---~----------~----------------------------

Neuroto~icity of Active Compounds -Establishment of hESC-Lines andProteomics Technologies for HumanEmbryo- and Neurotoxici7" Screeningand Biomarker IdentificationMartina Klemm and Andre SchrattenholzProteoSys AG, D-Mainz

SummaryPharmaceutical and chemical industries are facing newchallenges for hazard and risk assessment from regulatoryagencies. Especially for potential embryo toxicity of active com-pounds, conclusions from animal testing remain problematicdue to numerous species-specijic effects. Developmental toxici-ty screening preferentially should be performed with humanmaterial.Appropriate models are scarce or missing, and the developmentof a human in vitro model for the molecular characterisation ofembryo toxic effects appears to be highly desirable. Theoutstanding advantages of a human embryonic stem cell(hESC) based in vitro screening model for embryonic neuro-toxicity become clear from corresponding results from a murineESC-screening system. This in vitro test system is based onneuronal differentiated murine embryonic stem cells and quan-titative differential proteomic display techniques to identifybiomarkers for neurotoxicity. Results are superior to those ofconventional array technologies (nucleic acids), because theproteomic analysis covers posttranslational modijications.Under the new strict guidelines for stem cell importation of theGerman Ministry of Health and a Central Ethics Commissionfor Stem Cell Research, it is now possible for the first time toexploit the outstanding features of human embryonic stem cellsto establish an innovative screening method for embryo- andneurotoxicity and to identify toxicity biomarkers without usinganimal-based in vitro or in vivo systems.

Zusammenfassung: Neurotoxizitat von Arzneimitteln - embry-onale Stammzell-Modelle zur Identifikation diagnostischerBiomarker fur humane Embryo- und NeurotoxizitatIm Rahmen der Arzneimittelsicherheit miissen pharmakolo-gisch aktive Substanzen hinsichtlich ihres potenriellen embryo-toxischen und teratogenen Gefahrenpotenzials iiberpruftwerden. Da zahlreiche Chemikalien und Wirkstoffe gerade imBereich der Embryotoxizitdt spezies-spezifisch. wirken, erweistsich die Verwendung menschlicher Zellen zur Identifizierunghumaner Eniwicklungstoxizitat als zwingend notwendig.Die aussergewohnlichen Vorzuge eines auf humanen embryo-nalen Stammzellen (ES) basierenden in vitro screening Modellszur embryonalen Neurotoxizitdt wird aus den dargelegtenResultaten eines murinen ES-Zell-Screening Systems er-sichtlich. Dieses in vitro Testsystem zur Identifikation diagnos-tischer Neurotoxizitatsmarker basiert auf neuronal differen-zierten murinen embryonalen Stammzellen und quantitativerdifferentieller Proteinmusteranalyse. Das proteomische Ver-fahren ist den rein nukleinsdurebasierten Testsystemen an Aus-sagekraft weit iiberlegen, da posttranslationale Modijikationenerfasst werden.Durch die neuen gesetzlichen Regelungen (StZG) zu Import undUmgang mit humanen embryonalen Stammzellen (hES Zellen)ergibt sich jetzt auch in Deutschland erstmals die Moglichkeitdiagnostische Toxizitdtsbiomarker ohne in vitro und in vivoTierversuche unter Einsat; einer innovativen Screening-Methode for Embryo- und Neurotoxiritdt zu identifizieren.

Keywords: human embryonic stem (ES) cells, embryo toxicity, neurotoxicity, quantitative differential proteomics, biomarker

1 Introduction

Medical research is dedicated to the con-tinuous improvement of risk assessmentfor health protection. In the EuropeanCommission's regulatory framework ofdrug approval for treatment of patients, acertain set of standardised animal tests is

performed according to the EuropeanCouncil directives (EC 1999 and 2001)and the guidelines of the InternationalConference on Harmonisation (ICH,1993). The hierarchical decision-treeapproaches defined by the EuropeanCommission (EC, 1989) and the Organi-sation for Economic Cooperation and

Development (OECD, 1982 and 1996)for international standards of pharmaco-.logical and toxicological tests primarilyaddress the adult organism. Currentlycomprehensive multigeneration studiesare undertaken to provide information onall aspects of developmental toxicity. De-velopmental studies to address thespecific risks of the developing embryoare not regulatory requirements andReceived 27 November 2003; received in final form and accepted for publication 12 February 2004

