocytes. J Biol Chem, 1996. 271(44): p. 27802-9.

83. POE, M., BLAKE, J.T., BOULTON, D.A., GAMMON, M., SIGAL, N.H., WU, J.K., et al., Human cytotoxic lymphocyte

granzyme B. Its purification from granules and the characterization of substrate and inhibitor specificity. J Biol Chem, 1991. 266(1): p. 98-103.

84. HEDSTROM, L., Serine protease mechanism and specificity. Chem Rev, 2002. 102(12): p. 4501-24. 85. MARSZAL, E. and SHRAKE, A., Serpin crystal structure and serpin polymer structure. Arch Biochem Biophys, 2006.

453(1): p. 123-9. 86. SUN, J., WHISSTOCK, J.C., HARRIOTT, P., WALKER, B., NOVAK, A., THOMPSON, P.E., et al., Importance of the P4'

residue in human granzyme B inhibitors and substrates revealed by scanning mutagenesis of the proteinase

inhibitor 9 reactive center loop. J Biol Chem, 2001. 276(18): p. 15177-84. 87. YE, S., CECH, A.L., BELMARES, R., BERGSTROM, R.C., TONG, Y., COREY, D.R., et al., The structure of a Michaelis

serpin-protease complex. Nat Struct Biol, 2001. 8(11): p. 979-83. 88. HODGKIN, T., On some Morbid Appearances of the Absorbent Glands and Spleen. Med Chir Trans, 1832. 17: p. 68-

114.

Literature ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

118

89. REED, W., Recent Researches concerning the Etiology, Propagation, and Prevention of Yellow Fever, by the United

States Army Commission. J Hyg (Lond), 1902. 2(2): p. 101-19. 90. STERNBERG, C., Über eine eigenartige unter dem Bilde der Pseudoleukämie verlaufende Tuberkolose des

lymphatischen Apparates. Z. Heilkunde, 1898. 19: p. 21-90. 91. KUPPERS, R., The biology of Hodgkin's lymphoma. Nat Rev Cancer, 2009. 9(1): p. 15-27. 92. KUPPERS, R., Molecular biology of Hodgkin lymphoma. Hematology Am Soc Hematol Educ Program, 2009: p. 491-

6. 93. IZBAN, K.F., ERGIN, M., HUANG, Q., QIN, J.Z., MARTINEZ, R.L., SCHNITZER, B., et al., Characterization of NF-kappaB

expression in Hodgkin's disease: inhibition of constitutively expressed NF-kappaB results in spontaneous caspase-

independent apoptosis in Hodgkin and Reed-Sternberg cells. Mod Pathol, 2001. 14(4): p. 297-310. 94. AGGARWAL, B.B., Nuclear factor-kappaB: the enemy within. Cancer Cell, 2004. 6(3): p. 203-8. 95. PAHL, H.L., Activators and target genes of Rel/NF-kappaB transcription factors. Oncogene, 1999. 18(49): p. 6853-

66. 96. KHAN, G., Epstein-Barr virus, cytokines, and inflammation: a cocktail for the pathogenesis of Hodgkin's

lymphoma? Exp Hematol, 2006. 34(4): p. 399-406. 97. BOLL, B., HANSEN, H., HEUCK, F., REINERS, K., BORCHMANN, P., ROTHE, A., et al., The fully human anti-CD30

antibody 5F11 activates NF-{kappa}B and sensitizes lymphoma cells to bortezomib-induced apoptosis. Blood, 2005. 106(5): p. 1839-42.

98. HIRSCH, B., HUMMEL, M., BENTINK, S., FOULADI, F., SPANG, R., ZOLLINGER, R., et al., CD30-induced signaling is

absent in Hodgkin's cells but present in anaplastic large cell lymphoma cells. Am J Pathol, 2008. 172(2): p. 510-20. 99. HORIE, R., HIGASHIHARA, M., and WATANABE, T., Hodgkin's lymphoma and CD30 signal transduction. Int J

Hematol, 2003. 77(1): p. 37-47. 100. HORIE, R., WATANABE, T., MORISHITA, Y., ITO, K., ISHIDA, T., KANEGAE, Y., et al., Ligand-independent signaling by

overexpressed CD30 drives NF-kappaB activation in Hodgkin-Reed-Sternberg cells. Oncogene, 2002. 21(16): p. 2493-503.

101. ANAGNOSTOPOULOS, I., HANSMANN, M.L., FRANSSILA, K., HARRIS, M., HARRIS, N.L., JAFFE, E.S., et al., European

Task Force on Lymphoma project on lymphocyte predominance Hodgkin disease: histologic and immunohistologic

analysis of submitted cases reveals 2 types of Hodgkin disease with a nodular growth pattern and abundant

lymphocytes. Blood, 2000. 96(5): p. 1889-99. 102. CHANG, K.C., KHEN, N.T., JONES, D., and SU, I.J., Epstein-Barr virus is associated with all histological subtypes of

Hodgkin lymphoma in Vietnamese children with special emphasis on the entity of lymphocyte predominance

subtype. Hum Pathol, 2005. 36(7): p. 747-55. 103. ANSELL, S.M., Brentuximab vedotin: delivering an antimitotic drug to activated lymphoma cells. Expert Opin

Investig Drugs, 2011. 20(1): p. 99-105. 104. BRENNER, H., GONDOS, A., and ARNDT, V., Recent major progress in long-term cancer patient survival disclosed by

modeled period analysis. J Clin Oncol, 2007. 25(22): p. 3274-80. 105. GRUSS, H.J., PINTO, A., GLOGHINI, A., WEHNES, E., WRIGHT, B., BOIANI, N., et al., CD30 ligand expression in

nonmalignant and Hodgkin's disease-involved lymphoid tissues. Am J Pathol, 1996. 149(2): p. 469-81. 106. SCHWAB, U., STEIN, H., GERDES, J., LEMKE, H., KIRCHNER, H., SCHAADT, M., et al., Production of a monoclonal

antibody specific for Hodgkin and Sternberg-Reed cells of Hodgkin's disease and a subset of normal lymphoid cells. Nature, 1982. 299(5878): p. 65-7.

107. LOCKSLEY, R.M., KILLEEN, N., and LENARDO, M.J., The TNF and TNF receptor superfamilies: integrating

mammalian biology. Cell, 2001. 104(4): p. 487-501. 108. HANSEN, H.P., RECKE, A., REINEKE, U., VON TRESCKOW, B., BORCHMANN, P., VON STRANDMANN, E.P., et al., The

ectodomain shedding of CD30 is specifically regulated by peptide motifs in its cysteine-rich domains 2 and 5. FASEB J, 2004. 18(7): p. 893-5.

109. HUHN, M., SASSE, S., TUR, M.K., MATTHEY, B., SCHINKOTHE, T., RYBAK, S.M., et al., Human angiogenin fused to

human CD30 ligand (Ang-CD30L) exhibits specific cytotoxicity against CD30-positive lymphoma. Cancer Res, 2001. 61(24): p. 8737-42.

