Upload
doquynh
View
222
Download
0
Embed Size (px)
Citation preview
Novel Biomarkers in DLBCL
Colm Keane MB BCh BAO MSc MBA
School of Medical ScienceHealth Group Griffith University
Submitted in fulfillment of the requirements of the degree ofDoctor of Philosophy
August 2014
1/17
Abstract
Diffuse large B cell lymphoma (DLBCL) is the commonest aggressive lymphoma.
Despite the advent of combined chemo-‐immunotherapy, one third of patients
still die from their disease. Prognostication of the disease still relies on a clinical
scoring system known as the International Prognostic Index (IPI). This divides
patients into risk categories. However marked heterogeneity within IPI sub-‐
categories persist. The IPI is a clinical score based predominantly on estimates of
patient fitness and tumour burden, but does not utilize information regarding
the biology of the tumour cell or the immune tumour microenvironment (TME),
in which the malignant B cells reside. The latter is the focus of this thesis.
The anti-‐CD20 monoclonal antibody rituximab has improved the outcome for
patients with DLBCL, however the impact of host genetics on its effectiveness is
still unclear. The mechanisms of action for rituximab include antibody
dependent cytotoxicity (ADCC) and complement mediated cytotoxicity (CDC).
Recent reports suggest genetic polymorphisms in the FCGR3A receptor
(expressed on NK-‐cells and monocytes which mediate ADCC) may be a predictor
of event free and overall survival in B-‐cell lymphoma. Data also implicates the
same polymorphism in the susceptibility to rituximab induced late-‐onset
neutropenia (LON). There remains no data on the impact of genetic
polymorphisms on either outcome or LON in genes involved in the CDC pathway
such as C1qA. One hundred and fifteen DLBCL patients treated with ‘Ru CHOP’
(rituximab/cyclophosphamide/vincristine/doxorubicin/prednisolone)
chemou immunotherapy were compared with 105 healthy Caucasian controls
with regards to FCGR3Aq V158F and C1qAq A276G polymorphisms. Event free
and overall survival (EFS and OS) and LON incidence were analysed for linkage
to either polymorphism. The FCGR3Aq V158F but not the C1qAq A276G
polymorphism influenced the risk of developing LON. 50% of FCGR3A-‐158V/V
patients experienced LON. In contrast, only 7% V/F and 2% F/F experienced
LON. The FCGR3A-‐158V/V genotype was associated with LON compared to V/F
(p=0.028) and F/F genotypes (p=0.005). Although no patients with either LON or
FCGR3A-‐158V homozygosity relapsed compared to 33% FCGR3A-‐158F/F and
21% non-‐LON, this did not translate into improved EFS or OS and results are
2/17
likely to be influenced by lead-‐time bias. Polymorphic analysis may be a
predictive tool to identify those at high-‐risk of LON. Larger prospective studies
are required to definitively establish if LON or FCGR3A-‐158V/V genotype
influences outcome.
Host immune status has consistently been shown to play an important role in
DLBCL. The manipulation of the host immune environment has shown promise
in solid and lymphoid tumours and identification of patients benefiting from
immune based therapy may have therapeutic relevance in DLBCL. The impact of
circulating monocytes and circulating and intratumoural lymphocytes on
outcome were assessed in a cohort of 122 patients with DLBCL. All were treated
with R-‐CHOP, and median follow up was 48 months. Lymphocyte and monocyte
counts were assessed prior to commencement of therapy, and in addition the
majority had data from flow cytometric immunophenotyping on lymphocyte (but
not monocyte) subsets available on fresh diagnostic lymphoid tissue.
The circulating lymphocyte to monocyte ratio (LMR) was a significant predictor
of outcome with patients with high LMR having an estimated five-‐year survival
of 86% compared to 63% (p=0.01) in patients with low LMR. This finding was
independent of the IPI. Low (0,1) and intermediate IPI (2,3) did not predict a
significantly different survival from each other in our cohort, with both having
excellent outcome (83% estimated 5 year survival in these combined groups).
However when the 90 patients in these low risk IPI groups were split by the
LMR, there was a significant overall survival advantage (estimated 5 year OS
93% vs. 72%, p=0.01) for patients with a higher LMR. Amongst intratumoural
lymphocyte subsets, CD4+T cell infiltration was the most significant predictor of
improved outcome. Those with high intratumoural CD4+ T cells had a superior
EFS (p=0.009) and OS (p=0.006) compared to those with low values. CD3+T cell
infiltration was also associated with improved EFS (P=0.01) and OS (p=0.01),
whereas CD8+T cell infiltration did not predict outcome. CD4+T cells were
independent of IPI and LMR for both EFS and OS. Importantly when
low/intermediate IPI groups were analyzed, a high CD4+T cell infiltration
remained a striking predictor of OS (5 year OS 93% vs. 60%, p=0.004). These
results confirm the importance of the local immune environment in patients with
3/17
DLBCL treated with chemo-‐immunotherapy, and indicate that peripheral
immune subsets are surrogate markers of intratumoural immunity.
Based on these findings, it was envisaged that detailed functional and
quantitative assessment of blood would enable identification of the optimal
diagnostic tissue based TME immune-‐effector and monocyte/macrophage-‐
checkpoints to assist sub-‐stratification of conventional prognosticators. Blood
from 140 R-‐CHOP treated DLBCL patients in the NHL21 Australasian Leukaemia
and Lymphoma Group trial were prospectively analysed. A circulating immune-‐
effector: monocyte-‐checkpoint signature segregating interim-‐PET/CT-‐positivity
was identified. Intratumoural applicability was tested in two independent R-‐
CHOP treated DLBCL cohorts, with cell-‐of-‐origin (COO) and international
prognostic index (IPI) as co-‐variates. CD163+CD14+HLA-‐DRlo blood monocytes
were immunosuppressive. CD8+:CD163+CD14+HLA-‐DRlo ratios were highest in
interim-‐PET/CT-‐ve patients (P≤0.0001). Digital multiplexed gene expression
(DMGE) in 191 DLBCL tissues demonstrated co-‐clustering of CD8 with immune-‐
checkpoints (CD163/PD1/PDL1/PDL2/TIM3/LAG3, all P<0.001), indicating an
adaptive immune-‐checkpoint response to immune-‐effector activation. In
multivariate analysis of 128 R-‐CHOP treated DLBCL patients, CD8:CD163 ratios
(net anti-‐tumoral immunity) were prognostic independent of IPI and COO.
Combining CD8:CD163 to the germinal centre B-‐cell marker LMO2
(LMO2/CD8:CD163) strengthened the predictive ability. Results were externally
validated in 233 patients, separating good-‐risk IPI (0-‐2) into two categories of
85% and 48% (P=0.0003) 4 year survival after R-‐CHOP. Similarly, poor-‐risk IPI
(3-‐5) stratified into two survival groupings of 66% and 36% (P=0.0007). In
patients with DLBCL, a measure of net anti-‐tumoral immunity within the TME is
a powerful new prognosticator that is independent of IPI and COO.
LMO2/CD8:CD163 adds to the predictive ability of IPI.
4/17
Statement of Originality
This work has not previously been submitted for a degree or diploma in any university. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made in the thesis itself.
(Signed)_____________________________
5/17
Acknowledgements
I would firstly like to thank Maher Gandhi for his amazing support throughout
this PhD. He has helped me become a much better researcher, medical writer
and doctor. His dedication and hard work are an inspiration to everyone in the
laboratory. I would also like to thank the support I have received from Griffith
University and in particular my supervisors Rod Lea and Lyn Griffiths.
I would like to thank everyone at the Gandhi Lab who helped during my research
in particular Jamie Nourse, Frank Vari and Pauline Crooks. They showed great
patience and support in helping someone with limited laboratory experience to
get through this research. I would also like to thank Frank in particular for his
hard work and insights to our co-‐authored paper.
I would like to thank my funding sources and in particular the Leukaemia
Foundation, without whom I would have been unable to perform any of this
research. They should be acclaimed for not only work they perform in
supporting research but also the tireless work they perform in trying to make life
better for patients with haematological malignancy.
I would like to thank my work colleagues at the Princess Alexandra Hospital who
have been supportive throughout my research and in particular Devinder Gill
who is always supportive of translational research in the unit. I wish to
acknowledge Dipti Talaulikar for her help in gathering additional patient
samples which have had a huge impact on the quality of my work. I would also
like to thank Mark Hertzberg, who as the clinical lead on the ALLG study has
been so supportive in trying to get as many samples and clinical data for all our
patients. I want to thank all the patients who participate in all our research
including NHL21. One hopes that we can do their hard work and trust in us
justice, and hopefully contribute in some small way to improving the length and
quality of life for patients with lymphoma in the future.
I would like to thank my wife Sharon and three children who have put up with a
lot over the last few years. Their understanding in allowing me to spend the last
few years being a student again has meant huge sacrifices on their behalf and I
will forever be grateful for their support.
6/17
Publications Arising from this Thesis(*indicates papers directly forming part of thesis)
*1.Measures of net anti-‐tumoral immunity add to the predictive power of conventional prognostic factors in diffuse large B cell lymphoma (DLBCL). Colm Keane*, Frank Vari*, Mark Hertzberg, John Seymour, Rodney Hicks, Devinder Gill, Pauline Crooks, Kimberly Jones, Erica Han, Rod Lea, Lyn Griffiths, Maher Gandhi. Submitted to Cancer Discovery May 2014 *Co-‐Authors
*2.CD4(+) tumor infiltrating lymphocytes are prognostic and independent of R-‐IPI in patients with DLBCL receiving R-‐CHOP chemo-‐immunotherapy. Keane C, Gill D, Vari F, Cross D, Griffiths L, Gandhi M. Am J Hematol. 2013 Apr;88(4):273-‐6. Epub 2013 Mar 5.
3.Plasma MicroRNA Are Disease Response Biomarkers in Classical HodgkinLymphoma. Jones K, Nourse JP, Keane C, Bhatnagar A, Gandhi MK. Clin Cancer Res. 2014 Jan 1;20(1):253-‐64. Epub 2013 Nov 12.
4.Serum CD163 and TARC are disease response biomarkers in classicalHodgkin lymphoma. Jones K, Vari F, Keane C, Crooks P, Nourse JP, Seymour LA, Gottlieb D, Ritchie D, Gill D, Gandhi MK. Clin Cancer Res. 2013 Feb 1;19(3):731-‐42.
5.High-‐resolution loss of heterozygosity screening implicates PTPRJ as apotential tumor suppressor gene that affects susceptibility to Non-‐Hodgkin's lymphoma. Aya-‐Bonilla C, Green MR, Camilleri E, Benton M, Keane C, Marlton P, Lea R, Gandhi MK, Griffiths LR. Genes Chromosomes Cancer. 2013 May;52(5):467-‐79. doi: 10.1002/gcc.22044. Epub 2013 Jan 23.
6.Tumor-‐specific but not non-‐specific cell-‐free circulating DNA can be usedto monitor disease response in lymphoma. Kimberley Jones, Jamie P. Nourse, Colm Keane, Pauline Crooks, David Gottlieb, David S. Ritchie, Devinder Gill and Maher K. Gandhi American Journal of Haematology
*7.Homozygous FCGR3A-‐158V alleles predispose to late onset neutropenia after CHOP-‐R for Diffuse Large B-‐cell Lymphoma Colm Keane, Jamie P. Nourse, Pauline Crooks, Do Nguyen-‐Van, Howard Mutsando, Peter Mollee, Rod A. Lea, Maher K. Gandhi Intern Med J. 2011 Sep 1. doi: 10.1111/j.1445-‐5994.2011.02587.x.
7/17
8.Epstein-‐Barr virus-‐positive diffuse large B-‐cell lymphoma of the elderlyexpresses EBNA3A with conserved CD8+ T-‐cell epitopes Do Nguyen-‐Van, Colm Keane, Erica Han, Kimberley Jones, Jamie P. Nourse, Frank Vari, Nathan Ross, Pauline Crooks, Olivier Ramuz, Michael Green, Lyn Griffith, Ralf Trappe, Andrew Grigg, Peter Mollee, Maher K. Gandhi Am J Blood Res 2011;1(2):146-‐159
Book Chapters *“Rituximab induced Late-‐onset Neutropenia” for the book 'Rituximab: Pharmacology, Clinical Uses and role in Investigating B cell immunology in Man"published by Novus in 2012 Colm Keane, Jamie Nourse, Maher K. Gandhi
Presentations Arising from this Thesis
1.Net antitumoral immunity and the predictive power of conventionalprognosticators in diffuse large B-‐cell lymphoma Poster Highlights Session Merit Award Winner American Society of Clinical Oncology, Annual Meeting 2014 Colm Keane, Frank Vari, Mark S. Hertzberg, Michael R Green, Erica Han, John Francis Seymour, Rodney J Hicks, Devinder Singh Gill, Pauline Crooks, Clare Gould, Kimberley Jones, Kristen Radford, Lyn Griffiths, Dipti Talaulikar, Sanjiv Jain, Josh Tobin, Maher K. Gandhi
2.Noninvasive monitoring of cellular versus acellular tumor DNA fromimmunoglobulin genes for DLBCL Oral Presentation American Society of Clinical Oncology, Annual Meeting 2014 David Matthew Kurtz, Michael R Green, Scott Victor Bratman, Chih-‐Long Liu, Cynthia Glover, Colm Keane, Katie Kong, Malek Faham, David Bernard Miklos, Ranjana H. Advani, Ronald Levy, Mark S. Hertzberg, Maher K Gandhi, Maximilian Diehn, Ash A. Alizadeh
3.“Utility Of Non-‐Invasive Monitoring Of Circulating Tumor DNA At Diagnosis, Interim Therapy, and Relapse Of DLBCL Using High-‐Throughput Sequencing Of Immunoglobulin Genes” Poster Presentation American Society of Haematology 2013 Michael R Green, Scott Bratman , Chih Long Liu, Kazuhiro Takahashi, Cynthia Glove*, Colm Keane, Shingo Kihira, Katie Kong, Malek Faham, MD, PhD, Corbelli Karen, David B. Miklos, Ranjana H. Advani, , Ronald Levy, Mark S. Hertzberg, Maher K Gandhi, Maximilian Diehn, Ash A. Alizadeh,
8/17
4."CD163+ Identifies A Highly Immunosuppressive Subset Of Monocytic-‐Myeloid Derived Suppressor Cells (moMDSCs) In Poor-‐Risk Diffuse Large B-‐Cell Lymphoma (DLBCL): An ALLG Laboratory Sub-‐Study Of NHL21." Oral Presentation International Conference on Malignant Lymphoma, Lugano 2013 Colm Keane, Mark Hertzberg, John Seymour, Rodney Hicks, Devinder Gill, Frank Vari, Pauline Crooks, Kimberly Jones, Erica Han, Rod Lea, Lyn Griffiths, Maher Gandhi.
5.Flow Cytometric analysis demonstrates prognostic significance oftumour-‐infiltrating CD4+T lymphocytes in patients with diffuse large B cell lymphoma receiving R-‐CHOP chemoimmunotherapy. Poster Presentation American Association of Cancer Researchers Tumour Immunology Meeting 2012 Colm Keane, Frank Vari, Lynn Griffiths, Devinder Gill, Peter Mollee, Rod A. Lea, Maher K. Gandhi
5.Serum CD163 and TARC in Combination As Disease Response Biomarkersin Classical Hodgkin Lymphoma Oral Presentation American Society of Haematology 2012 Maher K Gandhi, Frank Vari, Pauline Crooks, Colm Keane, Jamie P Nourse, Louise A Seymour, David Ritchie, David Gottlieb, Devinder Gill, and Kimberley Jones.
6.Circulating microRNAs as prognostic and disease response biomarkers inpatients with high-‐risk diffuse large b-‐cell lymphoma (DLBCL): a prospective Australasian leukaemia & lymphoma group study Poster Presentation European Haematology Association Annual Meeting 2012 Colm Keane, Mark Hertzberg, John Seymour, Rodney Hicks, Devinder Gill, Frank Vari, Pauline Crooks, Kimberly Jones, Erica Han, Rod Lea, Lyn Griffiths, Maher Gandhi.
7.Tissue Microarray in Patients with DLBCL Receiving R-‐CHOP Chemo-‐immunotherapy Shows Survival Benefit for Coexpression of LMO2/BCL6 Poster Presentation American Society of Haematology 2011 Colm Keane, Linda Shen, Jamie Nourse, Erica Han, Rod Lea, Peter Mollee, Devinder Gill, Maher Gandhi
8.Monocytes are Associated with Impaired T-‐cell Immunity and ResidualInterim-‐ PET/CT Avidity after 4 Cycles of CHOP-‐R In Patients with High-‐Risk DLBCL Poster Presentation American Society of Haematology 2011 Frank Vari, Mark Hertzberg, Erica Han, John F Seymour, Rodney Hicks, Devinder Gill, Colm Keane, Pauline Crooks, Kristen Radford, Maher K. Gandhi
9/17
9.Tissue Microarray in Patients with DLBCL Receiving R-‐CHOP Chemo-‐immunotherapy Shows Survival Benefit for Coexpression of LMO2/BCL6 Oral Presentation HSANZ 2011 Colm Keane, Linda Shen, Jamie Nourse, Erica Han, Rod Lea, Peter Mollee, Devinder Gill, Maher Gandhi
10.Monocytes are Associated with Impaired T-‐cell Immunity and ResidualInterim-‐ PET/CT Avidity After 4 Cycles of CHOP-‐R In Patients With High-‐Risk DLBCL Oral Presentation Haematology Society of Australia and New Zealand Annual Meeting 2011 Frank Vari, Mark Hertzberg, Erica Han, John F Seymour, Rodney Hicks, Devinder Gill, Colm Keane, Pauline Crooks, Kristen Radford, Maher K. Gandhi
11.Lymphoma-‐Specific But Not Non-‐Specific Cell-‐Free Circulating DNA CanBe Used to Monitor Disease Response in Lymphoma Oral Presentation Haematology Society of Australia and New Zealand Annual Meeting 2011 Kimberley Jones, Jamie P Nourse, Colm Keane, Pauline Crooks, David Gottlieb, David S Ritchie, Devinder Gill, Maher K Gandhi
12.Tissue Microarray in DLBCL patients receiving CHOP-‐R chemo-‐immunotherapy shows survival benefit for coexpression of LMO2/BCL6 and poor outome for EBER-‐ISH positive patients. Oral Presentation International Conference on Malignant Lymphoma, Lugano June 2011 Colm Keane, Linda Shen, Jamie Nourse, Erica Han, Kimberley Jones,Maher Gandhi
13.Homozygous FCGR3A-‐158V alleles predispose to late onset neutropeniaafter CHOP-‐R for Diffuse Large B-‐cell Poster Presentation International Conference on Malignant Lymphoma meeting, Lugano June 2011 Colm Keane, Jamie Nourse, Pauline Crooks, Do Nguyen Van, Howard Mutsando, Maher Gandhi
14.The kinetics of circulating immunosuppressive monocytes in patientswith poor-‐risk diffuse large B-‐cell lymphoma treated with ‘CHOP-‐R’. Poster Presentation European Haematology Association meeting, London, June 2011 Frank Vari, Mark Hertzberg, Erica Han, Colm Keane, J. Alejandro Lopez, Kristen Radford, John F Seymour, Devinder Gill, Maher K Gandhi.
10/17
15.EBV-‐positive DLBCL of the elderly is a distinct clinico-‐biological entity with poor outcome in CHOP-‐R treated patients, with properties likely amenable to anti-‐EBV targeting. Oral Presentation. The 14th Biennial Conference of the International Association for Research on Epstein -‐ Barr virus & Associated Diseases, Birmingham, September 2010. Do Nguyen-‐Van, Colm Keane, Jamie P.Nourse, Erica Han, Nathan Ross, Kimberley Jones, Pauline Crooks, and Maher K. Gandhi. 16.Epstein-‐Barr virus microRNAs are differentially expressed during EBV infection and EBV-‐driven differentiation of naïve B-‐cells. The 14th Biennial Conference of the International Association for Research on Epstein-‐Barr Virus & Associated Diseases, Birmingham, September 2010. Oral Presentation Jamie P Nourse , Pauline Crooks , Colm Keane, Do Nguyen-‐Van,Sally Mujaj, Nathan Ross, Kimberley Jones , Frank Vari , Erica Han , and Maher K. Gandhi.
11/17
Abbreviations Full Title ADCC Antibody dependent cellular cytotoxicity ALC Absolute Lymphocyte Count ALLG Australasian Leukaemia Lymphoma Group AMC Absolute Monocyte Count APC Antigen Presenting Cell ASCT Autologous stem cell transplant BCL2 B-‐cell CLL/lymphoma 2 BCL6 B-‐cell CLL/lymphoma 6 BD Becton Dickenson BMDCs Bone marrow-‐derived dendritic cells BTK Bruton's tyrosine kinase CBC Complete Blood Count CCL3 Chemokine (C-‐C motif) ligand 3 CCND2 G1/S-‐Specific Cyclin-‐D2 CD Cluster Differentiaition CDC Complement dependent cytotoxicity cDNA Complementary Deoxyribonucleic acid CNS Central Nervous System COO Cell-‐of-‐Origin CT Computed Tomography CTLA4 Cytotoxic T-‐lymphocyte-‐associated protein 4 CTLs Cytotoxic T lymphocytes DA-‐EPOCH-‐R Dose-‐adjusted etoposide / prednisone / vincristine /
cyclophosphamide, hydroxyduanorubicin / rituximab DLBCL Diffuse Large B cell Lymphoma DMSO Dimethyl sulfoxide DNA Deoxyribonucleic acid EBV Epstein Barr Virus EFS Event Free Survival ELISA Enzyme-‐linked immunosorbent assay FBC Full Blood Count FCG3A FCGammaReceptor 3A FDA Food and Drug Administration FFPE Formalin-‐fixed paraffin-‐embedded tissue FITC Fluorescein isothiocyanate FN1 Fibronectin 1 FOXP1 Forkhead box P1 GA101 Other name for Obinotuzumab GCB Germinal Cell B cell like GCET1 Germinal center B-‐cell expressed transcript 1 H2O2 Hydrogen peroxide
12/17
HIV Human Immunodeficiency virus HLA Human Leukoctye Antigen IgG Immunoglobulin G IHC Immunohistochemistry IL1 Interleukin 1 IL10 Interleukin 10 IL12 Interleukin 12 IL13 Interleukin 13 IL4 Interleukin 4 iNOS Inducible nitric oxide synthase IPI International Prognostic Index LAG3 Lymphocyte-‐activation gene 3 LDH Lactate Dehydrogenase LMO2 LIM domain only 2 LMR Lymphocyte to Monocyte ratio M1 Macrophage Type I M2 Macrophage Type II MDSC Myeloid-‐derived suppressor cell mg Milligram MHC Major histocompatibility complex ml Milliliter mM Millimolar moMDSC Monocytic Myeloid-‐derived suppressor cell MUM1 Melanoma associated antigen (mutated) 1 NHL Non -‐Hodgkins Lymphoma NK Natural Killer Non-‐GCB Non-‐Germinal Cell B cell like Ob-‐ADCC Obinotuzumab-‐Antibody dependent cellular cytotoxicity OS Overall Survival PBMC Peripheral blood mononucleated cell PBS Phosphate buffered saline PCR Polymerase Chain Reaction PD-‐1 Programmed cell death 1 PD-‐L1 Programmed cell death ligand 1 PD-‐L2 Programmed cell death ligand 2 pDCs Plasmacytoid derived dendritic cells PE Phycoerythrin peg-‐G-‐CSF Pegalated granulocytic colony stimulating factor PerCP Peridinin chlorophyll protein PET Positron Emission Tomography PRDM1 PR domain containing 1, with ZNF domain PTLD Post-‐Transplant Lymphoproliferative Disease R-‐ADCC Rituximab-‐Antibody dependent cellular cytotoxicity R-‐CHOP Cyclophosphamide / doxorubicin /vincristine / prednisone / rituximab. RNA Ribonucleic acid ROS Reactive oxygen species
13/17
rpm Revolutions per minute SCYA3 Alternate name for CCL3 SDF-‐1 Stromal cell-‐derived factor 1 SPSS Statistical Package for the Social Sciences STAT1 Signal transducer and activator of transcription 1 TAM Tumour Associated Macrophages TBE Tris/Borate/EDTA TGF Transforming growth factor Th1 T Helper cell I Th2 T Helper cell II TILs Tumour infiltrating lymphocytes TIM3 T-‐Cell Membrane Protein 3 TME Tumour Microenvironment TNF Tumour Necrosis Factor Treg T regulatory cell Z-‐BEAM Zevalin-‐ Carmustine, Etoposide, Cytarabine, Melphalan µg Microgram µl Microliter µM Micromolar
14/17
ALL PAPERS INCLUDED ARE CO-AUTHORED
Acknowledgement of Papers included in this Thesis
Section 9.1 of the Griffith University Code for the Responsible Conduct of Research (“Criteria for Authorship”), in accordance with Section 5 of the Australian Code for the Responsible Conduct of Research, states:
To be named as an author, a researcher must have made a substantial scholarly contribution to the creative or scholarly work that constitutes the research output, and be able to take public responsibility for at least that part of the work they contributed. Attribution of authorship depends to some extent on the discipline and publisher policies, but in all cases, authorship must be based on substantial contributions in a combination of one or more of:
• conception and design of the research project
• analysis and interpretation of research data
• drafting or making significant parts of the creative or scholarly workor critically revising it so as to contribute significantly to the finaloutput.
Section 9.3 of the Griffith University Code (“Responsibilities of Researchers”), in accordance with Section 5 of the Australian Code, states:
Researchers are expected to:
• Offer authorship to all people, including research trainees, who meetthe criteria for authorship listed above, but only those people.
• accept or decline offers of authorship promptly in writing.
• Include in the list of authors only those who have acceptedauthorship
• Appoint one author to be the executive author to record authorshipand manage correspondence about the work with the publisher andother interested parties.
• Acknowledge all those who have contributed to the research,facilities or materials but who do not qualify as authors, such asresearch assistants, technical staff, and advisors on cultural orcommunity knowledge. Obtain written consent to name individuals.
Included in this thesis are papers in Chapters 3, 4 and 5 which are co-authored with other researchers. My contribution to each co-authored paper is outlined at the front of the relevant chapter. The bibliographic details (if published or accepted for publication)/status (if prepared or submitted for publication) for these papers including all authors, are:
(Where a paper(s) has been published or accepted for publication, you must also include a statement regarding the copyright status of the paper(s).
15/17
Chapter 3: Homozygous FCGR3A-‐158V alleles predispose to late onset neutropenia after CHOP-‐R for Diffuse Large B-‐cell Lymphoma Colm Keane, Jamie P. Nourse, Pauline Crooks, Do Nguyen-‐Van, Howard Mutsando, Peter Mollee, Rod A. Lea, Maher K. Gandhi Intern Med J. 2011 Sep 1. doi: 10.1111/j.1445-‐5994.2011.02587.x. Rituximab induced Late-‐onset Neutropenia” for the book 'Rituximab: Pharmacology, Clinical Uses and role in Investigating B cell immunology in Man"published by Novus in 2012 Colm Keane, Jamie Nourse, Maher K. Gandhi
Chapter 4: CD4(+) tumor infiltrating lymphocytes are prognostic and independent of R-‐IPI in patients with DLBCL receiving R-‐CHOP chemo-‐immunotherapy. Keane C, Gill D, Vari F, Cross D, Griffiths L, Gandhi M. Am J Hematol. 2013 Apr;88(4):273-‐6. Epub 2013 Mar 5.
Chapter 5: The immunobiological score: a robust 3-‐gene assay that segregates the international prognostic index into disparate survival categories in aggressive B-‐cell lymphoma Colm Keane*, Frank Vari*, Mark Hertzberg, John Seymour, Rodney Hicks, Devinder Gill, Pauline Crooks, Kimberly Jones, Erica Han, Rod Lea, Lyn Griffiths, Maher Gandhi. *Joint Authorship Submitted to Cancer Discovery May 2014
Appropriate acknowledgements of those who contributed to the research but did not qualify as authors are included in each paper.
(Signed) _________________________________ (Date)______________
Name of Student
(Countersigned) ___________________________ (Date)______________
Supervisor: Name of Supervisor tures should be included for each paper).
16/17
Table of Contents
Chapter 1 Introduction 1.1 Background....................................................................................................................................2 1.2 Molecular classification of DLBCL........................................................................................5 1.3 Clinical utility of molecular classifications.......................................................................6 1.4 FCG3A and C1qA polymorphisms and their impact in DLBCL................................8 1.5 Tumour Infiltrating Lymphocytes.....................................................................................10 1.6 CD8 cells in DLBCL...................................................................................................................14 1.7 Tumour associated Macrophages/Monocytes.............................................................15 1.8 Myeloid Derived Suppressor Cells....................................................................................18 1.9 Digital multiplex gene expression (DMGE) by nanoString nCounter................18
Chapter 2 Research Design and Methods 2.1 Clinical Sample Accrual ..........................................................................................................32 2.2 Patient Cohorts..............................................................................................................................33 2.3 PET/CT analysis.........................................................................................................................36 2.4 RNA/DNA extraction..................................................................................................................36 2.5 FCG3A polymorphism PCR....................................................................................................37 2.6 C1QA Polymorphism PCR......................................................................................................37 2.7 Flow cytometric analysis of retrospective samples...................................................38 2.8 NHL21 Blood processing..........................................................................................................39 2.9 Flow Cytometry NHL21 Study.............................................................................................40 2.10 Effector lymphocyte assays..................................................................................................41 2.11 Enzyme linked immuno-‐absorbent assays (ELISA)........................................................42 2.12 Digital multiplex gene expression by NanoString nCounter.................................42 2.13 Statistics and analysis.............................................................................................................43
Chapter 3 Homozygous FCGR3A-‐158V alleles predispose to late onset neutropenia after CHOP-‐R for Diffuse Large B-‐cell Lymphoma.........................................................................47 Rituximab induced Late-‐onset Neutropenia...........................................................…………54
Chapter 4 CD4 (+) tumor infiltrating lymphocytes are prognostic and independent of R-‐IPI in patients with DLBCL receiving R-‐CHOP chemo-‐immunotherapy...................69
Chapter 5 The immune-‐biological score: a robust 3 gene assay that segregates the international prognostic index into disparate survival categories in aggressive B-‐cell lymphoma.......................................................................................................................................79
Chapter 6 Discussion 6.1 Immuno-‐genetic polymorphisms ..................................................................................140 6.2 Immune Microenvironment................................................................................................145 6.3 Net Tumoral Immunity..........................................................................................................149 6,4 Future directions..................................................................................................................... 155
17/17
CHAPTER 1 Introduction
1
Introduction
Brief Aim1
Examine the influence of C1qA-A276G and FCGR3A-V158F polymorphisms on
survival in patients with DLBCL treated with R-CHOP chemo-immunotherapy. I will
also investigate if these polymorphisms influence the risk of rituximab induced late
onset neutropenia.
Brief Aim2
Examine if absolute lymphocyte and monocytes and their interactions predict
outcome in patients with DLBCL treated with standard R-CHOP chemo-
immunotherapy.
Investigate if the level of T-cell infiltration in DLBCL tumours, as assessed by flow
cytometry, predicts outcome in patients with DLBCL
Brief Aim 3
Using a novel investigative platform with digital bar-coding to obtain highly accurate
gene expression data on paraffin based tissue, an extensive investigation of 191
DLBCL patient samples will be performed examining the role of immune effectors
and checkpoints with regards to expression and outcomes in patients treated with
standard chemo-immunotherapy.
1.1 Background
Diffuse large B cell lymphoma is a malignancy of B-lymphocytes. DLBCL is the
most common aggressive subtype and accounts for approximately 40% of all patients
with non-Hodgkin’s lymphoma.[1] The median age of presentation is usually in the
7th decade but it is not uncommon in younger adults and can occur in children. The
incidence of this disease has risen rapidly over the last 4 decades but has now
plateaued.[2] The reasons for this are unclear, however it is likely that better
diagnostics and improved life expectancy likely contribute to this increase. In addition
there has been increased levels of immune-impairment seen in this time period,
secondary to the rise of HIV in the 1980s, but also due to the increasing levels of
immune suppression used to treat autoimmune conditions or protect tissue and bone
marrow grafts in patients undergoing transplantation.[3-5] Primary and acquired
2
impaired immunity are well-recognised strong risk factors for the development of B
cell lymphoma. However at present these reasons seem insufficient to account for the
marked increase in lymphomas, and there may be additional environmental factors of
importance.[6, 7]
Clinically patients with DLBCL can present with fatigue, painless adenopathy or
symptoms related to tumour bulk such as abdominal pain. It can occur as a primary
lesion in a nodal or extranodal site (e.g. bone, liver, lung etc.). Primary central
nervous system DLBCL, testicular DLBCL, primary DLBCL of the bone or of the
skin though less common, are well described. Patients can also present with so called
“B –Symptoms” such as night sweats, weight loss >5% or unexplained fevers.
Paraneoplastic syndromes such as hypercalcaemia can also occur.
Biopsy specimens from patients with DLBCL show a diffuse proliferation of large
centroblast like cells that completely disrupt normal lymph node architecture but the
predominant cell can also be immunoblastic or anaplastic.[1] By flow cytometry the
malignant B cells typically express the surface markers CD19/20/22/79A with another
50% of tumours also expressing CD10.[8] The proliferation marker Ki-67 is usually
high in DLBCL varying from 40-90%. Tumours expressing greater than 90% should
raise suspicion of a myc gene translocation which if present is generally classified as a
tumour intermediate between DLBCL and Burkitt Lymphoma and is associated with
poorer outcome.[9] In general a myc gene rearrangement is found in approximately
10% of diagnosed DLBCL and is associated with poor outcome.[10-12] Clonally
rearranged immunoglobulin heavy and light chains are usually detected at diagnosis.