ALTEX 21, Supp!. Linz 03/2004 41

KLEMM AND SCHRATTENHOLZ ~---------------------------------------~---

appropriate methods for the screening oftoxic substance effects during embryonicdevelopment, especially on the part ofdevelopmental neurotoxicity, are essen-tially lacking.The interpretation for human embryo-

toxicity remains problematic. There havebeen cases of toxic side effects withpathological consequences during devel-opment, which were only discoveredafter approval of the drug for therapeuticapplication in patients (dire case of Con-tergan). More than thirty compoundswith highly toxic potential had to bewithdrawn from the market by the FDA(Food and Drug Administration) between1998 and 200l. In each case this toxicitywas not detected in the required set ofprior animal tests. One of the most recentexamples is Lipobay (A. P. Li, Phase 1Molecular Toxicology, Inc., Santa Fe,New Mexico). Among these 30, eightdrugs had substantial side effects espe-cially in women. Current animal tests aresimply not sufficient to detect gender-specific predispositions in humans(Dr. K. Olden, Director of NIEHS). Inthe time after Contergan, only about 15%of those cases of drugs with highly toxicpotential could be predicted safely byanimal tests (Heywood, 1990 und Parke,1994).With the establishment of human

embryonic stem cells, there now existsfor the first time the chance to developinnovative in vitro analytical methods forearly detection of human-specific embryo-toxic risks of compounds and chemicals.The combination of hESC-based in vitrotest models with cutting edge molecularanalysis holds the promise to deliver withextreme sensitivity highly relevant func-tional and molecular marker end pointsfor the early detection of embryo toxiceffects of drugs or chemicals. In terms ofrelevance, proteomics is in principlefar superior to RNAIDNA-based arraytechnologies, because the relativelystatic and small human genome of ESC-differentiationapprox. 30,000 genes is translated into ahighly dynamic and complex proteomeof several million protein molecules ofposttranslationally modified isoforms. It Embryoid bodies

is on this latter level where dynamicfunctional aspects of cellular events takeplace (Schrattenholz, 2001; Kitano,2001; Lowe and Marth, 2003). Neuro- Fig. 1: Establishment of human and murine embryonic stem cells.

42

and embryotoxic substances bind to pro-teins, changing interactions in complexfunctional networks, sometimes on veryfast time scales (e.g. phosphorylation),and the changes induced by this type ofintervention are ideally analysed byproteomics technologies. Emergingmolecular signatures are derived fromfunctionally controlled hES cells differ-entiating towards mature neurons, keybiomarker proteins are reliably identifiedand quantified, thus enabling very earlyunderstanding of side effects and innova-tive strategies for acceleration in drugdevelopment, validation and toxicitytesting. After identification of develop-mental neurotoxicity markers on proteinbasis, affinity molecule-based proteinchips for common readers can be devel-oped for high throughput screeningassays in industry.

2 Analysis of developmentalneurotoxic substances- humanembryonic or adult stem cells?

Stem cells have basically two character-istic properties: they can be maintained

undifferentiated in culture for long peri-ods of time, and they can be induced tospecific organotypic differentiation.There exist embryonic and adult tissue-specific stem cells.Embryonic stem cells are established

from the inner cell mass (ICM) of 5-7day old blastocysts (Evans and Kaufman,1981; Martin, 1981) (Fig. 1).For long-term cultivation of undiffer-

entiated cells, embryonic stem cell cul-tures are maintained in co-culture withso-called "feeder" cells, i.e. mouse fibro-blasts (Robertson, 1987) or in cell-freeconditioned media (Smith and Hooper,1987). The major inhibitory factor fordifferentiation in these cell culture sys-tems is the cytokine LIF (Leukemia In-hibitory Factor) (Nichols et al., 1990).Embryonic stem cells display far-reach-ing pluripotent properties, i.e. they candifferentiate into virtually every organ-specific human cell type, when culturedas embryoid bodies. Differentiation intodiverse cell types of endodermal (e.g.pancreatic and hepatic cells), mesoder-mal (e.g. bones, muscle and blood cells,cardiomyocytes) or ectodermal origin(e.g. neurons and glial cells) is induced

4-cell stage

Cell dissociation and seeding

Blastocyst stage

:--'\'s'- Inner cell mass (ICM).". ..•..:ES-cell colony on top of murine fibroblastfeeder cells(irradiated or mitomycin C treated)

t-

ALTEX 21, Suppl. Linz 03/2004

m..... KLEMM AND SCHRATTENHOLZ---~----------~---------------------------

by LIF deprivation and the respectivecombination of growth factors.