110. HIRSCH, B., BRAUER, J., FISCHDICK, M., LODDENKEMPER, C., BULFONE-PAUS, S., STEIN, H., et al., Anti-CD30

human IL-2 fusion proteins display strong and specific cytotoxicity in vivo. Curr Drug Targets, 2009. 10(2): p. 110-7. 111. HORN-LOHRENS, O., TIEMANN, M., LANGE, H., KOBARG, J., HAFNER, M., HANSEN, H., et al., Shedding of the

soluble form of CD30 from the Hodgkin-analogous cell line L540 is strongly inhibited by a new CD30-specific

antibody (Ki-4). Int J Cancer, 1995. 60(4): p. 539-44. 112. BARTH, S., HUHN, M., MATTHEY, B., TAWADROS, S., SCHNELL, R., SCHINKOTHE, T., et al., Ki-4(scFv)-ETA', a new

recombinant anti-CD30 immunotoxin with highly specific cytotoxic activity against disseminated Hodgkin tumors

in SCID mice. Blood, 2000. 95(12): p. 3909-14. 113. VARDIMAN, J.W., IMBERT, M., PIERRE, R., BRUNNING, R.D., BAIN, B., and FLANDRIN, G., Chronic myelomonocytic

leukaemia. In: Jaffe ES, Harris NL, Sten H, Vardiman JW (eds). World Health Organization Classification of Tumors: Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues., 2001: p. IARC Press: Lyon, pp 49-52.

114. VARDIMAN, J.W., HARRIS, N.L., and BRUNNING, R.D., The World Health Organization (WHO) classification of the

myeloid neoplasms. Blood, 2002. 100(7): p. 2292-302. 115. ORAZI, A. and GERMING, U., The myelodysplastic/myeloproliferative neoplasms: myeloproliferative diseases with

dysplastic features. Leukemia, 2008. 22(7): p. 1308-19.

Literature ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

119

116. XU, Y., MCKENNA, R.W., KARANDIKAR, N.J., PILDAIN, A.J., and KROFT, S.H., Flow cytometric analysis of monocytes

as a tool for distinguishing chronic myelomonocytic leukemia from reactive monocytosis. Am J Clin Pathol, 2005. 124(5): p. 799-806.

117. MESAROSOVA, A., HRIVNAKOVA, A., KLOBUSICKA, M., and BABUSIKOVA, O., Chronic myeloid leukemia:

correlation between purine metabolism enzyme activities and membrane immunophenotype. Neoplasma, 1995. 42(1): p. 9-14.

118. ROLLINS-RAVAL, M.A. and ROTH, C.G., The value of immunohistochemistry for CD14, CD123, CD33,

myeloperoxidase and CD68R in the diagnosis of acute and chronic myelomonocytic leukaemias. Histopathology, 2012. 60(6): p. 933-42.

119. ORAZI, A., CHIU, R., O'MALLEY, D.P., CZADER, M., ALLEN, S.L., AN, C., et al., Chronic myelomonocytic leukemia: The

role of bone marrow biopsy immunohistology. Mod Pathol, 2006. 19(12): p. 1536-45. 120. FENAUX, P., BEUSCART, R., LAI, J.L., JOUET, J.P., and BAUTERS, F., Prognostic factors in adult chronic

myelomonocytic leukemia: an analysis of 107 cases. J Clin Oncol, 1988. 6(9): p. 1417-24. 121. ONIDA, F., BALL, G., KANTARJIAN, H.M., SMITH, T.L., GLASSMAN, A., ALBITAR, M., et al., Characteristics and

outcome of patients with Philadelphia chromosome negative, bcr/abl negative chronic myelogenous leukemia. Cancer, 2002. 95(8): p. 1673-84.

122. GERMING, U. and NEUKIRCHEN, J., How to treat patients with CMML? Leuk Res, 2013. 37(6): p. 605-6. 123. BACHER, U., HAFERLACH, T., SCHNITTGER, S., KREIPE, H., and KROGER, N., Recent advances in diagnosis,

molecular pathology and therapy of chronic myelomonocytic leukaemia. Br J Haematol, 2011. 153(2): p. 149-67. 124. DUNPHY, C.H. and TANG, W., The value of CD64 expression in distinguishing acute myeloid leukemia with

monocytic differentiation from other subtypes of acute myeloid leukemia: a flow cytometric analysis of 64 cases. Arch Pathol Lab Med, 2007. 131(5): p. 748-54.

125. THEPEN, T., HUHN, M., MELMER, G., TUR, M.K., and BARTH, S., Fcgamma receptor 1 (CD64), a target beyond

cancer. Curr Pharm Des, 2009. 15(23): p. 2712-8. 126. FANGER, M.W., SHEN, L., GRAZIANO, R.F., and GUYRE, P.M., Cytotoxicity mediated by human Fc receptors for IgG.

Immunol Today, 1989. 10(3): p. 92-9. 127. THEPEN, T., VAN VUUREN, A.J.H., KIEKENS, R.C.M., DAMEN, C.A., VOOIJS, W.C., and VAN DE WINKEL, J.G.J.,

Resolution of cutaneous inflammation after local elimination of macrophages. Nature Biotechnology, 2000. 18(1): p. 48-51.

128. HRISTODOROV, D., MLADENOV, R., HUHN, M., BARTH, S., and THEPEN, T., Macrophage-Targeted Therapy: CD64-

Based Immunotoxins for Treatment of Chronic Inflammatory Diseases. Toxins (Basel), 2012. 4(9): p. 676-94. 129. ZHONG, R.K., VAN DE WINKEL, J.G., THEPEN, T., SCHULTZ, L.D., and BALL, E.D., Cytotoxicity of anti-CD64-ricin a

chain immunotoxin against human acute myeloid leukemia cells in vitro and in SCID mice. J Hematother Stem Cell Res, 2001. 10(1): p. 95-105.

130. VAN ROON, J.A., VAN VUUREN, A.J., WIJNGAARDEN, S., JACOBS, K.M., BIJLSMA, J.W., LAFEBER, F.P., et al., Selective elimination of synovial inflammatory macrophages in rheumatoid arthritis by an Fcgamma receptor I-

directed immunotoxin. Arthritis Rheum, 2003. 48(5): p. 1229-38. 131. DE KRUIF, J., TIJMENSEN, M., GOLDSEIN, J., and LOGTENBERG, T., Recombinant lipid-tagged antibody fragments

as functional cell-surface receptors. Nat Med, 2000. 6(2): p. 223-7. 132. GRAZIANO, R.F., TEMPEST, P.R., WHITE, P., KELER, T., DEO, Y., GHEBREMARIAM, H., et al., Construction and

characterization of a humanized anti-gamma-Ig receptor type I (Fc gamma RI) monoclonal antibody. J Immunol, 1995. 155(10): p. 4996-5002.

133. TUR, M.K., HUHN, M., THEPEN, T., STOCKER, M., KROHN, R., VOGEL, S., et al., Recombinant CD64-specific single

chain immunotoxin exhibits specific cytotoxicity against acute myeloid leukemia cells. Cancer Res, 2003. 63(23): p. 8414-9.

134. BOGAN, A.A. and THORN, K.S., Anatomy of hot spots in protein interfaces. J Mol Biol, 1998. 280(1): p. 1-9. 135. CLACKSON, T. and WELLS, J.A., A hot spot of binding energy in a hormone-receptor interface. Science, 1995.