Both BCL6 and BCL2 rearrangements are commonly found in DLBCL.[1]
There has been remarkable progress made in this disease over the last decade with the
addition of the anti-CD20 monoclonal antibody rituximab to standard
chemotherapy.[13] Prior to the introduction of rituximab, 5-year survival was
approximately 45-50%.[14, 15] The five-year overall survival is now between 55-
75%. Despite these improvements, one third of patients will still die from their
disease.[16, 17]
DLBCL is staged using the Ann Arbor classification with Cotswold’s modification. In
1993, a seminal paper was published that attempted to identify the most important
prognostic factors in order to derive a score that might help to predict survival.[18]
This scoring system is called the International Prognostic Index (IPI) which predicts
survival accurately in patients with DLBCL.A patient is scored according to Age>60,
3
stage >II, LDH level>normal, Eastern Cooperative Group performance status >1 and
number of extra-nodal sites >1. A patient receives a score of one if any of the
prognostic features is positive and zero if absent. The score thus ranges from zero to
five, with five having poor survival and patients with no factors present having
excellent outcome approaching five-year survival of 95%. Despite its usefulness,
heterogeneity of outcome within IPI sub-groups persists. For example, although 30%
of patients are known to relapse within 2 years, the IPI fails to accurately identify
which patients these will be. For example poor risk IPI (IPI 3-5) categories still have
cure rates of greater than 55%.[18, 19] Despite the improved outcome in DLBCL with
the introduction of rituximab, the IPI is still highly predictive of outcome but it now
splits patients into 3 rather than 5 groups, as was the case prior to its introduction to
front-line therapy. An IPI score of zero predicts excellent outcome, IPI score of 1 or 2
is associated with good outcome but patients scoring >3 have poor outcome. [19]
Given the IPI’s relative inability to predict very poor outcome, recently some of the
parameters of the IPI have been expanded to provide better stratification.[20]
Importantly, the IPI tells us little about potential therapeutic targets or mechanisms of
disease resistance. Despite the relative accuracy of the IPI in predicting outcome,
most patients still receive 6-8 cycles of standard chemo-immunotherapy irrespective
of their IPI score. There is thus, a pressing need to identify the significant clinical and
pathological prognostic factors that may better guide therapy for patients who are
unlikely to benefit from standard therapy
This would allow physicians to not only investigate new therapies for these poor risk
patients, but would also allow confidence to identify patients with excellent outcome
with current standard therapy who do not require alternative strategies.
IPI Factors (Score 1 for each factor present)
Stage>2
LDH>N
Extranodal sites>1
Performance Status (ECOG>2)
Age>60
Table 1. IPI Factors
4
1.2 Molecular Classification of DLBCL
It has been known for some time that there are distinct molecular subgroups of
DLBCL. Alizadeh and colleagues in their seminal paper from 2000, showed three
distinct molecular groupings in DLBCL. These were classified as germinal centre,
activated B cell and a third group termed unclassifiable that lay in between these two
groups.[21] What he showed, and what has been shown consistently since then is that
dividing patients into two groupings, germinal centre-like B cell (GCB) and non-GCB
could separate patients into two distinct groups with very different outcome after
standard chemotherapy. [21-23] The prognosis of the unclassified grouping seems to
lie in-between these two groups but it is generally included in a poor risk group
incorporating the non-GCB subtype.[24, 25]
These gene expression findings have been backed up by similar findings from other
groups who found that even low IPI scoring patients do relatively poorly if they are in
a poor gene expression subclass such as non-GCB type.[26] The non-GCB subtype
seems to have higher proliferation with a reduced immune reaction against it, and
have a signature very different from normal lymph node signals. However, even in
studies showing cell of origin as predictive of outcome the survival differences in
patients treated with chemo-immunotherapy is modest between GCB and non-GCB
with a 15-35% improvement in outcome for GCB patients. Prognosis based on COO
based on gene expression from frozen tissue appears most accurate.[20, 27, 28] The
data, although not conclusive, indicates that rituximab is particularly beneficial in the
non-GCB group rather than the GCB classification.[28] However not all gene
expression data published at this time is consistent or confirms that GCB and non-
GCB subtypes are prognostic.[29] This shows the problems with performing analysis
on different platforms, using different analysis methods and heterogeneous patient
cohorts.[30] Because of these varied results, Lossos et al. set out to identify genes
found in all the three previously described gene expression array studies performed at
that time to find common genes that might allow development of a robust small
number of genes to predict outcome and be more user friendly in the
diagnostic/prognostic setting.[31] This group tested all 36 prognostic genes from gene
expression studies and also any genes described as being prognostic from a number of
other studies using real time-PCR in their patients’ samples. They found six genes
that added additional prognostic value to the IPI. The six genes were LMO2, BCL6,
FN1, CCND2, BCL2 and SYCA3. Of note two of these genes were associated with
5
the GCB phenotype (LMO2, BCL6) while FN1 is a normal lymph node signature.
The three genes (CCND2, BCL2 and SYCA3) associated with poor prognosis are all
associated with the non-GCB (ABC) subtype. While this data has subsequently been
replicated by this group it required frozen tissue and it has not been widely applied.
However in 2008 the authors published data confirming prognostic usefulness in
paraffin based tissue samples from an R-CHOP tested cohort.[32] However these
methods are still relatively complex for routine clinical use and have not translated
into general clinical practice.
One of the challenges for applying these and associated classifications is the failure of
any of these poor prognostic groups to benefit from alternate therapy to the current
standard (R-CHOP). However a number of studies in recent years have shown that
there may well be some therapeutic strategies that might specifically target tumours of
the ABC subtype. Targeted therapies have been developed based on gene expression
data such as Bortezomib or dose modified etoposide, doxorubicin, vincristine,
prednisolone, cyclophosphamide and rituximab that may be more effective against the
phenotype with a poor outcome (non-GCB).[33, 34]. Newer strategies such as the use
of BTK inhibitors seem to preferentially target ABC DLBCL and current clinical
trials are awaited to assess their use in DLBCL.[35, 36] However it is imperative that
new technologies are validated and tested to ensure accurate classification so that the
right patients receive the correct drug.
1.3 Clinical utility of Molecular Classifications
Because of the cost of DNA microarray technology, it is not routinely used in
lymphoma diagnosis. It also requires frozen tissue, which is very difficult to acquire,
as this requires the diagnosis to be known before biopsy, (unless procedures are in
place at an Institute to snap freeze all potential new lymphoma diagnosis) and special
storage facilities are required. It is only large tumour banks that have this tissue
available.[30] Immunohistochemistry on formalin fixed paraffin embedded tissue
(FFPET) is the standard test for most pathology departments for the diagnosis and
investigation of malignancy and it is still the gold standard for diagnosis of diffuse
large cell lymphoma. An attempt was made by Hans et al. in 2004 to establish an
immunohistochemical algorithm that might approximate gene expression
findings.[24] This algorithm used the immunohistochemical markers CD10, MUM1,
6
BCL6 in a sequential arrangement to divide patients into GC and non-GC subtypes
and these groupings did have significantly different outcome. As would be expected
BCL6 and CD10 conferred good outcome whereas MUM1 conferred poor outcome as
single markers. The 5-year OS rate for GC patients was 76% compared to 34% for the
non-GCB rates. This is similar to previous described cDNA analysis. They also had
cDNA data on a number of patient samples and these showed reasonable correlation
with IHC with the algorithm sensitive in picking up 74% of GCB subtype and 87% of
non-GCB subtype. This shows that there are still significant issues with this algorithm
with approximately one fifth of patients being misclassified. This is perhaps not
surprising since it is unlikely that 3 proteins would provide similar results to 36 genes,
nor does this algorithm identify the unclassifiable cases that are identified by gene
expression profiling. In addition, a number of large studies have subsequently shown
no survival difference between immunohistochemical-defined cell of origin including
a sub-study of the large RICOVER-60 trial, which had full data on over 500 patients
treated with R-CHOP.[37-39] However other research has shown that the GCB /non-
GCB prognostic divide is still relevant in the rituximab era.[40, 41]. The Choi
algorithm, which uses all the Hans antibodies but includes two additional antibodies
to identify GCB and non-GCB, has also been found to be effective at predicting
outcome based on cell of origin. [25] The additional antibodies are GCET1 and
FOXP1. This study also had gene expression data and the IHC had a 91%
concordance with it. A number of studies have confirmed the accuracy of this Choi
algorithm compared to the other described algorithms but there are still large studies
finding no prognostic impact for it. [39, 42]This heterogeneity certainly restricts the
clinical use of these classifications.
There is also concern about reproducibility of IHC staining between different labs
with the Lunenburg Consortium suggesting that a number of antibodies used in the
Hans algorithm gave variable results between labs, with BCL6 (a nuclear staining
antibody) in particular demonstrating marked variability.[43, 44] In most algorithms
BCL6 positivity is a ‘deciding’ antibody for classification. For example, in the Hans
classification, CD10 negative DLBCL biopsies are classified as non-GCB if BCL6
negative. Therefore inter-laboratory variability with this antibody is a particular
concern. BCL6 is a germinal centre marker that also appears in a number of gene
expression and IHC models that predict outcome with expression of BCL6 usually
associated with improved outcome. In addition the importance of prognostic markers
7
such as BCL6 may have changed with the improvement in outcome seen with
rituximab. In numerous pre-rituximab studies it has been shown to be associated with
good prognosis. The addition of rituximab to standard chemotherapy appears to have
improved outcome in BCL6 negative patients in particular, so that the prognostic
significance of BCL6 expression in patients treated with rituximab now appears less
clear.
In contrast is BCL2, which is an ABC marker. Prior to introduction of rituximab this
IHC marker always predicted poor outcome however the utilization of rituximab
seems to have overcome the poor prognosis associated with this protein.[45, 46]
There is currently no substitute for gene expression for defining cell of origin but
there is still a paucity of gene expression data in the rituximab era. For routine use
immunohistochemical methods for classification would still appear to be the most
practical, with newer algorithms such as those described by Choi and Meyer likely to
replace the Han’s classifications.
New technologies such as digital gene multiplex hybridization (DGME) hold promise
for more clinically applicable use of cell of origin classifications. One such platform
is NanoString nCounter ® that accurately assess gene expression from FFPE. This
technology has promise for defining important subgroups based on cell of origin and
requires minimal technical and bioinformatics expertise and therefore is transferrable
to the routine diagnostic laboratory.[47]
1.4 FCG3A and C1qA polymorphisms and their impact in DLBCL
The anti-CD20 monoclonal antibody rituximab is a standard component of front-line
DLBCL chemo-immunotherapy, typically as part of ‘R-CHOP’. It has resulted in a
marked improvement in response and survival.[16] The importance of host genetics
on the mode of action of rituximab in DLBCL is unclear.
Two principle mechanisms of action for rituximab are postulated. The first is antibody
dependent cellular cytotoxicity (ADCC) whereby rituximab binds to
FCGammaReceptor (FCGR) bearing Natural Killer (NK) cells, resulting in
destruction of CD20+ normal and malignant B-cells by the reticulo-endothelial
system.[48-51] The other is direct lysis via complement dependent cytotoxicity
(CDC).[52-56] There is also data to support direct apoptosis of malignant cells on
exposure to rituximab.[57-59]
8
Rituximab is a chimeric murine-human monoclonal antibody directed against the
CD20 antigen. Its broad efficacy and attractive toxicity profile has resulted in its use
for the treatment of a variety of malignant and autoimmune disorders. It is generally
well tolerated with most of its side effects occurring during the first infusion, typically
brief fever and rigors.[60, 61] Since rituximab targets CD20, expression of which is
restricted to benign and malignant B-cells, it is not associated with the acute myelo-
suppression that is commonly seen following cytotoxic agents. Intriguingly however,
it is associated with a phenomenon termed ‘late-onset neutropenia’ (LON), an
idiosyncratic and relatively rare but well-recognized late complication of rituximab
containing therapy.[62] LON is defined as neutropenia occurring after neutrophil
recovery, at least 4 weeks from last therapy, and in the absence of other causes. It is
therefore distinct from the occasional episodes of neutropenia that have been reported
during the administration of rituximab monotherapy (neutropenia that developed
within 30 days after completion of treatment). The latter has been postulated as due to
the accelerated destruction of neutrophils caused by binding of rituximab–antigen
complexes to neutrophil Fc receptors.[63] LON is relatively benign phenomenon with
spontaneous neutrophil recovery being the norm, however severe neutropenic sepsis
has occurred.
Due to the lack of controlled studies, the incidence of LON has yet to be adequately
defined. Genentech, the manufacturer of rituximab, in post-marketing surveillance,
declared an overall rate of 0.02% in more than 300, 000 patients.[62] However, given
that many cases may not have been reported via this mechanism, this figure is almost
certainly an under-estimation. The majority of patients followed up post therapy have
a routine blood test only every 3 months and unless the patient has a fever related to
neutropenia, any LON occurring in this period will not be detected. LON has been
reported in association with rituximab use in a variety of pathologies, particularly
lymphoproliferative disorders but also stem cell transplantation and autoimmune
diseases.[64-68] In these reports, LON was attributed to rituximab, but the results are
difficult to interpret because of variable use of concomitant treatments and differing
rituximab regimens. There is also inconsistency in neutropenic cut-off points.
Furthermore, incidence of LON may vary between individual disorders treated with
rituximab, and disease specific estimates might provide a more accurate assessment.
9
FCGR3A is a low-affinity receptor capable of binding the FC portion of complexed
but not monomeric IgG. A polymorphism, alternatively encoding for a valine (V) or
phenylalanine (F), has been identified. FCGR3A-V158 has a higher binding affinity
for IgG1 than FCGR3A-F158.[69] There is evidence that this polymorphism is
important in the treatment of colon and breast cancers with monoclonal antibodies
such as cetuximab and transtuzumab respectively.[70, 71] Data on the impact of the
FCGR3A-V158F polymorphism in DLBCL treated with R-CHOP is conflicting.[72,
73] There are also reports that the polymorphism may contribute to the development
of late-onset neutropenia (LON).[74, 75] Although not designed for survival analysis,
it is notable that in previous case series lymphoma patients with LON have strikingly
low incidences of disease progression. It has therefore been proposed that
development of LON is a measure of good outcome, perhaps reflecting enhanced
potency of rituximab.[76] However lead-time bias needs to be considered carefully
when interpreting such data. It is important to note that many patients with poor
outcome in DLBCL are either refractory to initial therapy or relapse shortly after
completing therapy, and thus may not develop LON. It is likely that the neutropenia
caused by rituximab will be masked by salvage therapy or progressive disease.
Binding of C1q to the Fc portion of immune complexes activates CDC through
initiation of the complement cascade. C1q is encoded by C1qA, whose sole coding
polymorphism is at position 276, coding for adenine (C1qA-A276) or guanine (C1qA-
G276). C1qA-A276 results in lower C1q protein levels than the C1qA-G276
polymorphism. Breast cancer patients heterozygous or homozygous for the C1qA-
G276 genotype have a higher rate of metastasis.[77] In a study of 133 patients with
follicular lymphoma treated with single agent rituximab, possession of the C1qA-
A276 allele was associated with increased response rates and prolonged response
duration, even after adjusting for FCGR3A-V158F polymorphisms.[78] The role of
C1qA polymorphisms in DLBCL has not yet been evaluated.
Immune Parameters in DLBCL
1.5 Tumour Infiltrating Lymphocytes
There remains a critical need for identification of simple reproducible prognostic
factors in DLBCL that are capable of identifying the approximately one-third of
patients that will go on to have refractory disease or relapse early. In addition such
10
factors would be inherently more valuable if effective therapeutic modalities could be
instigated to negate the adverse outcome associated with these factors. Recently, in
addition to new therapies that target tumours directly, a number of therapies have
emerged that alter host immune status in order to improve tumour surveillance.[79-
81].
Tumor infiltration by non-malignant T-cells has been demonstrated in multiple
lymphoma histologies and a variety of solid-organ non-haematopoietic tumours.
There is heterogeneity with regards to prognostic importance of specific T cell subset
in B cell lymphomas with CD4 T-cells, CD8 T-cells and CD4 T-regulatory cells all
appearing to be of importance in predicting survival.[82-86] Even within subsets of
CD4 such as TH1 and TH2 subtypes, there is heterogeneous data with regards to
outcome. Not only is there heterogeneity within the subsets and number of these cells,
but the location of these T cells in different tumours appears important such as
whether the tumour is infiltrated at the centre of the tumour, tumour edges or tumour
stroma all giving variable results.[87, 88] Key studies have shown that culture and re-
infusion of tumour infiltrating lymphocytes (removed from the tumour site of
patients) from melanoma tumours induces effective response rates in approximately
50% of patients with advanced disease indicating importance of these cells in
cancer.[89-91]
Despite the findings in solid tumours, the data in DLBCL is surprisingly sparse with
regards to immune cell infiltration. Immune parameters have been shown to be
predictive of outcome in DLBCL.[26, 92-97] Iatrogenic immunosuppression given to
prevent rejection post organ transplantation or acquired immunosuppression induced
by HIV infection leads to high rates of lymphoma development. [98] The risk of
lymphoma development is directly related to the intensity of immune-suppression
with transplant recipients that require intense immune suppression, such as those
undergoing small bowel and multi-organ transplants, having the highest rates of
lymphoma development.[4, 99] Renal transplant patients have generally modest
immune suppression but represent the largest number of PTLD patients seen, as this
would be the most common transplantation procedure performed. Simply removing
immune suppression can lead to complete remission in some cases despite
morphology consistent with DLBCL.[100]
In an attempt to identify simple clinical markers of the immune microenvironment,
two recent studies have used circulating absolute lymphocyte counts (ALC) and
11
absolute monocyte counts (AMC) as surrogate markers of host immune-effectors and
immune-checkpoints to successfully predict outcome.[101, 102] These parameters are
an inexpensive simple assessment of immune status that are routinely performed, and
that are available to all treating physicians. While these markers assess circulating
immunity in DLBCL, little is known about how reflective they are of immune cells
within the tumour microenvironment (TME) and in particular the tumour infiltrating
lymphocytes (TILs) in DLBCL. Prior to the introduction of rituximab, it had been
shown that low levels of CD4+ T cells in the diagnostic biopsy as assessed by flow
cytometric immunophenotyping was associated with inferior outcome.[82, 83] Recent
studies have confirmed the importance of T cell activation in the tumour
microenvironment with the T cell activation marker CD137 predicting outcome in a
large DLBCL cohort.[95] This is of particular interest given an agonist antibody to
CD137 may lead to improved immune responses against lymphoma.[95, 103]
High levels of CD4+ T cells are associated with improved outcome in many
malignancies.[84] This is however tempered by the fact that CD4 T cells are
heterogeneous and are composed of subsets with variable functions such as TH1, TH2
and Tregs. There is no clear evidence about which CD4 subset contributes to improved
outcome and certainly this needs to be addressed in prospective trials. The Treg subset
are identified as CD4+FOXP3+, (or CD4+CD25hiCD127-). They are associated with
poor outcome in epithelial cancers but paradoxically appear to be associated with
improved outcome in some B cell lymphomas and within subsets of epithelial
cancers.[85, 86, 104] One explanation is that increased numbers of Tregs reflects a
relatively intact host immune system. Alternatively it has also been postulated that
Tregs have a direct negative effect on proliferation of B cells.[105] In many studies
immunohistochemical staining with FOXP3 is used to define a Treg subset and in two
reported DLBCL cohorts, higher levels of FOXP3 lymphocytes are associated with
improved outcome.[106, 107]
The second potentially important subset of CD4+T cells of interest is T helper cells. T
helper cells can be divided into two subgroups. TH1 subtype (which express the
transcription factor T-bet, and secrete interferon-γ and TNF-α) is universally
associated with good outcome in malignancy as they assist immunological clearance
of tumours. The TH2 subtype (which express the transcription factor GATA-3, and
secrete IL-4, IL-5, IL-10 and IL-13) is associated with immunosuppression, and is
12
generally associated with inferior outcome.[84, 108] There is very limited data on the
role of these various CD4 subsets in DLBCL. Interestingly there is evidence obtained
prior to the introduction of rituximab that circulating lymphocytes of patients with
DLBCL before treatment are skewed to a TH2 phenotype and revert to a TH1
phenotype with successful treatment.[109]
A number of recent animal models in B cell lymphoma have shown that CD4+ T cells
are key cells in creating an anti-tumour microenvironment.[110] TH1 cells in
particular stimulate and up-regulate antigen presentation and tumour clearance.[108]
Improved antigen presentation has been shown to be a key survival determinant in
patients with DLBCL.[37, 96, 111, 112] These animal models have shown that
cytokines derived from TH1 cells such as Interferon gamma, IL1 and TNF alpha are
implicated in stimulating macrophage mediated malignant B cell clearance.
Mouse models have also shown that PD-1 which is present on many CD4 T cells may
eventually down regulate CD4+T cell tumour surveillance leading to relapse.[110]
PD-1 has been found expressed in 27% of DLBCL tumours, but when one looks at
the tumour microenvironment 38% of PD1 non-expressing DLBCL have PD1
positive histiocytes surrounding the tumours.[113] It is possible that PD-1 could be
responsible for down-regulating CD4+ T cells in a large number of DLBCLs. This is
of particular interest as two highly successful trials targeting PD-1 in advanced
cancers have recently been described in solid tumours.[79, 80] To date two trials
using an antibody to PD-1 have been described in B cell lymphomas.[114, 115] CT-
011 (anti-PD1) directly increases the numbers of CD4+T cells post autograft for
relapsed disease with survival higher than historical controls.[115] In another study
using Ipilumimab (an anti-CTLA-4 antibody) in a range of advanced relapsed and
refractory lymphomas the antibody produced an ongoing complete remission in one
of the three patients with DLBCL.[116] Recent publication of results of blocking PD-
1 in DLBCL and follicular lymphoma show the feasibility and potential excellent
response rates using immune checkpoint blockade in B cell lymphoma however larger
studies are required to confirm these published findings.[117, 118] The study in
follicular lymphoma seems to indicate enhanced T and NK cell function as a result of
effective treatment with the anti-PD1 antibody.
Many of these molecules target immune checkpoints that are present in healthy
immune responses to restrict excessive tissue damage and contain the immune
response to where it is needed most. In some malignancies these checkpoints are
13
expressed at high levels within the tumour milieu and by tumour cells as an adaption
to escape immune attack.[119]
1.6 CD8 cells in DLBCL
CD8 T cells are the end effector cells for the immune system and are directly
cytotoxic to cells. CD8 cells with specificity for EBV epitopes have been shown in
numerous studies to elicit responses in EBV positive lymphomas.[120-122] One of
the commonest mutations found in DLBCL relates to loss of the key MHC I related
protein beta-2-microglobulin with a recent study showing this molecule mutated in
29% of DLBCL cases.[123] This study also identified deletions in CD58 that can also
affect CD8 recognition of antigen in 21% of cases, with many cases showing
mutations in both genes. In addition alternate mechanisms of aberrant expression of
MHC Class I and CD58 were frequent, indicating that cumulatively antigen
presentation to CD8 cells is defective in >60% of patients with DLBCL. Prior IHC
based studies had also shown a direct correlation between absence of immune
recognition proteins such as HLA Class I and II molecules and CD80/CD86 that was
associated with inferior outcome as well as reduced levels of CD4 and CD8
lymphoma infiltration in chemotherapy treated patients[124] However the data is
sparse with regards to the impact of CD8 infiltration and outcome in DLBCL in the
chemo-immunotherapy era. Early studies (pre rituximab) showed that poor outcome
was associated with high levels of activated CD8 cells in the TME. The authors
postulated that this led to killing of HLA-Class I expressing malignant B-cells
allowing progression of tumours lacking HLA expression.[125] In particular large
numbers of testicular and CNS lymphomas have loss of HLA expression but high
numbers of CD8 T cells and have more aggressive disease.[126] However this data is
conflicting with other studies showing high levels of CD8 infiltration in tumours with
high levels of CD40 postulated to lead to consequent enhanced autologous tumour
clearance and improved outcome.[127] It should also be noted that there is a subtype
of DLBCL associated with extensive CD8 and to a lesser extent CD4 infiltration and
relatively sparse CD20 malignant B cell population. These tumours appear to behave
more aggressively with small studies indicating that the CD8 T-cells infiltrating these
tumours are anergic and have reduced cytotoxic activity.[120] In contrast, data in an
indolent B cell lymphoma termed follicular lymphoma shows enhanced outcome
when tumours infiltrated by CD8 T cells were treated with single agent
14
rituximab[128, 129] DLBCL is an aggressive B cell lymphoma in which rituximab is
only administered in combination with anthracyline based combination chemotherapy
such as R-CHOP. It is unknown whether CD8 T cell infiltration is a predictor of
outcome of patients treated with R-CHOP in DLBCL.
1.7 Tumour associated Macrophages/Monocytes
Much data regarding the role of tumour-associated macrophages was described in
Hodgkin Lymphoma. A seminal paper by Stiedl et al. looked at two groups of
Hodgkin Lymphoma patients with differing outcomes using gene expression, and
found that those genes associated with the tumour stroma/environment were strongly
associated with poor outcome.[130] CD68 enumeration by immunohistochemistry
seemed to act as a surrogate of the adverse macrophage gene expression signature.
However not all studies have been able to replicate the IHC findings. [131, 132]
A seminal paper from Lenz et al demonstrated that gene expression related to tumour
microenvironment was predictive of outcome in DLBCL.[27] Poor outcome was
defined by a particular stromal signature associated with angiogenesis. A second
stromal signature was associated with improved outcome, and this was associated
with monocytic and histiocytic infiltration with expression of CD68 noted to be
elevated in this subgroup. In this study and in a subsequent analysis the
immunohistochemical stain for SPARC acted as a basic surrogate of this complex
gene expression signature.[133] Thus increased immune cell infiltration was
associated with improved outcome, however in this study T cell infiltration was not
prognostic and a strong MHC II immune signature was only prognostic in the cohort
of patients not treated with rituximab.
There is added complexity to the importance of tumour-associated macrophages in
DLBCL given the improved survival and standard adoption of rituximab in
combination chemotherapy for the disease. ADCC is likely one of the main effector
mechanisms causing rituximab induced death of tumour cells.[134] Macrophages are
a main contributor to this mode of cell death (along with NK cells). One of the key
reasons for the heterogeneous results with regard to macrophages in DLBCL may
relate to overreliance on CD68 staining to classify macrophages. This stain is not
specific to macrophages and other stromal cells can stain positively.[135] In addition
there are two main subtypes of macrophages found in tumour microenvironments.
These are categorized as type I and type II or M1 or M2 subtype.[136] These two
15
subtypes have distinct functions and macrophage sub-types are best viewed as a
continuum, with M1 and M2 as opposite ends of a spectrum. The M1 subtype is
associated with a pro-inflammatory response and might be considered potentially
beneficial in tumour clearance. The M2 subtype is anti-inflammatory. There is strong
evidence that many tumors manipulate their local environment to create an immune
suppressing microenvironment that contains high levels of M2 macrophages to
circumvent the natural immune response. In addition it is possible that the initial
immune effector response against a tumour is strong with little immune suppression,
but in the majority of cases this response is eventually counter-balanced by increased
immune checkpoints/suppressors that down-regulate immune-effectors so as to
prevent damage of normal tissue. Alternatively, or in addition, it is possible that
tumours manipulate their local environment to counter effective immune responses by
up-regulating immune checkpoints or down-regulating immune recognition
mechanisms. Thus a large M1 macrophage response or a strong T cell response would
be rendered ineffective if there is a large M2 macrophage response to counter it.
CD68 alone is unable to distinguish between M1 and M2 macrophages. A
combination of 2/3 markers will provide more information regarding macrophage
status although this is inconsistently performed.[136] The combination of
CD68hi/HLA-DRhi positive cells favours an M1 phenotype whereas a CD68hi/CD163hi
phenotype favours an M2 phenotype.[137, 138] HLA-DR is associated with antigen
recognition and is associated with an active immune response. CD163 is a heme
scavenger molecule, and high expression is specific for M2 macrophages.[137, 139]
However there is no definitive cut-off to distinguish between these two subtypes
using these markers, and given that there is a spectrum of macrophages from M1 to
M2, it would be extremely difficult to classify macrophages in the middle of these
two subtypes. In addition it is felt that macrophages can reverse function and
transition from M2 to M1 phenotype and vice versa depending on local cytokine
signaling.[3] The cytokines produced by the subtypes of macrophages are very
distinctive and represent the most accurate assessment of status but is relatively
difficult to measure unless one has specific cytokine arrays and fresh cells. The M1
phenotype is associated with IL1, IL6, IL12 and TNF-alpha whereas M2 is associated
with IL4, IL10, IL13 and TGF-beta production.[4, 10] These cytokines appear to
mimic those associated with CD4 T cells where CD4 TH1 are associated with immune
activation and the CD4 TH2 subtype appears similar to the M2 type macrophage.
16
Early studies in DLBCL seem to indicate that the M2 macrophage as detected by
CD163 is associated with inferior outcome and that an increased ratio of CD68 to
CD163 may indicate improved outcome in patients receiving chemo-
immunotherapy.[140] A small study from Japan showed that specific M1/M2
subtyping using CD68/HLA-DR and CD163/CD68 was also prognostic in
DLBCL.[141] However these findings are not always consistent and do not appear to
be comparable to findings prior to the introduction of rituximab.[142, 143] A large
study of R-CHOP treated patients found no prognostic impact for expression of CD68
in 262 patients, however a macrophage marker consistent with a stromal 1 signature
(SPARC) as described in Lenz et. al did predict outcome but had no correlation with
CD68.[92, 133] These findings indicate that CD68 is not specific enough for
assessment of macrophage subtype nor survival outcome, as there is likely to be
significant prognostic differences between tumours infiltrated by an M1 or M2
phenotype of macrophage, that cannot be identified using a single
immunohistochemical stain.[133, 143]
Whilst using CD163 as a more specific marker of the M2 Macrophage using
immunohistochemistry has been widely published, there is relatively little known
about the role that cells expressing CD163 play in the circulation. In agreement with
tissue expression of CD163, it is now felt that circulating cells bearing CD163
contribute to an immune-suppressive environment in many cancers. A number of
studies indicate a direct inhibition of lymphocytes by CD163. For example, in a
published study from my research group in Hodgkin Lymphoma, high circulating
levels of soluble CD163 (sCD163) were associated with significantly lower
lymphocyte count at diagnosis.[144, 145] It is felt that increased levels of these
markers may indicate systemic immune suppression triggered by the tumour cells.
There is no data of sCD163 in DLBCL. An initial study in melanoma used an ELISA
based assay to measure sCD163 levels in over 200 patients with early stage
disease.[146] High levels of serum sCD163 were associated with inferior outcome
though not necessarily disease specific survival. Interestingly, in this study levels of
CD163 in patients were not consistently higher than healthy control participants. In
Hodgkin Lymphoma, circulating sCD163 is markedly elevated at diagnosis compared
to controls and acts as a disease response biomarker.[145]
17
1.8 Myeloid Derived Suppressor Cells
Recent studies in NHL indicate that particular subsets of monocytes may have a
phenotype consistent with that of an M2 macrophage. Lin et al described a myeloid
suppressor subset that was associated with immune dysfunction and suppression that
seemed to be associated with inferior outcome in a heterogeneous B-NHL
population.[147] These cells that have been described in a number of malignancies
are characterised by a phenotype of high CD14+ and low HLA-DR expression.[148,
149] Circulating monocytes infiltrate tissue such as lymph nodes to become
macrophages so their phenotype may be critical in determining the tumour
microenvironment. Whilst there is data supporting inferior outcome in DLBCL
relating to high CD163 levels as discussed, this is not always consistent. There is no
data investigating the relative interactions of CD163 in tissue or the circulation with
immune effectors such as CD4 and CD8 in DLBCL. CD163 is only one of many
molecules that may counter effective immune responses.
In addition there is minimal data on the relative impact on expression of key immune
checkpoints in DLBCL. As described, targeting the immune checkpoints CTLA4,
PD-1, PDL-1 and PDL-2 have all shown promise in the treatment of melanoma, renal
cancer, lung cancer and there is emerging data on the targeting of these pathways in
lymphoma.[84, 150, 151] In DLBCL responses have been seen in patients treated
with an anti-CTLA4 antibody and adjuvant use of an anti-PD1 antibody post stem
autograft in relapsed patients appears to improve outcome.[116, 118, 152] These
therapies are relatively well tolerated and early evidence suggests excellent activity
and safety in combining these agents with anti-CTLA4 and anti-PD1 therapy
combined therapy giving remarkable responses in metastatic melanoma.[153] In
addition, targeting of checkpoints such as LAG3 and TIM3 will soon enter clinical
trials.[150] There are many other molecules of interest in this field which have yet to
be fully investigated.[154] Little is known on how tumour infiltration by cells of the
immune system might dictate response not only to current standard therapy but also
the new wave of immune modifying agents.
1.9 Digital multiplex gene expression (DMGE) by nanoString nCounter.
NanoString is a new novel technology first described in 2009.[155] It captures and
measures mRNA transcripts without any amplification or enzymatic steps. Each gene
is detected by a probe with a particular color based barcoding system which then
18
attaches to a capture probe which allows the reporter probe to be immobilized. Excess
probes and capture are washed off and a digital reader calculates the relative
expression of the mRNA transcript based on the colour code of the reporter probe.
Given this unique mechanism the bioinformatics load is relatively mild and data can
easily be generated via a simple analysis program provided by the company. The
correlation between runs is highly reproducible but more importantly data from
matched frozen and paraffin tissue samples shows exceptionally high correlation
which has now been replicated in a number of published studies.[156-158] The
quality of data from paraffin tissues opens up much larger banks of lymphoma
samples from retrospective databases as good quality gene expression data no longer
requires snap frozen tissue. Indeed an early study in DLBCL has shown that
NanoString will likely be highly sensitive at accurately determining cell of origin in
DLBCL.[159] Early results from a small series of aggressive large cells lymphomas
indicated that Nanostring technology could be effective at distinguishing between
standard DLBCL and aggressive subtypes more closely related to Burkitt Lymphoma,
however this study was small.[160] To date these two relatively small studies are the
only research using digital gene barcoding in DLBCL. Nanostring is now an accepted
diagnostic platform approved by the FDA for assessment of prognosis in Breast
Cancer patients with recent approval of the PAM50 gene signature.[161]
References:
1. Swerdlow, S.H., ed. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. 2008.
2. Garber, K., Lymphoma rate rise continues to baffle researchers. Journal of the National Cancer Institute, 2001. 93(7): p. 494-‐6.
3. Zimmermann, H. and R.U. Trappe, EBV and posttransplantation lymphoproliferative disease: what to do? Hematology / the Education Program of the American Society of Hematology. American Society of Hematology. Education Program, 2013. 2013: p. 95-‐102.
4. van Leeuwen, M.T., et al., Immunosuppression and other risk factors for early and late non-‐Hodgkin lymphoma after kidney transplantation. Blood, 2009. 114(3): p. 630-‐7.
5. Guiguet, M., et al., Effect of immunodeficiency, HIV viral load, and antiretroviral therapy on the risk of individual malignancies (FHDH-‐ANRS CO4): a prospective cohort study. The lancet oncology, 2009. 10(12): p. 1152-‐9.
19
6. Bulka, C., et al., Residence proximity to benzene release sites is associatedwith increased incidence of non-‐Hodgkin lymphoma. Cancer, 2013.119(18): p. 3309-‐17.
7. De Roos, A.J., et al., Residential proximity to industrial facilities and risk ofnon-‐Hodgkin lymphoma. Environmental research, 2010. 110(1): p. 70-‐8.
8. Craig, F.E. and K.A. Foon, Flow cytometric immunophenotyping forhematologic neoplasms. Blood, 2008. 111(8): p. 3941-‐67.
9. Perry, A.M., et al., B-‐cell lymphoma, unclassifiable, with featuresintermediate between diffuse large B-‐cell lymphoma and burkitt lymphoma:study of 39 cases. British journal of haematology, 2013. 162(1): p. 40-‐9.