Tissue- or organ-specific stem cells fromadult tissues such as epidermis, hair, in-testine, liver, haematopoietic system,brain or bone marrow have limited possi-bilities of proliferation and differentia-tion ("developmental restriction"). Dueto their multipotent properties, thesestem cells are mostly restricted to onetissue only. Plasticity of the adult stemcells, which means the ability to differ-entiate into cell types characteristic ofanother organ, was only demonstratedfor haematopoietic stem cells, stromalcells of the bone marrow and multipotentadult precursors, which can be generatedin vitro from certain cells of the bonemarrow and certain neural cells (NIHstem cell report, 2001). These adult stemcells appear to have the capability to dif-ferentiate into tissues other than thosefrom which they originated, namely1) blood and bone marrow (unpurifiedhaematopoietic) stem cells differentiateinto the 3 major types of brain cells (neu-rons, oligodendrocytes, and astrocytes),skeletal muscle cells, cardiac musclecells, and liver cells;2) bone marrow (stromal) cells differen-tiate into cardiac muscle cells, skeletalmuscle cells, fat, bone, and cartilage; and3) brain stem cells differentiate intoblood cells and skeletal muscle cells.

The establishment of a screeningmethod for early developmental process-es related to human embryonic neurotox-icity is feasible with embryonic but notwith adult stem cells. Adult stem cells arederived from mature organs of adult indi-viduals and a profiling for early embry-onic differentiation and correspondinganalysis of toxic effects is simply not fea-sible. The question, whether adult stemcells can be artificially re-differentiatedis currently under intense investigation,but remains open for the time being. Afurther disadvantage of adult stem cells istheir limited ability to differentiate andproliferate, which goes together with de-creased lifetime in vitro. Moreover, onlyvery small numbers of these cells areavailable (decreasing further with age),they are difficult to distinguish from sur-rounding cells and their derivation is dif-ficult or impossible, depending on the or-gan. To date it remains unclear how adultstem cells function and how their poten-tial could best be exploited (G. Q. Daley,chairman of the Whitehead Institute TaskForce on Genetics and Public Policy).

In consequence, results with directrelevance for human developmental toxi-city can only be obtained by human em-bryonic stem cell approaches. Only suchmodels can provide information on theeffects of compounds on all aspects ofembryonic neuronal development.

3 Legal framework in Germanyfor import of human embryonicstem cells andstrictly defined use

Import and use of human embryonicstem cells is generally prohibited in Ger-many under a so-called "stem cell law"(Stammzellgesetz, StZG, § 4, 1) (seeTable 1, according to Commission staffworking paper, Brussel, 03.04.2003)

Permission can only be granted in ex-ceptional cases and after a thorough re-view procedure by the responsible regu-latory authority (Robert Koch Institut,RKl) in close agreement with a CentralEthics Commission. The following con-ditions have to be fulfilled: i) Stem cellsmust have been produced prior to the 1stof January 2002, according to the legalrequirements of the country in whichthey originate; ii) embryos were generat-ed for reproductive purposes and subse-quently abandoned; iii) embryos weresubmitted to stem cell derivation withoutany financial connotations or financialadvantage to the donors. Moreover, pro-jects involving hESC are only permittedfor basic research (§5) and under condi-tion of no available scientific alternatives(§5).

On the 9th of September, 2003, Pro-teoSys AG, a biotechnology company inMainz, obtained permission from RKl

AT BE OK DE ES FI FR GR IE IT LU NL PT SE UK

Allowance: Acquisition of human embryonic stem x x x x xcells from supernumerary embryos by law

Prohibition: Acquisition of human embryonic stem x x xcells from supernumerary embryos

Allowance: Importation of human embryonic stem xcell lines (StZG)

Prohibition: Acquisition of human embryonic stem x x xcells from supernumerary embryos

No specific legislation regarding human embryo x x x xresearch

Allowance: Derivation of human embryos for stem xcell procurement by law

Prohibition: Derivation of human embryos for x x x x x x x x x x xresearch purposes and for the acquisition of stem cellsby law or by ratification of the convention of thecouncil of Europe on human rights and biomedicinesigned in Oviedo on 4 April 1997

Tab. 1: Human embryonic stem cell research: regulations in EU member states (Commission staff working paper,Brussel, 3. April 2003).