267(5196): p. 383-6. 136. KORTEMME, T. and BAKER, D., A simple physical model for binding energy hot spots in protein-protein complexes.

Proceedings of the National Academy of Sciences of the United States of America, 2002. 99(22): p. 14116-21. 137. STOCKER, M., TUR, M.K., SASSE, S., KRUSSMANN, A., BARTH, S., and ENGERT, A., Secretion of functional anti-

CD30-angiogenin immunotoxins into the supernatant of transfected 293T-cells. Protein Expr Purif, 2003. 28(2): p. 211-9.

138. STAHNKE, B., THEPEN, T., STOCKER, M., ROSINKE, R., JOST, E., FISCHER, R., et al., Granzyme B-H22(scFv), a human

immunotoxin targeting CD64 in acute myeloid leukemia of monocytic subtypes. Molecular Cancer Therapeutics, 2008. 7(9): p. 2924-32.

139. BALL, E.D., MCDERMOTT, J., GRIFFIN, J.D., DAVEY, F.R., DAVIS, R., and BLOOMFIELD, C.D., Expression of the three

myeloid cell-associated immunoglobulin G Fc receptors defined by murine monoclonal antibodies on normal bone

marrow and acute leukemia cells. Blood, 1989. 73(7): p. 1951-6. 140. VON KALLE, C., WOLF, J., BECKER, A., SCKAER, A., MUNCK, M., ENGERT, A., et al., Growth of Hodgkin cell lines in

severely combined immunodeficient mice. Int J Cancer, 1992. 52(6): p. 887-91. 141. SAMBROOK, J. and RUSSELL, D.W., Molecular cloning: a laboratory manual. . 3rd ed. 2001, Cold Spring Harbor

(NY): Cold Spring Harbor Laboratory. 142. HANAHAN, D., Studies on transformation of Escherichia coli with plasmids. J Mol Biol, 1983. 166(4): p. 557-80.

Literature ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

120

143. LAEMMLI, U.K., Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 1970. 227(5259): p. 680-5.

144. BARTH, S., HUHN, M., MATTHEY, B., KLIMKA, A., GALINSKI, E.A., and ENGERT, A., Compatible-solute-supported

periplasmic expression of functional recombinant proteins under stress conditions. Appl Environ Microbiol, 2000. 66(4): p. 1572-9.

145. KAMPMEIER, F., RIBBERT, M., NACHREINER, T., DEMBSKI, S., BEAUFILS, F., BRECHT, A., et al., Site-specific, covalent

labeling of recombinant antibody fragments via fusion to an engineered version of 6-O-alkylguanine DNA

alkyltransferase. Bioconjug Chem, 2009. 20(5): p. 1010-5. 146. VON TRESCKOW, B., KALLEN, K.J., VON STRANDMANN, E.P., BORCHMANN, P., LANGE, H., ENGERT, A., et al.,

Depletion of cellular cholesterol and lipid rafts increases shedding of CD30. J Immunol, 2004. 172(7): p. 4324-31. 147. STOCKER, M., PARDO, A., HETZEL, C., REUTELINGSPERGER, C., FISCHER, R., and BARTH, S., Eukaryotic expression

and secretion of EGFP-labeled annexin A5. Protein Expr Purif, 2008. 58(2): p. 325-31. 148. GODAL, R., KEILHOLZ, U., UHAREK, L., LETSCH, A., ASEMISSEN, A.M., BUSSE, A., et al., Lymphomas are sensitive to

perforin-dependent cytotoxic pathways despite expression of PI-9 and overexpression of bcl-2. Blood, 2006. 107(8): p. 3205-11.

149. LOSASSO, V., SCHIFFER, S., BARTH, S., and CARLONI, P., Design of human granzyme B variants resistant to serpin

B9. Proteins, 2012. 80(11): p. 2514-22. 150. PARDO, A., STOCKER, M., KAMPMEIER, F., MELMER, G., FISCHER, R., THEPEN, T., et al., In vivo imaging of

immunotoxin treatment using Katushka-transfected A-431 cells in a murine xenograft tumour model. Cancer Immunol Immunother, 2012. 61(10): p. 1617-26.

151. SALOMON, D.S., BRANDT, R., CIARDIELLO, F., and NORMANNO, N., Epidermal growth factor-related peptides and

their receptors in human malignancies. Crit Rev Oncol Hematol, 1995. 19(3): p. 183-232. 152. TUR, M.K., NEEF, I., JAGER, G., TEUBNER, A., STOCKER, M., MELMER, G., et al., Immunokinases, a novel class of

immunotherapeutics for targeted cancer therapy. Curr Pharm Des, 2009. 15(23): p. 2693-9. 153. ROSENBLUM, M.G. and BARTH, S., Development of novel, highly cytotoxic fusion constructs containing granzyme

B: unique mechanisms and functions. Curr Pharm Des, 2009. 15(23): p. 2676-92. 154. SENNVIK, K., BOEKHOORN, K., LASRADO, R., TERWEL, D., VERHAEGHE, S., KORR, H., et al., Tau-4R suppresses

proliferation and promotes neuronal differentiation in the hippocampus of tau knockin/knockout mice. FASEB J, 2007. 21(9): p. 2149-61.

155. GRULLICH, C., FRITSCH, K., and FINKE, J., Constitutive Expression of the Granzyme B Inhibitor PI-9 Protects

Leukemic Cells from Granzyme B Induced Apoptosis. Blood (ASH Annual Meeting Abstracts), 2005. 106(Abstract 3030).

156. CLASSEN, C.F., USHMOROV, A., BIRD, P., and DEBATIN, K.M., The granzyme B inhibitor PI-9 is differentially

expressed in all main subtypes of pediatric acute lymphoblastic leukemias. Haematologica, 2004. 89(11): p. 1314-21.

157. SCHIFFER, S., HANSEN, H., HEHMANN-TITT, G., HUHN, M., FISCHER, R., BARTH, S., et al., Efficacy of an adapted

Granzyme B-based anti-CD30 cytolytic fusion protein against PI-9-positive classical Hodgkin lymphoma cells in a

murine model. Blood Cancer Journal, 2013. 3: p. e106. 158. LIU, Y., CHEUNG, L.H., HITTELMAN, W.N., and ROSENBLUM, M.G., Targeted delivery of human pro-apoptotic

enzymes to tumor cells: In vitro studies describing a novel class of recombinant highly cytotoxic agents. Mol Cancer Ther, 2003. 2(12): p. 1341-50.

159. LIU, Y., ZHANG, W., NIU, T., CHEUNG, L.H., MUNSHI, A., MEYN, R.E., JR., et al., Targeted apoptosis activation with

GrB/scFvMEL modulates melanoma growth, metastatic spread, chemosensitivity, and radiosensitivity. Neoplasia, 2006. 8(2): p. 125-35.

160. DALKEN, B., GIESUBEL, U., KNAUER, S.K., and WELS, W.S., Targeted induction of apoptosis by chimeric granzyme B

fusion proteins carrying antibody and growth factor domains for cell recognition. Cell Death Differ, 2006. 13(4): p. 576-85.