10. Perry, A.M., et al., MYC and BCL2 protein expression predicts survival inpatients with diffuse large B-‐cell lymphoma treated with rituximab. Britishjournal of haematology, 2014.
11. Tzankov, A., et al., Rearrangements of MYC gene facilitate risk stratificationin diffuse large B-‐cell lymphoma patients treated with rituximab-‐CHOP.Modern pathology : an official journal of the United States and CanadianAcademy of Pathology, Inc, 2013.
12. Savage, K.J., et al., MYC gene rearrangements are associated with a poorprognosis in diffuse large B-‐cell lymphoma patients treated with R-‐CHOPchemotherapy. Blood, 2009. 114(17): p. 3533-‐7.
13. Coiffier, B., et al., Rituximab (anti-‐CD20 monoclonal antibody) for thetreatment of patients with relapsing or refractory aggressive lymphoma: amulticenter phase II study. Blood, 1998. 92(6): p. 1927-‐32.
14. Armitage, J.O., My treatment approach to patients with diffuse large B-‐celllymphoma. Mayo Clinic proceedings, 2012. 87(2): p. 161-‐71.
15. Ziepert, M., et al., Standard International prognostic index remains a validpredictor of outcome for patients with aggressive CD20+ B-‐cell lymphomain the rituximab era. Journal of clinical oncology : official journal of theAmerican Society of Clinical Oncology, 2010. 28(14): p. 2373-‐80.
16. Coiffier, B., et al., CHOP chemotherapy plus rituximab compared with CHOPalone in elderly patients with diffuse large-‐B-‐cell lymphoma. The NewEngland journal of medicine, 2002. 346(4): p. 235-‐42.
17. Sehn, L.H., et al., Introduction of combined CHOP plus rituximab therapydramatically improved outcome of diffuse large B-‐cell lymphoma in BritishColumbia. Journal of clinical oncology : official journal of the AmericanSociety of Clinical Oncology, 2005. 23(22): p. 5027-‐33.
18. A predictive model for aggressive non-‐Hodgkin's lymphoma. TheInternational Non-‐Hodgkin's Lymphoma Prognostic Factors Project. TheNew England journal of medicine, 1993. 329(14): p. 987-‐94.
19. Sehn, L.H., et al., The revised International Prognostic Index (R-‐IPI) is abetter predictor of outcome than the standard IPI for patients with diffuselarge B-‐cell lymphoma treated with R-‐CHOP. Blood, 2007. 109(5): p. 1857-‐61.
20. Roschewski, M., L.M. Staudt, and W.H. Wilson, Diffuse large B-‐celllymphoma-‐treatment approaches in the molecular era. Nature reviews.Clinical oncology, 2014. 11(1): p. 12-‐23.
21. Alizadeh, A.A., et al., Distinct types of diffuse large B-‐cell lymphomaidentified by gene expression profiling. Nature, 2000. 403(6769): p. 503-‐11.
20
22. Vari, F. and M.K. Gandhi, Back to basics: the complete blood cell count adds to the ability of immunohistochemistry in diffuse large B-‐cell lymphoma prognosis. Leukemia & lymphoma, 2012. 53(11): p. 2097-‐8.
23. Wright, G., et al., A gene expression-‐based method to diagnose clinically distinct subgroups of diffuse large B cell lymphoma. Proceedings of the National Academy of Sciences of the United States of America, 2003. 100(17): p. 9991-‐6.
24. Hans, C.P., et al., Confirmation of the molecular classification of diffuse large B-‐cell lymphoma by immunohistochemistry using a tissue microarray. Blood, 2004. 103(1): p. 275-‐82.
25. Choi, W.W., et al., A new immunostain algorithm classifies diffuse large B-‐cell lymphoma into molecular subtypes with high accuracy. Clinical cancer research : an official journal of the American Association for Cancer Research, 2009. 15(17): p. 5494-‐502.
26. Rosenwald, A., et al., The use of molecular profiling to predict survival after chemotherapy for diffuse large-‐B-‐cell lymphoma. The New England journal of medicine, 2002. 346(25): p. 1937-‐47.
27. Lenz, G., et al., Stromal gene signatures in large-‐B-‐cell lymphomas. The New England journal of medicine, 2008. 359(22): p. 2313-‐23.
28. Dunleavy, K., et al., BCL-‐6 and rituximab in diffuse large B-‐cell lymphoma: where are we? Blood, 2007. 109(2): p. 843-‐4; discussion 844-‐5.
29. Shipp, M.A., et al., Diffuse large B-‐cell lymphoma outcome prediction by gene-‐expression profiling and supervised machine learning. Nature medicine, 2002. 8(1): p. 68-‐74.
30. Lossos, I.S., Diffuse large B cell lymphoma: from gene expression profiling to prediction of outcome. Biology of blood and marrow transplantation : journal of the American Society for Blood and Marrow Transplantation, 2008. 14(1 Suppl 1): p. 108-‐11.
31. Lossos, I.S., et al., Prediction of survival in diffuse large-‐B-‐cell lymphoma based on the expression of six genes. The New England journal of medicine, 2004. 350(18): p. 1828-‐37.
32. Malumbres, R., et al., Paraffin-‐based 6-‐gene model predicts outcome in diffuse large B-‐cell lymphoma patients treated with R-‐CHOP. Blood, 2008. 111(12): p. 5509-‐14.
33. Dunleavy, K., et al., Differential efficacy of bortezomib plus chemotherapy within molecular subtypes of diffuse large B-‐cell lymphoma. Blood, 2009. 113(24): p. 6069-‐76.
34. Wilson, W.H., et al., Phase II study of dose-‐adjusted EPOCH and rituximab in untreated diffuse large B-‐cell lymphoma with analysis of germinal center and post-‐germinal center biomarkers. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2008. 26(16): p. 2717-‐24.
35. Mathews Griner, L.A., et al., High-‐throughput combinatorial screening identifies drugs that cooperate with ibrutinib to kill activated B-‐cell-‐like diffuse large B-‐cell lymphoma cells. Proceedings of the National Academy of Sciences of the United States of America, 2014.
36. Choi, M.Y. and T.J. Kipps, Inhibitors of B-‐cell receptor signaling for patients with B-‐cell malignancies. Cancer journal, 2012. 18(5): p. 404-‐10.
21
37. Ott, G., et al., Immunoblastic morphology but not the immunohistochemicalGCB/nonGCB classifier predicts outcome in diffuse large B-‐cell lymphoma inthe RICOVER-‐60 trial of the DSHNHL. Blood, 2010. 116(23): p. 4916-‐25.
38. Seki, R., et al., Prognostic impact of immunohistochemical biomarkers indiffuse large B-‐cell lymphoma in the rituximab era. Cancer science, 2009.100(10): p. 1842-‐7.
39. Gutierrez-‐Garcia, G., et al., Gene-‐expression profiling and notimmunophenotypic algorithms predicts prognosis in patients with diffuselarge B-‐cell lymphoma treated with immunochemotherapy. Blood, 2011.117(18): p. 4836-‐43.
40. Fu, K., et al., Addition of rituximab to standard chemotherapy improves thesurvival of both the germinal center B-‐cell-‐like and non-‐germinal center B-‐cell-‐like subtypes of diffuse large B-‐cell lymphoma. Journal of clinicaloncology : official journal of the American Society of Clinical Oncology,2008. 26(28): p. 4587-‐94.
41. Nyman, H., et al., Prognostic impact of immunohistochemically definedgerminal center phenotype in diffuse large B-‐cell lymphoma patientstreated with immunochemotherapy. Blood, 2007. 109(11): p. 4930-‐5.
42. Meyer, P.N., et al., Immunohistochemical methods for predicting cell oforigin and survival in patients with diffuse large B-‐cell lymphoma treatedwith rituximab. Journal of clinical oncology : official journal of theAmerican Society of Clinical Oncology, 2011. 29(2): p. 200-‐7.
43. de Jong, D., et al., Immunohistochemical prognostic markers in diffuse largeB-‐cell lymphoma: validation of tissue microarray as a prerequisite for broadclinical applications-‐-‐a study from the Lunenburg Lymphoma BiomarkerConsortium. Journal of clinical oncology : official journal of the AmericanSociety of Clinical Oncology, 2007. 25(7): p. 805-‐12.
44. de Jong, D., et al., Immunohistochemical prognostic markers in diffuse largeB-‐cell lymphoma: validation of tissue microarray as a prerequisite for broadclinical applications (a study from the Lunenburg Lymphoma BiomarkerConsortium). Journal of clinical pathology, 2009. 62(2): p. 128-‐38.
45. Shivakumar, L. and J.O. Armitage, Bcl-‐2 gene expression as a predictor ofoutcome in diffuse large B-‐cell lymphoma. Clinical lymphoma & myeloma,2006. 6(6): p. 455-‐7.
46. Iqbal, J., et al., BCL2 expression is a prognostic marker for the activated B-‐cell-‐like type of diffuse large B-‐cell lymphoma. Journal of clinical oncology :official journal of the American Society of Clinical Oncology, 2006. 24(6):p. 961-‐8.
47. Kulkarni, M.M., Digital multiplexed gene expression analysis using theNanoString nCounter system. Current protocols in molecular biology /edited by Frederick M. Ausubel ... [et al.], 2011. Chapter 25: p. Unit25B10.
48. de Haij, S., et al., In vivo Cytotoxicity of Type I CD20 Antibodies CriticallyDepends on Fc Receptor ITAM Signaling. Cancer Research, 2010. 70(8): p.3209-‐3217.
49. Uchida, J., et al., The Innate Mononuclear Phagocyte Network Depletes BLymphocytes through Fc Receptor–dependent Mechanisms during Anti-‐CD20 Antibody Immunotherapy. The Journal of Experimental Medicine,2004. 199(12): p. 1659-‐1669.
22
50. Clynes, R.A., et al., Inhibitory Fc receptors modulate in vivo cytoxicityagainst tumor targets. Nat Med, 2000. 6(4): p. 443-‐6.
51. Minard-‐Colin, V., et al., Lymphoma depletion during CD20 immunotherapyin mice is mediated by macrophage FcgammaRI, FcgammaRIII, andFcgammaRIV. Blood, 2008. 112(4): p. 1205-‐13.
52. Di Gaetano, N., et al., Complement activation determines the therapeuticactivity of rituximab in vivo. J Immunol, 2003. 171(3): p. 1581-‐7.
53. Golay, J., et al., CD20 levels determine the in vitro susceptibility to rituximaband complement of B-‐cell chronic lymphocytic leukemia: further regulationby CD55 and CD59. Blood, 2001. 98(12): p. 3383-‐9.
54. Golay, J., et al., Biologic response of B lymphoma cells to anti-‐CD20monoclonal antibody rituximab in vitro: CD55 and CD59 regulatecomplement-‐mediated cell lysis. Blood, 2000. 95(12): p. 3900-‐8.
55. Kennedy, A.D., et al., An anti-‐C3b(i) mAb enhances complement activation,C3b(i) deposition, and killing of CD20+ cells by rituximab. Blood, 2003.101(3): p. 1071-‐9.
56. Kennedy, A.D., et al., Rituximab infusion promotes rapid complementdepletion and acute CD20 loss in chronic lymphocytic leukemia. J Immunol,2004. 172(5): p. 3280-‐8.
57. Byrd, J.C., et al., The mechanism of tumor cell clearance by rituximab in vivoin patients with B-‐cell chronic lymphocytic leukemia: evidence of caspaseactivation and apoptosis induction. Blood, 2002. 99(3): p. 1038-‐43.
58. Cardarelli, P.M., et al., Binding to CD20 by anti-‐B1 antibody or F(ab')(2) issufficient for induction of apoptosis in B-‐cell lines. Cancer immunology,immunotherapy : CII, 2002. 51(1): p. 15-‐24.
59. Shan, D., J.A. Ledbetter, and O.W. Press, Apoptosis of malignant human Bcells by ligation of CD20 with monoclonal antibodies. Blood, 1998. 91(5): p.1644-‐52.
60. Byrd, J.C., et al., Rituximab therapy in hematologic malignancy patientswith circulating blood tumor cells: association with increased infusion-‐related side effects and rapid blood tumor clearance. Journal of clinicaloncology : official journal of the American Society of Clinical Oncology,1999. 17(3): p. 791-‐5.
61. Winkler, U., et al., Cytokine-‐release syndrome in patients with B-‐cell chroniclymphocytic leukemia and high lymphocyte counts after treatment with ananti-‐CD20 monoclonal antibody (rituximab, IDEC-‐C2B8). Blood, 1999.94(7): p. 2217-‐24.
62. Voog, E., F. Morschhauser, and P. Solal-‐Celigny, Neutropenia in patientstreated with rituximab. The New England journal of medicine, 2003.348(26): p. 2691-‐4; discussion 2691-‐4.
63. Chaiwatanatorn, K., et al., Delayed-‐onset neutropenia associated withrituximab therapy. British journal of haematology, 2003. 121(6): p. 913-‐8.
64. Tesfa, D., et al., Late-‐onset neutropenia following rituximab therapy inrheumatic diseases: association with B lymphocyte depletion and infections.Arthritis and rheumatism, 2011. 63(8): p. 2209-‐14.
65. Besada, E., W. Koldingsnes, and J. Nossent, Characteristics of late onsetneutropenia in rheumatologic patients treated with rituximab: a casereview analysis from a single center. QJM : monthly journal of theAssociation of Physicians, 2012. 105(6): p. 545-‐50.
23
66. Rozman, S., M. Sonc, and B.J. Novakovic, Late-‐onset neutropenia followingprimary treatment of diffuse large B-‐cell lymphoma with rituximab-‐containing therapy. Leukemia & lymphoma, 2012. 53(10): p. 1945-‐8.
67. Nitta, E., et al., A high incidence of late-‐onset neutropenia followingrituximab-‐containing chemotherapy as a primary treatment of CD20-‐positive B-‐cell lymphoma: a single-‐institution study. Annals of oncology :official journal of the European Society for Medical Oncology / ESMO,2007. 18(2): p. 364-‐9.
68. Weng, W.K., et al., Immunoglobulin G Fc receptor FcgammaRIIIa 158 V/Fpolymorphism correlates with rituximab-‐induced neutropenia afterautologous transplantation in patients with non-‐Hodgkin's lymphoma.Journal of clinical oncology : official journal of the American Society ofClinical Oncology, 2010. 28(2): p. 279-‐84.
69. Koene, H.R., et al., Fc gammaRIIIa-‐158V/F polymorphism influences thebinding of IgG by natural killer cell Fc gammaRIIIa, independently of the FcgammaRIIIa-‐48L/R/H phenotype. Blood, 1997. 90(3): p. 1109-‐14.
70. Musolino, A., et al., Immunoglobulin G fragment C receptor polymorphismsand clinical efficacy of trastuzumab-‐based therapy in patients with HER-‐2/neu-‐positive metastatic breast cancer. J Clin Oncol, 2008. 26(11): p.1789-‐96.
71. Zhang, W., et al., FCGR2A and FCGR3A polymorphisms associated withclinical outcome of epidermal growth factor receptor expressing metastaticcolorectal cancer patients treated with single-‐agent cetuximab. J Clin Oncol,2007. 25(24): p. 3712-‐8.
72. Kim, D.H., et al., FCGR3A gene polymorphisms may correlate with responseto frontline R-‐CHOP therapy for diffuse large B-‐cell lymphoma. Blood, 2006.108(8): p. 2720-‐5.
73. Mitrovic, Z., et al., FC{gamma}RIIIA and FC{gamma}RIIA polymorphismsare not associated with response to rituximab and CHOP in patients withdiffuse large B-‐cell lymphoma. Haematologica, 2007. 92(7): p. 998-‐999.
74. Li, S.C., et al., Rituximab-‐induced late-‐onset neutropenia in newly diagnosedB-‐cell lymphoma correlates with Fc receptor FcgammaRIIIa 158(V/F)polymorphism. Am J Hematol, 2010. 85(10): p. 810-‐2.
75. Weng, W.K., et al., Immunoglobulin G Fc receptor FcgammaRIIIa 158 V/Fpolymorphism correlates with rituximab-‐induced neutropenia afterautologous transplantation in patients with non-‐Hodgkin's lymphoma. JClin Oncol, 2010. 28(2): p. 279-‐84.
76. Hincks, I., B.E. Woodcock, and J. Thachil, Is Rituximab-‐induced Late-‐onsetNeutropenia a good prognostic indicator in lymphoproliferatve disorders?Br J Haematol, 2011.
77. Racila, E., et al., The pattern of clinical breast cancer metastasis correlateswith a single nucleotide polymorphism in the C1qA component ofcomplement. Immunogenetics, 2006. 58(1): p. 1-‐8.
78. Racila, E., et al., A polymorphism in the complement component C1qAcorrelates with prolonged response following rituximab therapy of follicularlymphoma. Clin Cancer Res, 2008. 14(20): p. 6697-‐703.
79. Topalian, S.L., et al., Safety, activity, and immune correlates of anti-‐PD-‐1antibody in cancer. The New England journal of medicine, 2012. 366(26):p. 2443-‐54.
24
80. Brahmer, J.R., et al., Safety and activity of anti-‐PD-‐L1 antibody in patients with advanced cancer. The New England journal of medicine, 2012. 366(26): p. 2455-‐65.
81. Witzig, T.E., et al., An international phase II trial of single-‐agent lenalidomide for relapsed or refractory aggressive B-‐cell non-‐Hodgkin's lymphoma. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO, 2011. 22(7): p. 1622-‐7.
82. Xu, Y., et al., Prognostic significance of tumour-‐infiltrating T lymphocytes and T-‐cell subsets in de novo diffuse large B-‐cell lymphoma: a multiparameter flow cytometry study. British journal of haematology, 2001. 112(4): p. 945-‐9.
83. Ansell, S.M., et al., Cd4+ T-‐cell immune response to large B-‐cell non-‐Hodgkin's lymphoma predicts patient outcome. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2001. 19(3): p. 720-‐6.
84. Fridman, W.H., et al., The immune contexture in human tumours: impact on clinical outcome. Nature reviews. Cancer, 2012. 12(4): p. 298-‐306.
85. deLeeuw, R.J., et al., The prognostic value of FoxP3+ tumor-‐infiltrating lymphocytes in cancer: a critical review of the literature. Clinical cancer research : an official journal of the American Association for Cancer Research, 2012. 18(11): p. 3022-‐9.
86. Lee, A.M., et al., Number of CD4+ cells and location of forkhead box protein P3-‐positive cells in diagnostic follicular lymphoma tissue microarrays correlates with outcome. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2006. 24(31): p. 5052-‐9.
87. Mei, Z., et al., Tumour-‐infiltrating inflammation and prognosis in colorectal cancer: systematic review and meta-‐analysis. British journal of cancer, 2014.
88. Seo, A.N., et al., Tumour-‐infiltrating CD8+ lymphocytes as an independent predictive factor for pathological complete response to primary systemic therapy in breast cancer. British journal of cancer, 2013. 109(10): p. 2705-‐13.
89. Besser, M.J., et al., Adoptive transfer of tumor-‐infiltrating lymphocytes in patients with metastatic melanoma: intent-‐to-‐treat analysis and efficacy after failure to prior immunotherapies. Clinical cancer research : an official journal of the American Association for Cancer Research, 2013. 19(17): p. 4792-‐800.
90. Nguyen, L.T., et al., Expansion and characterization of human melanoma tumor-‐infiltrating lymphocytes (TILs). PloS one, 2010. 5(11): p. e13940.
91. Dudley, M.E., et al., Adoptive cell therapy for patients with metastatic melanoma: evaluation of intensive myeloablative chemoradiation preparative regimens. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2008. 26(32): p. 5233-‐9.
92. Lenz, G., et al., Molecular subtypes of diffuse large B-‐cell lymphoma arise by distinct genetic pathways. Proceedings of the National Academy of Sciences of the United States of America, 2008. 105(36): p. 13520-‐5.
93. Monti, S., et al., Molecular profiling of diffuse large B-‐cell lymphoma identifies robust subtypes including one characterized by host inflammatory response. Blood, 2005. 105(5): p. 1851-‐61.
25
94. Hasselblom, S., et al., The number of tumour-‐infiltrating TIA-‐1+ cytotoxic Tcells but not FOXP3+ regulatory T cells predicts outcome in diffuse large B-‐cell lymphoma. British journal of haematology, 2007. 137(4): p. 364-‐73.
95. Alizadeh, A.A., et al., Prediction of survival in diffuse large B-‐cell lymphomabased on the expression of 2 genes reflecting tumor and microenvironment.Blood, 2011. 118(5): p. 1350-‐8.
96. Rimsza, L.M., et al., Loss of MHC class II gene and protein expression indiffuse large B-‐cell lymphoma is related to decreased tumorimmunosurveillance and poor patient survival regardless of otherprognostic factors: a follow-‐up study from the Leukemia and LymphomaMolecular Profiling Project. Blood, 2004. 103(11): p. 4251-‐8.
97. Phipps-‐Yonas, H., et al., Low GILT Expression is Associated with PoorPatient Survival in Diffuse Large B-‐Cell Lymphoma. Frontiers inimmunology, 2013. 4: p. 425.
98. Tran, H., et al., Immunodeficiency-‐associated lymphomas. Blood reviews,2008. 22(5): p. 261-‐81.
99. Quinlan, S.C., et al., Risk factors for early-‐onset and late-‐onset post-‐transplant lymphoproliferative disorder in kidney recipients in the UnitedStates. American journal of hematology, 2011. 86(2): p. 206-‐9.
100. Porcu, P., et al., Successful treatment of posttransplantation lymphoproliferative disorder (PTLD) following renal allografting is associated with sustained CD8(+) T-‐cell restoration. Blood, 2002. 100(7): p. 2341-‐8.
101. Wilcox, R.A., et al., The absolute monocyte and lymphocyte prognostic score predicts survival and identifies high-‐risk patients in diffuse large-‐B-‐cell lymphoma. Leukemia : official journal of the Leukemia Society of America, Leukemia Research Fund, U.K, 2011. 25(9): p. 1502-‐9.
102. Li, Z.M., et al., Blood Lymphocyte-‐to-‐Monocyte Ratio Identifies High-‐Risk Patients in Diffuse Large B-‐Cell Lymphoma Treated with R-‐CHOP. PloS one, 2012. 7(7): p. e41658.
103. Houot, R., et al., Therapeutic effect of CD137 immunomodulation in lymphoma and its enhancement by Treg depletion. Blood, 2009. 114(16): p. 3431-‐8.
104. Droeser, R., et al., Differential pattern and prognostic significance of CD4+, FOXP3+ and IL-‐17+ tumor infiltrating lymphocytes in ductal and lobular breast cancers. BMC cancer, 2012. 12: p. 134.
105. Yang, Z.Z., et al., Intratumoral CD4+CD25+ regulatory T-‐cell-‐mediated suppression of infiltrating CD4+ T cells in B-‐cell non-‐Hodgkin lymphoma. Blood, 2006. 107(9): p. 3639-‐46.
106. Tzankov, A., et al., Correlation of high numbers of intratumoral FOXP3+ regulatory T cells with improved survival in germinal center-‐like diffuse large B-‐cell lymphoma, follicular lymphoma and classical Hodgkin's lymphoma. Haematologica, 2008. 93(2): p. 193-‐200.
107. Coutinho R, C.A., Owen D, Greaves P, Hallam S, Wilson A, Matthews J, Calaminici M, Gribben J, , Exploring the Immune Microenvironment of Diffuse Large B Cell Lymphoma in a Tissue Microarray: Predicting Survival with a Score That Incorporates Macrophages, Cytotoxic and Regulatory T Cells, in Oral Abstract 951 American Society of Haematology Annual Meeting2011: San Diego.
26
108. Haabeth, O.A., et al., Inflammation driven by tumour-‐specific Th1 cells protects against B-‐cell cancer. Nature communications, 2011. 2: p. 240.
109. Mori, T., et al., T-‐helper (Th)1/Th2 imbalance in patients with previously untreated B-‐cell diffuse large cell lymphoma. Cancer immunology, immunotherapy : CII, 2001. 50(10): p. 566-‐8.
110. Ding, Z.C., et al., Polyfunctional CD4+ T cells are essential for eradicating advanced B-‐cell lymphoma after chemotherapy. Blood, 2012.
111. Bernd, H.W., et al., Loss of HLA-‐DR expression and immunoblastic morphology predict adverse outcome in diffuse large B-‐cell lymphoma -‐ analyses of cases from two prospective randomized clinical trials. Haematologica, 2009. 94(11): p. 1569-‐80.
112. Pasqualucci, L., et al., Analysis of the coding genome of diffuse large B-‐cell lymphoma. Nature genetics, 2011. 43(9): p. 830-‐7.
113. Andorsky, D.J., et al., Programmed death ligand 1 is expressed by non-‐hodgkin lymphomas and inhibits the activity of tumor-‐associated T cells. Clinical cancer research : an official journal of the American Association for Cancer Research, 2011. 17(13): p. 4232-‐44.
114. Berger, R., et al., Phase I safety and pharmacokinetic study of CT-‐011, a humanized antibody interacting with PD-‐1, in patients with advanced hematologic malignancies. Clinical cancer research : an official journal of the American Association for Cancer Research, 2008. 14(10): p. 3044-‐51.
115. Rotem-‐Yehudar R, A.P., Devine S, et al., 1216 the human-‐ized anti PD-‐1 antibody, CT-‐011, increases specific CD4+ effector/memory and memory T lymphocytes in patients with diffuse large B cell lymphoma (DLBCL) following autologous stem cell transplantation (AuSCT). Blood (ASH Annual Meeting Abstracts). 2009;114:1216.
116. Ansell, S.M., et al., Phase I study of ipilimumab, an anti-‐CTLA-‐4 monoclonal antibody, in patients with relapsed and refractory B-‐cell non-‐Hodgkin lymphoma. Clinical cancer research : an official journal of the American Association for Cancer Research, 2009. 15(20): p. 6446-‐53.
117. Westin, J.R., et al., Safety and activity of PD1 blockade by pidilizumab in combination with rituximab in patients with relapsed follicular lymphoma: a single group, open-‐label, phase 2 trial. The lancet oncology, 2014. 15(1): p. 69-‐77.
118. Armand, P., et al., Disabling immune tolerance by programmed death-‐1 blockade with pidilizumab after autologous hematopoietic stem-‐cell transplantation for diffuse large B-‐cell lymphoma: results of an international phase II trial. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2013. 31(33): p. 4199-‐206.
119. Taube, J.M., et al., Colocalization of inflammatory response with B7-‐h1 expression in human melanocytic lesions supports an adaptive resistance mechanism of immune escape. Science translational medicine, 2012. 4(127): p. 127ra37.
120. Felgar, R.E., et al., T-‐cell-‐rich large-‐B-‐cell lymphomas contain non-‐activated CD8+ cytolytic T cells, show increased tumor cell apoptosis, and have lower Bcl-‐2 expression than diffuse large-‐B-‐cell lymphomas. The American journal of pathology, 1998. 153(6): p. 1707-‐15.
27
121. Bollard, C.M., L.J. Cooper, and H.E. Heslop, Immunotherapy targeting EBV-‐expressing lymphoproliferative diseases. Best practice & research. Clinical haematology, 2008. 21(3): p. 405-‐20.
122. Kanakry, J.A. and R.F. Ambinder, EBV-‐related lymphomas: new approaches to treatment. Current treatment options in oncology, 2013. 14(2): p. 224-‐36.
123. Challa-‐Malladi, M., et al., Combined genetic inactivation of beta2-‐Microglobulin and CD58 reveals frequent escape from immune recognition in diffuse large B cell lymphoma. Cancer cell, 2011. 20(6): p. 728-‐40.
124. List, A.F., et al., Deficient tumor-‐infiltrating T-‐lymphocyte response in malignant lymphoma: relationship to HLA expression and host immunocompetence. Leukemia : official journal of the Leukemia Society of America, Leukemia Research Fund, U.K, 1993. 7(3): p. 398-‐403.
125. Galand, C., et al., Influence of Tumor Location on the Composition of Immune Infiltrate and Its Impact on Patient Survival. Lessons from DCBCL and Animal Models. Frontiers in immunology, 2012. 3: p. 98.
126. Riemersma, S.A., et al., High numbers of tumour-‐infiltrating activated cytotoxic T lymphocytes, and frequent loss of HLA class I and II expression, are features of aggressive B cell lymphomas of the brain and testis. The Journal of pathology, 2005. 206(3): p. 328-‐36.
127. Rydstrom, K., et al., Gene expression profiling indicates that immunohistochemical expression of CD40 is a marker of an inflammatory reaction in the tumor stroma of diffuse large B-‐cell lymphoma. Leukemia & lymphoma, 2012. 53(9): p. 1764-‐8.
128. Wahlin, B.E., et al., CD8+ T-‐cell content in diagnostic lymph nodes measured by flow cytometry is a predictor of survival in follicular lymphoma. Clinical cancer research : an official journal of the American Association for Cancer Research, 2007. 13(2 Pt 1): p. 388-‐97.
129. Wahlin, B.E., et al., T cells in tumors and blood predict outcome in follicular lymphoma treated with rituximab. Clinical cancer research : an official journal of the American Association for Cancer Research, 2011. 17(12): p. 4136-‐44.
130. Steidl, C., et al., Tumor-‐associated macrophages and survival in classic Hodgkin's lymphoma. The New England journal of medicine, 2010. 362(10): p. 875-‐85.
131. Steidl, C., J.M. Connors, and R.D. Gascoyne, Molecular pathogenesis of Hodgkin's lymphoma: increasing evidence of the importance of the microenvironment. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2011. 29(14): p. 1812-‐26.
132. Greaves, P., et al., Expression of FOXP3, CD68, and CD20 at diagnosis in the microenvironment of classical Hodgkin lymphoma is predictive of outcome. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2013. 31(2): p. 256-‐62.
133. Meyer, P.N., et al., The stromal cell marker SPARC predicts for survival in patients with diffuse large B-‐cell lymphoma treated with rituximab. American journal of clinical pathology, 2011. 135(1): p. 54-‐61.
134. Keane, C., et al., Homozygous FCGR3A-‐158V alleles predispose to late onset neutropenia after CHOP-‐R for diffuse large B-‐cell lymphoma. Internal medicine journal, 2012. 42(10): p. 1113-‐9.
28
135. Muller, C., et al., The role of sex and weight on rituximab clearance and serum elimination half-‐life in elderly patients with DLBCL. Blood, 2012. 119(14): p. 3276-‐84.
136. Kristiansen, M., et al., Identification of the haemoglobin scavenger receptor. Nature, 2001. 409(6817): p. 198-‐201.
137. Schaer, D.J., et al., CD163 is the macrophage scavenger receptor for native and chemically modified hemoglobins in the absence of haptoglobin. Blood, 2006. 107(1): p. 373-‐80.
138. Natkunam, Y., et al., LMO2 protein expression predicts survival in patients with diffuse large B-‐cell lymphoma treated with anthracycline-‐based chemotherapy with and without rituximab. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2008. 26(3): p. 447-‐54.
139. Mikami, M., et al., Monocyte/Macrophage-‐specific marker CD163+ histiocytic sarcoma: case report with clinical, morphologic, immunohistochemical, and molecular genetic studies. International journal of hematology, 2004. 80(4): p. 365-‐9.
140. Nam, S.J., et al., An increase of M2 macrophages predicts poor prognosis in diffuse large B-‐cell lymphoma patients treated with Rituximab-‐CHOP. Leukemia & lymphoma, 2014.
141. Wada, N., et al., Tumour-‐associated macrophages in diffuse large B-‐cell lymphoma: a study of the Osaka Lymphoma Study Group. Histopathology, 2012. 60(2): p. 313-‐9.
142. Hasselblom, S., et al., Expression of CD68+ tumor-‐associated macrophages in patients with diffuse large B-‐cell lymphoma and its relation to prognosis. Pathology international, 2008. 58(8): p. 529-‐32.
143. Cai, Q.C., et al., High expression of tumor-‐infiltrating macrophages correlates with poor prognosis in patients with diffuse large B-‐cell lymphoma. Medical oncology, 2012. 29(4): p. 2317-‐22.
144. Timmermann, M., et al., Interaction of soluble CD163 with activated T lymphocytes involves its association with non-‐muscle myosin heavy chain type A. Immunology and cell biology, 2004. 82(5): p. 479-‐87.
145. Jones, K., et al., Serum CD163 and TARC as disease response biomarkers in classical Hodgkin lymphoma. Clinical cancer research : an official journal of the American Association for Cancer Research, 2013. 19(3): p. 731-‐42.
146. Jensen, T.O., et al., Macrophage markers in serum and tumor have prognostic impact in American Joint Committee on Cancer stage I/II melanoma. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2009. 27(20): p. 3330-‐7.
147. Lin, Y., et al., Immunosuppressive CD14+HLA-‐DR(low)/-‐ monocytes in B-‐cell non-‐Hodgkin lymphoma. Blood, 2011. 117(3): p. 872-‐81.
148. Jordan, K.R., et al., Myeloid-‐derived suppressor cells are associated with disease progression and decreased overall survival in advanced-‐stage melanoma patients. Cancer immunology, immunotherapy : CII, 2013. 62(11): p. 1711-‐22.
149. Zhang, B., et al., Circulating and tumor-‐infiltrating myeloid-‐derived suppressor cells in patients with colorectal carcinoma. PloS one, 2013. 8(2): p. e57114.
29
150. Pardoll, D.M., The blockade of immune checkpoints in cancer immunotherapy. Nature reviews. Cancer, 2012. 12(4): p. 252-‐64.
151. Myklebust, J.H., et al., High PD-‐1 expression and suppressed cytokine signaling distinguish T cells infiltrating follicular lymphoma tumors from peripheral T cells. Blood, 2013. 121(8): p. 1367-‐76.
152. Jacobsen, E.D., Restoring antitumor immunity via PD-‐1 blockade after autologous stem-‐cell transplantation for diffuse large B-‐cell lymphoma. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2013. 31(33): p. 4268-‐70.
153. Wolchok, J.D., et al., Nivolumab plus ipilimumab in advanced melanoma. The New England journal of medicine, 2013. 369(2): p. 122-‐33.
154. Greaves, P. and J.G. Gribben, The role of B7 family molecules in hematologic malignancy. Blood, 2013. 121(5): p. 734-‐44.
155. Geiss, G.K., et al., Direct multiplexed measurement of gene expression with color-‐coded probe pairs. Nature biotechnology, 2008. 26(3): p. 317-‐25.
156. Ramaswamy, V., et al., Recurrence patterns across medulloblastoma subgroups: an integrated clinical and molecular analysis. The lancet oncology, 2013. 14(12): p. 1200-‐7.
157. Kolbert, C.P., et al., Multi-‐platform analysis of microRNA expression measurements in RNA from fresh frozen and FFPE tissues. PloS one, 2013. 8(1): p. e52517.
158. Malkov, V.A., et al., Multiplexed measurements of gene signatures in different analytes using the Nanostring nCounter Assay System. BMC research notes, 2009. 2: p. 80.