ALTEX 21. Supp\. Linz 03/2004 43

KLEMM AND SCHRATTENHOLZ ~-------------~-

and the Central Ethics Commission toimport hESC for a basic research projectwith the following topic: Development ofan in vitro system for the analysis of neu-rotoxic effects of a selected set of com-pounds using differentiating human em-bryonic stem cells as a model.

4 Biomarker discoveryin neuronal embryotoxicity:proteosys' approach

4.1 ObjectivesA clear understanding of neurotoxic ef-fects of medical drugs on human em-bryos during pregnancy would constitutea major contribution towards improveddrug safety and preventive health care.The effects of neurotoxic substance ap-plication can be defined on the level ofdistinct molecular consequences in termsof immediate protein expressionchanges. Advanced proteomics technolo-gies in correlation with synchronisedfunctional/physiological measurementscan provide a comprehensive, preciseand quantitative molecular pattern analy-sis of the underlying mechanisms. Thereduction of the enormous complexity ofproteomic data, with exact quantificationof differentially expressed proteins, helpsto obtain meaningful interpretations ofrelevant events within complex and high-ly dynamic molecular interactions.Moreover the availability of humanbiomarkers for neuronal embryotoxicitycould drastically improve early screeningprocedures for embryo- and neurotoxicsubstances.Therefore, the main objective of Pro-

teoSys' project is the identification ofmolecular mechanisms underlying hu-man neurotoxicity of 20 selected sub-stances. The corresponding events takeplace in early embryonic and specificallyneural developmental stages and matura-tion processes, thus an in vitro test sys-tem based on hESC is ideally suited forfunctional and molecular analysis. Cellswill be exposed to various concentrationsof toxic compounds and analysed on themolecular level at various endpoints, rep-resenting precursors, early and maturestages of neurons. The differential pro-teomic analysis compares protein expres-sion of hESC after substance application

44

with untreated controls. The quantitativeand reliably differential Proteomics tech-nology applied by ProteoSys will gener-ate molecular signatures of toxic condi-tions based on human biomarkers forembryonic neurotoxicity (specific post-translational modifications of proteins).

4.2 Proteomicstechnologyplatforms4.2.1 FunctionalcontrolProteins are multifunctional and highlydynamic modules. Because several mil-lion protein isoforms and modificationsare generated from approx. 30,000human genes, the necessity of rigorouspattern control of actual protein expres-sion in a cellular system is evident. Pro-teoSys has established procedures for acorrelational analysis of ionic currents,concentration changes of intra- and ex-tracellular small molecules (like neuro-transmitters), as well as physicochemicalparameters (like temperature), which set.the context and functionally define themeaning of protein changes.

4.2.2 Quantitative anddiHerential protein detectionProteins exist in a linear dynamic rangeover more than eight orders of magnitude.Quantitative display of protein patterns isachieved by integrated analytical plat-forms applying isotopic techniques atthree levels of sensitivity. On the first lev-el, which is fully automated, proteins aredifferentially detected, quantified and sub-sequently identified down to a level of ap-prox. 1 femtomol (thousands of identifica-tions per week). Detection of proteinsdown to a range of 50 to 100 atto moles isequally fast, but current limitations ofmass spectrometry necessitate non-auto-mated enrichment procedures for the iden-tification. Very scarce proteins (approx. 1to 50 atto moles) need special radioactivetracer methods, applying the proprietaryMPD technology (Multi Photon Detec-tion). This ultrasensitive method is basedon protein labelling with so-called "elec-tron capture"-isotopes (e.g. 125-I, 131-I),which on decay emit coincident events,which enable detection without back-ground limitations.The following chart gives an example

from cancer research of quantitative dif-

ferential protein pattern display asachieved by labelling samples with ra-dioactive isotopes (Fig. 2):