161. KURSCHUS, F.C., KLEINSCHMIDT, M., FELLOWS, E., DORNMAIR, K., RUDOLPH, R., LILIE, H., et al., Killing of target

cells by redirected granzyme B in the absence of perforin. FEBS Lett, 2004. 562(1-3): p. 87-92. 162. LIU, Y., CHEUNG, L.H., THORPE, P., and ROSENBLUM, M.G., Mechanistic studies of a novel human fusion toxin

composed of vascular endothelial growth factor (VEGF)121 and the serine protease granzyme B: directed

apoptotic events in vascular endothelial cells. Mol Cancer Ther, 2003. 2(10): p. 949-59. 163. ZHANG, L., ZHAO, J., WANG, T., YU, C.J., JIA, L.T., DUAN, Y.Y., et al., HER2-targeting recombinant protein with

truncated pseudomonas exotoxin A translocation domain efficiently kills breast cancer cells. Cancer Biol Ther, 2008. 7(8): p. 1226-31.

164. WANG, T., ZHAO, J., REN, J.L., ZHANG, L., WEN, W.H., ZHANG, R., et al., Recombinant immunoproapoptotic

proteins with furin site can translocate and kill HER2-positive cancer cells. Cancer Res, 2007. 67(24): p. 11830-9. 165. KANATANI, I., LIN, X., YUAN, X., MANOREK, G., SHANG, X., CHEUNG, L.H., et al., Targeting granzyme B to tumor

cells using a yoked human chorionic gonadotropin. Cancer Chemother Pharmacol, 2011. 68(4): p. 979-90. 166. CAO, Y., MOHAMEDALI, K.A., MARKS, J.W., CHEUNG, L.H., HITTELMAN, W.N., and ROSENBLUM, M.G.,

Construction and Characterization of Novel, Completely Human Serine Protease Therapeutics Targeting Her2/neu. Mol Cancer Ther, 2013.

167. BUEHRING, G.C., EBY, E.A., and EBY, M.J., Cell line cross-contamination: how aware are Mammalian cell culturists

of the problem and how to monitor it? In Vitro Cell Dev Biol Anim, 2004. 40(7): p. 211-5.

Literature ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

121

168. BIRD, C.H., SUTTON, V.R., SUN, J., HIRST, C.E., NOVAK, A., KUMAR, S., et al., Selective regulation of apoptosis: the

cytotoxic lymphocyte serpin proteinase inhibitor 9 protects against granzyme B-mediated apoptosis without

perturbing the Fas cell death pathway. Mol Cell Biol, 1998. 18(11): p. 6387-98. 169. RAY, M., HOSTETTER, D.R., LOEB, C.R., SIMKO, J., and CRAIK, C.S., Inhibition of Granzyme B by PI-9 protects

prostate cancer cells from apoptosis. Prostate, 2012. 72(8): p. 846-55. 170. SORIANO, C., MUKARO, V., HODGE, G., AHERN, J., HOLMES, M., JERSMANN, H., et al., Increased proteinase

inhibitor-9 (PI-9) and reduced granzyme B in lung cancer: mechanism for immune evasion? Lung Cancer, 2012. 77(1): p. 38-45.

171. BLADERGROEN, B.A., MEIJER, C.J., TEN BERGE, R.L., HACK, C.E., MURIS, J.J., DUKERS, D.F., et al., Expression of the

granzyme B inhibitor, protease inhibitor 9, by tumor cells in patients with non-Hodgkin and Hodgkin lymphoma: a

novel protective mechanism for tumor cells to circumvent the immune system? Blood, 2002. 99(1): p. 232-7. 172. TIACCI, E., DORING, C., BRUNE, V., VAN NOESEL, C.J., KLAPPER, W., MECHTERSHEIMER, G., et al., Analyzing

primary Hodgkin and Reed-Sternberg cells to capture the molecular and cellular pathogenesis of classical Hodgkin

lymphoma. Blood, 2012. 120(23): p. 4609-20. 173. TEN BERGE, R.L., MEIJER, C.J., DUKERS, D.F., KUMMER, J.A., BLADERGROEN, B.A., VOS, W., et al., Expression levels

of apoptosis-related proteins predict clinical outcome in anaplastic large cell lymphoma. Blood, 2002. 99(12): p. 4540-6.

174. ROUSALOVA, I., KREPELA, E., PROCHAZKA, J., CERMAK, J., and BENKOVA, K., Expression of proteinase inhibitor-

9/serpinB9 in non-small cell lung carcinoma cells and tissues. Int J Oncol, 2010. 36(1): p. 275-83. 175. JIANG, X.G., PATTERSON, N.M., LING, Y., XIE, J.W., HELFERICH, W.G., and SHAPIRO, D.J., Low Concentrations of the

Soy Phytoestrogen Genistein Induce Proteinase Inhibitor 9 and Block Killing of Breast Cancer Cells by Immune Cells. Endocrinology, 2008. 149(11): p. 5366-73.

176. VAN HOUDT, I.S., OUDEJANS, J.J., VAN DEN EERTWEGH, A.J., BAARS, A., VOS, W., BLADERGROEN, B.A., et al., Expression of the apoptosis inhibitor protease inhibitor 9 predicts clinical outcome in vaccinated patients with

stage III and IV melanoma. Clin Cancer Res, 2005. 11(17): p. 6400-7. 177. OUDEJANS, J.J., HARIJADI, H., KUMMER, J.A., TAN, I.B., BLOEMENA, E., MIDDELDORP, J.M., et al., High numbers of

granzyme B/CD8-positive tumour-infiltrating lymphocytes in nasopharyngeal carcinoma biopsies predict rapid

fatal outcome in patients treated with curative intent. J Pathol, 2002. 198(4): p. 468-75. 178. CHAPPELOW, A.V. and KAISER, P.K., Neovascular age-related macular degeneration: potential therapies. Drugs,

2008. 68(8): p. 1029-36. 179. MADISON, E.L., GOLDSMITH, E.J., GERARD, R.D., GETHING, M.J., and SAMBROOK, J.F., Serpin-resistant mutants of

human tissue-type plasminogen activator. Nature, 1989. 339(6227): p. 721-4. 180. FRITSCH, K., PhD Thesis: Die Bedeutung des Proteinase Inhibitor 9 für die Apoptoseresistenz in leukämischen

Zelllinien, in Medicinal faculty 2008, Albert-Ludwigs-University, Freiburg im Breisgau. 181. BOSSARD, C., BELHADJ, K., REYES, F., MARTIN-GARCIA, N., BERGER, F., KUMMER, J.A., et al., Expression of the

granzyme B inhibitor PI9 predicts outcome in nasal NK/T-cell lymphoma: results of a Western series of 48 patients

treated with first-line polychemotherapy within the Groupe d'Etude des Lymphomes de l'Adulte (GELA) trials. Blood, 2007. 109(5): p. 2183-9.

182. KORTEMME, T., KIM, D.E., and BAKER, D., Computational alanine scanning of protein-protein interfaces. Sci STKE, 2004. 2004(219): p. pl2.