159. Scott, D.W., et al., Determining cell-‐of-‐origin subtypes of diffuse large B-‐cell lymphoma using gene expression in formalin-‐fixed paraffin embedded tissue. Blood, 2014.
160. Masque-‐Soler, N., et al., Molecular classification of mature aggressive B-‐cell lymphoma using digital multiplexed gene expression on formalin-‐fixed paraffin-‐embedded biopsy specimens. Blood, 2013. 122(11): p. 1985-‐6.
161. Dowsett, M., et al., Comparison of PAM50 risk of recurrence score with oncotype DX and IHC4 for predicting risk of distant recurrence after endocrine therapy. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2013. 31(22): p. 2783-‐90.
30
CHAPTER 2
Research Design and Methods
31
Research Design and Methods Introduction
Many of the research design and methods are described in the published
chapters making up this PhD. However in the following pages, I have tried to
expand some of the important methods in more detail. Published articles give a
basic review of methods so I felt it appropriate to include more detailed
explanations in this section.
2.1 Clinical Sample Accrual
My research questions required well annotated patient cohorts in addition to
tissue availability to produce meaningful results. I will describe the cohorts used
in each of my research questions below as there was minor changes in exact
numbers of patients used for each of these questions based on available tissue
and clinical information at the time of analysis.
Firstly, I acquired a clinical database from Dr. Peter Mollee at Princess Alexandra
Hospital which included all patients diagnosed with DLBCL between 2003-‐2009.
Using this database I was able to identify approximately 200 patients with
DLBCL. I personally continued this database and researched and obtained
additional samples occurring between 2009-2011 and other patients missing
from the database but treated at PA hospital. This time period was chosen as it
reflected the introduction of rituximab in Queensland. However not all patients
in this time period received rituximab as it only was initially available to patients
over the age of 60 until early 2004 at which stage government funding allowed
all patients with DLBCL to be treated with rituximab in association with their
standard chemotherapy.
A formal request was then made to Pathology Queensland for access to biopsy
specimens related to these patients. Approval for tissue collection was provided
by the Ethics Committee of the Princess Alexandra Hospital. Unfortunately,
biopsy specimens are only kept on-‐site for 6-‐9 months after which they are
stored at a commercial contractor by Pathology Queensland. Due to significant
costs in contracting this commercial company to find relevant biopsies for my
study ($50 per specimen), Pathology Queensland allowed me to get special
access to the offsite commercial storage facility to obtain these samples. I spent
32
12-‐15 hours reviewing boxes of archived FFPE tissue in order to obtain DLBCL
specimens.
Due to Pathology Queensland and legal guidelines, I was only allowed to obtain
material in samples with significant amounts of tissue as all specimens must be
kept for 10 years in case of requirement for review and sufficient tissue must
remain for re-‐analysis. In addition, large number of patients with DLBCL are
diagnosed without ever having an excisional biopsy, rather these patients have
core biopsies which are generally so small that is highly unlikely there will be
sufficient tissue for research processing. These factors had significant impact on
the number of patient samples obtained. Unsurprisingly, this is also reflected in
the NHL21 cohort in which at the time of analysis I received approximately 50%
of tissue samples compared to almost 87% of patients in whom we have received
blood samples on. Additional samples have been received but these were not
available during the time period of my laboratory analysis for this thesis. My
initial ideal requirements were to obtain 3x30 micron slices. Processing was
performed at Pathology Queensland, PAH and at the Pathology Department at
Queensland Institute of Medical Research. The tissue slices were used to extract
RNA and DNA required for PCR and gene expression based analysis.
The restrictions imposed by small diagnostic biopsies and institutional and legal
requirements has significant impacts on amount of tissue available for research
studies. My research was significantly helped when it became possible to access
further tissue samples from another Australian retrospective cohort based at
Canberra, courtesy of Dr Dipti Talaulikar. In combination with current tissue
processed, I have developed a database of 195 samples from patients with
DLBCL (63 retrospective from PAH, 65 retrospective from Canberra, 67 samples
from the prospective NHL21 cohort) all with tissue available and good clinical
data. However, at this stage there is no survival data available for the NHL21
cohort with a first survival analysis due to be performed in mid-‐2015.
2.2 Patient Cohorts
Retrospective Cohort for Chapter 3
One hundred and fifteen patients were identified. R-‐CHOP consisted of an
intravenous infusion of cyclophosphamide 750 mg/m2, adriamycin 50 mg/m2,
33
vincristine 1.4 mg/m2 (capped at 2 mg), oral administration of 100 mg prednisone
on days 1 to 5 (CHOP), and Rituximab 375 mg/m2 at day 1 before CHOP
chemotherapy began. Patients with stage I/II disease typically received 4 courses
of chemoc immunotherapy followed by involvedc field radiotherapy (30c 40
Gy), while patients with advanced stage disease received 6 to 8 cycles of chemoc
immunotherapy followed by radiotherapy to bulky sites. In patients receiving <8
cycles of combination therapy, further rituximab was administered as
monotherapy so that in total patients received 8 doses. Maintenance rituximab
after completion of Rc CHOP therapy was not administered. DNA extracted from
formalinc fixed paraffinc embedded tissue (FFPET) was of sufficient quality to
perform PCR analysis in 90 patients for FCGR3A2 V158F polymorphisms and 81
patients for C1qA2 A276G polymorphism. One hundred and five consenting
healthy adult volunteers served as controls. Controls specifically denied
haematological or autoimmune disorders of any kind.
Retrospective Cohort for Chapter 4
Analysis was restricted to DLBCL patients treated with R-‐CHOP between 1st
January 2003 – 1st January 2010 at the Princess Alexandra Hospital, Brisbane.
Inclusion criteria were age ≥18 years, histologically confirmed de novo
DLBCL (grade IIIB follicular lymphoma and transformed follicular
lymphoma were excluded) with full clinical annotation including Rc IPI
scores, CBC and survival data. All patients were assigned to Rc CHOP
chemoc immunotherapy, as per standard institutional practice. Therapy was
delivered as per cohort 1 described above. HIV positive patients or those postc
transplant were excluded.
This retropective cohort included patients treated at the Princess Alexandra
Hospital. Tissue was not required for assessment of absolute lymphocyte and
monocyte count and 122 patients were available for this assessement. Flow
cytometry data was available for 75 of these patients.
Prospective NHL21 Cohort for Chapter 5
My prospective cohort consists of patients treated on the ALLG NHL21 trial.
This was a planned prospective laboratory subc study sponsored by the ALLG
within the NHL21 clinical trial (Clinical Trials Identifier:
ACTRN12609001077257). Clinical parameters were source verified and disease
stage and IPI were central reviewed. Patients gave written informed consent to
34
provide blood samples for laboratory analysis. Where possible, tissue slices for
RNA/DNA and TMA blocks were mandated to be supplied to our lab if sufficient
tissue was available. In addition, thirty milliliters of blood were collected at pre-‐
therapy and day 21 after cycle four of R-‐CHOP (post-‐cycle 4). Peripheral blood
was also taken from healthy laboratory volunteers without prior diagnoses of
malignant, hematological or autoimmune disorders who also provided written
informed consent. The study was approved by all participating
Hospital/Research Institute Ethics Committees and performed in accordance
with the Declaration of Helsinki. Induction chemotherapy consisted of four
cycles of R-‐CHOP administered every 14 days (R-‐CHOP-‐14), supported with
pegylated granulocyte colony stimulating factor (peg-‐G-‐CSF). Post -‐cycle 4,
chemo-‐immunotherapy w a s d e layed a w e ek a n d a n i n terim-‐PET/CT scan
performed between days 17-‐20. R-‐CHOP consisted of: day 1 rituximab 375
mg/m2 intravenous (IV) infusion, cyclophosphamide 750 mg/m2 IV, doxorubicin
50 mg/m2 IV, vincristine 1.4 mg/m2 (maximum 2 mg) IV, and prednisone 100
mg/day orally for 5 days. Following interim-‐PET/CT, patients received risk-‐
stratified therapy.
Figure 2. Schema for ALLG NHL21 Study
35
2.3 PET/CT analysis Prec approved imaging centres performed dualc modality PET/CT,
with centralised review at the Peter MacCallum Cancer Centre, Melbourne. All
interim scans were reviewed centrally alongside diagnostic prec therapy scans
(blinded to the local PET/CT assessment), to verify residual abnormal FDG
uptake at sites of previously identified activity. Metabolic response was
categorised using the ‘International Harmonisation Project in
Lymphoma’ guidelines, on which Professor Rodney Hicks was a member
of the working party. Patients with a complete metabolic response were
classified as interimc PET/CTc ve, and those remaining PET FDGc avid
interimc PET/CT+ve
General Methods
2.4 RNA/DNA extraction
Nucleic acid extraction was performed using an Ambion Recover All Total
Nucleic acid kit as per manufacturer’s instructions. In total between
retrospective and prospective cohorts 195 FFPE samples had RNA and DNA
extraction performed. This does not include samples were extraction failed.
FFPE samples have extensive proteinc protein and proteinc nucleic acid
crossc linking due to the fixation process which reduces yields and quality of
extraction compared to freshly prepared tissue. Nucleic acid modification causes
fragmentation of RNA likely due to formaldehyde used in fixation process. In the
majority of cases simultaneous DNA and RNA extractions were performed with
an approximate 50/50 split of sample to each individual DNA/RNA pathway
after initial deparaffisation processing steps were performed. DNA was used in
PCRs for the FCG3A and C1Q work. In general good extractions were obtained,
but as expected DNA yields tended to be higher due to relative age of specimens
and increased level of RNA degradation. DNA/RNA quantity and quality were
assesses using Nanodrop. Nanodrop uses spectrophotometric measurement to
assess quality/purity. The A260/A280 ratio was measured with majority of
specimens being within or close to target ranges of 1.8-‐2.1 for RNA and 1.7-‐1.9
for DNA when 1.5ul volumes are measured.
36
2.5 FCG3A polymorphism PCR
DNA was extracted from FFPE tissue (patients) or buccal scrapes (controls)
using standard procedures with analysis performed in batches. FCGR3A-‐V/F158
genotyping was performed using allele-‐specific PCR based on a previously
described protocol. This consisted of the following 20 µl reaction (made up to
this level with nuclease free water): 0.2 µmol/l each primer (common forward:
5'-‐TCCAAAAGCCACACTCAAAGT-‐3', F-‐allele-‐specific
reverse:5'GCGGGCAGGGCGGCGGGGGCGGGGCCGGTGATGTTCACAGTCTCGTAAGA
CACATTTTTACTCCCAGA-‐3' and V-‐allele-‐specific reverse: 5'-‐
TGAAGACACATTTTTACTCCCATCc 3'), 0.4 µl Accuprime Taq DNA polymerase
(Invitrogen), 1× Accuprime buffer I, 1 mmol/l MgCl and 10 ng genomic DNA.
Reactions were cycled using 94°C for 2 min followed by 35 cycles of 94°C for 15 s
and 60°C for 15 s. Bands were visualized on agarose gels with the F-‐allele
resulting in an 118 bp and the V-‐allele in a 73 bp band.
Figure 3. Electrophoresis for FCG3A polymorphism
2.6 C1QA Polymorphism PCR
For C1QA polymorphism general procedures such as DNA extraction were
performed as above for FCG3A. C1QA-‐A/G276 genotyping was performed using
allele-‐specific PCR based on a previously described protocol. This consisted of
the following 20 µl reaction (made up with nuclease free water:
0.4 µmol/l each primer C1qA -‐F1 5-‐GGGGGAACCTGGGCCCTCTGG-‐3 ,
F (118bp)
V (73bp)
VV FFVFVV VF VF FF VV -ve
Patients Controls
37
C1qAc R1 5 CGGCCGGAGTGGTTCTGGTACGGc 3, 0.2 µl AmpliTaq Gold DNA
polymerase (Invitrogen), 2 µl 10× PCR buffer I, 1.2 µl 25 mmol/l MgCl, 0.4 µl
10mM dNTPs and 50 ng genomic DNA. Reactions were cycled using 95°C for 10
min followed by 40 cycles of 94°C for 15 s and 60°C for 30 s. Bands were
visualized on 12% acrylamide/TBE gels with the G-‐allele resulting in an 170 bp
and the A-‐allele in a 189 bp band. With this PCR there was significant difficulties
in getting good bands with the FFPE extracted DNA which was improved by
adding 50 ng DNA to each PCR reaction, whereas for fresh bucccal swabs only
10 ng was required.
AG GG AA AA AG AA AG AA AA AG AA AG AG -‐ve
Figure 4. Electrophoresis for C1QA polymorphism
2.7 Flow cytometric analysis of retrospective samples The following protocol was the standard protocol used at the diagnostic
flow laboratory at the Princess Alexandra Hospital. Mononuclear cell
preparations were made using diagnostic fresh tissue by mechanical disruption
in HANKS containing 2% bovine calf serum. Aggregates and large particles were
removed by filtration through nylon gauze. A 100 ul aliquot of cells were stained
with the appropriate concentration of antibody for ten minutes in the dark at
room temperature. Red cells were removed using ammonium choride lysis
buffer and the cells fixed in FACS fixative.
Mononuclear cells were stained with a panel of antibodies against T-‐, and B-‐cell
markers, including CD3 (APC BD Clone SK7), CD4 (FITC BD Clone SK3), CD5 (CD5
A (189 bp) G (170 bp)
38
PE BD Clone L17F12), CD8 (PE BD Clone SK1), CD19 (APC BD Clone SJ25C1),
CD20 (APC BD Clone L27) and CD45 (PerCP BD Anti-‐HLe-‐1 Clone 2D1), Flow
cytometric data was acquired on the FACS Calibur 4 colour bench-‐top analyser,
(Becton Dickenson, New Jersey, USA) and analysed using BD CellQuest Pro
Software. A minimum of 10,000 cells were analysed per tube. An isotype control
was used to set the control levels of fluorescence for each fluorchrome.
A B
Figure 5. (A) Gating of Lymphocytes (Nodal Biopsy) (B) CD4 Gating in high
expressing patient (>20%)
2.8 NHL21 Blood processing
The received tubes of blood were spun at 1180 rpm with no brake to separate
plasma. Plasma was collected into 10 to 50 ml tubes and spun at 3000 rpm for
10 mins to pellet platelets. Plasma aliquots of 1ml was then placed into Nunc
cryovials and stored at c 80*C. The remaining blood was transferred into 50ml
tubes with a maximum of 15 mls of blood diluted up to 35mls with RPMI. These
samples were mixed well. Ficoll of 10 ml volumes was then gently added as an
underlay with care taken not to disturb the interface. These tubes were then
centrifuged at 1500 rpm for 20 minutes. The Buffy coat was then harvested into
10 ml tubes and pelleted after centrifugation at 1180 rpm. The supernatant was
then aspirated and removed from each tube. Each pellet was then resuspended
into a single tube with a small volume of RPMI and then made up to the 10 ml
mark with further RPMI. Further centrifugation was performed to derive a final
39
pellet. The PBMC pellet was resuspended in 10ml of RPMI and 20% fetal calf
serum with DMSO as a cryoprotectant. Sample were placed in controlled
Mr.FrostyTM containers for gradual freezing over 2/3 days prior to transfer to
liquid nitrogen.
2.9 Flow Cytometry NHL21 Study
Between 1 x105 and 5X105 PBMC were diluted in 100 ul of PBS + 2% foetal
bovine serum and added to a well in a 96 well U bottom tissue culture plate. A
pre-determined amount of the fluorophore labelled antibody was added
according to the phenotype of the cells being stained. The tray was placed in a
refrigerator for 15c 30 minutes. At the completion of the incubation a further
100 ul of cool PBS was added to each well. The plate was centrifuged for 5 min at
400 g.The supernatant was carefully removed and replaced with 200 ul of cool
PBS. The plate was centrifuged for 5 min at 400 g.The supernatant was carefully
removed and replaced with 200 ul of cool PBS + 2% FBS. The supernatant was
transferred to a FACS tube and a further 200ul of PBS +2% FBS was added to
each. The tube was either refrigerated or placed on ice and subject to analysis
within 4 hours of staining. The flow cytometry data was acquired using a BD
LSRII flow cytometer controlled by FACSDiva software(BD, Australia). Data
analysis was performed on compensated data using FlowJo 4.2 or 9.x software
(Tree Star, USA).
40
Figure 6. CD163 expression on moMDSCs from healthy, pre-‐treatment
sample and post treatment sample (peripheral blood)
2.10 Effector lymphocyte assays
The antiCD20-‐ADCC assay measures killing of a carboxyfluorescein
succinimidyl ester (CFSE)-‐labeled CD20-‐positive DLBCL cell-‐line (SU-‐DHL4)
incubated for 4-‐6 hours +/-‐ PBMC and +/-‐ rituximab/obinutuzumab. CHL cell-‐
lines were used as CD20-‐ve control targets. Subtraction of the number of target
cells lysed by addition of rituximab/obinutuzumab alone from the total number
lysed by antiCD20 mAb with PBMC, enabled enumeration of antiCD20-‐ADCC
mediated killing. CD107ab de-‐granulation of CD56+ and CD14+ cells, and CD137
activation of CD56+ was measured by flow cytometry. For Tc cell proliferation, CFSE stained mononuclear cells were seeded in
96c well roundc bottom plates at 2c 5x105 cells/well in RPMI 1640 with
10% FBS and 1×P/S, with 10 u/ml interleukinc 2. Antic CD3/CD28/
CD137 beads (Invitrogen) at a 1:10 bead: Cell ratios were added to
stimulate polyclonal expansion. Cells were cultured for 96c 120 hours and
assessed by flow cytometry.
Monocytes were depleted using an immunomagnetic CD14 selection
procedure (EasySep, Stemcell Technologies) as per manufacturer
recommendations achieving >90% monocyte depletion.
Healthy Pre-Therapy Post-Cycle 4 CD
14
HLA-DR HLA-DR HLA-DR
41
2.11 Enzyme linked immuno-‐absorbent assays (ELISA)
Plasma arginase I (BioVendor) and CD163 (R&D Systems) were quantified by
ELISA according to the manufacturer’s instructions (at 1:20 CD163 dilution). I
have summarized the protocol for CD163 assay below.
Basically, 100 ul of assay diluent was added to each ELISA well. 50 ul of
standard control or sample were added to each well and incubated for two
hours at room temperature. Each well was then aspirated and washed repeated
three times for a total of 4 washes. 200 ul of CD163 conjugate was then added
and the wells incubated for a further 2 hours. The wash steps were repeated
once more. After this wash, 200 ul of substrate solution was added, incubated
for 30minutes and protected from light. 50 ul of stop solution was added was
added to each well. The optical density of the wells were read at 450 nm using a
microplate reader. Samples were run in duplicate with results and data
extrapolated from a standard curve.
2.12 Digital multiplex gene expression by NanoString nCounter.
Nucleic acid was extracted from tumour biopsies using RecoverAll total nucleic
acid extraction kit for FFPE (Life Technologies, Carlsbad, CA, USA). Using the
nCounter platform (NanoString Technologies, Seattle, WA, USA) gene expression
profiling was performed on 191 DLBCL samples in total. An initial pilot project
was performed on 83 DLBCL samples and this was extended to 191 DLBCL
samples based on significant results garnered from this study. For an initial pilot
project, the minimum number of genes that could be tested on each sample was
one hundred. This allowed me to assess important genes in DLBCL. Given my
earlier work I was most interested in including immune based genes, but I also
included cell of origin genes based on the Wright algorithm, and a number of
genes found on next generation sequencing to be mutated in DLBCL. At that time,
there was no physical machine in Australia, and to my knowledge this is the first
large Nanostring project performed in Australia, and the largest study of DLBCL
using this technology anywhere. All samples were sent to Nanostring in Seattle,
USA for testing. All analysis was performed by myself with self-‐learning of the
Nanostring provided NCounter software program. The success of this project has
42
now led to my laboratory becoming the first in Queensland to acquire this
technology. Hybridisations were carried out according to the NanoString Gene Expression
Assay Manual. Five microliters of each RNA sample (100 ng) was mixed with 20
μl of nCounter Reporter probes in hybridisation buffer and 5 μl of nCounter
Capture probes for a total reaction volume of 30 μl. The hybridisations incubated
at 65°C for approximately 16c 20 hours. For this digital gene expression, two
separate runs were performed requiring the production of two identical code
sets. Run 1 was an initial pilot study of 97 samples expanded to 191 patient
samples with run 2. Expression counts were normalised between code sets based
on relative differences between duplicate samples in both runs which allowing us
to develop a correction factor for each individual gene between the runs.
Raw data (RCC files) was imported and analyzed in the NanoString® data
analysis tool nSolver. For normalization, gene expression data was internally
controlled to the mean of the positive control probes to account for interc assay
variability. Gene normalisation was then performed using the geometric mean of
four housekeeper gene to account for factors that affect RNA quality and quantity
(PGK1, GAPDH, PGAM1, OAZ1) Housekeeping genes were selected as previously
described and as per manufacturer recommendation. After this normalisation
procedure 12 of 97 samples from run 1 and 11 of 191 samples from run 2 did not
pass QC and were excluded from further analysis.
2.13 Statistics and analysis
I performed the statistical analysis for the majority of the above work. Of note
my co-‐author Frank Vari performed statistical analysis for flow cytometry data
related to moMDSC research in this currently submitted paper. I used Graphpad
Prism for simple Mann-‐Whitney, Student –T tests and Kaplan Meier survival
testing. I performed multivariate survival analysis using SPSS software.
Nanostring data was analysed firstly using nSolver analysis and then data was
transferred to Prism based analysis. All my data is stored in excel based
spreadsheets. Categorical data was compared using Fisher’s exact test or Chi
squared test as appropriate. For non-‐parametric variables and analysis of factors
43
influencing outcome, Mann-‐Whitney U test were applied. Event free survival
(EFS) was measured from time of diagnostic biopsy to the date of disease
progression, relapse or death as a result of any cause, or the date of last follow up.
Overall survival was measured from diagnosis to date of last follow up or death.
Survival analysis was performed using Kaplan Meier curves and the log rank test.
Multivariate analysis was performed using Cox regression. All tests were two
sided at the threshold of P=0.05. All statistical analysis were prepared using
Graphpad Prism 6 (Graphpad, La Jolla, CA) or SPSS (Statistical Package for the
Social Sciences, IBM, NY, USA) version 13.0 for Windows.
44
CHAPTER 3
1. Homozygous FCGR3A-‐158V alleles predispose to lateonset neutropenia after CHOP-‐R for Diffuse Large B-‐cell Lymphoma Colm Keane, Jamie P. Nourse, Pauline Crooks, Do Nguyen-‐Van, Howard Mutsando, Peter Mollee, Rod A. Lea, Maher K. Gandhi Intern Med J. 2011 Sep 1
2. “Rituximab induced Late-‐onset Neutropenia” for thebook 'Rituximab: Pharmacology, Clinical Uses and role in Investigating B cell immunology in Man"published by Novus in 2012 Colm Keane, Jamie Nourse, Maher K. Gandhi
45
Homozygous FCGR3A-‐158V alleles predispose to late onset neutropenia after CHOP-‐R for Diffuse Large B-‐cell Lymphoma Colm Keane, Jamie P. Nourse, Pauline Crooks, Do Nguyen-‐Van, Howard Mutsando, Peter Mollee, Rod A. Lea, Maher K. Gandhi Intern Med J. 2011 Sep 1.
Author Keane, C Nourse, J Crooks, P Nguyen, D Mutsando, H Mollee, P Lea, R Gandhi, M
Project Conception 40% 30%
30%
Experimental Design 45% 45%
10%
Sample Collection and Processing 100% Data Acquisition/Lab Work 85%
15%
Analysis and Interpretation of work 40% 20% 5% 5% 10% 20%
Manuscript Preparation 50%
50%
Reviewed and edited manuscript Yes Yes Yes Yes Yes Yes Yes Yes
Rituximab induced Late-‐onset Neutropenia” for the book 'Rituximab: Pharmacology, Clinical Uses and role in Investigating B cell immunology in Man"published by Novus in 2012 Colm Keane, Jamie Nourse, Maher K. Gandhi
Author Keane, C Nourse, J Gandhi, M Project Conception 40% 20% 40%
Manuscript Preparation 40% 20% 40% Reviewed and edited manuscript Yes Yes Yes
46
CHAPTER 4 CD4(+) tumor infiltrating lymphocytes are prognostic and independent of R-‐IPI in patients with DLBCL receiving R-‐CHOP chemo-‐immunotherapy. Keane C, Gill D, Vari F, Cross D, Griffiths L, Gandhi M. Am J Hematol. 2013 Apr;88(4):273-‐6.
67
CD4(+) tumor infiltrating lymphocytes are prognostic and independent of R-‐IPI in patients with DLBCL receiving R-‐CHOP chemo-‐immunotherapy. Keane C, Gill D, Vari F, Cross D, Griffiths L, Gandhi M. Am J Hematol. 2013 Apr;88(4):273-‐6
Author Keane, C Gill, D Vari, F Cross, F Griffith, L Gandhi, M Project Conception 70% 30% Experimental Design 70% 30%
Sample Collection and Processing 90% 10% Data Acquisition/Lab Work 90% 10%
Analysis and Interpretation of work 80% 10% 10%
Manuscript Preparation 50% 50% Reviewed and edited manuscript Yes Yes Yes Yes Yes Yes
68
CHAPTER 5
Measures of net anti-‐tumoral immunity add to the predictive power of conventional prognostic factors in diffuse large B cell lymphoma (DLBCL). Colm Keane*, Frank Vari*, Mark Hertzberg, John Seymour, Rodney Hicks, Devinder Gill, Pauline Crooks, Kimberly Jones, Erica Han, Rod Lea, Lyn Griffiths, Maher Gandhi. Submitted to Cancer Discovery May 2014 *Co-‐Authors
77
The immunobiological score: a robust 3-‐gene assay that segregates the international prognostic index into disparate survival categories in aggressive B-‐cell lymphoma Colm Keane*, Frank Vari*, Mark Hertzberg, John Seymour, Rodney Hicks, Devinder Gill, Pauline Crooks, Kimberly Jones, Erica Han, Rod Lea, Lyn Griffiths, Maher Gandhi. *Joint Authorship Submitted to Cancer Discovery May 2014
Project Conception Keane Vari Hertzberg Green Han Seymour Hicks Gill Crooks Gould Jones Radford Griffiths Jain Talaulikar Tobin Gandhi
Experimental Deseign 25% 25% 10%
5%
5%
5% 5%
30% Sample Collection and
Processing 25% 25%
10%
10% 5%
5% 5% 5%
Data Acquisition/Lab Work 30% 30%
10%
10% 5% 5% Analysis and Interpretation of
work 30% 30%
5%
5%
25%
Manuscript Preparation 30% 30% 5% 5%
40%
Reviewed and edited manuscript Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
78
1
The immunobiological score: a robust 3-gene assay that segregates the
international prognostic index into disparate survival categories in aggressive
B-cell lymphoma
1,2Frank Vari*, 1,2,3,4Colm Keane*, 5Mark Hertzberg, 6Michael R. Green, 1,2Erica Han,
7John F. Seymour, 7Rodney J. Hicks, 4Devinder Gill, 1,2Pauline Crooks, 1,2Clare
Gould, 1,2Kimberley Jones, 8Kristen J. Radford, 4Lyn R. Griffiths, 9,10Dipti Talaulikar,
10Sanjiv Jain, 9Josh Tobin, 1,2,3Maher K. Gandhi.
1Experimental Haematology, School of Medicine, Translational Research Institute,
University of Queensland, Australia; 2Queensland Institute of Medical Research,
Brisbane, Queensland, Australia; 3Princess Alexandra Hospital, Brisbane,
Queensland, Australia; 4Genomics Research Centre, Griffith University, Queensland,
Australia. 5Department of Haematology, Westmead Hospital, Sydney, New South
Wales, Australia; 6Division of Oncology, School of Medicine, Stanford University,
Stanford, California, USA. 7Peter MacCallum Cancer Centre and University of
Melbourne, Melbourne, Victoria, Australia; 8Mater Medical Research Institute,
Translational Research Institute, Brisbane, Queensland, Australia; 9Canberra
Hospital, Canberra, Australian Central Territory; 10Australian National University
Medical School, Australian Central Territory.
*These authors contributed equally to the manuscript.
Running title. LMO2 +/- CD8:CD163 in DLBCL
Keywords: lymphoma; LMO2; CD8; CD163; gene expression
79
2
Funding relevant to this project was provided by the Leukaemia Foundation (C.K.), a
bequest from the Kasey-Anne Oklobdzijato Memorial fund, a donation from Malcolm
Broomhead via the ALLG, the Cancer Council of Queensland and the Health and
Medical Research Queensland Office (M.K.G).
Corresponding Author: Maher K. Gandhi, Experimental Haematology, School of
Medicine, Translational Research Institute, University of Queensland, Diamantina
Road West, Brisbane, Queensland, 4102, Australia.
Tel +617 3443 8026; Fax +617 3443 7779; [email protected]
Conflict of Interest
Roche provided funding towards the NHL21 clinical trial but not the laboratory
study.
Word count: 5520
Figures 7, Tables 0. Supplemental Figures 3; Supplemental Tables 3.
80
3
Abstract
Diffuse large B-cell lymphoma (DLBCL) is a common and aggressive lymphoma with
approximately 30% mortality. Risk-stratification requires prognosticators to identify
poor outcome patients in whom investigational therapeutic intervention is justified.
Circulating lymphocyte:monocyte ratios are prognostic, implicating them as
surrogate immune-effectors and monocyte/macrophage-checkpoint within the tumor
microenvironment. Blood from 140 ‘R-CHOP’ chemo-immunotherapy treated DLBCL
patients from an Australasian Leukaemia and Lymphoma Group trial was analysed.
Detailed functional and quantitative assessment enabled identification of the optimal
immune-effector and monocyte/macrophage-checkpoint molecules to interrogate
within the tissue. CD163 identified a highly immunosuppressive subset of
CD14+HLA-DRlo monocytoid-myeloid-derived-suppressor cells ‘moMDSC’. Ratios of
various immune-effectors to CD163himoMDSC were used as a measure of total anti-
tumoral immunity: i.e. the net balance between the antagonistic forces of immune-
effectors and monocyte/macrophage-checkpoints. All ratios were higher in early R-
CHOP responders compared to delayed responders, with CD8:CD163himoMDSC the
most discriminatory. To test for intratumoral applicability, genes were quantified by
digital hybridization in an independent cohort of 128 R-CHOP treated DLBCL
patients. Co-clustering of CD8 with CD163 was observed, consistent with an
adaptive immune-checkpoint response to immune-effector activation. CD8:CD163
ratios were prognostic independent of cell-of-origin and international prognostic
index (IPI). An immunobiological score combining CD8:CD163 to the germinal-centre
marker LMO2 strengthened the predictive ability, identifying 24% at risk of very poor
outcome. It separated low-risk IPI into 91% and 44%, and high-risk IPI into 76% and
26% 4-year survivals. Results were externally validated in 233 patients. The
81
4
immunobiological score is a powerful new 3-gene assay that segregates IPI into
markedly disparate survival categories.
250 words.
82
5
Introduction
Diffuse large B-cell lymphoma (DLBCL) is a common and aggressive form of
B-cell lymphoma for which approximately only 70% of patients will be cured (1). The
majority of those who die from their lymphoma, either display refractoriness to first-
line therapy, or relapse within 24 months (2). A number of promising new therapeutic
strategies are at various stages of clinical development (3). However, the application
of any novel therapeutic strategy requires accurate identification of those patients
likely to die despite current therapies. In these patients, prolonged exposure to
conventional first-line agents may contribute to the induction of chemo-resistance as
well as unnecessary toxicity. Consequently in refractory/early relapsing patients,
alternate strategies should be instituted early. However, despite the use of
conventional pre-treatment prognosticators such as the international prognostic
index (IPI), very considerable heterogeneity of outcome persists (1). Therefore there
is a pressing need to develop tools to more accurately predict response to initial
therapy.
It has recently been established that patients with low peripheral blood
absolute lymphocyte:monocyte ratios (LMR) have inferior outcomes (4-6). It is known
that circulating lymphocyte and monocyte subsets each have active roles in DLBCL
control (7, 8). For example recent data strongly implicates CD8+ T-cells in the
prevention of murine models of lymphoma (9), and adoptive T-cell therapy has
shown clinical benefit for the treatment of immunosuppression related lymphomas
(10). Within peripheral blood monocytes, the CD14+HLA-DRlo ‘monocytoid-myeloid-
derived-suppressor cells’ (moMDSC) subset is associated with higher rates of
disease progression in patients with B-cell lymphoma (11-13).
83
6
We have established that circulating immune subsets reflect immunity within
the DLBCL tissue tumor microenvironment (TME) (4). A simple, robust and easily
standardized tissue biomarker, applicable at diagnosis, that provided additional
information to conventional prognosticators, would be a highly useful tool for risk-
stratification. With new digital multiplexed gene expression (DMGE) platforms
applicable to formalin-fixed, paraffin-embedded tissues (FFPET)(14), simultaneous
interrogation of a range of aspects of DLBCL biology, that include the TME, is
achievable by the diagnostic laboratory (15, 16).
Our aim was to develop a gene expression score applicable to FFPET that
reliably identified those patients with DLBCL at high-risk of treatment failure within 24
months following treatment with conventional ‘R-CHOP’ chemo-immunotherapy. A
composite score, incorporating assays of malignant B-cell biology with the balance of
the antagonistic forces of immune-effectors and monocyte/macrophage-checkpoints,
would likely have more prognostic value than a measure of B-cell biology alone. We
rationalized that detailed functional and quantitative assessment of circulating
immune-effector and monocyte-checkpoints would enable identification of the
optimal immune molecules to incorporate within the composite gene expression
score. To develop such a prognostic tool, we utilized sequential blood samples taken
as part of an Australasian Leukaemia and Lymphoma Group (ALLG) clinical trial in
poor-risk DLBCL. Immune-effector and monocyte/macrophage molecules were
identified that were applied to diagnostic tissues. Candidate molecules were
quantified by DMGE in DLBCL tissues in an independent Australian R-CHOP cohort.
The resultant algorithm was then externally validated in an independent international
R-CHOP DLBCL cohort.
84
7
Materials and Methods
ALLGNHL21
Details of the study and eligibility criteria are provided with the supplemental
methods. Induction chemo-immunotherapy comprised three cycles of R-CHOP
administered every 14 days (R-CHOP-14). After a fourth cycle of R-CHOP, chemo-
immunotherapy was delayed a week and an interim-PET/CT scan performed
between days 17-20. Dual modality PET/CT was performed in pre-approved imaging
centres, with centralized review (R.H.) at the Peter MacCallum Cancer Centre,
Melbourne. All interim scans were reviewed centrally alongside diagnostic pre-
therapy scans (blinded to the local PET/CT assessment), to verify residual abnormal
FDG uptake at sites of previously identified involvement. Blood assays were planned
prospectively within the ALLGNHL21 clinical trial. Thirty millilitres of blood were
collected pre-therapy and on day 21 post-cycle 4 of R-CHOP. Investigators were
blinded to PET/CT results. Blood was also taken from 23 healthy participants without
prior diagnoses of malignant, hematological or autoimmune disorders. All
participants gave written informed consent. The study was approved by responsible
Ethics Committees at participating sites and performed in accordance with the
Declaration of Helsinki.