183. ANDRUSIER, N., MASHIACH, E., NUSSINOV, R., and WOLFSON, H.J., Principles of flexible protein-protein docking. Proteins, 2008. 73(2): p. 271-89.

184. GEHRMANN, M., DOSS, B.T., WAGNER, M., ZETTLITZ, K.A., KONTERMANN, R.E., FOULDS, G., et al., A novel

expression and purification system for the production of enzymatic and biologically active human granzyme B. J Immunol Methods, 2011. 371(1-2): p. 8-17.

185. GRIFFITHS, G.M. and ISAAZ, S., Granzymes A and B are targeted to the lytic granules of lymphocytes by the

mannose-6-phosphate receptor. J Cell Biol, 1993. 120(4): p. 885-96. 186. BRETTHAUER, R.K. and CASTELLINO, F.J., Glycosylation of Pichia pastoris-derived proteins. Biotechnol Appl

Biochem, 1999. 30 ( Pt 3): p. 193-200. 187. KLIMKA, A., MATTHEY, B., ROOVERS, R.C., BARTH, S., ARENDS, J.W., ENGERT, A., et al., Human anti-CD30

recombinant antibodies by guided phage antibody selection using cell panning. Br J Cancer, 2000. 83(2): p. 252-60. 188. HULETT, M.D. and HOGARTH, P.M., The second and third extracellular domains of FcgammaRI (CD64) confer the

unique high affinity binding of IgG2a. Mol Immunol, 1998. 35(14-15): p. 989-96. 189. NIKCEVIC, G., KOVACEVIC-GRUJICIC, N., and STEVANOVIC, M., Improved transfection efficiency of cultured human

cells. Cell Biol Int, 2003. 27(9): p. 735-7. 190. WOOD, K.M., ROFF, M., and HAY, R.T., Defective IkappaBalpha in Hodgkin cell lines with constitutively active NF-

kappaB. Oncogene, 1998. 16(16): p. 2131-9. 191. GRUSS, H.J., BOIANI, N., WILLIAMS, D.E., ARMITAGE, R.J., SMITH, C.A., and GOODWIN, R.G., Pleiotropic effects of

the CD30 ligand on CD30-expressing cells and lymphoma cell lines. Blood, 1994. 83(8): p. 2045-56. 192. KRAPPMANN, D., EMMERICH, F., KORDES, U., SCHARSCHMIDT, E., DORKEN, B., and SCHEIDEREIT, C., Molecular

mechanisms of constitutive NF-kappaB/Rel activation in Hodgkin/Reed-Sternberg cells. Oncogene, 1999. 18(4): p. 943-53.

193. PAVLINKOVA, G., BERESFORD, G.W., BOOTH, B.J., BATRA, S.K., and COLCHER, D., Pharmacokinetics and

biodistribution of engineered single-chain antibody constructs of MAb CC49 in colon carcinoma xenografts. J Nucl Med, 1999. 40(9): p. 1536-46.

Literature ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

122

194. KAMPMEIER, F., NIESEN, J., KOERS, A., RIBBERT, M., BRECHT, A., FISCHER, R., et al., Rapid optical imaging of EGF

receptor expression with a single-chain antibody SNAP-tag fusion protein. Eur J Nucl Med Mol Imaging, 2010. 37(10): p. 1926-34.

195. BECK, R., PEDROSA, R.C., DEJEANS, N., GLORIEUX, C., LEVEQUE, P., GALLEZ, B., et al., Ascorbate/menadione-

induced oxidative stress kills cancer cells that express normal or mutated forms of the oncogenic protein Bcr-Abl.

An in vitro and in vivo mechanistic study. Invest New Drugs, 2011. 29(5): p. 891-900. 196. PATHAK, A.P., ARTEMOV, D., WARD, B.D., JACKSON, D.G., NEEMAN, M., and BHUJWALLA, Z.M., Characterizing

extravascular fluid transport of macromolecules in the tumor interstitium by magnetic resonance imaging. Cancer Res, 2005. 65(4): p. 1425-32.

197. WIIG, H., RUBIN, K., and REED, R.K., New and active role of the interstitium in control of interstitial fluid pressure:

potential therapeutic consequences. Acta Anaesthesiol Scand, 2003. 47(2): p. 111-21. 198. CHAKROBORTY, D., SARKAR, C., YU, H., WANG, J., LIU, Z., DASGUPTA, P.S., et al., Dopamine stabilizes tumor blood

vessels by up-regulating angiopoietin 1 expression in pericytes and Kruppel-like factor-2 expression in tumor

endothelial cells. Proc Natl Acad Sci U S A, 2011. 108(51): p. 20730-5. 199. GEHRMANN, M., STANGL, S., KIRSCHNER, A., FOULDS, G.A., SIEVERT, W., DOSS, B.T., et al., Immunotherapeutic

targeting of membrane Hsp70-expressing tumors using recombinant human granzyme B. PLoS One, 2012. 7(7): p. e41341.

200. CULLEN, S.P., ADRAIN, C., LUTHI, A.U., DURIEZ, P.J., and MARTIN, S.J., Human and murine granzyme B exhibit

divergent substrate preferences. J Cell Biol, 2007. 176(4): p. 435-44. 201. SCHIFFER, S., HRISTODOROV, D., MLADENOV, R., ASLANIAN, E., HUHN, M., R., F., et al., Species-Dependent

Functionality of the Human Cytolytic Fusion Proteins Granzyme B-H22(scFv) and H22(scFv)-Angiogenin in

Macrophages. Antibodies, 2013. 2(1): p. 9-18. 202. BIRD, C.H., SUN, J., UNG, K., KARAMBALIS, D., WHISSTOCK, J.C., TRAPANI, J.A., et al., Cationic sites on granzyme B

contribute to cytotoxicity by promoting its uptake into target cells. Mol Cell Biol, 2005. 25(17): p. 7854-67. 203. BUZZA, M.S., ZAMURS, L., SUN, J., BIRD, C.H., SMITH, A.I., TRAPANI, J.A., et al., Extracellular matrix remodeling by

human granzyme B via cleavage of vitronectin, fibronectin, and laminin. J Biol Chem, 2005. 280(25): p. 23549-58. 204. LANG, D.S., DROEMANN, D., SCHULTZ, H., BRANSCHEID, D., MARTIN, C., RESSMEYER, A.R., et al., A novel human ex

vivo model for the analysis of molecular events during lung cancer chemotherapy. Respir Res, 2007. 8: p. 43. 205. CHRISTIANSEN, J. and RAJASEKARAN, A.K., Biological impediments to monoclonal antibody-based cancer

immunotherapy. Mol Cancer Ther, 2004. 3(11): p. 1493-501. 206. FEURING-BUSKE, M., FRANKEL, A.E., ALEXANDER, R.L., GERHARD, B., and HOGGE, D.E., A diphtheria toxin-

interleukin 3 fusion protein is cytotoxic to primitive acute myeloid leukemia progenitors but spares normal

progenitors. Cancer Res, 2002. 62(6): p. 1730-6. 207. VALLERA, D.A., JIN, N., BALDRICA, J.M., PANOSKALTSIS-MORTARI, A., CHEN, S.Y., and BLAZAR, B.R., Retroviral

immunotoxin gene therapy of acute myelogenous leukemia in mice using cytotoxic T cells transduced with an

interleukin 4/diphtheria toxin gene. Cancer Res, 2000. 60(4): p. 976-84. 208. TEN CATE, B., BREMER, E., DE BRUYN, M., BIJMA, T., SAMPLONIUS, D., SCHWEMMLEIN, M., et al., A novel AML-

selective TRAIL fusion protein that is superior to Gemtuzumab Ozogamicin in terms of in vitro selectivity, activity

and stability. Leukemia, 2009. 23(8): p. 1389-97. 209. SCHWEMMLEIN, M., PEIPP, M., BARBIN, K., SAUL, D., STOCKMEYER, B., REPP, R., et al., A CD33-specific single-

chain immunotoxin mediates potent apoptosis of cultured human myeloid leukaemia cells. Br J Haematol, 2006. 133(2): p. 141-51.

210. SANTOS, I.M., FRANZON, C.M., and KOGA, A.H., Laboratory diagnosis of chronic myelomonocytic leukemia and

progression to acute leukemia in association with chronic lymphocytic leukemia: morphological features and

immunophenotypic profile. Rev Bras Hematol Hemoter, 2012. 34(3): p. 242-4. 211. TUR, M.K., HUHN, M., JOST, E., THEPEN, T., BRUMMENDORF, T.H., and BARTH, S., In vivo efficacy of the

recombinant anti-CD64 immunotoxin H22(scFv)-ETA' in a human acute myeloid leukemia xenograft tumor model. Int J Cancer, 2011. 129(5): p. 1277-82.

212. NORDIGARDEN, A., ZETTERBLAD, J., TRINKS, C., GREEN, H., ELIASSON, P., DRUID, P., et al., Irreversible pan-ERBB

inhibitor canertinib elicits anti-leukaemic effects and induces the regression of FLT3-ITD transformed cells in mice. Br J Haematol, 2011. 155(2): p. 198-208.

213. RIBBERT, T., THEPEN, T., TUR, M.K., FISCHER, R., HUHN, M., and BARTH, S., Recombinant, ETA'-based CD64

immunotoxins: improved efficacy by increased valency, both in vitro and in vivo in a chronic cutaneous

inflammation model in human CD64 transgenic mice. Br J Dermatol, 2010. 163(2): p. 279-86. 214. SILVERMAN, G.A., BIRD, P.I., CARRELL, R.W., CHURCH, F.C., COUGHLIN, P.B., GETTINS, P.G., et al., The serpins are

an expanding superfamily of structurally similar but functionally diverse proteins. Evolution, mechanism of

inhibition, novel functions, and a revised nomenclature. J Biol Chem, 2001. 276(36): p. 33293-6. 215. JEDEMA, I., BARGE, R.M., FRANKEL, A.E., WILLEMZE, R., and FALKENBURG, J.H., Acute myeloid leukemia cells in G0

phase of the cell cycle that are unresponsive to conventional chemotherapy are sensitive to treatment with

granulocyte-macrophage colony-stimulating factor/diphtheria toxin fusion proteins. Exp Hematol, 2004. 32(2): p. 188-94.

216. DRUKER, B.J., GUILHOT, F., O'BRIEN, S.G., GATHMANN, I., KANTARJIAN, H., GATTERMANN, N., et al., Five-year

follow-up of patients receiving imatinib for chronic myeloid leukemia. N Engl J Med, 2006. 355(23): p. 2408-17. 217. ARMSTRONG, P.B. and QUIGLEY, J.P., Alpha2-macroglobulin: an evolutionarily conserved arm of the innate

immune system. Dev Comp Immunol, 1999. 23(4-5): p. 375-90.

Literature ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

123

218. ROBERTS, R.C., Protease inhibitors of human plasma. Alpha-2-macroglobulin. J Med, 1985. 16(1-3): p. 129-224. 219. REHMAN, A.A., AHSAN, H., and KHAN, F.H., Alpha-2-macroglobulin: A physiological guardian. J Cell Physiol, 2012. 220. BRUELL, D., BRUNS, C.J., YEZHELYEV, M., HUHN, M., MULLER, J., ISCHENKO, I., et al., Recombinant anti-EGFR

immunotoxin 425(scFv)-ETA' demonstrates anti-tumor activity against disseminated human pancreatic cancer in

nude mice. Int J Mol Med, 2005. 15(2): p. 305-13. 221. KHURSHID, R., SALEEM, M., AKHTAR, M.S., and SALIM, A., Granzyme M: characterization with sites of post-

translational modification and specific sites of interaction with substrates and inhibitors. Mol Biol Rep, 2011. 38(5): p. 2953-60.

222. BARTH, S., HUHN, M., MATTHEY, B., SCHNELL, R., TAWADROS, S., SCHINKOTHE, T., et al., Recombinant anti-CD25

immunotoxin RFT5(SCFV)-ETA' demonstrates successful elimination of disseminated human Hodgkin lymphoma in

SCID mice. Int J Cancer, 2000. 86(5): p. 718-24. 223. DE KONING, P.J., KUMMER, J.A., DE POOT, S.A., QUADIR, R., BROEKHUIZEN, R., MCGETTRICK, A.F., et al.,

Intracellular serine protease inhibitor SERPINB4 inhibits granzyme M-induced cell death. PLoS One, 2011. 6(8): p. e22645.

224. JABULOWSKY, R.A., OBEROI, P., BAHR-MAHMUD, H., DALKEN, B., and WELS, W.S., Surface Charge-Modification

Prevents Sequestration and Enhances Tumor-Cell Specificity of a Recombinant Granzyme B-TGF alpha Fusion

Protein. Bioconjugate Chemistry, 2012. 23(8): p. 1567-76.

Appendix ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

I

8 Appendix

8.1 DNA AND AMINO ACID SEQUENCE OF GBR201K

The wild-type DNA sequence triplet cga was exchanged by aag in the modified GbR201K (position

601-603).

atcatcgggg gacatgaggc caagccccac tcccgcccct acatggcttt tcttatgatc 60 tgggatcaga agtctctgaa gaggtgcggt ggcttcctga tacgagacga cttcgtgctg 120 acagctgctc actgttgggg aagctccata aatgtcacct tgggggccca caatatcaag 180 gaacaggagc cgacccagca gtttatccct gtgaaaagag ccatccccca tccagcctat 240 aatcctaaga acttctccaa tgacatcatg ctactgcagc tggagagaaa ggccaagcgg 300 accagagctg tgcagcccct caggctacct agcaacaagg cccaggtgaa gccagggcag 360 acatgcagtg tggccggctg ggggcagacg gcccccctgg gaaaacactc acacacacta 420 caagaggtga agatgacagt gcaggaagat cgaaagtgcg aatctgactt acgccattat 480 tacgacagta ccattgagtt gtgcgtgggg gacccagaga ttaaaaagac ttcctttaag 540 ggggactctg gaggccctct tgtgtgtaac aaggtggccc agggcattgt ctcctatgga 600 aagaacaatg gcatgcctcc acgagcctgc accaaagtct caagctttgt acactggata 660 aagaaaacca tgaaacgcta c 681

In accordance to that, the modified GbR201K amino acid sequence comprises a lysine (K) at position

201 instead of wild-type arginine (R).