Tissue cohorts
Tissue cohort one comprised 128 patients with histologically confirmed
DLBCL: 66 from Princess Alexandra Hospital, Queensland; and 62 from Canberra
Hospital, Australian Capital Territory. All patients received R-CHOP, and were
selected on the basis of tissue and outcome data availability. These tissues were
supplemented with 63 FFPET from the ALLG Tissue Bank from whom survival data
85
8
was not available for tests of co-clustering (191 FFPET). Only de-novo cases were
included. Grade IIIB follicular lymphoma, transformed follicular lymphoma, HIV-
positive and post-transplant lymphoproliferative disorder patients were excluded. The
external validation tissue cohorts utilized a publicly available data-set (17). Patients
were divided into two groups based on treatment regimen (CHOP 181 patients
versus R-CHOP 233 patients). Details of RNA quantification of DLBCL samples is
provided with supplemental methods.
Statistical analysis
Values between groups of data were tested for statistical significance using
the 2-tailed paired (e.g. between time-points) or where appropriate non-paired tests.
Categorical data were compared using Fisher’s exact test or Chi-squared test as
appropriate. Overall survival (OS) was measured from diagnosis to date of last
follow-up or death. Survival analysis was performed using Kaplan–Meier curves and
the log-rank test. Multivariate analysis was performed using Cox regression. All tests
were two sided at the threshold of P=0.05. All analyses were prepared using
GraphPad Prism platform (version 6, GraphPad Software, La Jolla California USA)
and Statistical Package for the Social Sciences SPSS version 13 (International
Business Machines Corporation, New York USA).
86
9
Results
Patient characteristics.
Of 161 patients with DLBCL accrued to NHL21, 154 had interim-PET/CT. The
140 patients in whom blood samples were obtained (91%) were included in this
study. As anticipated from the inclusion criteria, patients had poorer-risk features
(82% stage III-IV, 43% mass ≥7.5cm) than an unselected DLBCL population. There
were 23 healthy participants. There was no significant difference in age/sex between
healthy participants (median age 49, range 31-68 years, 39% female) and patients
(56 years, range 26-70, 34% female). Patient characteristics for NHL21 and the
Australian tissue cohort are outlined in supplemental Table S1. Details of the
external validation tissue cohorts are as previously published (17).
Monocytes suppress CD8+ and CD4+ T-cell proliferation in poor-risk DLBCL.
Monocytes were increased as a proportion of the mononuclear cells among
the DLBCL patients compared to healthy participants (P<0.0001, Figure 1A). The
impact of monocytes upon immune-effectors was tested. It has previously been
shown that total T-cell proliferation is impaired by monocytes in B-cell lymphomas
(13). However as that study did not interrogate individual T-cell subsets, we tested
the effect of monocyte depletion in order to assess the differential effect upon CD4+
and CD8+ T-cells. Peripheral blood mononuclear cells (PBMC) from randomly
selected patients and healthy participants were tested for T-cell proliferation.
Although total CD3+ T-cells were reduced, the proportion of total and CD8+ and CD4+
T-cell subsets in patients that proliferated in response to in-vitro stimulation were not
different to healthy participants at both time-points (P=NS). However, monocyte
87
10
depletion enhanced T-cell proliferation in pre-therapy samples for total CD3+, CD4+
and CD8+ T-cell subsets, conversely monocyte depletion from healthy participant
samples did not (Figure 1B-D).
Monocytes of poor-risk DLBCL patients suppress rituximab but not
obinutuzumab mediated antibody-dependent cell-mediated cytotoxicity.
AntiCD20 monoclonal antibodies (mAB) have improved outcome in DLBCL
(1). However, the impact of circulating monocytes upon antibody-dependent cell-
mediated cytotoxicity (ADCC) in B-cell lymphoma patients has not previously been
explored. CD107ab degranulation by flow cytometry demonstrated that antiCD20-
ADCC was overwhelmingly mediated by NK-cells (as opposed to monocyte
mediated ADCC). We assessed the effect of monocyte depletion of PBMC from
randomly selected poor-risk DLBCL patients and healthy participants. In healthy
participants monocyte depletion did not impact the levels of ADCC mediated by
either rituximab or the type-II antiCD20 mAB obinutuzumab (Figure 1E-F, R- and Ob-
ADCC, respectively). However, monocyte depletion in patients did enhance R-ADCC
(P=0.01), which indicates that any benefit of monocyte mediated antiCD20-ADCC is
masked by monocytes mediated immunosuppression. Interestingly, Ob-ADCC was
not altered by monocyte depletion (P=NS). Comparing healthy participants and pre-
therapy patients, when monocytes were intact, R-ADCC was significantly lower in
patients compared to healthy participants (P=0.01) but became equivalent (P=NS)
when monocytes were depleted (Figure 1G). Ob-ADCC was equivalent between
patients and healthy participants with monocytes intact and depleted (both P=NS).
CD137 (TNFRSF9) is an inducible cell-surface co-stimulatory receptor and
immune-effector activation marker. NK-cell activation was measured upon
88
11
encountering antiCD20 coated DLBCL cell-lines, by expression of CD137 during R-
or Ob- ADCC (Figure 1H). In R-ADCC, NK-cell activation was markedly reduced pre-
therapy in monocyte replete poor-risk DLBCL patients compared with healthy
participants. However, no difference in NK-cell activation was observed between
patients and healthy participants for Ob-ADCC. These results indicate that in poor-
risk DLBCL monocytes suppress rituximab- but not obinutuzumab mediated ADCC
activity.
CD163 expression identifies a highly immunosuppressive subset of moMDSC
(CD14+HLA-DRlo) in DLBCL.
In contrast to healthy participants, PBMC from DLBCL patients collected pre-
therapy consisted almost exclusively of CD14+CD16- classical monocytes, and had
lower levels of HLA-DR (Figure 2A). The CD14+HLA-DRlo moMDSC subset were ~2-
fold elevated (Figure 2B, P<0.0001) compared to healthy participants (representative
plot shown in Figure 2C). The absolute number of CD163+monocytes was also
increased in patients (P=0.0054). Absolute moMDSC and CD163+monocytes values
were correlated (Figure 2D, r=0.41, P<0.0001), and pre-therapy CD163+moMDSC
were ~4-fold raised relative to healthy participants (Figure 2E, P=0.001). Arginase
production by moMDSC depletes arginine and impairs immune-effector signal
transduction and function (18). Consistent with such an effect mediated by the
monocytes of patients with DLBCL, arginase activity was elevated compared to
healthy participants (Figure 2F, P=0.0005). CD163 mean fluorescent intensity (MFI)
on moMDSC showed moderate correlation with arginase activity in patients and
healthy participants (Figure 2G, r=0.45, P=0.002).
89
12
To characterize CD163himoMDSC versus CD163lomoMDSC, CD14+HLA-DRlo
cells with the highest and lowest (top and bottom thirds) CD163 expression were
analysed (Figure 3A-I). A higher proportion of CD163himoMDSC expressed lymphoid
migratory markers CD62L (P<0.0001) and CD11c (P=0.035). CD163himoMDSC
expressed more CD120b (a marker associated with MDSC survival, P=0.001) (19).
Cell surface expression of the myeloid specific inhibitory receptor CD33
(P=0.0108),(20) colony stimulating factor-1 (CSF-1R: a monocyte/macrophage
trafficking, differentiation and survival factor, P<0.0001),(21) and CD80 (a T-cell
regulatory receptor, P=0.0088) were also higher in CD163himoMDSC. Levels of the
monocyte/macrophage marker CD68, the integrin alpha M marker CD11b, and the
co-stimulatory molecule CD86 were equivalent.
In aggregate, these results indicate that CD163himoMDSC have distinct
features, including a migratory and regulatory phenotype and elevated arginase
activity.
CD8:CD163+moMDSC ratios are elevated in patients with poor-risk DLBCL that
become interim-PET/CT-ve after post-cycle 4 R-CHOP.
All patients received uniform R-CHOP chemo-immunotherapy for four cycles,
after which therapeutic response was assessed with an interim-PET/CT. The rate of
interim-PET/CT-positivity for the 140 patients was 29%. Neither individual IPI
parameters nor the combined IPI were associated with interim-PET/CT treatment
response (each P=NS).
We then tested for associations of immune-effectors and
monocyte/macrophage-checkpoints with treatment response. Neither pre-therapy
peripheral blood lymphocytes nor any lymphocyte subset were associated with
90
13
interim-PET/CT response (P=NS), nor were absolute monocytes (Figure 4A left). It is
known that peripheral blood absolute lymphocyte:monocyte ratios (LMR) are
predictive of overall survival (4-6), but their association with interim-PET/CT
response has not previously been evaluated. There was no difference in LMR
between interim-PET/CT+ve and interim-PET/CT-ve patients (Figure 4B left). Pre-
therapy ratios of total lymphocytes (and NK-cells or CD3+ T-cells) to monocytes were
not associated with interim-PET/CT (Figure 4B) or the individual components of the
IPI and the combined IPI score (P=NS).
Next, based on the elevation of immunosuppressive monocyte subsets
observed, we evaluated the association of absolute moMDSC and CD163+moMDSC
with treatment response. Interim-PET/CT+ve patients had ~1.5-fold and ~3-fold higher
levels of moMDSC and CD163+moMDSC than interim-PET/CT-ve patients (Figure 4A
middle and right panels, P<0.001 and P<0.0001 respectively).
Ratios of various immune-effectors with moMDSC and then CD163+moMDSC
ratios were then evaluated for possible association with treatment response.
Interestingly, ratios of total lymphocytes (and NK-cells or T-cells) to moMDSC were
all ~2-fold elevated in patients that achieved interim-PET/CT negativity (Figure 4C).
When CD163+moMDSC was used as the denominator, ratios to total lymphocytes,
NK and CD3+ T-cells were ~4-fold (P<0.0001), ~6-fold (P=0.0002) and ~6-fold
(P<0.0001) elevated respectively in patients that became interim-PET/CT-ve (Figure
4D). To compare results in specific CD3+ T-cell subsets, ratios of CD8+ T-cells and
CD4+ T-cells divided by CD163+moMDSC were tested for association. Both ratios of
CD8+ T-cell and CD4+ T-cell to CD163+moMDSC were higher in interim-PET/CT-ve
patients (each P<0.0001), at ~10-fold and ~4-fold respectively. Plasmacytoid
dendritic cells (pDC) are circulating poly-functional innate immune-effectors. Ratios
91
14
of pDC to CD163+moMDSC were also ~6-fold higher in interim-PET/CT-ve patients
(P<0.0001).
Plasma soluble CD163 is associated with lower lymphocytes and higher IPI.
After metalloproteinase-mediated cleavage, CD163 is released from
macrophage or monocyte cell membranes and the extracellular portion of CD163
circulates in blood as a soluble protein (22). Soluble CD163 (sCD163) was
measured in healthy participants and pre-therapy and post-cycle 4 plasma of
patients with poor-risk DLBCL to further establish its ability to serve as a
monocyte/macrophage-checkpoint marker.
Pre-therapy patient sCD163 levels were higher (Figure 5A, P<0.0001) and
receiver operator curve analysis was highly discriminatory versus healthy
participants (Figure 5B, area under curve 0.94, P<0.0001). Pre-therapy sCD163
levels were associated with lower lymphocytes (P=0.004), advanced stage
(P=0.0072), older age (P=0.0143) and higher IPI (P=0.0003), but no other clinical
variables (Figure 5C-F). However, pre-therapy sCD163 levels were not significantly
different in those becoming interim-PET/CT-ve to those remaining interim-PET/CT+ve
(P=NS). Similarly, post-cycle 4 sCD163 was not associated with interim-PET/CT
status (P=NS). Soluble CD163 levels reduced by post-cycle 4 (P<0.0001), but were
still raised relative to healthy participants (P<0.0001). Reduction in sCD163 between
pre-therapy and post-cycle 4 was similar irrespective of whether the patient became
interim-PET/CT-ve or remained positive. Finally, tissue CD163 mRNA expression was
significantly but only modestly correlated with sCD163 (r=0.4, P=0.018).
Monocytes are ‘reset’ post-cycle 4 of R-CHOP.
92
15
There is minimal data on the kinetics of circulating immune-effectors and
monocyte/macrophage-checkpoints during therapy for DLBCL. By paired analysis,
absolute monocyte counts were not reduced post-cycle 4 relative to pre-therapy
(P=NS). The MFI of HLA-DR expression on monocytes increased between time-
points (P<0.0001). Consistent with this moMDSC numbers fell between time-points
(Figure 6A, P=0.0001) and CD163+moMDSC were lower post-cycle 4 (Figure 6B,
P=0.0002). Plasma arginase, a hallmark of moMDSC, was reduced post-cycle 4
relative to pre-therapy (pre: mean 513ng/ml, 56-733 ng/ml, post: 378 ng/ml, 69-709
ng/ml, P=0.004). Gene expression microarray on monocytes isolated from healthy
participants and paired pre-therapy-post-cycle 4 monocytes showed distinct
clustering of healthy/post-cycle 4 monocytes versus pre-therapy, with up-regulation
of antigen presentation and down-regulation of TH2 cytokine and tumor-associated
macrophage (TAM) genes in healthy/post-cycle 4 monocytes (Figure 6C). In line with
monocytes being ‘reset’ to a healthy profile, there was 2-3 fold up-regulation of
STAT1 and associated genes at post-cycle 4.
Monocytes are a source of blood myeloid dendritic cells (BMDCs), which are
circulating antigen presenting cells central to the development of innate and adaptive
immunity (23). Ex-vivo BMDCs in healthy participants were higher relative to patients
at both time-points (both P≤0.004), and increased between time-points (Figure 6D,
P=0.0002). In parallel with BMDC, there were fewer pDC in patients compared with
healthy participants prior to therapy (P<0.0001). CpG-DNA induced interferon-α
production from pDC in patients was reduced relative to healthy participants (Figure
6E, P=0.006), indicating a functional as well as numerical deficit in pDC in DLBCL
patients. However pDC increased post-cycle 4 (Figure 6F, P=0.0001) so that they
were equivalent to healthy levels.
93
16
As expected, chemo-immunotherapy resulted in a marked reduction in
lymphocyte subsets post-cycle 4. Ratios of all immune-effectors to moMDSC or
CD163+moMDSC performed post-cycle 4 were not associated with interim-PET/CT
(all P=NS).
Construction of a tissue-based immune-effector:monocyte/macrophage-
checkpoint ratio.
PBMC assays identified a variety of immune-effectors and
monocyte/macrophage-checkpoints that were associated with differential post-cycle
4 treatment response. Of these, CD8 and CD163 were highly discriminatory. To test
the applicability of these immune-effectors and monocyte/macrophage-checkpoints
as prognosticators within the diagnostic biopsy, FFPET from an independent
Australian cohort of 128 patients treated with R-CHOP were utilized. As the median
follow-up of the Australian tissue cohort was 3.9 years, the outcome measure
chosen was 4-year overall survival. IPI and cell-of-origin (COO) were used as co-
variates.
NanoString nCounter was used to perform digital multiplex gene expression
(DMGE). There was a strong correlation in three paired DLBCL frozen-FFPET (all
r>0.94, P<0.0001).
Initially, the predictive ability of multiple immune-effectors was compared. In
addition to CD8/CD4/CD56/CD137 (all of which had been shown to be impaired in
peripheral blood taken pre-therapy), we tested for TNFα, a cytokine known to be
secreted by T-cells, NK-cells and M1 macrophages. Patients were divided as being
above or below the median value for the relevant molecule. For all immune-effectors
except CD4 and CD56, values above the median segregated patients with superior
94
17
survival, compared to those with values below the median (Table 1). CD8, CD137
and TNFα were selected for further evaluation; of these CD8 was marginally the
most discriminatory (greatest percent difference in 4-year survival).
As CD163 identified a highly immunosuppressive subset of moMDSC, it was
therefore used for further evaluation as a monocyte/macrophage-checkpoint within
the lymphomatous tissue (Table 1). It did not stratify patient outcomes on its own.
However combining immune-effectors with CD163 as a ratio to measure net anti-
tumoral immunity, was highly discriminatory, and a better discriminator than using
immune-effectors on alone. For CD8:CD163, those with high ratios (i.e. above the
median) had superior survival to those with ratios below the median (P=0.0007). We
then tested the remaining immune-effectors as numerators in ratios with CD163 as
the denominator. CD137 but not TNFα in ratio with CD163 stratified patients into two
distinct groupings (Table 1). However CD8:CD163 remained the best discriminatory
ratio with markedly different 4-year survivals for high and low ratios of 89% and 59%
respectively (Table 1). This ratio was therefore chosen for additional appraisal as a
measure of net anti-tumoral immunity.
Interestingly, DMGE in 191 DLBCL tissues found co-clustering of CD8 with
CD163 within the TME (r=0.32, P<0.001). To see if similar findings were present in
blood, correlations in pre-therapy NHL21 blood samples were performed. Consistent
with the tissue findings, CD8+ T-cells and CD163+moMDSC modestly but
significantly correlated (r=0.42, P=0.0016). The blood and tissue data combined
suggest that an adaptive immune response is present, with the
monocyte/macrophage immune-checkpoint CD163+moMDSC activated in response
to CD8+ T-cell anti-tumoral immunity.
95
18
CD8:CD163 adds to the predictive ability of IPI and COO prognosticators.
In the Australian tissue cohort, IPI as well as a measure of malignant B-cell
biology, the COO (stratified as GCB and non-GCB), stratified patients into high and
low-risk survival categories as expected (supplemental Figure 1, P=0.002 and
P=0.018 and 4-year survivals of 83% versus 61%, and 82% versus 60%
respectively). As expected, few deaths occurred after 24 months.
Median CD8:CD163 ratios were then tested for their ability to sub-stratify IPI
and COO. Firstly, CD8:CD163 ratios were used to sub-divide patients with low-risk
(0-2) and high-risk (3-5) IPI’s (supplemental Figure 2). This showed that survival was
superior in those with high CD8:CD163 ratios for high-risk IPI (P=0.03), with a similar
but non-significant trend seen for low-risk IPI (P=0.07). Next, CD8:CD163 ratios were
used to stratify COO. Patients with high CD8:CD163 ratios had a higher probability
of survival than low CD8:CD163 ratio patients in both GCB and non-GCB groupings
(P=0.047 and P=0.011). For both IPI and COO, the discriminatory value of
CD8:CD163 ratios were most pronounced in those at risk of poorer outcome, i.e.
high-risk IPI and non-GCB (supplemental Figure 2B and 2D), compared to low-risk
IPI and GCB (supplemental Figure 2A and 2C).
Cox Regression multivariate analysis was consistent with the Kaplan-Meier
analysis. This showed that IPI, COO and CD8:CD163 were independently predictive
of OS (P=0.046, P=0.038 and P=0.015 respectively), whereas CD137:CD163 was
not (P=NS). We then performed Cox Regression in an external validation cohort,
comprising 181 patients treated with CHOP chemotherapy-alone and 233 patients
with R-CHOP chemo-immunotherapy. This enabled a comparison of the relative
importance of CD8:CD163 within the TME, with and without the addition of rituximab.
By multivariate analysis IPI, COO and CD8:CD163 remained independently
96
19
predictive of OS (P<0.0001, P=0.0006 and P=0.0006 respectively) in R-CHOP
treated patients. Interestingly, although CD8:CD163 was prognostic in the CHOP
chemotherapy-alone cohort (P=0.0007), only IPI was independently predictive in
multivariate analysis.
The 3-gene composite immunobiological score ‘LMO2+/-CD8:CD163’ adds to
the predictive ability of IPI and COO in R-CHOP treated patients.
LIM domain only 2 (LMO2) is a germinal-centre B-cell marker associated with
improved outcome after R-CHOP (24, 25). In keeping with this patients above the
median cut-off for LMO2 typed as GCB in 88% of cases, but only 13% in non-GCB
cases. As expected, in the Australian tissue cohort, patients with higher than median
expression of the germinal-centre B-cell marker LMO2 had superior outcome to
patients with low LMO2 (Figure 7A). The discriminatory value was similar to that of
IPI and COO, at 85% (LMO2 high) versus 62% (LMO2 low) 4-year survival
(P=0.0046).
We hypothesized that a composite score of LMO2 (as a marker of malignant
B-cell biology) and CD8:CD163 as a measure of net anti-tumoral immunity would
provide a simple 3-gene prognosticator with increased discriminator value over either
parameter alone. A binary score was tested in which high LMO2 and/or low LMO2
with a high CD8:CD163 ratio was designated ‘positive’, against a ‘double-negative’ of
low LMO2 and low CD8:CD163. As this measured both malignant B-cell biology and
net anti-tumoral immunity, it was termed the ‘immunobiological’ score. This identified
two groups with markedly different 4-year survivals of 86% (positive) and 33%
(double-negative, P=0.002, Figure 7B). Seventy six percent of patients were positive,
and 24% double-negative. Most deaths (>80%) occurred within 24 months.
97
20
The new score was compared to a previously proposed two-gene-scoring
weighted algorithm that used LMO2 and CD137 (26). Although less effective at
identifying a group at risk of a particularly poor outcome than the immunobiological
score, the two-gene-score did also effectively stratify patients (4-year survival 89%
for above median versus 58% below median, P<0.001).
For the immunobiological score to confer additional value to conventional
prognosticators, it was important to ascertain whether it was capable of sub-
stratifying IPI and COO. Critically, for each of low-risk and high-risk IPI (Figure 7C-
D), and GCB and non-GCB (Figure 6E-F), the composite score very strikingly added
to the IPI and COO’s ability to sub-stratify patient survival. With IPI, 4-year survival
was 91% and 44% (low-risk IPI, P<0.0001) and 76% and 26% (high-risk IPI,
P=0.002), and for COO 4-year survival was 89% and 27% (GCB, P<0.0001) and
79% and 36% (non-GCB, P=0.007).
External validation of the immunobiological score in R-CHOP treated DLBCL
patients.
The immunobiological score was then applied to the independent Affymetrix
tested gene-expression cohort of 181 CHOP treated and 233 R-CHOP treated
patients. The score segregated patients into groupings with different survival (Figure
7A-B, P=0.0007 and P<0.0001 respectively). As with the Australian R-CHOP treated
tissue cohort, for the Affymetrix R-CHOP cohort the score was additive to IPI and
COO (Figure 7D,F,H,J). For IPI, it separated the 79% 4-year survival of low-risk IPI
into two categories of 85% and 48% (Figure 7D, P<0.0001). Similarly, the 53% 4-
year survival of high-risk IPI was stratified into two groupings of 66% and 36%
(Figure 7F, P=0.0007). This confirmed that the immunobiological score had
98
21
predictive power and could sub-stratify IPI and COO in an external independent
cohort, and that DMGE gave similar findings to Affymetrix tested DLBCL patients.
Interestingly with the CHOP chemotherapy-alone patients, although LMO2+/-
CD8:CD163 had predictive power, this was not additive to non-GCB or low-risk IPI. It
did stratify high-risk IPI (P=0.006) patients and there was a non-significant trend with
GCB (P=0.056).
99
22
Discussion
It is established that in patients with DLBCL treated with CHOP-R chemo-
immunotherapy, low peripheral blood absolute lymphocyte:monocyte ratios confer
inferior outcomes (4-6, 27). This implicates circulating lymphocytes and monocytes
as surrogate markers for further analysis of intratumoral immunity. The present study
confirms that detailed functional assessment of blood immune-effector and
monocyte/macrophage-checkpoints, permits the rational identification of the optimal
immune molecules to incorporate within a lymphoma-tissue based prognostic gene
expression score. Furthermore, it demonstrates that the net balance of immune-
effectors and monocyte/macrophage-checkpoints are critical to outcome.
DLBCL biopsies are enriched in TAMs (17, 28-31). TAMs are thought to be
immunosuppressive ‘M2’ macrophages. TAMs derived from primary tumors are
believed to facilitate circulating tumor cell seeding of distant metastases in breast,
pancreatic and prostate cancer (32). Although it is known that monocytes can
migrate to lymph nodes and differentiate into macrophages, the relationship between
moMDSC and TAMs is incompletely understood. Following adoptive transfer of
MDSC into tumor-bearing mice, cells with the characteristics of TAMs can be
recovered from the tumor microenvironment (33). In other murine models TAMs
were shown to express relatively low levels of major histocompatibility class II
molecules, and tumor progression is positively correlated with increasing infiltration
of the tumor tissues by MHC class IIlo TAMs (34). However data in humans are
sparse.
TAMs have high expression of the scavenger receptor CD163, whereas pro-
inflammatory M1 macrophages are CD163lo. CD163 is also expressed on ‘classical’
(CD14+CD16-) monocytes. CD163+monocytes rise with aging and during HIV (35),
100
23
and are implicated in the lymphopenia of Hodgkin Lymphoma (36). Tissue CD163 is
an adverse prognosticator in melanoma and breast cancer (37, 38). We found that
addition of CD163 to conventional moMDSC indicators identified a highly
immunosuppressive subset of moMDSCs in DLBCL. MoMDSC and
CD163+monocytes values correlated and CD163+moMDSC were enriched within
circulating monocytes. The correlation between the M2 TAM marker CD163 and
HLA-DRlo on circulating CD14+ monocytes suggests a link between circulating
moMDSC and M2 TAMs within the malignant lymph node. This is supported by
CD163himoMDSC expressing markers that permit migration into lymphoid tissues.
Similar to M2 macrophages (M1 are CD68+CD163lo and M2 CD68+CD163hi),
expression of CD68 was equivalent irrespective of CD163.
Within the circulation, the immune-effector:CD163+moMDSC ratios were
lower in those remaining interim-PET/CT+ve. Pre-therapy patient sCD163 levels were
higher and reduced at post-cycle 4, but unlike our observations in Hodgkin
Lymphoma (36), did not become differentially reduced between interim-PET/CT+ve/-ve
patients. As repeat biopsy was not performed, no definitive conclusion can be made
as to whether pre-therapy CD163+moMDSC associate with residual DLBCL versus
inflammation within sites of interim-PET/CT-FDG-avidity. Rather the aim was to
functionally characterize circulating immune-effector and monocyte/macrophage-
checkpoints for further investigation of their expression within the diagnostic FFPET
biopsy, with survival as the outcome and IPI and COO as co-variates. New digital
hybridization technologies that are specifically applicable to FFPET are now
available. In keeping with previous reports, using a DMGE platform we observed
strong correlation between gene expression in paired frozen/paraffin samples across
101
24
a range of genes (15, 16, 39), indicating that this approach does accurately quantify
RNA on FFPET in R-CHOP treated DLBCL patients.
Interestingly, intratumoral CD163 expression alone was not prognostic.
However the CD8:CD163 ratio found that those with a lower ratio had inferior
outcome. This may explain why results of intratumoral CD163 alone as a
prognosticator are inconsistent (40, 41), and emphasises the importance of
measuring several markers to more accurately reflect the balance of TME immunity.
Similarly, CD8:CD163 ratios as a measure of net anti-tumoral immunity predicted
outcome more effectively than CD8 alone. CD8:CD163 was independent of IPI and
COO. Results were validated in an external R-CHOP gene-expression cohort.
Interestingly, although CD8:CD163 was predictive in a CHOP cohort, this was not
independent of conventional prognosticators. This may reflect the relatively
increased importance of TAMs in those treated with chemo-immunotherapy versus
chemotherapy-alone, and is consistent with our findings that patient monocytes
suppress R-ADCC.
Gene expression has identified distinct molecular subtypes of DLBCL, based
on the putative cell-of-origin of the malignant B-cell, termed ‘germinal-centre B-cell
(GCB)’ and ‘activated B-cell like (ABC)’ (28, 42, 43). LMO2, a GCB marker, alone is
also an independent predictor of survival (24, 25). In R-CHOP treated patients,
CD8:CD163 ratios were independent of COO, and enhanced the prognostic ability of
LMO2. The combination of a marker of B-cell differentiation with net anti-tumoral
immunity appears to better distinguish patients (than IPI or COO alone) with
markedly different survival. Furthermore LMO2 combined with CD8:CD163
successfully allowed segregation within low and high IPI groups. IPI is influenced by
patient fitness, age and tumor burden, which are non-over-lapping with CD8:CD163.
102
25
For the R-CHOP Australian tissue cohort, the application of the immunobiological
score to low and high-risk IPI segregated patients into groupings with four distinct,
well-spaced survival outcomes of 91%, 76%, 44% and 26%. Similar results were
seen with the external validation cohort. Although it is known that approximately 30%
of patients will remain refractory or relapse following conventional first-line chemo-
immunotherapy, upfront identification of this group (as candidates for alternative
induction therapy) has remained problematic. A combined clinical and
immunobiological score may permit rational selection of patients in whom novel
therapies should be tested.
CD137 expression on NK-cells is an important marker of mAB-mediated anti-
cancer cell ADCC (44). Interestingly we found that within ex-vivo blood taken from
patients enrolled into the NHL21 trial, CD137 up-regulation was reduced in NK-cells
relative to healthy participants upon exposure to rituximab coated DLBCL targets.
CD137 has also been proposed as a biomarker of tumor-reactive T-cells (45). We
observed that within the diagnostic tissue, CD137 gene expression alone had
equivalent prognostic ability to CD8. However, CD137:CD163 was less
discriminatory than CD8:CD163, and unlike CD8:CD163 was not prognostic by
multivariate analysis. A two-gene-score that combined LMO2 with the immune-
effector molecule CD137 has previously been recognized to be prognostic (26).
Notably, the immunobiological scoring system had markedly greater discriminatory
ability than the two-gene-score, and was especially effective in identifying a group at
risk of a particularly poor outcome.
These findings have other therapeutic implications. Firstly, monocyte
depletion in patients enhanced NK-cell mediated R-ADCC but not Ob-ADCC,
indicating that the immunosuppressive effects of monocytes in DLBCL may be
103
26
overcome by obinutuzumab (a type II antiCD20 monoclonal antibody). Clinical trials
comparing R-CHOP versus Ob-CHOP are ongoing and translational studies in that
population could further explore this hypothesis. Another notable finding was the
striking co-clustering of CD8 and CD163 within the tissue and circulation. This is in
line with emerging data that up-regulation of immune-checkpoints is an adaptive
(rather than constitutive) immune-checkpoint response to regulate immune-effector
activation (46). The correlations were significant but modest, reflecting the variable
success of the host to counter anti-tumoral immunity within the TME. Blockade of
other immune-checkpoints is a promising therapeutic approach (47), and strategies
that target CD163 may be similarly beneficial (48).
Immune-based strategies are gaining ground in DLBCL (49). The kinetics of
immune cells will likely impact the efficacy of these approaches, and may influence
dose-scheduling. There is minimal data on circulating moMDSC kinetics in DLBCL,
with a small study finding moMDSC returned to normal after therapy (12). We found
CD163+moMDSC reduced post-cycle 4, accompanied by an increase in ex-vivo
BMDCs and pDCs (i.e. antigen-presenting cells that orchestrate adaptive and innate
immunity). Monocytes are an important source of BMDCs in-vivo (23). Reduction in
moMDSC may be associated with an enhanced ability of monocytes differentiating
into BMDCs. The immunosuppressive profile of monocytes reduced by post-cycle 4,
including down-regulation of TH2 cytokines and TAM associated genes, and up-
regulation of STAT1 (18, 50). Similarly plasma CD163 reduced by post-cycle 4. As
with cHL, plasma CD163 levels associated with stage and reduced lymphocytes
(36).
Our data emphasizes the importance of capturing net anti-tumoral immunity
within the TME, by measuring the relative balance of immune-effector to
104
27
monocyte/macrophage-checkpoints. The immunobiological score is a robust, easily
standardized 3-gene assay applicable to FFPET that segregates IPI into markedly
disparate survival categories. The data indicates a link between human M2 TAMs
and moMDSC, and demonstrate that CD163 identifies a highly immunosuppressive
subset of moMDSC in DLBCL. Further investigation of CD163+moMDSC as a
therapeutic target is warranted, particularly in those with adverse CD8:CD163 ratios.
105
28
Acknowledgements
The authors would like to thank Bala Shankar and Ruth Columbus from the
ALLG, the ALLG Tissue Bank, all participating patients, healthy participants,
hospitals and PET/CT centres. The monoclonal antibody Obinutuzumab (GA101)
was provided by Roche Glycart AG (Schlieren, Switzerland).
106
29
References: 1. Coiffier, B. 2005. State-of-the-art therapeutics: diffuse large B-cell
lymphoma. J Clin Oncol 23:6387-6393. 2. Maurer, M.J., Ghesquieres, H., Jais, J.P., Witzig, T.E., Haioun, C.,
Thompson, C.A., Delarue, R., Micallef, I.N., Peyrade, F., Macon, W.R., et al. 2014. Event-free survival at 24 months is a robust end point for disease-related outcome in diffuse large B-cell lymphoma treated with immunochemotherapy. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 32:1066-1073.
3. Roschewski, M., Dunleavy, K., and Wilson, W.H. 2014. Moving beyond rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone for diffuse large B-cell lymphoma. Leukemia & lymphoma.
4. Keane, C., Gill, D., Vari, F., Cross, D., Griffiths, L., and Gandhi, M. 2013. CD4(+) tumor infiltrating lymphocytes are prognostic and independent of R-IPI in patients with DLBCL receiving R-CHOP chemo-immunotherapy. American journal of hematology 88:273-276.
5. Li, Z.M., Huang, J.J., Xia, Y., Sun, J., Huang, Y., Wang, Y., Zhu, Y.J., Li, Y.J., Zhao, W., Wei, W.X., et al. 2012. Blood Lymphocyte-to-Monocyte Ratio Identifies High-Risk Patients in Diffuse Large B-Cell Lymphoma Treated with R-CHOP. PLoS One 7:e41658.
6. Porrata, L.F., Ristow, K., Habermann, T.M., Ozsan, N., Dogan, A., Macon, W., Colgan, J.P., Witzig, T.E., Inwards, D.J., Ansell, S.M., et al. 2012. Absolute monocyte/lymphocyte count prognostic score is independent of immunohistochemically determined cell of origin in predicting survival in diffuse large B-cell lymphoma. Leuk Lymphoma.
7. Coussens, L.M., and Werb, Z. 2002. Inflammation and cancer. Nature 420:860-867.
8. Wilcox, R.A., Wada, D.A., Ziesmer, S.C., Elsawa, S.F., Comfere, N.I., Dietz, A.B., Novak, A.J., Witzig, T.E., Feldman, A.L., Pittelkow, M.R., et al. 2009. Monocytes promote tumor cell survival in T-cell lymphoproliferative disorders and are impaired in their ability to differentiate into mature dendritic cells. Blood 114:2936-2944.