IIGGHVAKPH SRPYMAYLMI WDQKSLKRCG GFLIRDDFVL TAAHCWGSSI NVTLGAHNIK 60 EQEPTQQFIP VKRAIPHPAY NPKNFSNDIM LLQLERKAKR TRAVQPLRLP SNKAQVKPGQ 120 TCSVAGWGQT APLGKHSHTL QEVKMTVQED RKCESDLRHY YDSTIELCVG DPEIKKTSFK 180 GDSGGPLVCN KVAQGIVSYG KNNGMPPRAC TKVSSFVHWI KKTMKRY 227

8.2 SYNTHETIC DNA SEQUENCE OF GM-H22(SCFV) IN PMS VECTOR

The sequence of Gm-H22(scFv) is shown exemplary for all granzyme M-based constructs generated

in this study. The single chain sequences were exchanged accordingly via appropriate restriction

sites.

Appendix ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

II

Appendix ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

III

Index of Figures and Tables ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

IV

9 Index of Figures and Tables

TABLES

Table 1: List of oligo nucleotides used for cloning and DNA sequencing. ............................................. 21

Table 2: Origin and specificity of antibody fragments. ......................................................................... 23

Table 3: Mammalian cell lines. .............................................................................................................. 24

Table 4: Buffers and media.................................................................................................................... 25

Table 5: Reagents for physical or chemical transfection of mammalian cell lines. .............................. 31

Table 6: Overview of the PI-9 expression profile of different cancer cell lines. ................................... 43

Table 7: Affinity constants and relative receptor expression level of CD30-positive cell lines............. 55

Table 8: Serum stability of Gb-H22(scFv) and GbR201K-H22(scFv) in human patient serum (CMML III)

or PBS. ................................................................................................................................................... 77

Table 9: Overview of different granzyme B-based cytolytic fusion proteins (CFPs) published so far. . 88

Table 10: Overview of PI-9 expression in different carcinomas / melanomas. .................................... 91

Table 11: Overview of PI-9 expression in different lymphomas. .......................................................... 93

FIGURES

Figure 1: X-ray crystal structure of the non-covalent Michaelis complex of A353K Manduca sexta

serpin B1 (purple) and S195A trypsin (cyan) (PDB file 1K9O [87]). ....................................................... 11

Figure 2: Workflow of this study. .......................................................................................................... 18

Figure 3: Expression of endogenous PI-9 in HL cell lines L428, L1236, L540cy and CML cell line K562.44

Figure 4: Overview of initial granzyme B constructs (A) and SDS-PAGE of the purified fusion proteins

before and after Enterokinase cleavage (B). ......................................................................................... 46

Figure 5: List of generated point-mutated variants (2.2.2) arised from literature (#: [87]; ##: [86]) or

molecular modeling (3.2). ..................................................................................................................... 47

Figure 6: Preparation of recombinant PI-9 and functionality determination via complex formation

with wild-type Gb-H22(scFv). ................................................................................................................ 49

Figure 7: Influence of endogenous PI-9 on apoptotic activity of Gb-Ki4(scFv). .................................... 51

Figure 8: Enzymatic activity of Gb-Ki4(scFv) and GbR201K-Ki4(scFv). .................................................. 52

Figure 9: Enzymatic activity of Gb-H322(scFv) wild-type and its mutants in the presence and absence

of PI-9. ................................................................................................................................................... 53

Figure 10: Binding analysis of granzyme B-based CFPs. ........................................................................ 54

Figure 11: Internalization analysis of Ki4(scFv)-SNAP-BG-647 and H22(scFv)-SNAP-BG-647 into

corresponding target cell lines. ............................................................................................................. 56

Index of Figures and Tables ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

V

Figure 12: Cytotoxic effects of Gb-Ki4(scFv) and Gb-H22(scFv) wild-type and their mutants

respectively on PI-9-negative target cell lines. ..................................................................................... 57

Figure 13: Apoptotic effects of Gb-Ki4(scFv) or Gb-H22(scFv) wild-type and their mutants respectively

on PI-9-negative target cell lines. .......................................................................................................... 58

Figure 14: Cytotoxic effect of Gb-Ki4(scFv) and its mutants on PI-9-positive target cell lines. ............ 60

Figure 15: Apoptotic effect of Gb-Ki4(scFv) and GbR201K-Ki4(scFv) on PI-9 positive target cell lines. 61

Figure 16: Serum stability of Gb-Ki4(scFv) and GbR201K-Ki4(scFv). ..................................................... 62

Figure 17: Evaluation of the established L540cy and L428 tumor models. .......................................... 64

Figure 18: Evaluation of the in vivo anti-tumor activity against L540cy-induced tumors. .................... 66

Figure 19: Evaluation of the in vivo anti-tumor activity against L428-induced tumors. ....................... 67

Figure 20: Expression of endogenous PI-9 (2.4.7) in primary leukemic cells derived from peripheral

blood of CMML or AMML patients (2.1.8). ........................................................................................... 69

Figure 21: Phenotyping of primary leukemic cells. ............................................................................... 70

Figure 22: Binding and internalization analysis with primary leukemic cells (2.1.8). ........................... 71

Figure 23: Cytotoxic and apoptotic effects of Gb-H22(scFv), GbR201K-H22(scFv) and H22(scFv)-ETA’

on primary leukemic cells. ..................................................................................................................... 72

Figure 24: Cytotoxic effects of Gb-H22(scFv) and GbR201K-H22(scFv) on primary leukemic cells of

AMML I. ................................................................................................................................................. 74

Figure 25: Specific reduction of target cancer cell population by Gb-H22(scFv), GbR201K-H22(scFv)

and H22(scFv)-ETA’ detected in FSC/SSC dotplot after flow cytometry. .............................................. 75

Figure 26: Reduction of target cancer cell population by Gb-H22(scFv) and GbR201K-H22(scFv)

detected with specific antibodies (2.1.5). ............................................................................................. 76

Figure 27: Overview of granzyme M-based constructs (A) and SDS-PAGE and western blot analysis of

Gm-H22(scFv) (B). .................................................................................................................................. 78

Figure 28: Enzymatic activity of Gm-H22(scFv) tested by cleavage of PI-9 or a colorimetric substrate.

............................................................................................................................................................... 79

Figure 29: Binding analysis of Gm-H22(scFv). ....................................................................................... 80

Figure 30: Cytotoxic and apoptotic effects of Gm-H22(scFv)................................................................ 82

Figure 31: Apoptotic effect of Gm-Ki4(scFv) compared to Gb-Ki4(scFv)............................................... 83

Figure 32: Cytotoxic and apoptotic effects of Gm-H22(scFv) on primary leukemic cells. .................... 85

Publications ________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

VI

10 Publications

Publications (*,# equal contribution)

Schiffer S, Rosinke R, Jost E, Huhn M, Fischer R, Thepen T*, Barth S*. Targeted ex vivo reduction of

CD64-positive monocytes in CMML and AMML using human granzyme B-based cytolytic fusion

proteins. [in revision]

Schiffer S, Letzian S, Jost E, Mladenov R, Hristodorov D, Huhn M, Fischer R, Barth S*. Thepen T*.