9. Afshar-Sterle, S., Zotos, D., Bernard, N.J., Scherger, A.K., Rodling, L., Alsop, A.E., Walker, J., Masson, F., Belz, G.T., Corcoran, L.M., et al. 2014.
107
30
Fas ligand-mediated immune surveillance by T cells is essential for the control of spontaneous B cell lymphomas. Nature medicine.
10. Ahmed, N., Heslop, H.E., and Mackall, C.L. 2010. T-cell-based therapies for malignancy and infection in childhood. Pediatric clinics of North America 57:83-96.
11. Obermajer, N., Muthuswamy, R., Lesnock, J., Edwards, R.P., and Kalinski, P. 2011. Positive feedback between PGE2 and COX2 redirects the differentiation of human dendritic cells towards stable myeloid-derived suppressor cells. Blood.
12. Tadmor, T., Fell, R., Polliack, A., and Attias, D. 2012. Absolute monocytosis at diagnosis correlates with survival in diffuse large B-cell lymphoma-possible link with monocytic myeloid-derived suppressor cells. Hematol Oncol.
13. Lin, Y., Gustafson, M.P., Bulur, P.A., Gastineau, D.A., Witzig, T.E., and Dietz, A.B. 2011. Immunosuppressive CD14+HLA-DR(low)/- monocytes in B-cell non-Hodgkin lymphoma. Blood 117:872-881.
14. Geiss, G.K., Bumgarner, R.E., Birditt, B., Dahl, T., Dowidar, N., Dunaway, D.L., Fell, H.P., Ferree, S., George, R.D., Grogan, T., et al. 2008. Direct multiplexed measurement of gene expression with color-coded probe pairs. Nature biotechnology 26:317-325.
15. Masque-Soler, N., Szczepanowski, M., Kohler, C.W., Spang, R., and Klapper, W. 2013. Molecular classification of mature aggressive B-cell lymphoma using digital multiplexed gene expression on formalin-fixed paraffin-embedded biopsy specimens. Blood 122:1985-1986.
16. Scott, D.W., Wright, G.W., Williams, P.M., Lih, C.J., Walsh, W., Jaffe, E.S., Rosenwald, A., Campo, E., Chan, W.C., Connors, J.M., et al. 2014. Determining cell-of-origin subtypes of diffuse large B-cell lymphoma using gene expression in formalin-fixed paraffin embedded tissue. Blood.
17. Lenz, G., Wright, G., Dave, S.S., Xiao, W., Powell, J., Zhao, H., Xu, W., Tan, B., Goldschmidt, N., Iqbal, J., et al. 2008. Stromal gene signatures in large-B-cell lymphomas. N Engl J Med 359:2313-2323.
108
31
18. Gabrilovich, D.I., and Nagaraj, S. 2009. Myeloid-derived suppressor cells as regulators of the immune system. Nature reviews. Immunology 9:162-174.
19. Zhao, X., Rong, L., Zhao, X., Li, X., Liu, X., Deng, J., Wu, H., Xu, X., Erben, U., Wu, P., et al. 2012. TNF signaling drives myeloid-derived suppressor cell accumulation. The Journal of clinical investigation 122:4094-4104.
20. Paul, S.P., Taylor, L.S., Stansbury, E.K., and McVicar, D.W. 2000. Myeloid specific human CD33 is an inhibitory receptor with differential ITIM function in recruiting the phosphatases SHP-1 and SHP-2. Blood 96:483-490.
21. DeNardo, D.G., Brennan, D.J., Rexhepaj, E., Ruffell, B., Shiao, S.L., Madden, S.F., Gallagher, W.M., Wadhwani, N., Keil, S.D., Junaid, S.A., et al. 2011. Leukocyte complexity predicts breast cancer survival and functionally regulates response to chemotherapy. Cancer discovery 1:54-67.
22. Moller, H.J. 2012. Soluble CD163. Scandinavian journal of clinical and laboratory investigation 72:1-13.
23. Mellman, I., and Steinman, R.M. 2001. Dendritic cells: specialized and regulated antigen processing machines. Cell 106:255-258.
24. Natkunam, Y., Farinha, P., Hsi, E.D., Hans, C.P., Tibshirani, R., Sehn, L.H., Connors, J.M., Gratzinger, D., Rosado, M., Zhao, S., et al. 2008. LMO2 protein expression predicts survival in patients with diffuse large B-cell lymphoma treated with anthracycline-based chemotherapy with and without rituximab. J Clin Oncol 26:447-454.
25. Natkunam, Y., Zhao, S., Mason, D.Y., Chen, J., Taidi, B., Jones, M., Hammer, A.S., Hamilton Dutoit, S., Lossos, I.S., and Levy, R. 2007. The oncoprotein LMO2 is expressed in normal germinal-center B cells and in human B-cell lymphomas. Blood 109:1636-1642.
26. Alizadeh, A.A., Gentles, A.J., Alencar, A.J., Liu, C.L., Kohrt, H.E., Houot, R., Goldstein, M.J., Zhao, S., Natkunam, Y., Advani, R.H., et al. 2011. Prediction of survival in diffuse large B-cell lymphoma based on the expression of 2 genes reflecting tumor and microenvironment. Blood 118:1350-1358.
109
32
27. Wilcox, R.A., Ristow, K., Habermann, T.M., Inwards, D.J., Micallef, I.N., Johnston, P.B., Colgan, J.P., Nowakowski, G.S., Ansell, S.M., Witzig, T.E., et al. 2012. The absolute monocyte count is associated with overall survival in patients newly diagnosed with follicular lymphoma. Leukemia & lymphoma 53:575-580.
28. Alizadeh, A.A., Eisen, M.B., Davis, R.E., Ma, C., Lossos, I.S., Rosenwald, A., Boldrick, J.C., Sabet, H., Tran, T., Yu, X., et al. 2000. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403:503-511.
29. Monti, S., Savage, K.J., Kutok, J.L., Feuerhake, F., Kurtin, P., Mihm, M., Wu, B., Pasqualucci, L., Neuberg, D., Aguiar, R.C., et al. 2005. Molecular profiling of diffuse large B-cell lymphoma identifies robust subtypes including one characterized by host inflammatory response. Blood 105:1851-1861.
30. Marmey, B., Boix, C., Barbaroux, J.B., Dieu-Nosjean, M.C., Diebold, J., Audouin, J., Fridman, W.H., Mueller, C.G., and Molina, T.J. 2006. CD14 and CD169 expression in human lymph nodes and spleen: specific expansion of CD14+CD169- monocyte-derived cells in diffuse large B-cell lymphomas. Hum Pathol 37:68-77.
31. Challa-Malladi, M., Lieu, Y.K., Califano, O., Holmes, A.B., Bhagat, G., Murty, V.V., Dominguez-Sola, D., Pasqualucci, L., and Dalla-Favera, R. 2011. Combined genetic inactivation of beta2-Microglobulin and CD58 reveals frequent escape from immune recognition in diffuse large B cell lymphoma. Cancer Cell 20:728-740.
32. Adams, D.L., Martin, S.S., Alpaugh, R.K., Charpentier, M., Tsai, S., Bergan, R.C., Ogden, I.M., Catalona, W., Chumsri, S., Tang, C.M., et al. 2014. Circulating giant macrophages as a potential biomarker of solid tumors. Proceedings of the National Academy of Sciences of the United States of America 111:3514-3519.
33. Kusmartsev, S., Nagaraj, S., and Gabrilovich, D.I. 2005. Tumor-associated CD8+ T cell tolerance induced by bone marrow-derived immature myeloid cells. J Immunol 175:4583-4592.
34. Wang, B., Li, Q., Qin, L., Zhao, S., Wang, J., and Chen, X. 2011. Transition of tumor-associated macrophages from MHC class II(hi) to
110
33
MHC class II(low) mediates tumor progression in mice. BMC immunology 12:43.
35. Tippett, E., Cheng, W.J., Westhorpe, C., Cameron, P.U., Brew, B.J.,Lewin, S.R., Jaworowski, A., and Crowe, S.M. 2011. Differentialexpression of CD163 on monocyte subsets in healthy and HIV-1 infectedindividuals. PloS one 6:e19968.
36. Jones, K., Vari, F., Keane, C., Crooks, P., Nourse, J.P., Seymour, L.A.,Gottlieb, D., Ritchie, D., Gill, D., and Gandhi, M.K. 2013. Serum CD163and TARC as Disease Response Biomarkers in Classical HodgkinLymphoma. Clin Cancer Res 19:731-742.
37. Jensen, T.O., Schmidt, H., Moller, H.J., Hoyer, M., Maniecki, M.B.,Sjoegren, P., Christensen, I.J., and Steiniche, T. 2009. Macrophagemarkers in serum and tumor have prognostic impact in American JointCommittee on Cancer stage I/II melanoma. J Clin Oncol 27:3330-3337.
38. Medrek, C., Ponten, F., Jirstrom, K., and Leandersson, K. 2012. Thepresence of tumor associated macrophages in tumor stroma as aprognostic marker for breast cancer patients. BMC cancer 12:306.
39. Scott, M.A., Gandhi, M.K., Jestice, H.K., Mahendra, P., Bass, G., andMarcus, R.E. 1998. A trend towards an increased incidence of chronicgraft-versus-host disease following allogeneic peripheral bloodprogenitor cell transplantation: a case controlled study. Bone MarrowTransplant 22:273-276.
40. Wada, N., Zaki, M.A., Kohara, M., Ogawa, H., Sugiyama, H., Nomura, S.,Matsumura, I., Hino, M., Kanakura, Y., Inagaki, H., et al. 2012. Diffuselarge B cell lymphoma with an interfollicular pattern of proliferationshows a favourable prognosis: a study of the Osaka Lymphoma StudyGroup. Histopathology 60:924-932.
41. Yoshida, N., Oda, M., Kuroda, Y., Katayama, Y., Okikawa, Y., Masunari,T., Fujiwara, M., Nishisaka, T., Sasaki, N., Sadahira, Y., et al. 2013.Clinical Significance of sIL-2R Levels in B-Cell Lymphomas. PloS one8:e78730.
42. Rosenwald, A., Wright, G., Chan, W.C., Connors, J.M., Campo, E., Fisher,R.I., Gascoyne, R.D., Muller-Hermelink, H.K., Smeland, E.B., Giltnane,J.M., et al. 2002. The use of molecular profiling to predict survival after
111
34
chemotherapy for diffuse large-B-cell lymphoma. N Engl J Med 346:1937-1947.
43. Wright, G., Tan, B., Rosenwald, A., Hurt, E.H., Wiestner, A., and Staudt, L.M. 2003. A gene expression-based method to diagnose clinically distinct subgroups of diffuse large B cell lymphoma. Proc Natl Acad Sci U S A 100:9991-9996.
44. Kohrt, H.E., Houot, R., Weiskopf, K., Goldstein, M.J., Scheeren, F., Czerwinski, D., Colevas, A.D., Weng, W.K., Clarke, M.F., Carlson, R.W., et al. 2012. Stimulation of natural killer cells with a CD137-specific antibody enhances trastuzumab efficacy in xenotransplant models of breast cancer. The Journal of clinical investigation 122:1066-1075.
45. Ye, Q., Song, D.G., Poussin, M., Yamamoto, T., Best, A., Li, C., Coukos, G., and Powell, D.J., Jr. 2014. CD137 accurately identifies and enriches for naturally occurring tumor-reactive T cells in tumor. Clinical cancer research : an official journal of the American Association for Cancer Research 20:44-55.
46. Taube, J.M., Anders, R.A., Young, G.D., Xu, H., Sharma, R., McMiller, T.L., Chen, S., Klein, A.P., Pardoll, D.M., Topalian, S.L., et al. 2012. Colocalization of inflammatory response with B7-h1 expression in human melanocytic lesions supports an adaptive resistance mechanism of immune escape. Science translational medicine 4:127ra137.
47. Westin, J.R., Chu, F., Zhang, M., Fayad, L.E., Kwak, L.W., Fowler, N., Romaguera, J., Hagemeister, F., Fanale, M., Samaniego, F., et al. 2013. Safety and activity of PD1 blockade by pidilizumab in combination with rituximab in patients with relapsed follicular lymphoma: a single group, open-label, phase 2 trial. The lancet oncology.
48. Graversen, J.H., Svendsen, P., Dagnaes-Hansen, F., Dal, J., Anton, G., Etzerodt, A., Petersen, M.D., Christensen, P.A., Moller, H.J., and Moestrup, S.K. 2012. Targeting the hemoglobin scavenger receptor CD163 in macrophages highly increases the anti-inflammatory potency of dexamethasone. Molecular therapy : the journal of the American Society of Gene Therapy 20:1550-1558.
112
35
49. Elstrom, R.L., Martin, P., and Leonard, J.P. 2008. New biologic agents and immunologic strategies. Hematol Oncol Clin North Am 22:1037-1049, x-xi.
50. Yu, H., Pardoll, D., and Jove, R. 2009. STATs in cancer inflammation and immunity: a leading role for STAT3. Nature reviews. Cancer 9:798-809.
113
36
Figure 1: Depletion of monocytes enhances T-cell proliferation and rituximab
but not obinutuzumab mediated ADCC.
A. Monocytes are increased in poor-risk DLBCL patients compared to healthy
participants;
B-D. CD3+/CD4+/and CD8+ T-cell proliferation with (depleted) and without (intact)
monocyte depletion in healthy participants and pre-therapy patient blood samples;
E. R-ADCC measured in healthy participants (left) and pre-therapy DLBCL patients
(right) with and without monocyte depletion;
F. Ob-ADCC measured in healthy participants (left) and pre-therapy DLBCL patients
(right) with and without monocyte depletion;
G. Monocytes impair R-ADCC but not Ob-ADCC in pre-therapy DLBCL patients;
H. Reduced activation of NK-cells during R- but not Ob- ADCC in healthy
participants and pre-therapy DLBCL patients.
Figure 2: Characterization of CD163+moMDSC.
A. Representative flow cytometry plots showing ‘classical’ (CD14+CD16-),
‘intermediate’ (CD14+CD16+) and ‘non-classical’ (CD14dimCD16+) monocytes in
healthy PBMC, with enrichment of CD14+HLA-DRlo (moMDSC) in classical
monocytes in a pre-therapy DLBCL patient;
B. MoMDSCs are elevated in pre-therapy DLBCL patients compared to healthy
participants;
C. HLA-DR expression on CD14+ monocytes is lower at pre-therapy compared to
the same participant at post-cycle 4 and to a healthy participant. The CD163
expression on DRlo monocytes (moMDSC) is raised pre-therapy. CD163 is indicated
by the shaded histogram, the open histogram is the control;
114
37
D. Pre-therapy DLBCL moMDSC and CD163+CD14+ cells are correlated;
E. CD163+moMDSC are elevated in pre-therapy DLBCL patients compared to
healthy participants;
F Arginase activity was elevated compared to healthy participants;
G CD163 mean fluorescent intensity (MFI) on moMDSC correlated with arginase
activity in patients and healthy participants.
Figure 3: Cell-surface phenotype of CD163himoMDSC versus CD163lomoMDSC.
Proportion of cell surface markers expressed in CD163hi and CD163lo moMDSC
compared for (A) CD11b, (B) CD11c, (C) CD68, (D) CD120b, (E) CD62L, (F) CD33,
(G) CD80 (H) CD86 and (I) CSF-R1.
Figure 4: Pre-therapy monocyte subsets are elevated in DLBCL patients
remaining interim-PET/CT+ve after 4 cycles of R-CHOP chemo-immunotherapy.
A. Pre-therapy absolute (i) monocyte (ii) moMDSC and (iii) CD163+moMDSC subset
counts in DLBCL patients, grouped by post-cycle 4 interim-PET/CT results.
B. Pre-therapy lymphocyte subset:monocyte ratios in DLBCL patients, grouped by
post-cycle 4 interim-PET/CT results. Numerators: panel (i) total lymphocytes; (ii) NK-
cells; (iii) T-cells.
C. Pre-therapy lymphocyte:CD14+HLA-DRlo ratios in DLBCL patients, grouped by
post-cycle 4 interim-PET/CT results. Numerators: panel (i) total lymphocytes; (ii) NK-
cells; (iii) T-cells.
D. Pre-therapy lymphocyte subset:CD163+moMDSC ratios in DLBCL patients,
grouped by post-cycle 4 interim-PET/CT results. Numerators: panel (i) total
lymphocytes; (ii) NK-cells; (iii) T-cells.
115
38
Figure 5: Kinetics of circulating monocytes and dendritic cells in poor-risk
DLBCL patients during R-CHOP.
A. Paired moMDSC counts in blood samples taken pre-therapy and post-cycle 4;
B. Paired CD163+moMDSC counts in peripheral blood samples taken pre-therapy
and post-cycle 4. Arrows denote mean values;
C. Heat map showing gene expression microarray analysis on circulating monocytes
isolated from 5 healthy participants and 6 DLBCL patients taken pre-therapy and
post-cycle 4, showing up-regulation of antigen-presentation genes, and down-
regulation of TAM associated and TH2 cytokine genes at post-cycle 4;
D. Paired BMDC at pre-therapy and post-cycle 4;
E. CpG-DNA induced interferon-α production from pDC in patients and healthy
participants;
F. Paired pDC at pre-therapy and post-cycle 4.
Arrows denotes mean values.
Figure 6: A binary composite score of LMO2+/-CD8:CD163 added to the ability
of IPI and COO to stratify patient survival.
Kaplan–Meier estimates of OS are shown. A. Patient tissues were divided by median
LMO2 into LMO2 high and low; B. A binary composite score of high LMO2 and/or
high CD8:CD163 (‘positive’), versus low LMO2 and low CD8:CD163 (‘double-
negative’). C. In low-risk (0-2) IPI patients, composite score positive patients had the
higher probability of survival; D. For high-risk IPI, double-negative score patients had
the lower probability of survival; E. GCB positive patients had the higher probability
of survival than GCB double-negative; F. Non-GCB positive patients had the higher
probability of survival than non-GCB double-negative.
116
39
Figure 7: External validation of the immunobiological score in 199 CHOP and
233 R-CHOP treated DLBCL patients.
The immunobiological score was tested using Kaplan–Meier estimates of survival in
a chemotherapy-alone (CHOP) and a chemo-immunotherapy (R-CHOP) treated
cohort, in the left and right columns respectively. All CHOP (A) and R-CHOP (B)
patients; Low-risk IPI CHOP (C) and R-CHOP (D) patients; High-risk IPI (>2) CHOP
(E) and R-CHOP (F) patients; GCB CHOP (G) and R-CHOP (H) patients; non-GCB
CHOP (I) and R-CHOP (J) patients.
117
40
Supplemental Table S1. Optimization of immune-effector and
monocyte/macrophage-checkpoint combinations. Significant P values are in
bold.
Supplemental Table S2. Characteristics of the blood and tissue patient cohorts.
Supplemental Table S3. Flow cytometry antibodies used.
Supplemental Figure S1: Plasma CD163 is elevated in poor-risk DLBCL.
A. Plasma CD163 is elevated in pre-therapy DLBCL patients compared to post-cycle
4 (paired analysis) and both pre-therapy and post-cycle 4 plasma CD163 are
elevated relative to healthy participants.
B. ROC analysis of plasma CD163 levels in pre-therapy DLBCL patients and healthy
participants.
C. Plasma CD163 was inversely associated with absolute lymphocyte counts (using
a median cut-off for 120/ml).
D. Plasma CD163 was associated with advanced stage.
E. Plasma CD163 was associated with age >60 years.
F. Plasma CD163 was associated with higher IPI.
Panels A, D-F show mean and SEM.
Supplemental Figure S2. IPI and COO stratification in the Australian tissue
cohort. A. Low (0-2) and high (3-5) risk IPI; B GCB and non-GCB.
118
41
Supplemental Figure S3. CD8:CD163 ratio adds to the predictive ability of
clinical and malignant B-cell biology prognosticators.
A. Low-risk IPI; B. B. High-risk IPI; C.GCB; D. Non-GCB.
119
Figure 1.
20
40
60
80 P
% c
lass
ical
mon
ocyt
es li
ve < 0.0001
Intact Depleted Intact Depleted
0
20
40
60
P=NS
Healthy Pre-Therapy
P=0.03
% C
FSEl
o C
D8
T-ce
ll
Intact Depleted Intact Depleted0
10
20
30
40
P=NS
Healthy Healthy
Pre-Therapy
P=0.03
% C
FSEl
o C
D3
T-ce
ll
Intact Depleted Intact Depleted-10
0
10
20
30
40
50P=NS
Healthy Pre-Therapy
P=0.027
% C
FSEl
o C
D4
T-ce
ll
A) B) C)Pre-Therapy
D)
Intact Depleted Intact Depleted0
20406080
100
Healthy Pre-Therapy
% A
DC
C
RituximabP=NS P=0.01
E)Intact Depleted Depleted
020406080
100 P=NS
Healthy Pre-Therapy
P=NS
% A
DC
C
Obinutuzumab
F) H)
Healthy Healthy0
20
40
60
80
100 P=0.02
ObinutuzumabRituximab
% C
D13
7 N
K-ce
lls
Pre-Therapy Pre-Therapy
P=NS
G)
0
20
40
60
80
%A
DC
C
P=0.01
Healthy HealthyPre-Therapy Pre-Therapy
Healthy HealthyPre-Therapy Pre-Therapy
DepletedIntactP=NS P=NS P=NS
Rituximab Obinutuzumab Rituximab Obinutuzumab
120
CD16
HLA-DR HLA-DR
Healthy Pre
1
231 23
1
221
CD14 CD14
CD16
A) C)B)
Healthy Pre-Therapy Post-Cycle 4
CD14
HLA-DR HLA-DR HLA-DR
CD163 CD163 CD163
Healthy Pre0
10
20
30
40
50 P
% m
o MD
SC
<0.0001
8Arginase
(milli units per 100,000 monocytes)
CD
163
MFI
on
moM
DSC
0 2 4 60
1000
2000
3000r =0.45P=0.002
D) F) G)Healthy Pre
0
10
20
30P=0.001
%C
D16
3+m
oMDS
C
E)Healthy Pre
0
2
4
6
8
Arg
inas
e (m
illi u
nits
per
100
,000
mon
o)
P=0.0005
CD163 counts per µl
0 100 200 300 400 15000
100
200
300
400
1000
1500
moM
DSC
cou
nts
per µ
l
r =0.41P<0.0001
Figure 2.121
B)A) C)CD163hi CD163 lo
0
20
40
60
80
100
%C
D11
b of
moM
DSC
P=NS
CD163hi CD163 lo0
50
100
%C
D68
of m
oMD
SC
P=NS
CD163hi CD163lo0
20
40
60
80
100
%C
D11
c of
moM
DSC
P=0.035
CD163hi CD163 lo0
50
100
%C
D12
0b o
f moM
DSC
=0P .001
CD163hi CD163lo0
5
10
1520406080
100
%C
SF-1
R o
f moM
DSC
<0.0001P
CD163hi CD163 lo0
50
100
%C
D33
hi o
f moM
DSC
P=0.011
CD163hi CD163 lo0
50
100
%C
D80
of
moM
DSC
P<0.009
CD163hi CD163 lo0
50
100
%C
D86
of m
oMD
SC
P=NS
CD163hi CD163 lo0
50
100
%C
D62
L of
moM
DSC
P<0.0001
E)D) F)
H)G) I)
Figure 3.122
PET -VE PET +VE0
20
40
60
400
800 P<0.0001
PET -VE PET +VE0
5
10
50
100
150 P=0.0001
NK
-cel
/ C
D16
3+ m
oMD
SC
3
PET-ve PET +VE 0
50
100
150
200400600
P=0.0002
T-ce
ll / C
D16
3+ m
oMD
SC
PET -VE PET +VE0
500
1000
1500
2000
2500 P=NS
mon
ocyt
e pe
r L
PET -VE PET +VE0
500
1000
1500 P=0.004
moM
DSC
C
per
L
A)
PET -VE PET +VE0
2
4
6
8
10
lym
phoc
yte
/ mon
ocyt
e
ePET -VE PET +VE
0
200
400
600
800 P<0.0001
CD
163+
moM
DSC
per
L
PET -VE PET +VE0.0
0.5
1.0
1.5N
K-c
ell /
mon
ocyt
e
e
PET -VE PET +VE0
2
4
6
8
10
T-ce
ll / m
onoc
yte
e
P=NS
PET -VE PET +VE0
1
2
3
468
10 P=0.01
NK
-cel
l / m
oMD
SC
C
PET -VE PET +VE0
10
20
30
4050
100150
P=0.03T-
cell
/ moM
DSC
lym
pho
/ CD
163
/
+ moM
DSC
150 P=0.02
PET -VE PET +VE0
50
100
Lym
phoc
yte
/ moM
DSC
B)
C)
P=NS P=NS
(i) (ii) (iii)
(i) (ii) (iii)
D)
(i) (ii) (iii)
(i) (ii) (iii)
Figure 4. 123
A)PRE POST
0
500
1000
1500
P=0.0001m
oMD
SC/m
L
PRE POST0
200
400
600
800 P=0.0002
CD
163+
moM
DSC
/mL
B)
Pre
Pre
Pre
Pre
Pre
Pre
Post
Post
Post
Post
Post
Hea
lthy
Hea
lthy
Hea
lthy
Hea
lthy
Hea
lthy
Post
HLA-DRA
HLA-DRB4
HLA-DMBHLA-DOA
HLA-DRB3
HLA-DRB6HLA-DRB1
MFGE8IL4IL10IL1R2IL13CD163
Hea
lthy
Post
Post
Post
Post
Hea
lthy
Hea
lthy
Post
Pre
Pre
Post
Hea
lthy
Pre
Pre
Hea
lthy
Pre
Pre
D)0
5
10
15
20406080
1000.0002P=
BM
DC
cou
nt/m
L
POSTPRE 0
10
20
30
40
% IF
N a
lpha
PD
C
P=0.0006
E) Healthy PRE
F) PRE POST0
5
10
15
400
800
PDC
cou
nts/m
L P<0.0001
Figure 5.124
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
LMO2>MedianLMO2<Median
P = 0.0046
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negativeP <0.0001
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negativeP < 0.0001
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negativeP <0.0001
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negativeP = 0.002
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negativeP = 0.007
A B
C D
E F
Figure 6.125
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble Negative
P = 0.0007
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negative
P = 0.13
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negative
P = 0.006
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negativeP = 0.056
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble Negative
P = <0.0001
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negativeP = 0.0003
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negative
P = 0.0007
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negative
P = 0.0007
A B
C D
E F
G H
I J
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negativeP = 0.07
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
PositiveDouble negative
P = 0.008
Figure 7.126
Supplemental Methods
ALLGNHL21 Clinical study
Eligible patients were aged between 18-70 years with CD20+ DLBCL with IPI
2-5 or IPI 0-1 with bulky tumor (≥7.5 cm), and FDG-PET–positive evaluable disease.
Patients with non-bulky IPI 0-1 disease were excluded. Patients with non-bulky IPI 0-
1 disease were excluded. Those with previously treated lymphoma, primary central
nervous system (CNS) lymphoma, transformed lymphoma or follicular lymphoma
grade 3B patients were ineligible. Patients had to have an Eastern Cooperative
Oncology Group (ECOG) performance status between 0-3 and be considered
suitable for R-CHOP-14, and high-dose chemotherapy with autologous stem-cell
rescue. Clinical parameters were source verified and IPI centrally reviewed. All
patients had serum creatinine ≤ 150 µmol/L, total bilirubin level ≤ 30 mmol/L,
transaminases ≤ 2.5 maximum normal level, neutrophils ≥1.5 x 109/L or platelets ≥
100 x 109/L unless due to lymphoma. Patients had to be HIV-negative, hepatitis B
virus (HBV) surface antigen negative, and/or HBV core antibody negative with HBV
surface antibody titre <100iu/ml unless clearly due to prior vaccination. Patients with
uncompensated cardiac failure, chronic lung disease with hypoxia, severe
psychiatric disease or any history of cancer during the last 5 years (with the
exception of non-melanoma skin tumours or in situ cervical carcinoma were
excluded.
All interim-PET/CT scans were reviewed centrally alongside diagnostic pre-
therapy scans (blinded to the local PET/CT assessment), to verify residual abnormal
FDG uptake at sites of previously identified activity. Metabolic response
categorisation used a 5-point scale (1). Patients with a complete metabolic response
(≤3 on 5-point scale) were classified as interim-PET/CT-ve, whereas those remaining
127
PET-FDG-avid (>3) were classified as interim-PET/CT positive. R-CHOP was given
as follows: day 1 rituximab 375 mg/m2 intravenous (IV) infusion, cyclophosphamide
750 mg/m2 IV, doxorubicin 50 mg/m2 IV, vincristine 1.4 mg/m2 (maximum 2 mg) IV,
and prednisone 100 mg/day orally for 5 days. Further therapy was dependent on the
outcome of the interim-PET/CT scan and is not the subject of this study (further
details are available under Clinical Trials Identifier: ACTRN12609001077257).
Immuno-phenotyping The antibodies used are outlined in supplemental Table S2. The flow
cytometry data was acquired using a BD LSRII flow cytometer controlled by
FACSDiva software (BD, Australia). Data analysis was performed on compensated
data using FlowJo 4.2 or 9.x software (Tree Star, USA). MoMDSC were defined as
CD14+ monocytes which were DRlo, i.e. with DR expression lower than the median
DR expression of B-cells in the PBMC population. Blood myeloid dendritic cells
(BMDCs) and plasmacytoid dendritic cells (pDC) were analysed as previously
outlined (2). Briefly, among the lineage negative DRhi cells, pDC were
CD123+CD11c- cells while BMDC were CD123-CD11c+. Intratumoral flow cytometry,
was performed as previously described (3).
Functional assays
T-cell proliferation and antiCD20-ADCC assays were performed with and
without monocyte depletion. Monocytes were depleted using an immunomagnetic
CD14 selection procedure (EasySep, Stemcell Technologies) as per manufacturer
recommendations achieving >90% monocyte depletion.
The antiCD20-ADCC assay measures killing of a carboxyfluorescein
succinimidyl ester (CFSE)-labelled CD20-positive DLBCL cell-line (SU-DHL4)
128
incubated for 4-6 hours +/- PBMC and +/- rituximab/obinutuzumab. Classical
Hodgkin Lymphoma (cHL) cell-lines (to confirm specificity for CD20) were used as
CD20-ve control targets. The absolute change in targets was measured by flow
cytometry. Subtraction of the number of target cells lysed by addition of
rituximab/obinutuzumab alone (direct lysis) from the total number lysed by antiCD20
monoclonal antibody with PBMC, enabled enumeration of antiCD20-ADCC mediated
killing. CD107ab de-granulation of CD56+ and CD14+ cells was measured by flow
cytometry.
For T-cell proliferation, CFSE stained mononuclear cells were seeded in 96-
well round-bottom plates at 2-5x105 cells/well in RPMI 1640 (Invitrogen) with 10%
foetal calf serum (Invitrogen) supplemented with 2 mmol/l L-glutamine and
1×Penicilin/Streptomycin (Invitrogen), termed ‘R-10’, with 10 u/ml interleukin-2. Anti-
CD3/CD28/CD137 beads (Invitrogen) at a 1:10 bead:cell ratio was added to
stimulate polyclonal expansion. Cells were cultured for 96-120 hours and assessed
by flow cytometry.
Isolated monocytes were tested for arginase activity using the Urea assay kit
(Abnova Urea Assay ABIN1082256 Taiwan). This measured the metabolite urea, a
by-product of arginine degradation.
For pDC function, PBMC in R-10 were cultured at 2 x 106 cells/ml in 96-well
culture plates and stimulated with type A (CpGA) or type B (CpGB) unmethylated
CPG oligodeoxynucleotides (ODN) (5 mM, CpG 2216 and CpG 2016, respectively,
InVivoGen, San Diego, USA) for 6-8 hours in a CO2 incubator. Cells were stained
(CD123, CD11c, lineage-cocktail, HLA-DR) for surface staining of pDC, and fixed
and permeabilized according to the manufacturer’s instructions (Fix and Perm, BD
Bio-sciences). Anti–IFN-alpha (Miltenyi Biotec) was used for intracellular staining.
129
Enzyme linked immuno-absorbent assays (ELISA)
Plasma levels of arginase I was determined using a human arginase I
sandwich ELISA (BioVendor) according to the manufacturer’s instructions. CD163
was quantified using the Quantikine® Human CD163 ELISA kit (R&D Systems) as
previously outlined (4).
RNA quantification
For DMGE, RNA was extracted from tumor biopsies using RecoverAll total
nucleic acid extraction kit for FFPET (Ambion, Life Technologies, Carlsbad, CA,
USA) as per manufacturer’s instructions and stored at -80oC. CD163 mRNA was
quantified by real-time RT-PCR as previously described using a Rotorgene 3000
real-time PCR machine (Corbett Research) as previously described (4).
Genes for DMGE using the nCounter platform (NanoString Technologies,
Seattle, WA, USA) were chosen to permit COO categorization,(5) and analysis of
immune-effectors and immune-checkpoints. Hybridizations were carried out
according to the NanoString Gene Expression Assay Manual. Five microliter of each
RNA sample (100 ng) was mixed with 20 µl of nCounter Reporter probes in
hybridization buffer and 5 µl of nCounter Capture probes for a total reaction volume
of 30 µl. The hybridizations incubated at 65°C for approximately 16-20 hours. Two
separate runs were performed requiring the production of two identical codesets.
Expression counts were normalized between codesets based on relative differences
between duplicate samples in both runs which allowing us to develop a correction
factor for each individual gene between the runs. Raw data was imported and
130
analysed in the NanoString® data analysis tool nSolver. For normalization, gene
expression data was internally controlled to the mean of the positive control probes
to account for inter-assay variability. Gene normalization was then performed using
the geometric mean of four housekeeper gene to account for factors that affect RNA
quality and quantity (PGK1, GAPDH, PGAM1, OAZ1). Housekeeping genes were
selected as previously described and as per manufacturer recommendation (6).
Gene expression data from frozen tissues utilized by Lenz et al. was taken
from the NCBI Gene Expression Omnibus database (GEO accession GSE10846)
(7). Normalisation was as described in the original published paper. In instances
where genes on the Affymetrix platform had multiple expression probes, we chose
the single probe that most accurately reflect the actual gene expression as per a
recently described scoring method (8).
For blood, RNA was extracted from FACS-sorted CD14+ monocytes, and
analysed using Illumina Human HT12v4 Bead Array for whole genome expression.
The data was extracted and pre-processed using Illumina’s Genome Studio.
Analysis was performed using Genespring GX11 (Agilent Technologies) with all data
quantile normalised. Clustering was performed using Genesis software (Genomics
and Bioinformatics Graz) (9).