Granzyme M as a novel effector molecule for human cytolytic fusion proteins: CD64-specific

cytotoxicity of Gm-H22(scFv) against leukemic cells. Cancer Letters 2013; 341, 178-185.

Schiffer S, Hansen H, Hehmann-Titt G, Huhn M, Fischer R, Barth S*, Thepen T*. Efficacy of an adapted

granzyme B-based anti-CD30 cytolytic fusion protein against PI-9-positive classical Hodgkin

lymphoma cells in a murine model. Blood Cancer Journal 2013; 3: e106.

Hehmann-Titt G*, Schiffer S*, Berges N*, Melmer G, Barth S. Improving the Therapeutic Potential of

Human Granzyme B for Targeted Cancer Therapy. Antibodies 2013; 2(1): 19-49.

Schiffer S*, Hristodorov D*, Mladenov R, Huhn M, Fischer R, Barth S*, Thepen T*. Species-dependent

functionality of the human cytolytic fusion proteins Granzyme B-H22(scFv) and H22(scFv)-

Angiogenin. Antibodies 2013; 2(1): 9-18.

Losasso V*, Schiffer S*, Barth S#, Carloni P#. Design of Human Granzyme B Variants resistant to Serpin

B9. Proteins 2012; 80(11): 2514-22.

Patent applications

Barth S, Schiffer S (2012), International patent application, Novel immunoproteases – EP12191711.6-

2401

Barth S, Schiffer S, Hehmann-Titt G (2011), International patent application, Novel serine protease

variants - EP11007764.1-2401, US61/538,638,PCT/EP2012/068607

Poster presentations

Schiffer S, Hehmann-Titt G, Fischer R, Barth S. Novel approach to circumvent the inhibitory potential

of Serpin B9 in cancer therapy. Biomedica Liège, Belgium (2012)

Schiffer S, Hehmann-Titt G, Thepen T, Huhn M, Fischer R, Barth S. Optimized human Granzyme B

based immunotoxins to circumvent the inhibitory potential of Serpin B9 in cancer therapy. CIMT

Mainz, Germany (2012)

Danksagung

An dieser Stelle möchte ich allen danken, die mich während meiner Dissertation fachlich oder

persönlich unterstützt haben. Mein besonderer Dank gilt:

Prof. Dr. Rainer Fischer für die Möglichkeit diese Arbeit im Fraunhofer Institut für Molekularbiologie

und angewandte Ökologie anzufertigen, sowie die Korrektur dieser Arbeit. Außerdem gilt ihm mein

Dank für die Übernahme des Referats.

Prof. Dr. Dr. Stefan Barth für die Möglichkeit diese Arbeit in seiner Abteilung durchzuführen, seine

wissenschaftliche Betreuung, sein Vertrauen, sowie die gewährte Freiheit bei der Gestaltung meiner

Projekte.

Dr. Grit Hehmann-Titt für die Betreuung dieser Arbeit, alle wertvollen Tipps und Anregungen und die

Korrektur dieser Arbeit. Besonders gilt ihr mein Dank für ihre Aufmunterungen in schwierigeren

Phasen sowie für wertvolle Diskussionen.

Dr. Michael Huhn für seine unermüdliche Unterstützung während dieser Arbeit, die nie abreißende

Flut an neuen Ideen und dass er immer ein offenes Ohr für mich hatte.

Dr. Theo Thepen für die Korrekturen meiner Veröffentlichungen und dieser Arbeit, seine

Unterstützung bei den in vivo Experimenten und seine wertvollen Tipps und Anregungen.

Dr. Edgar Jost für die Bereitstellung des primären Probenmaterials.

Dr. Hinrich Hansen, der mich maßgeblich bei der Verfassung einer meiner Publikationen unterstützt

hat und mir viele motivierende Worte aus Köln geschickt hat.

Prof. Dr. Paolo Carloni, Dr. Valeria Losasso und Dr. Heinrich Delbrück für ihre Hilfe bei der

Modellierung der Mutanten und die gute Zusammenarbeit.

Der ganzen PPD und Pharmedartis Gruppe für die tolle Atmosphäre in der Arbeitsgruppe, durch die

das Lachen nie zu kurz kam. Besonders danke ich hier Nora Frohnecke und Vivian Dietrich für ihre

Freundschaft sowie ihre fachliche und persönliche Unterstützung; Sarah Kirchhoff für ihre Loyalität

und große Unterstützung im Labor; Judith Niesen und Lea Hein für ihr treues Mitleiden und

Mitfreuen; Dmitrij Hristodorov und Radoslav Mladenov für ihr großes Engagement; Jens Bührmann,

Torsten Sieben und Anh-Tuan Pham für ihre sehr verlässliche Hilfe im Labor; besonders Reinhard

Rosinke für seine unermüdliche Unterstützung, seine immer guten Ideen, sowie seine exzellente Hilfe

bei der Proteinaufreinigung, sowie bei der Arbeit mit Primärmaterial.

Meinen Eltern, die mich immer unterstützt haben, mich zu der Person gemacht haben, die ich bin,

und ohne die ich niemals so weit gekommen wäre.

Meinem Mann, der auch in schwierigen Zeiten immer hinter mir stand und meine Höhen und Tiefen

während dieser Arbeit geduldig ertragen hat.

Lebenslauf

Persönliche Daten

Sonja Schiffer

geboren 1984 in Neuss

Ausbildung

2003 Erwerb der Allgemeinen Hochschulreife, Pascal-Gymnasium, Grevenbroich

2003 - 2004 Auslandsaufenthalt in Chicago, USA

2004 – 2009

2009

Studium des Bioingenieurwesens, TU Dortmund; Schwerpunkt Biotechnologie

(Abschluss: Dipl. Ing.)

2007: Auslandssemester an der TH Lund, Schweden

2008: Auslandspraktikum bei „Sumitomo Chemical“ in Osaka, Japan

Diplomarbeit am Max-Planck-Institut für molekulare Physiologie, Dortmund,

Abteilung Strukturbiologie

Betreuung: Prof. Dr. Alfred Wittinghofer (MPI) /

Prof. Dr. Andreas Schmid (TU Dortmund)

Titel: “Screening for functional ‘GAPettes’ directed against constitutive active

Ras proteins”

2009-2013

Promotion am Fraunhofer Institut für Molekularbiologie und angewandte

Ökologie IME, Abteilung Pharmazeutische Produktentwicklung / RWTH Aachen

Betreuung: Prof. Dr. Rainer Fischer /

Prof. Dr. Dr. Stefan Barth

Titel: “Improving the therapeutic potential of human granzyme B and

evaluation of granzyme M as novel effector molecules in cytolytic fusion

proteins for the treatment of Serpin B9-positive cancer“

Documents

Improving the therapeutic potential of human granzyme B and … · Improving the therapeutic potential of human granzyme B and evaluation of granzyme M as novel effector molecules