1. Meignan M, Gallamini A, Haioun C, Polliack A. Report on the Second International Workshop on interim positron emission tomography in lymphoma held in Menton, France, 8-‐9 April 2010. Leuk Lymphoma. 2010;51:2171-‐80. 2. Jongbloed SL, Kassianos AJ, McDonald KJ, Clark GJ, Ju X, Angel CE, et al. Human CD141+ (BDCA-‐3)+ dendritic cells (DCs) represent a unique myeloid DC subset that cross-‐presents necrotic cell antigens. J Exp Med. 2010;207:1247-‐60. 3. Keane C, Gill D, Vari F, Cross D, Griffiths L, Gandhi M. CD4(+) tumor infiltrating lymphocytes are prognostic and independent of R-‐IPI in patients with DLBCL receiving R-‐CHOP chemo-‐immunotherapy. Am J Hematol. 2013;88:273-‐6.
131
4. Jones K, Vari F, Keane C, Crooks P, Nourse JP, Seymour LA, et al. Serum CD163 and TARC as Disease Response Biomarkers in Classical Hodgkin Lymphoma. Clin Cancer Res. 2013;19:731-‐42. 5. Wright G, Tan B, Rosenwald A, Hurt EH, Wiestner A, Staudt LM. A gene expression-‐based method to diagnose clinically distinct subgroups of diffuse large B cell lymphoma. Proc Natl Acad Sci U S A. 2003;100:9991-‐6. 6. de JHJ, Fehrmann RS, de BES, Hofstra RM, Gerbens F, Kamps WA, et al. Evidence based selection of housekeeping genes. PloS one. 2007;2:e898. 7. Lenz G, Wright G, Dave SS, Xiao W, Powell J, Zhao H, et al. Stromal gene signatures in large-‐B-‐cell lymphomas. N Engl J Med. 2008;359:2313-‐23. 8. Li Q, Birkbak NJ, Gyorffy B, Szallasi Z, Eklund AC. Jetset: selecting the optimal microarray probe set to represent a gene. BMC Bioinformatics. 2011;12:474. 9. Sturn A, Quackenbush J, Trajanoski Z. Genesis: cluster analysis of microarray data. Bioinformatics. 2002;18:207-‐8.
132
1
Supplemental Tables
% 4 year Survival (% difference)
P value % 4 year Survival
(% difference)
P value
Immune-effectors CD8>Median CD8<Median
CD4>Median CD4<Median
CD56>Median CD56<Median
CD137>Median CD137<Median
TNFα>Median TNFα<Median
85 60 (15)
77 69 (8)
78 71 (7)
85 63 (14)
85 63 (12)
0.006
0.22
0.26
0.006
0.004
Monocyte/Macrophage-checkpoints CD163>Median CD163<Median
Effector:Checkpoint ratios CD8:CD68>Median CD8:CD68<Median
CD8:CD163>Median CD8:CD163<Median
CD137:CD163>Median CD137:CD163<Median
TNFα:CD163>Median TNFα:CD163<Median
71 77 (6)
85 62 (13)
89 59 (30)
84 64 (20)
80 60 (20)
0.5
0.02
0.0007
0.02
0.15
Table 1. Optimization of immune-effector and monocyte/macrophage-checkpoint combinations.
Significant P values are in bold.
133
Characteristic NHL21 blood cohort Australian Tissue cohort*
Median age (range) 56 years (26-70) 62 years (27-87)
Female F: 34% F: 41%
Interim-PET/CT+ve 29% N/A
IPI 0-1: 17% 0-1: 28%
2: 28% 2: 30%
3: 35% 3: 24%
4,5: 20% 4,5: 19%
Supplemental Table S2. Characteristics of the blood and tissue patient
cohorts.
*Age, sex and IPI were unavailable in 5, 2 and 6 patients respectively.
134
Specificity (anti-human)
Clone Fluorochrome Manufacturer
CD11b Mac-1 PE BD Pharmingen
CD11c B-ly6 V450 BD Horizon
CD123 9F5 PE BD
CD127 HIL-7R-M21 PE BD Pharmingen
CD137 4B4-1 PE BioLegend
CD14 M5E2 FITC, PerCP-Cy5.5 BD Pharmingen
CD16 3G8 Pacific Blue Invitrogen
CD16 3G8 PE-CY7 BD Pharmingen
CD163 RM3/1 Alexa647, PE BD Pharmingen
CD19 HIB19 Alexa700 BD Pharmingen
CD25 M-2A51 APC BD Pharmingen
CD3 UCHT1 Alexa700, FITC, APC
BD Pharmingen
CD33 WM53 V450 BD Horizon
CD4 RPA-T4 PE BD Pharmingen
CD54 HA-58 PE BD Pharmingen
CD56 AF12-7H3 APC BD
CD56 B159 PE-CY7 BD Pharmingen
CD62L DREG-56 APC eBioscience
CD68 Ki-M7 FITC Molecular Probes
CD8 RPA-T8 PerCP-Cy 5.5 BD Pharmingen
CD80 L307.4 FITC BD Pharmingen
CD83 HB15 FITC BD Pharmingen
CD86 2331 APC BD Pharmingen
CD120b hTNFR-M1 PE BD Pharmingen
HLA-DR L243 APC, PerCP, V500 BD
Lineage cocktail (CD3,CD14, CD16,CD19, CD20,CD56)
SK3,MΦP9,3G8,SJ25C1, L27,NCAM16.2
FITC BD Pharmingen
Supplemental Table S3. Flow cytometry antibodies used.
135
0
1000
200030004000
CD
163
(ng/
l of p
lasm
a)P<0.0001
P<0.0001 P<0.0001
PRE POST Healthy
A)0 20 40 60 80 100
0
20
40
60
80
100
AUC=0.94P<0.0001
B)
0
1000
200030004000
0
1000
200030004000
0
1000
200030004000
0
1000
200030004000
P=0.004 P=0.007
P=0.014 P=0.0003
1.2 >1.2 I-II III-IV
0-1 260 > 60
CD
163
(ng/
l of p
lasm
a)C
D16
3 (n
g/l o
f pla
sma)
CD
163
(ng/
l of p
lasm
a)C
D16
3 (n
g/l o
f pla
sma)
Lymphocyte count Stage
Age IPI
C) D)
E) F)
% S
ensi
tivity
% Specificity
Figure 1 Supplemental.136
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
Low IPIHigh IPIP = 0.002
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
GCBNon-GCBP= 0.018
A B
Figure 2 Supplemental.
137
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
CD8/CD163 HighCD8/CD163 Low
P = 0.07
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
CD8:CD163 HighCD8:CD163 Low
P = 0.047
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
CD8:CD163 HighCD8:CD163 Low
P= 0.03
0 1 2 3 4 50
50
100
Time (years)
Ove
rall
Surv
ival
(%)
CD8:CD163 HighCD8:CD163 Low
P = 0.011
A B
C D
Figure 3 Supplemental.138
CHAPTER 6
DISCUSSION
139
Discussion
DLBCL is one of the most common aggressive B cell lymphomas. Despite
increased understanding and advances in therapeutic modalities such as the
introduction of anti-‐CD20 monoclonal antibody therapy, up to one third of
patients still die from their disease. [1] Prognostic outlook is guided by the IPI,
however considerable heterogeneity of outcome exists within IPI groupings. In
spite of its accuracy, it is limited by the fact that the poorest scoring patients still
have survival rates greater than 50%.[2] In addition the IPI gives no information
on the biological prognostic factors that relate to the malignant B cell or the TME.
New therapeutic options are emerging, however it is still not clear which
patients would most likely benefit from newer regimens. It is important to
ascertain immune correlates that might better prognosticate in patients with
DLBCL treated with R-‐CHOP. It is also important to gain insight into how these
factors might guide therapy.
Primarily, this research focused on the role of the TME in predicting outcome for
patients with DLBCL treated with chemo-‐immunotherapy. Initially the research
focused on the role of host genetics and how this affects immune cells and their
interaction with rituximab. It is unclear how these interactions contribute to the
effectiveness of the drug and the risk of LON. The research progressed focusing on
the immune macroh and microh environment in DLBCL. This initially involved
looking at simple measures of immune response in patients using a full blood
count differential. This progressed to examining flow cytometric identification of
immune cells in tumour biopsies. Finally, I directly analysed circulating immune
cells in detail, and applied these findings to the immune TME using DMGE. The
DMGE enabled direct digital analysis of immune genes in tumour biopsies in the
largest lymphoma dataset to be analysed using this novel technology.
6.1 Immuno-‐genetic polymorphisms
The likely principle mechanism of action of rituximab is antibody-‐dependent
cytotoxicity whereby rituximab binds to FCG3A receptors on NK Cells and
macrophages, resulting in tumour cell lysis by the reticulo-‐endothelial system.[3-‐
6] Another mechanism of action is complement mediated activation, leading to
direct lysis of tumour cells via the complement cascade.[7, 8] Finally, rituximab
140
has a direct apoptotic effect on lymphoma cells.[9-‐11] The relative importance of
these three mechanisms on the lysis of the malignant B cells in DLBCL remains
unclear.
I investigated if polymorphisms in the FCG3A and complement systems could
impact on response of patients’ tumour to rituximab. These polymorphisms are
not the only factors impacting on rituximab response. It has been speculated that
relative levels of CD20 expression are important in responsive lymphomas with
high expression conferring good outcome.[12, 13] In addition prior exposure to rituximab can also promote down regulation of CD20 and subsequent resistance to
therapy.[14, 15] There is also accumulating evidence for improved response
rates in women compared to men in DLBCL treated with
chemoh immunotherapy.[16] Male patients and an increased body weight in
patients with DLBCL appear to correlate with increased clearance of rituximab.
[17, 18] FCG3A is a low affinity receptor capable of binding to complexed
but not monomorphic IgG. A polymorphism at position 158 substituting
valine for phenylalanine results in stronger binding to IgG. In DLBCL this
polymorphism has been shown to predict enhanced responses in a Korean
population.[19] Binding of C1q to the Fc portion of immune complexes
activates CDC through initiation of the complement cascade. C1q is encoded by
C1qA, whose sole coding polymorphism is at position 276, coding for adenine
(C1qA-A276) or guanine (C1qA-G276). C1qA-A276 results in lower C1q protein
levels than the C1qA-G276 polymorphism. The role of this polymorphism in the
complement cascade and its impact on patient outcome is not well understood at
present.
At the time of publication, my study was the largest assessment of the impact of
FCG3A and C1qA on outcome and incidence of LON as a late side effect. In a
uniformly treated DLBCL population, there was no difference in the occurrence of
the polymorphisms between healthy controls and lymphoma patients. Only
small numbers of patients were homozygous for the strong binding
valine FCG3A158 allele (6%). Higher occurrence of valine FCG3A158
allele homozygosity is more frequent in Southh East Asia and may account for a
survival benefit seen in some Asian studies in DLBCL. [19] The low
frequency of VV homozygotes observed in the Queensland cohort
prevents any definitive
141
conclusions regarding the impact of FCG3A158 on outcome for patients with
DLBCL treated with R-‐CHOP.
The frequency of VV homozygosity that I observed is consistent with other
studies. These find rates of 6-‐15% in Caucasian populations. However a recent
German DLBCL study found a rate of 28% for the VV polymorphism. [20] In this
study possession of the polymorphism was not predictive of OS but there was a
trend towards improved PFS (non-‐significant) in patients with VV
polymorphism. The authors summarise that although FCG3A-‐V158
polymorphisms may have a role in predicting outcome in DLBCL treated with
chemo-‐immunotherapy, this cannot be definitively confirmed or excluded
without a much larger cohort. They estimate this to be over 2000 patients. The
authors postulated that the recent increased intensity of rituximab in 14-‐day
regimens may reduce the influence of FCG3A polymorphisms on outcome
parameters. In this scenario, the higher concentrations of rituximab that occur as
a result of the 14-‐day administration may negate the disadvantage of a low
binding FCGR3A allele. The retrospective cohort used in my study had patients
treated with both 14 and 21 day R-‐CHOP regimens, however this data was not
collected so definitive conclusions cannot be made with regards to timing of
rituximab in this cohort.
Interestingly, although there was no statistical association between FCG3A-‐V158
polymorphism and outcome in patients with DLBCL treated with R-‐CHOP, none
of the six patients homozygous for valine have relapsed. Studies have shown that
at high concentrations rituximab is equally effective in VV and FF subtypes but
that the threshold for effective ADCC in VV patients is four times lower than
those for FF patients.[3] The VV polymorphism has been associated with
improved response in other settings with a recent large study in rheumatoid
arthritis showing improved ongoing response to rituximab and that integration
with other RA clinical factors was very predictive of disease response to
rituximab.[21, 22]
Other polymorphisms have been described such as FCG2A that may impact on
rituximab effect, however it appears these are in fact strongly linked to FCG3A
due to lineage disequilibrium and I did not investigate this further.[23] These
relatively consistent findings of differing trends of improved outcome related to
142
NK and macrophage cell binding has led to the development of new monoclonal
antibodies with glycol-‐engineering to ensure optimal engagement of FCG3A
receptors. This has led to the development of GA-‐101 which is a Type-‐II antibody
undergoing clinical trials with much stronger affinity for FCG receptors and
improved responses in early trials.[24, 25]
In spite of the lack of association of survival with regards to the VV
polymorphism, I was able to demonstrate a striking association between this
allele and the occurrence of the established rituximab side-‐effect LON. Despite
the low occurrence of valine homozygosity, this was significantly associated with
the development of LON. We observed a 6% incidence of LON in DLBCL patients
treated with R-‐CHOP chemo-‐immunotherapy however 50% of patients
homozygous for valine developed LON in long term follow up. Previous reports
of LON after induction chemo-‐immunotherapy for B-‐cell lymphoma have not
been restricted to a single histological sub-‐type or treatment regimen. In
addition the neutrophil cut-‐off used to identify LON has varied between grades 2
to grade 4.[26-‐28] Once studies are restricted to those using a definition of grade
3/4 neutropenia, then irrespective of ethnicity the incidence of LON (range 7-‐
13%) is of a similar order of magnitude to our series.[29-‐32] In these, as with
our study, incidence may be under-‐estimated as patients are frequently
asymptomatic. [33] The incidence of LON appears to rise markedly in studies
that include patients undergoing high-‐dose therapy with autologous stem cell
rescue.[33-‐35] Put together, it may be that intensity of treatment regimen (use of
high-‐dose over conventional dose chemo-‐immunotherapy) is the principal
determinant of LON incidence rather than lymphoma histology. The incidence of
LON in non-‐malignant states is well described and appears to be lower than
levels seen in haematological malignancy.[36-‐38]
The mechanism behind LON is yet to be established.[33, 38] Notably,
bone marrow histology at the time of LON has been variously reported as
maturation arrest or myeloid hypoplasia, indicating that several mechanisms
may be responsible. Mechanisms postulated include anti-‐neutrophil antibodies
or increased large granular lymphocytes in the absence of B-‐cells leading to FAS
ligand mediated destruction of neutrophils. [39, 40] However this has not been
consistently observed. A report of six cases of aggressive lymphoma (including
143
two with AIDS related lymphoma) treated with DA-‐EPOCH-‐R (dose-‐adjusted
etoposide/prednisone/Oncovin[vincristine]/cyclophosphamide,
hydroxyduanorubicin / rituximab), implicated perturbations of stromal derived
factor-‐1 (SDF-‐1) / CXCL12 during B-‐cell recovery as a potential aetiology.[26]
The SDF1 chemokine is important for granulocyte egress from the bone marrow
and also in B-‐cell development. A shift of SDF-‐1 towards B-‐cell recovery over
granulocyte homeostasis may result in LON due to neutrophil maturation arrest
whilst B-‐cell recovery occurs. Cytokine kinetics were evaluated in detail in a
single case of LON associated with granulocytic hypoplasia following
fludarabine, cyclophosphamide and rituximab for Waldenstroms
macroglobulinaemia. In this patient LON was associated with markedly raised
levels of serum B-‐cell activating factor (BAFF).[41] However, many of the above
findings are in small cohorts or single patients and are hypothesis generating
only at the current time.
It may be that the pharmacokinetics of rituximab are different in patients
homozygous for the FCGR3A-‐V158 polymorphism. FCGR3A-‐V158 has a higher
binding affinity for IgG1 antibodies (such as rituximab) than FCGR3A-‐F158.[42]
Thus FCGR3A-‐V158 may result in enhanced clearing of CD20 expressing cells by
ADCC by NK cells. FCGR3A transcripts are higher in NK cells from subjects with
FCGR3A-‐V/V158 versus the V/F or F/F genotype and V/V homozygotes have
enhanced in-‐vitro ADCC activity. In-‐vivo this may result in more profound B-‐cell
depletion.[43] Upon B-‐cell recovery, heightened stimulation of lymphopoiesis
may result in temporary imbalance of cytokines and transient ineffective
granulopoiesis.
In contrast with our findings regarding FCGR3A-‐V158F, there was no significant
association between LON and the C1qA-‐A276G polymorphisms. Furthermore, the
combination of C1qA-‐A276G and FCGR3A-‐V158F did not appear to expose any
linkage disequilibrium and did not appear to influence outcome or development
of LON. These findings suggest that C1qA is not implicated in the pathogenesis of
LON. We did not observe any difference in EFS or OS related to the C1q-‐A276G
polymorphism. The literature in this area is sparse and conflicting with a recent
study showing no effect of this polymorphism in a small follicular lymphoma
population.[44] Another study in low-‐grade lymphoma showed a prolonged
144
progression free survival in patients receiving single agent rituximab but not in
overall survival.[45] A single study in breast cancer using a monoclonal antibody
to Herceptin demonstrated reduced metastases associated with this
polymorphism, but a more recently published large study of over 2000 breast
cancer patients suggests no impact on outcome for this polymorphism in breast
cancer patients.[46, 47] There is one recent study described of 129 patients with
DLBCL which does show prolonged PFS and OS for patients homozygous for the
A allele in a study of a population from Beijing.[48] This study reported almost
30% of patients homozygous for G allele compared to 6% in our population
which could make the significance of the A allele more significant in the Chinese
population studied. The G allele is associated with worse outcome despite
increased and more active complement lysis of cells.[49] It has been postulated
that increased complement activity may reduce ADCC and adaptive immunity
and that enhanced complement activity rapidly clears tumour cells allowing
inadequate time for an effective immune response to develop.[50, 51]
In conclusion, in our uniformly treated group of DLBCL patients who received
rituximab, 6% of patients developed LON. The FCGR3A-‐158 V/V genotype was
significantly associated with development of LON. Polymorphic analysis may be
a predictive tool to identify those at high-‐risk of LON. Although no patients with
either LON or FCGR3A-‐158V homozygosity relapsed, neither were associated
with improved EFS or OS after R-‐CHOP. Large prospective studies are required
to establish if FCGR3A-‐V158F polymorphisms have a bearing on response rates in
DLBCL, and whether either LON or FCGR3A-‐V158F polymorphisms are
predictors of outcome.
6.2 Immune Microenvironment
There is substantial evidence demonstrating the key role of host immunity in
DLBCL. Numerous gene expression studies have identified immune signatures
that are prognostic in the disease.[52-‐54] There have also been a number of
immunohistochemical studies confirming the importance of the tumour
microenvironment in DLBCL.[55-‐58] However, both gene expression studies and
IHC have significant drawbacks. Gene expression is expensive, requiring fresh
frozen tissue and is available only in a limited number of centres. While IHC is
145
relatively cheap there is marked heterogeneity in sample preparation, stains
used and interpretations.
I set out to identify immune markers that would be relatively simple to use,
inexpensive and available to most diagnostic centres. For this reason I looked at
circulating immune cells obtained from full blood counts and also flow immune-‐
phenotyping taken from fresh diagnostic tissue. All lymphoma diagnostic
samples at the Princess Alexandra Hospital have a standard B and T cell panel
performed to assess cell phenotype at diagnosis. In B cell lymphomas, the
information with regards to T cell infiltration is generally ignored as the test is
performed for diagnostic information only. I was interested to identify if T cell
infiltration might predict survival in patients with DLBCL. In addition I had the
ability to look at circulating immune parameters prior to therapy and in
particular, absolute lymphocyte and monocyte counts at diagnosis to see if
circulating immune status was also important. My study of 122 patients with
DLBCL confirms the significance of host immune status in predicting outcome. I
have shown for the first time that CD4+ T cell infiltration of fresh tumour (as
assessed by flow cytometric immune-‐phenotyping) is a very strong predictor of
OS and EFS when patients with DLBCL are treated with standard chemo-‐
immunotherapy. This was independent of the IPI. The two groups of patients
separated by a 20% CD4 cut-‐off did not differ for any of the single parameters
that make up the IPI nor were the groups significantly different in their IPI
scores. Interestingly, CD8 TILs were not associated with any outcome endpoints.
Importantly the level of CD4+TILs was a significant predictor of outcome in
patients with good prognosis disease as assessed by IPI (0, 1, 2, 3). In addition I
have confirmed the prognostic significance of the absolute lymphocyte to
monocyte ratio with this ratio able to predict outcome in the whole cohort but
also predicting poor outcome for patients in good risk IPI categories. Our data
does indicate however, that the CD4 TILs are the strongest predictors of outcome
in this cohort and were independent of IPI and LMR.
My findings are consistent with two previous studies described prior to the
introduction of rituximab.[59, 60] The two smaller studies described, in 55 and
72 patients respectively, demonstrated a benefit for improved outcome in
patient samples with higher number of CD4+T lymphocytes prior to the
146
introduction of rituximab therapy. In the larger study a cut-‐off of 20% CD4+T
cells was also the most predictive of outcome. Interestingly, in agreement with
our results, neither of these studies identified CD8+ TILs as being prognostic.
Recent studies have confirmed the importance of T cell activation in the tumour
microenvironment with the T cell activation marker CD137 predicting outcome
in a large DLBCL cohort.[56] This is of particular interest given an agonist
antibody to CD137 may lead to improved immune responses against
lymphoma.[56, 61] As discussed in the introductory chapter, CD4 T cells are
relatively heterogeneous with variable and diverse functions such as enhancing
inflammation to overt immunosuppressive effects. In general TH1 cells have anti-‐
tumour effects and TH2 may well have an opposite effect by enhancing immune
suppression. [62, 63] T regulatory cells in malignancies are generally associated
with inferior outcome, however for unknown reasons at present this may not be
the case in B cell lymphomas where high levels of these cells in the TME can be
associated with improved outcome. It is felt that increased numbers of Tregs could
reflect a more normal host immune system, while it has also been postulated
that Tregs may have a direct negative effect on proliferation of B cells. [64-‐67] My
study gives no information on the CD4 T cell subsets in the tumour which will be
key in future research to more adequately classify these cells. There is still
limited data on the role of these various CD4 subsets in DLBCL. Interestingly
there is evidence from prior to the introduction of rituximab that circulating
lymphocytes of patients with DLBCL before treatment are skewed to a TH2
phenotype and revert to a TH1 phenotype with successful treatment.[68] There
is also emerging evidence that T helper cells are key to inducing an anti-‐tumour
effect post vaccination with DLBCL related peptides.[69]Animal models in B cell
lymphoma have shown that CD4 T cells are key cells in creating an anti-‐tumour
microenvironment.[70] TH1 cells in particular stimulate and up regulate antigen
presentation and tumour clearance.[63] Improved antigen presentation has been
shown to be a key survival determinant in patients with DLBCL.[57, 71-‐73]These
animal models have shown that cytokines derived from TH1 cells such as
Interferon gamma, IL1 and TNF alpha are the key to stimulating tumour-‐killing
macrophages. It is possible that patients with low CD4+TILs could benefit from
interferon therapy. Interferon has been shown to be effective in the treatment of
147
many lymphomas but its use has been restricted due to short half-‐life and side
effects, however methods of new local tumour delivery systems look promising
in lymphomas.[74]
Mouse models have also shown that PD-‐1 which is present on many CD4 T cells
may eventually down regulate CD4+T cell tumour surveillance leading to
relapse.[70] PD-‐1 has been found expressed in 27% of DLBCL tumours, but when
one looks at the local tumour microenvironment 38% of PD1 non-‐expressing
DLBCL have PD1 positive histiocytes surrounding the tumours. It is possible that
PD-‐1 could be responsible for down-‐regulating CD4 T cells in a large number of
DLBCLs. This is of particular interest as two highly successful trials targeting PD-‐
1 in advanced cancers have recently been described in solid tumours.[75, 76] To
date two trials using an antibody to PD-‐1 have been described in B cell
lymphomas.[77, 78] CT-‐011 (anti-‐PD1) directly increases the numbers of CD4+T
cells post autograft for relapsed disease with survival higher than historical
controls.[78] As of yet it is unclear from these studies what exact role T cells and
other immune cells such as macrophages play in eliciting a response with
immune checkpoint modulation, and it will be important to determine the
influence of PD-‐1 expression on not only the numbers of TILs but the activity of
these key cells in future trials.
With regard to simple immune parameters garnered from the full blood count, I
have confirmed the prognostic significance of pre-‐diagnostic lymphocyte and
monocyte counts incorporated into the LMR ratio and also the recently
described algorithm from Wilcox et al.[79, 80] I have also shown that a low
lymphocyte count on its own is a predictor of poor outcome. This confirms the
findings of a number of recent studies that show that a low lymphocyte count at
diagnosis and also post therapy is associated with poor outcome.[81-‐84] There is
also emerging data that the AMC may predict outcome in DLBCL.[85] In addition
there is evidence that specific subsets of monocytes called monocyte derived
myeloid suppressor cells are increased in patients with DLBCL and contribute to
marked immune suppression and poor outcome.[67, 85] The importance of
these cells are discussed further below.
My data identified the LMR as being a significant prognosticator independent of
IPI. In addition LMR could identify poor risk patients in groups of patients with
148
good risk IPI scores. It is likely that these combined scores reflect the poor
outcome associated with systemic immunosuppression. It is likely the combined
LMR measures the potential level of immune suppression dictated by the
monocyte count and the subsequent effects of this in a low lymphocyte count.
Further studies are required to identify the subsets within these two cell types
that are contributing to outcome. Interestingly there did not seem to be any
correlation between circulating and tumour infiltrating lymphocytes (as
assessed from my flow cytometry data). This may reflect the difference in
retention and recruitment between nodal tissue and the circulation.[80]
In summary, I have shown that CD4+TILs appear to be very strong predictors of
outcome in DLBCL treated with chemo-‐immunotherapy independent of IPI.
Further studies are required to analyse the influence of other local
microenvironment factors such as macrophages and histiocytes on CD4 function
and numbers. In addition we need to identify the particular CD4 subsets that
may contribute to improved survival and thus investigate the specifics of why
the lymphocyte to monocyte ratio is prognostic in DLBCL.
6.3 Net Tumoral Immunity
My findings described above confirm that circulating lymphocyte:monocyte
ratios are prognostic, implicating them as surrogate immune-‐effectors and
monocyte/macrophage-‐checkpoints within the tumor microenvironment. I
hypothesised that detailed functional and quantitative assessment would enable
identification of the optimal immune-‐effector and monocyte/macrophage-‐
checkpoint molecules to interrogate within the tissue. Blood from 140 ‘R-‐CHOP’
chemo-‐immunotherapy treated DLBCL patients from a prospective Australasian
Leukaemia and Lymphoma Group trial was analysed. PBMC from the patients
enrolled on the NHL21 clinical study were fully characterised so that the
lymphocyte and monocyte subsets that may contribute to outcome were
identified.
Amongst blood monocyte immune-‐checkpoint markers, addition of CD163 to
conventional moMDSC indicators (CD14+HLA-‐DRlo) identified a highly
immunosuppressive subset of moMDSCs in DLBCL. MoMDSCs and
CD163+monocytes values correlated and CD163+moMDSC were enriched within
149
circulating monocytes. CD163 is a heme scavenger molecule felt to be specific for
the immunosuppressive M2 Macrophage in tissue.[86, 87] CD163 is a surrogate
of the M2 macrophage and is felt to predict inferior outcome when expressed in
tumours in DLBCL and Hodgkin Lymphoma, but also in most solid cancers.[66,
88, 89] Effective treatment of primary tumour leads to reduced levels of these
cells.[90] High levels of M2 macrophages around the tumour prevents an
effective anti-‐tumour response. The correlation between the M2 TAM marker
CD163 and HLA-‐DRlo on circulating CD14+ monocytes suggests a link between
circulating moMDSC and M2 TAMs within the malignant lymph node. This is
supported by CD163himoMDSC expressing CD62L and CD11c, enabling migration
into secondary lymphoid tissues.
MoMDSCs seem capable of causing immune suppression in the circulation but
also capable of migration to lymph nodes and exerting local effects in the tumour
bed.[91] They are cells in a state of maturation block with an inability to
progress to a normal mature myeloid/monocyte cell. Increased levels of these
cells in the circulation have been described in a number of cancers.[92, 93] Some
of these studies showed a direct link between numbers of moMDSCs/MDSCs and
tumour bulk and risk of metastases. MoMDSCs/MDSC related
immunosuppression seems to be mediated through multiple mechanisms
including arginase, iNOS, peroxynitrite, ROS and H2O2 metabolism.[94-‐98]
These cells also seem to increase the production of other immune suppressive
cells such as Tregs via secretion of chemokines such as CCL3. [99]MDSCs can also
effect non immune parameters contributing to increased angiogenesis and
metastases of tumours.[100] My findings make it likely that moMDSC migration
to tissue results in accumulation of M2 macrophages. An M2 macrophage is
characterised by positivity for CD68 and CD163 whereas M1 pro-‐inflammatory
macrophage has expression of CD68 and high levels of HLA-‐DR.
Within the circulation of patients enrolled on the ALLG NHL21 study, the
immune-‐effector: CD163+moMDSC ratios were more informative than the LMR,
with ratios lower in those remaining interim-‐PET/CT+ve. Interim-‐PET/CT
imaging was delayed to day 17-‐20 post-‐cycle 4 and scans reviewed centrally. In
particular CD4:CD163+moMDSC and CD8: CD163+moMDSC ratios were
predicative of early lymphoma clearance as assessed by PET/CT negativity.
150
However given repeat biopsy was not performed, no definitive conclusion can be
made as to whether pre-‐therapy CD163+moMDSC associate with residual DLBCL
versus inflammation within sites of interim-‐PET/CT-‐FDG-‐avidity.
My core study findings from the NHL21 study indicate that in DLBCL, the balance
of immune-‐effectors and immune-‐checkpoints are critical to interim-‐PET/CT
outcome. Given the uncertainty regarding the prognostic significance of interim-‐
PET/CT on overall survival, I went on to validate the markers that I had
identified in the peripheral blood in an independent DLBCL cohort. Here, the
molecules were applied to the diagnostic tissue, and overall survival was the
primary end-‐point.
128 R-‐CHOP treated DLBCL patients with full clinical annotation and long-‐term
(median 4 year) survival data derived from two Australian centres (Canberra
Hospital and the Princess Alexandra Hospital) were tested. We used the DMGE
platform NanoString nCounter™ to permit accurate mRNA quantification on
FFPE tissues.[101-‐103] In keeping with previous reports, we observed strong
correlation between gene expression in paired frozen/paraffin samples across a
range of genes.
DMGE identified CD8 as the strongest single immune predictor of outcome in
DLBCL. [104] CD8 T cells are the end effector cells for the immune system and
are directly cytotoxic to tumour cells. Autologous and third-‐party T cells
generated against EBV have shown excellent responses in patients with EBV+ B
cell lymphomas in particular.[105-‐107] One of the commonest mutations found
in DLBCL relates to loss of the key MHC I related protein beta-‐2-‐microglobulin
with a recent study showing this molecule mutated in 29% of DLBCL cases.[108]
This study also identified deletions in CD58 which can also affect CD8
recognition of antigen. In addition to gene mutations, antigen presentation is
impaired by other mechanisms so that overall 60% of patients with DLBCL have
reduced ability to present antigen. Loss of function of the tumor suppressor gene
PRDM1 or over-‐expression of the oncogene BCL6 occurs in a large proportion of
DLBCL cases. However, BCL6 is also frequently mutated in activated B cells of
healthy individuals. A recent study using PRDM1-‐deficient and/or BCL6 over-‐
expressing transgenic mice, observed that lymphoma development was delayed
and relatively rare. However in the context of polyclonal T cell impairment, the
151
transgenic mice exhibited a high incidence of rapid-‐onset aggressive B cell
lymphomas.[109] The implication is that development of these lymphoma
associated mutations is relatively common but that a healthy immune system
eradicates cells containing these mutations to prevent overt lymphoma. Prior
IHC based studies have also shown that absence of immune recognition proteins
such as HLA Class I and II molecules and CD80/CD86 is associated with inferior
outcome and reduced levels of CD4 and CD8 tumour infiltration.[110]It is
interesting that CD8 was not prognostic in my flow cytometry model but it
should be noted that flow cytometry reflects the proportion of CD4 and CD8 T
cells, not their absolute number. Furthermore, mRNA levels reflect the turnover
of the gene signal in a sample. Therefore quantification of CD8 gene expression
within the malignant node and CD8 T cell flow cytometry can not be considered
strictly comparable.
By DMGE, all markers of T cell infiltration predicted improved outcome.
Interestingly, neither intratumoral CD163 nor CD68 expression alone was
prognostic. However the CD68:CD163 ratio as an estimate of the relative
proportion of macrophage sub-‐types found that those with a higher proportion
of M2 macrophages (low CD68:CD163) had inferior outcome. This may be one
explanation why results of intratumoral CD163 alone by IHC as a prognosticator
are inconsistent, emphasizing the importance of measuring several markers to
more accurately reflect aspects of TME immunity. [89, 111-‐113] Similarly,
CD8:CD163 ratios as a measure of net anti-‐tumoral immunity was more
discriminatory in predicting outcome than CD8 alone. Whereas the contribution
of immune-‐effector molecules within the TME has previously been recognized to
be prognostic, this is the first report of immune-‐effector: immune-‐checkpoint
ratios. CD8:CD163 was independent of IPI and COO. Results were validated in an
external R-‐CHOP-‐like (233 patients) gene-‐expression cohort.
IPI is influenced by factors such as patient fitness, age and tumour burden, which
might reasonably be expected to be non-‐over-‐lapping with CD8:CD163. COO
reflects the postulated B-‐cell differentiation stage at which malignant
transformation occurred. One interpretation is that CD8:CD163 ratios within the
TME are not influenced by differentiation stage of the malignant B-‐cell.
Interestingly, although CD8:CD163 was predictive in the CHOP-‐like cohort, this
152
was not independent of conventional prognosticators. This may reflect the
relatively increased importance of tumour-‐associated macrophages within the
TME in those treated with immuno-‐chemotherapy versus chemotherapy alone,
and is consistent with our findings that CD163+moMDSC suppress R-‐ADCC.
Although COO did prognosticate, it was an inferior predictor compared to the
CD8:CD163 immune signature.[114] Of the 18 genes used to determine COO,
only LMO2 as a stand-‐alone gene was predictive for survival.. This confirms
findings from previous gene and immunohistochemical studies.[99, 114, 115] In
R-‐CHOP treated patients, CD8:CD163 ratios enhanced the prognostic ability of
this GCB marker. Furthermore LMO2 combined with CD8:CD163 successfully
allowed segregation within low and high IPI patient groupings. The combination
of a marker of B-‐cell differentiation with net anti-‐tumoral immunity appears to
capture more patients (than predicted by IPI or COO alone) at low and high-‐risk.
This may assist risk-‐stratification to select patients in whom novel therapies
should be tested, and those in whom R-‐CHOP is sufficient. Patients who have a
poor immune ratio and have low LMO2 expression (dual negative) are
approximately one quarter of patients. This group of patients do very poorly
with R-‐CHOP and my work allows identification of these patients failing current
therapy who need alternate strategies to improve their outcome. It also shows
that the other 75% of patients have cure rates approaching 90% with current
standard therapy. This differentiation will be very important given that the cost
of new emerging therapies will be significant. These compounds will need to be
targeted to those patients who will derive most benefit.
These findings have other therapeutic implications. Firstly, monocyte depletion
in patients enhanced NK-‐cell mediated R-‐ADCC but not Ob-‐ADCC, indicating that
the immunosuppressive effects of monocytes in DLBCL may be overcome by
obinutuzumab (a type II antiCD20 monoclonal antibody). Another notable
finding was the striking co-‐clustering of immune-‐effectors and immune-‐
checkpoints within the tissue and circulation, including CD8-‐CD163. This is in
line with emerging data that up-‐regulation of immune-‐checkpoints is an adaptive
immune-‐checkpoint response to immune-‐effector activation.[116] The
correlations were significant but modest, reflecting the variable success of the
host to counter anti-‐tumoral immunity within the TME. However, an alternative
153
explanation is possible, with the high levels of checkpoints associated with
increasing CD8 and other effectors reflecting a healthy immune checkpoint
response to prevent local tissue injury in the setting of an effective immune
response. Blockade of immune-‐checkpoints is a promising therapeutic approach,
with a variety of antagonistic monoclonal antibodies or small inhibitory
molecules in development.[117, 118]
Immune based strategies are gaining ground in DLBCL.[70, 119-‐121] The
kinetics of systemic host immune cells will likely impact the efficacy of these
approaches, and may influence dose-‐scheduling. However there is minimal data
on circulating moMDSC kinetics in DLBCL, with a small study finding moMDSC
returned to normal once therapy was complete.[122] We found post-‐cycle 4 that
CD163+moMDSC were reduced (accompanied by an increase in numbers of the
antigen-‐presenting cells BMDCs and pDCs). The immunosuppressive profile of
monocytes had reduced by post-‐cycle 4, including down-‐regulation of TH2
cytokines and TAM associated genes, and up-‐regulation of STAT1.[91, 109, 123]
Similarly elevated plasma CD163 was reduced by post-‐cycle 4. In line with cHL,
higher plasma CD163 levels were associated with advanced stage and reduced
lymphocytes.[124]
My data emphasizes the importance of capturing the net anti-‐tumoral immunity
within the DLBCL TME, by measuring the relative balance of immune-‐effector to
tumour-‐associated macrophages. Addition of LMO2 as a marker of germinal
centre B-‐cells to CD8:CD163 was powerfully prognostic. The work provides a
link between M2 TAMs and moMDSC, and demonstrate that CD163 identifies a
highly immunosuppressive subset of moMDSC in DLBCL. Further investigation of
CD163+moMDSC as a therapeutic target is warranted, particularly in those with
adverse CD8:CD163 ratios. Emerging data from mouse models indicate the
possibility of switching macrophage phenotype from M2 to M1 and consequent
improved immune clearance of the lymphomas using immune modulating drugs
such as Pomalidomide.[125] In addition, direct targeting of CD163 using a
monoclonal antibody has shown promise as an anti-‐inflammatory agent and
could have the potential for activity in the setting of cancer as a means of
targeting the tumour microenvironment.[126]
154
To my knowledge, this data is the first to establish the importance of the relative
balance of immune activation as being key to outcome in DLBCL. Neither
intratumoral immune-‐effectors nor immune-‐monocyte/macrophage checkpoints
should be considered in isolation. A high level of immune effectors does not
necessarily predict good outcome, if countered by an equally robust effective
immune-‐monocyte/macrophage-‐checkpoint response. This data provides a
strong rationale for the use of checkpoint inhibitors in the treatment of DLBCL.
6.4 Future directions
A pressing issue from my findings is the identification of patients who could be
targeted by prospective immune checkpoint therapy, targeting molecules such as
PD-‐1 and CTLA-‐4 or by using other immune-‐modulating agents such as
lenalidomide given our findings are based on common, standardised diagnostic
tests available to many laboratories. Biomarkers that might predict response to
new therapies such as immune checkpoint blockade in lymphoma are urgently
required. This will be a complex process. In solid tumours such as melanoma
while expression of PDL-‐1 does predict response to anti-‐PD1 therapy, up to 10%
of patients with no PDL-‐1/PDL-‐2 expression in their tumours still responded to
anti-‐PD1 therapy. This indicates the complexity of predicting responses in this
new era of immune checkpoint blockade. This needs to be investigated further in
prospective studies.
The intratumoural LMO2/CD8:CD163 score was independent of additive to the
widely used clinical prognosticator IPI. One future direction would be to develop
and prospectively validate an integrated clinical and biological score, that might
enhance stratification of DLBCL patients. Importantly, this would allow
identification of patients that might benefit from novel agents (preferably given
within the context of a clinical trial), and equally importantly predict patients
highly likely to be cured with conventional chemo-‐immunotherapy alone.
Should sufficient NHL21 tissue samples become available (acquisition is ongoing,
with approximately 67% of tissues available to date), this cohort will be used to
prospectively validate LM02/CD8:CD163. Three year EFS analysis (the clinical
studies primary end-‐point) is be performed in 2015. It will also be possible to
155
examine if the strong association of circulating immune-‐effector cells:moMDSC
ratios with tumour clearance as assessed by interim PET/CT results in a
significant EFS benefit. The confounding factor for both these analyses is that
following interim-‐PET/CT, patients had risk-‐adapted therapy and hence were no
longer uniformly treated.
My intention is also to extend investigation of the immune TME into other
lymphomas, including Hodgkin Lymphoma and Follicular Lymphoma, and
contrast this with post transplant lymphoproliferative disorder (PTLD). The
latter occurs as a consequence of iatrogenic immunosuppression, and
CD14+CD163+ monocytes are implicated in reducing the risk of graft
rejection.[90].
References:
1. Coiffier, B., et al., CHOP chemotherapy plus rituximab compared with CHOP alone in elderly patients with diffuse large-‐B-‐cell lymphoma. The New England journal of medicine, 2002. 346(4): p. 235-‐42.
2. Sehn, L.H., et al., The revised International Prognostic Index (R-‐IPI) is a better predictor of outcome than the standard IPI for patients with diffuse large B-‐cell lymphoma treated with R-‐CHOP. Blood, 2007. 109(5): p. 1857-‐61.
3. Dall'Ozzo, S., et al., Rituximab-‐dependent cytotoxicity by natural killer cells: influence of FCGR3A polymorphism on the concentration-‐effect relationship. Cancer research, 2004. 64(13): p. 4664-‐9.
4. Uchida, J., et al., The innate mononuclear phagocyte network depletes B lymphocytes through Fc receptor-‐dependent mechanisms during anti-‐CD20 antibody immunotherapy. The Journal of experimental medicine, 2004. 199(12): p. 1659-‐69.
5. Minard-‐Colin, V., et al., Lymphoma depletion during CD20 immunotherapy in mice is mediated by macrophage FcgammaRI, FcgammaRIII, and FcgammaRIV. Blood, 2008. 112(4): p. 1205-‐13.
6. Boross, P. and J.H. Leusen, Mechanisms of action of CD20 antibodies. American journal of cancer research, 2012. 2(6): p. 676-‐90.
7. Di Gaetano, N., et al., Complement activation determines the therapeutic activity of rituximab in vivo. Journal of immunology, 2003. 171(3): p. 1581-‐7.
8. Golay, J., et al., Biologic response of B lymphoma cells to anti-‐CD20 monoclonal antibody rituximab in vitro: CD55 and CD59 regulate complement-‐mediated cell lysis. Blood, 2000. 95(12): p. 3900-‐8.
9. Byrd, J.C., et al., The mechanism of tumor cell clearance by rituximab in vivo in patients with B-‐cell chronic lymphocytic leukemia: evidence of caspase activation and apoptosis induction. Blood, 2002. 99(3): p. 1038-‐43.
156
10. Cardarelli, P.M., et al., Binding to CD20 by anti-‐B1 antibody or F(ab')(2) is sufficient for induction of apoptosis in B-‐cell lines. Cancer immunology, immunotherapy : CII, 2002. 51(1): p. 15-‐24.
11. Shan, D., J.A. Ledbetter, and O.W. Press, Apoptosis of malignant human B cells by ligation of CD20 with monoclonal antibodies. Blood, 1998. 91(5): p. 1644-‐52.
12. Horvat, M., et al., Predictive significance of the cut-‐off value of CD20 expression in patients with B-‐cell lymphoma. Oncology reports, 2010. 24(4): p. 1101-‐7.
13. van Meerten, T., et al., Complement-‐induced cell death by rituximab depends on CD20 expression level and acts complementary to antibody-‐dependent cellular cytotoxicity. Clinical cancer research : an official journal of the American Association for Cancer Research, 2006. 12(13): p. 4027-‐35.
14. Cartron, G., et al., Interindividual variability of response to rituximab: from biological origins to individualized therapies. Clinical cancer research : an official journal of the American Association for Cancer Research, 2011. 17(1): p. 19-‐30.
15. Olejniczak, S.H., et al., Acquired resistance to rituximab is associated with chemotherapy resistance resulting from decreased Bax and Bak expression. Clinical cancer research : an official journal of the American Association for Cancer Research, 2008. 14(5): p. 1550-‐60.
16. Pfreundschuh, M., et al., Suboptimal dosing of rituximab in male and female patients with DLBCL. Blood, 2014. 123(5): p. 640-‐6.
17. Muller, C., et al., The role of sex and weight on rituximab clearance and serum elimination half-‐life in elderly patients with DLBCL. Blood, 2012. 119(14): p. 3276-‐84.
18. Arpon, D.R., M.K. Gandhi, and J.H. Martin, A new frontier in haematology -‐ combining pharmacokinetics with pharmacodynamics factors to improve choice and dose of drug. British journal of clinical pharmacology, 2014.
19. Kim, D.H., et al., FCGR3A gene polymorphisms may correlate with response to frontline R-‐CHOP therapy for diffuse large B-‐cell lymphoma. Blood, 2006. 108(8): p. 2720-‐5.
20. Ahlgrimm, M., et al., The impact of Fc-‐gamma receptor polymorphisms in elderly patients with diffuse large B-‐cell lymphoma treated with CHOP with or without rituximab. Blood, 2011. 118(17): p. 4657-‐62.
21. Quartuccio, L., et al., The 158VV Fcgamma receptor 3A genotype is associated with response to rituximab in rheumatoid arthritis: results of an Italian multicentre study. Annals of the rheumatic diseases, 2013.
22. Ruyssen-‐Witrand, A., et al., Fcgamma receptor type IIIA polymorphism influences treatment outcomes in patients with rheumatoid arthritis treated with rituximab. Annals of the rheumatic diseases, 2012. 71(6): p. 875-‐7.
23. Lejeune, J., et al., Evidence for linkage disequilibrium between Fcgamma RIIIa-‐V158F and Fcgamma RIIa-‐H131R polymorphisms in white patients, and for an Fcgamma RIIIa-‐restricted influence on the response to therapeutic antibodies. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2008. 26(33): p. 5489-‐91; author reply 5491-‐2.
157
24. Alduaij, W., et al., Novel type II anti-‐CD20 monoclonal antibody (GA101) evokes homotypic adhesion and actin-‐dependent, lysosome-‐mediated cell death in B-‐cell malignancies. Blood, 2011. 117(17): p. 4519-‐29.
25. Salles, G., et al., Phase 1 study results of the type II glycoengineered humanized anti-‐CD20 monoclonal antibody obinutuzumab (GA101) in B-‐cell lymphoma patients. Blood, 2012. 119(22): p. 5126-‐32.
26. Dunleavy, K., et al., B-‐cell recovery following rituximab-‐based therapy is associated with perturbations in stromal derived factor-‐1 and granulocyte homeostasis. Blood, 2005. 106(3): p. 795-‐802.
27. Kim, M., et al., [Late-‐onset neutropenia following rituximab therapy as a treatment of diffuse large B-‐cell lymphoma: a single institution study]. Korean J Lab Med, 2010. 30(6): p. 575-‐9.
28. Chaiwatanatorn, K., et al., Delayed-‐onset neutropenia associated with rituximab therapy. Br J Haematol, 2003. 121(6): p. 913-‐8.
29. Tesfa, D., et al., Late-‐onset neutropenia associated with rituximab therapy: evidence for a maturation arrest at the (pro)myelocyte stage of granulopoiesis. Med Oncol, 2008. 25(4): p. 374-‐9.
30. Li, S.C., et al., Rituximab-‐induced late-‐onset neutropenia in newly diagnosed B-‐cell lymphoma correlates with Fc receptor FcgammaRIIIa 158(V/F) polymorphism. Am J Hematol, 2010. 85(10): p. 810-‐2.
31. Lai, G.G., et al., Late-‐onset neutropenia following RCHOP chemotherapy in diffuse large B-‐cell lymphoma. Am J Hematol, 2009. 84(7): p. 414-‐7.
32. Rozman, S., M. Sonc, and B.J. Novakovic, Late-‐onset neutropenia following primary treatment of diffuse large B-‐cell lymphoma with rituximab-‐containing therapy. Leukemia & lymphoma, 2012. 53(10): p. 1945-‐8.
33. Cairoli, R., et al., High incidence of neutropenia in patients treated with rituximab after autologous stem cell transplantation. Haematologica, 2004. 89(3): p. 361-‐3.
34. Weng, W.K., et al., Immunoglobulin G Fc receptor FcgammaRIIIa 158 V/F polymorphism correlates with rituximab-‐induced neutropenia after autologous transplantation in patients with non-‐Hodgkin's lymphoma. J Clin Oncol, 2010. 28(2): p. 279-‐84.
35. Nitta, E., et al., A high incidence of late-‐onset neutropenia following rituximab-‐containing chemotherapy as a primary treatment of CD20-‐positive B-‐cell lymphoma: a single-‐institution study. Ann Oncol, 2007. 18(2): p. 364-‐9.
36. Besada, E., W. Koldingsnes, and J. Nossent, Characteristics of late onset neutropenia in rheumatologic patients treated with rituximab: a case review analysis from a single center. QJM : monthly journal of the Association of Physicians, 2012. 105(6): p. 545-‐50.
37. Tesfa, D., et al., Late-‐onset neutropenia following rituximab therapy in rheumatic diseases: association with B lymphocyte depletion and infections. Arthritis and rheumatism, 2011. 63(8): p. 2209-‐14.
38. Tesfa, D. and J. Palmblad, Late-‐onset neutropenia following rituximab therapy: incidence, clinical features and possible mechanisms. Expert review of hematology, 2011. 4(6): p. 619-‐25.
39. Papadaki, T., et al., Evidence for T-‐large granular lymphocyte-‐mediated neutropenia in Rituximab-‐treated lymphoma patients: report of two cases. Leuk Res, 2002. 26(6): p. 597-‐600.
158
40. Voog, E., et al., Factors that predict chemotherapy-‐induced myelosuppression in lymphoma patients: role of the tumor necrosis factor ligand-‐receptor system. J Clin Oncol, 2000. 18(2): p. 325-‐31.
41. Terrier, B., et al., Late-‐onset neutropenia following rituximab results from a hematopoietic lineage competition due to an excessive BAFF-‐induced B-‐cell recovery. Haematologica, 2007. 92(2): p. e20-‐3.
42. Koene, H.R., et al., Fc gammaRIIIa-‐158V/F polymorphism influences the binding of IgG by natural killer cell Fc gammaRIIIa, independently of the Fc gammaRIIIa-‐48L/R/H phenotype. Blood, 1997. 90(3): p. 1109-‐14.
43. Hatjiharissi, E., et al., Increased natural killer cell expression of CD16, augmented binding and ADCC activity to rituximab among individuals expressing the Fc{gamma}RIIIa-‐158 V/V and V/F polymorphism. Blood, 2007. 110(7): p. 2561-‐4.
44. Cornec, D., et al., Identification of patients with indolent B cell lymphoma sensitive to rituximab monotherapy. Annals of hematology, 2012. 91(5): p. 715-‐21.
45. Racila, E., et al., A polymorphism in the complement component C1qA correlates with prolonged response following rituximab therapy of follicular lymphoma. Clinical cancer research : an official journal of the American Association for Cancer Research, 2008. 14(20): p. 6697-‐703.
46. Azzato, E.M., et al., Common germ-‐line polymorphism of C1QA and breast cancer survival. British journal of cancer, 2010. 102(8): p. 1294-‐9.
47. Racila, E., et al., The pattern of clinical breast cancer metastasis correlates with a single nucleotide polymorphism in the C1qA component of complement. Immunogenetics, 2006. 58(1): p. 1-‐8.
48. Jin, X., et al., Homozygous A polymorphism of the complement C1qA276 correlates with prolonged overall survival in patients with diffuse large B cell lymphoma treated with R-‐CHOP. Journal of hematology & oncology, 2012. 5: p. 51.
49. Racila, D.M., et al., Homozygous single nucleotide polymorphism of the complement C1QA gene is associated with decreased levels of C1q in patients with subacute cutaneous lupus erythematosus. Lupus, 2003. 12(2): p. 124-‐32.
50. Castellano, G., et al., Maturation of dendritic cells abrogates C1q production in vivo and in vitro. Blood, 2004. 103(10): p. 3813-‐20.
51. Wang, S.Y., et al., NK-‐cell activation and antibody-‐dependent cellular cytotoxicity induced by rituximab-‐coated target cells is inhibited by the C3b component of complement. Blood, 2008. 111(3): p. 1456-‐63.
52. Lenz, G., et al., Molecular subtypes of diffuse large B-‐cell lymphoma arise by distinct genetic pathways. Proceedings of the National Academy of Sciences of the United States of America, 2008. 105(36): p. 13520-‐5.
53. Monti, S., et al., Molecular profiling of diffuse large B-‐cell lymphoma identifies robust subtypes including one characterized by host inflammatory response. Blood, 2005. 105(5): p. 1851-‐61.
54. Rosenwald, A., et al., The use of molecular profiling to predict survival after chemotherapy for diffuse large-‐B-‐cell lymphoma. The New England journal of medicine, 2002. 346(25): p. 1937-‐47.
159
55. Hasselblom, S., et al., The number of tumour-‐infiltrating TIA-‐1+ cytotoxic T cells but not FOXP3+ regulatory T cells predicts outcome in diffuse large B-‐cell lymphoma. British journal of haematology, 2007. 137(4): p. 364-‐73.
56. Alizadeh, A.A., et al., Prediction of survival in diffuse large B-‐cell lymphoma based on the expression of 2 genes reflecting tumor and microenvironment. Blood, 2011. 118(5): p. 1350-‐8.
57. Rimsza, L.M., et al., Loss of MHC class II gene and protein expression in diffuse large B-‐cell lymphoma is related to decreased tumor immunosurveillance and poor patient survival regardless of other prognostic factors: a follow-‐up study from the Leukemia and Lymphoma Molecular Profiling Project. Blood, 2004. 103(11): p. 4251-‐8.
58. Phipps-‐Yonas, H., et al., Low GILT Expression is Associated with Poor Patient Survival in Diffuse Large B-‐Cell Lymphoma. Frontiers in immunology, 2013. 4: p. 425.
59. Xu, Y., et al., Prognostic significance of tumour-‐infiltrating T lymphocytes and T-‐cell subsets in de novo diffuse large B-‐cell lymphoma: a multiparameter flow cytometry study. British journal of haematology, 2001. 112(4): p. 945-‐9.
60. Ansell, S.M., et al., Cd4+ T-‐cell immune response to large B-‐cell non-‐Hodgkin's lymphoma predicts patient outcome. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2001. 19(3): p. 720-‐6.
61. Houot, R., et al., Therapeutic effect of CD137 immunomodulation in lymphoma and its enhancement by Treg depletion. Blood, 2009. 114(16): p. 3431-‐8.
62. Fridman, W.H., et al., The immune contexture in human tumours: impact on clinical outcome. Nature reviews. Cancer, 2012. 12(4): p. 298-‐306.
63. Haabeth, O.A., et al., Inflammation driven by tumour-‐specific Th1 cells protects against B-‐cell cancer. Nature communications, 2011. 2: p. 240.
64. Tzankov, A., et al., Correlation of high numbers of intratumoral FOXP3+ regulatory T cells with improved survival in germinal center-‐like diffuse large B-‐cell lymphoma, follicular lymphoma and classical Hodgkin's lymphoma. Haematologica, 2008. 93(2): p. 193-‐200.
65. Coutinho R, C.A., Owen D, Greaves P, Hallam S, Wilson A, Matthews J, Calaminici M, Gribben J, , Exploring the Immune Microenvironment of Diffuse Large B Cell Lymphoma in a Tissue Microarray: Predicting Survival with a Score That Incorporates Macrophages, Cytotoxic and Regulatory T Cells, in Oral Abstract 951 American Society of Haematology Annual Meeting2011: San Diego.
66. Nam, S.J., et al., An increase of M2 macrophages predicts poor prognosis in diffuse large B-‐cell lymphoma patients treated with Rituximab-‐CHOP. Leukemia & lymphoma, 2014.
67. Lin, Y., et al., Immunosuppressive CD14+HLA-‐DR(low)/-‐ monocytes in B-‐cell non-‐Hodgkin lymphoma. Blood, 2011. 117(3): p. 872-‐81.
68. Mori, T., et al., T-‐helper (Th)1/Th2 imbalance in patients with previously untreated B-‐cell diffuse large cell lymphoma. Cancer immunology, immunotherapy : CII, 2001. 50(10): p. 566-‐8.
160
69. Ait-‐Tahar, K., et al., CD4-‐positive T-‐helper cell responses to the PASD1 protein in patients with diffuse large B-‐cell lymphoma. Haematologica, 2011. 96(1): p. 78-‐86.
70. Ding, Z.C., et al., Polyfunctional CD4+ T cells are essential for eradicating advanced B-‐cell lymphoma after chemotherapy. Blood, 2012.
71. Bernd, H.W., et al., Loss of HLA-‐DR expression and immunoblastic morphology predict adverse outcome in diffuse large B-‐cell lymphoma -‐ analyses of cases from two prospective randomized clinical trials. Haematologica, 2009. 94(11): p. 1569-‐80.
72. Ott, G., et al., Immunoblastic morphology but not the immunohistochemical GCB/nonGCB classifier predicts outcome in diffuse large B-‐cell lymphoma in the RICOVER-‐60 trial of the DSHNHL. Blood, 2010. 116(23): p. 4916-‐25.
73. Pasqualucci, L., et al., Analysis of the coding genome of diffuse large B-‐cell lymphoma. Nature genetics, 2011. 43(9): p. 830-‐7.
74. Rossi, E.A., et al., Preclinical studies on targeted delivery of multiple IFNalpha2b to HLA-‐DR in diverse hematologic cancers. Blood, 2011. 118(7): p. 1877-‐84.
75. Topalian, S.L., et al., Safety, activity, and immune correlates of anti-‐PD-‐1 antibody in cancer. The New England journal of medicine, 2012. 366(26): p. 2443-‐54.
76. Brahmer, J.R., et al., Safety and activity of anti-‐PD-‐L1 antibody in patients with advanced cancer. The New England journal of medicine, 2012. 366(26): p. 2455-‐65.
77. Berger, R., et al., Phase I safety and pharmacokinetic study of CT-‐011, a humanized antibody interacting with PD-‐1, in patients with advanced hematologic malignancies. Clinical cancer research : an official journal of the American Association for Cancer Research, 2008. 14(10): p. 3044-‐51.
78. Rotem-‐Yehudar R, A.P., Devine S, et al., 1216 the human-‐ized anti PD-‐1 antibody, CT-‐011, increases specific CD4+ effector/memory and memory T lymphocytes in patients with diffuse large B cell lymphoma (DLBCL) following autologous stem cell transplantation (AuSCT). Blood (ASH Annual Meeting Abstracts). 2009;114:1216.
79. Li, Z.M., et al., Blood Lymphocyte-‐to-‐Monocyte Ratio Identifies High-‐Risk Patients in Diffuse Large B-‐Cell Lymphoma Treated with R-‐CHOP. PLoS One, 2012. 7(7): p. e41658.
80. Wilcox, R.A., et al., The absolute monocyte and lymphocyte prognostic score predicts survival and identifies high-‐risk patients in diffuse large-‐B-‐cell lymphoma. Leukemia : official journal of the Leukemia Society of America, Leukemia Research Fund, U.K, 2011. 25(9): p. 1502-‐9.
81. Kim, D.H., et al., Absolute lymphocyte counts predicts response to chemotherapy and survival in diffuse large B-‐cell lymphoma. Leukemia : official journal of the Leukemia Society of America, Leukemia Research Fund, U.K, 2007. 21(10): p. 2227-‐30.
82. Song, M.K., et al., Influence of low absolute lymphocyte count of patients with nongerminal center type diffuse large B-‐cell lymphoma with R-‐CHOP therapy. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO, 2010. 21(1): p. 140-‐4.
83. Porrata, L.F., et al., New-‐onset lymphopenia assessed during routine follow-‐up is a risk factor for relapse postautologous peripheral blood
161
hematopoietic stem cell transplantation in patients with diffuse large B-‐cell lymphoma. Biology of blood and marrow transplantation : journal of the American Society for Blood and Marrow Transplantation, 2010. 16(3): p. 376-‐83.
84. Gergely, L., et al., Pretreatment T lymphocyte numbers are contributing to the prognostic significance of absolute lymphocyte numbers in B-‐cell non-‐Hodgkins lymphomas. Pathology oncology research : POR, 2011. 17(2): p. 249-‐55.
85. Tadmor, T., et al., Absolute monocytosis at diagnosis correlates with survival in diffuse large B-‐cell lymphoma-‐possible link with monocytic myeloid-‐derived suppressor cells. Hematological oncology, 2012.
86. Kristiansen, M., et al., Identification of the haemoglobin scavenger receptor. Nature, 2001. 409(6817): p. 198-‐201.
87. Schaer, D.J., et al., CD163 is the macrophage scavenger receptor for native and chemically modified hemoglobins in the absence of haptoglobin. Blood, 2006. 107(1): p. 373-‐80.
88. Steidl, C., et al., Tumor-‐associated macrophages and survival in classic Hodgkin's lymphoma. The New England journal of medicine, 2010. 362(10): p. 875-‐85.
89. Yamamoto, W., et al., HLA-‐DR expression on flow cytometry and tumor-‐associated macrophages in diffuse large B cell lymphoma treated by R-‐CHOP therapy: Retrospective cohort study. Leukemia & lymphoma, 2014.
90. Sekerkova, A., et al., CD14+CD16+ and CD14+CD163+ monocyte subpopulations in kidney allograft transplantation. BMC immunology, 2014. 15(1): p. 4.
91. Gabrilovich, D.I. and S. Nagaraj, Myeloid-‐derived suppressor cells as regulators of the immune system. Nature reviews. Immunology, 2009. 9(3): p. 162-‐74.
92. Almand, B., et al., Increased production of immature myeloid cells in cancer patients: a mechanism of immunosuppression in cancer. Journal of immunology, 2001. 166(1): p. 678-‐89.
93. Diaz-‐Montero, C.M., et al., Increased circulating myeloid-‐derived suppressor cells correlate with clinical cancer stage, metastatic tumor burden, and doxorubicin-‐cyclophosphamide chemotherapy. Cancer immunology, immunotherapy : CII, 2009. 58(1): p. 49-‐59.
94. Movahedi, K., et al., Identification of discrete tumor-‐induced myeloid-‐derived suppressor cell subpopulations with distinct T cell-‐suppressive activity. Blood, 2008. 111(8): p. 4233-‐44.
95. Bronte, V. and P. Zanovello, Regulation of immune responses by L-‐arginine metabolism. Nature reviews. Immunology, 2005. 5(8): p. 641-‐54.
96. Rodriguez, P.C. and A.C. Ochoa, Arginine regulation by myeloid derived suppressor cells and tolerance in cancer: mechanisms and therapeutic perspectives. Immunological reviews, 2008. 222: p. 180-‐91.
97. Szuster-‐Ciesielska, A., et al., Reactive oxygen species production by blood neutrophils of patients with laryngeal carcinoma and antioxidative enzyme activity in their blood. Acta oncologica, 2004. 43(3): p. 252-‐8.
98. Kusmartsev, S., et al., Antigen-‐specific inhibition of CD8+ T cell response by immature myeloid cells in cancer is mediated by reactive oxygen species. Journal of immunology, 2004. 172(2): p. 989-‐99.
162
99. Lossos, I.S., et al., Prediction of survival in diffuse large-‐B-‐cell lymphoma based on the expression of six genes. The New England journal of medicine, 2004. 350(18): p. 1828-‐37.
100. Murdoch, C., et al., The role of myeloid cells in the promotion of tumour angiogenesis. Nature reviews. Cancer, 2008. 8(8): p. 618-‐31.
101. Masque-‐Soler, N., et al., Molecular classification of mature aggressive B-‐cell lymphoma using digital multiplexed gene expression on formalin-‐fixed paraffin-‐embedded biopsy specimens. Blood, 2013. 122(11): p. 1985-‐6.
102. Scott, D.W., et al., Gene expression-‐based model using formalin-‐fixed paraffin-‐embedded biopsies predicts overall survival in advanced-‐stage classical hodgkin lymphoma. J Clin Oncol, 2013. 31(6): p. 692-‐700.
103. Scott, D.W., et al., Determining cell-‐of-‐origin subtypes of diffuse large B-‐cell lymphoma using gene expression in formalin-‐fixed paraffin embedded tissue. Blood, 2014.
104. Plonquet, A., et al., Peripheral blood natural killer cell count is associated with clinical outcome in patients with aaIPI 2-‐3 diffuse large B-‐cell lymphoma. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO, 2007. 18(7): p. 1209-‐15.
105. Kanakry, J.A. and R.F. Ambinder, EBV-‐related lymphomas: new approaches to treatment. Current treatment options in oncology, 2013. 14(2): p. 224-‐36.
106. Jones, K., et al., Expansion of EBNA1-‐specific effector T cells in posttransplantation lymphoproliferative disorders. Blood, 2010. 116(13): p. 2245-‐52.
107. Nguyen-‐Van, D., et al., Epstein-‐Barr virus-‐positive diffuse large B-‐cell lymphoma of the elderly expresses EBNA3A with conserved CD8 T-‐cell epitopes. American journal of blood research, 2011. 1(2): p. 146-‐59.
108. Challa-‐Malladi, M., et al., Combined genetic inactivation of beta2-‐Microglobulin and CD58 reveals frequent escape from immune recognition in diffuse large B cell lymphoma. Cancer cell, 2011. 20(6): p. 728-‐40.
109. Afshar-‐Sterle, S., et al., Fas ligand-‐mediated immune surveillance by T cells is essential for the control of spontaneous B cell lymphomas. Nature medicine, 2014.
110. List, A.F., et al., Deficient tumor-‐infiltrating T-‐lymphocyte response in malignant lymphoma: relationship to HLA expression and host immunocompetence. Leukemia : official journal of the Leukemia Society of America, Leukemia Research Fund, U.K, 1993. 7(3): p. 398-‐403.
111. Wada, N., et al., Tumour-‐associated macrophages in diffuse large B-‐cell lymphoma: a study of the Osaka Lymphoma Study Group. Histopathology, 2012. 60(2): p. 313-‐9.
112. Wada, N., et al., Diffuse large B cell lymphoma with an interfollicular pattern of proliferation shows a favourable prognosis: a study of the Osaka Lymphoma Study Group. Histopathology, 2012. 60(6): p. 924-‐32.
113. Yoshida, N., et al., Clinical Significance of sIL-‐2R Levels in B-‐Cell Lymphomas. PloS one, 2013. 8(11): p. e78730.
114. Wright, G., et al., A gene expression-‐based method to diagnose clinically distinct subgroups of diffuse large B cell lymphoma. Proceedings of the National Academy of Sciences of the United States of America, 2003. 100(17): p. 9991-‐6.
163
115. Natkunam, Y., et al., LMO2 protein expression predicts survival in patients with diffuse large B-‐cell lymphoma treated with anthracycline-‐based chemotherapy with and without rituximab. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 2008. 26(3): p. 447-‐54.
116. Taube, J.M., et al., Colocalization of inflammatory response with B7-‐h1 expression in human melanocytic lesions supports an adaptive resistance mechanism of immune escape. Science translational medicine, 2012. 4(127): p. 127ra37.
117. Westin, J.R., et al., Safety and activity of PD1 blockade by pidilizumab in combination with rituximab in patients with relapsed follicular lymphoma: a single group, open-‐label, phase 2 trial. The lancet oncology, 2013.
118. Pardoll, D.M., The blockade of immune checkpoints in cancer immunotherapy. Nature reviews. Cancer, 2012. 12(4): p. 252-‐64.
119. Elstrom, R.L., P. Martin, and J.P. Leonard, New biologic agents and immunologic strategies. Hematol Oncol Clin North Am, 2008. 22(5): p. 1037-‐49, x-‐xi.
120. Goldstein, M.J., et al., A CpG-‐loaded tumor cell vaccine induces antitumor CD4+ T cells that are effective in adoptive therapy for large and established tumors. Blood, 2011. 117(1): p. 118-‐27.
121. Wiernik, P.H., et al., Lenalidomide monotherapy in relapsed or refractory aggressive non-‐Hodgkin's lymphoma. J Clin Oncol, 2008. 26(30): p. 4952-‐7.
122. Tadmor, T., et al., Absolute monocytosis at diagnosis correlates with survival in diffuse large B-‐cell lymphoma-‐possible link with monocytic myeloid-‐derived suppressor cells. Hematol Oncol, 2012.
123. Jinushi, M., et al., Tumor-‐associated macrophages regulate tumorigenicity and anticancer drug responses of cancer stem/initiating cells. Proceedings of the National Academy of Sciences of the United States of America, 2011. 108(30): p. 12425-‐30.
124. Jones, K., et al., Serum CD163 and TARC as Disease Response Biomarkers in Classical Hodgkin Lymphoma. Clin Cancer Res, 2013. 19(3): p. 731-‐42.
125. Li, Z., et al., Pomalidomide shows significant therapeutic activity against CNS lymphoma with a major impact on the tumor microenvironment in murine models. PloS one, 2013. 8(8): p. e71754.
126. Graversen, J.H., et al., Targeting the hemoglobin scavenger receptor CD163 in macrophages highly increases the anti-‐inflammatory potency of dexamethasone. Molecular therapy : the journal of the American Society of Gene Therapy, 2012. 20(8): p. 1550-‐8.
164