4.2.3 Complexity reduction byfractionationAffinity-based fractionations of phos-phorylated, glycosylated and other func-tionally related posttranslational modifi-cations have become important tools ofsystem biology. These subproteomesconstitute fractions (0.5 to 30%) of totalprotein content of samples and thus focuson crucial aspects of fast cellular sig-nalling. Together with the mentioned dis-play technologies and highly automated,reproducible procedures with reasonablethroughput, we have established a flexi-ble analytical basis for comprehensivemolecular characterisation of mode ofaction and side effects.A further, even more focussed ap-

proach, called chemical proteomics, usesimmobilised ligands for specific affinitycapture of membrane receptors and inter-acting proteins from cellular compart-ments otherwise considered to be diffi-cult in analytical terms.The corresponding suite of proprietary

protein analytical technologies andknow-how enables ProteoSys AG toperform the comprehensive moleculardescription of the embryonic stem cellmodels, with the ultimate goal of obtain-ing signatures of biomarkers for develop-mental neurotoxicity.

5 "System biology andtoxicoproteomics" using ahuman in vitro ES-cellscreeningsystem

In detail the research project granted iscomposed of two parts (Fig. 3).In the first part of the project human

embryonic stem cells will be cultivatedin vitro and differentiated to functionalneurons according to published proce-dures (Thomson et al., 1998; Itskovitz-Eldor et al., 2000; Pera et al., 2000;Richards et al., 2002). UndifferentiatedhESC's will be characterised by surfaceantigens like SSEA-3, SSEA-4 (Solterand Damjanow, 1979), TRA-I-60, TRA-1-81, Oct-4 (Scholer et al., 1989) andGCTM-2 or by enzyme activities (alka-

ALTEX 21, Supp!. Linz 0312004

~ KLEMM AND SCHRATTENHOLZ--~--------------------------

(pH 4-7, 18cm)

Fig. 2: Quantitative differential pattern control of neuronaldifferentiation with undifferentiated murine stem cells, early andlate neurons. Small aliquots of three different biological samples werelabelled with three different radioactive iodine isotopes (125-1,123-1 and131-1), mixed together and jointly ana lysed in one 2D gel. Theradioactive labelling enables a very reliable and exact quantification anddifferential detection of even subtle changes of protein expressionpatterns. The table shows differential expression levels for selectedspots. Selection is based on statistical significance calculated usinganalysis of variance (ANOVA).

line phosphatase). Neurons differentiatedfrom precursors are characterised bymorphological criteria and immunohisto-chemistry (antibodies for neuron-specificproteins MAP-2 and synaptophysin).Subsequently and more importantly, afunctional and physiological control ofneurons based on their response to vari-ous neuron-specific stimuli (neurotrans-mitter, depolarisation, etc.) is performed,e.g. by calcium-imaging.Together with specific agonists and

antagonists these measurements allowfairly precise pharmacological character-isation of the major ionotropic andmetabotropic receptors (for glutamate,GABA, acetylcholine, etc.) and voltage-dependent ion channels. This fine-tuningof physiological responses of the cellularsystem is a prerequisite for definingappropriate amplitudes of conditions,which later form the basis for the sub-sequent molecular analysis.The exposition to 20 selected toxic

test substances will be investigatedduring the stage of hESC differentiationinto neuronal cells, as well as afterreaching postmitotic mature status. Theresults from functional measurementswill define the endpoints for generationof samples submitted to differentialproteomics technologies. The tight func-tional control of the biological material

ALTEX 21, Suppl. Linz 03/2004

is one of the most crucial steps of com-plexity reduction.In the second part of the project, neu-

ro-embryotoxic and neurotoxic effects oftest substances will be analysed on themolecular level (Fig. 4).Dose-response relationships of toxic

effects induced by substances are essen-tially quantified by the influence of thesecompounds on normal functional signalsmeasured by calcium imaging. Evensubtle influences of toxic compounds onintracellular transient calcium concentra-tion changes due to normal reactionscaused by neurotransmitters or otherphysiological parameters will becomeapparent on a statistically significantlevel.Based on the functional analysis (pro-

ject part 1), protein pellets will be gener-ated from treated and untreated neuronaldifferentiated human stem cells at appro-priate end points, and will subsequentlybe submitted to a quantitative and differ-ential protein pattern analysis (Fig. 5).Small aliquots of up to three different

biological samples will be labelled withthree different radioactive iodineisotopes (125-1, 123-1 and l31-D, mixedtogether and jointly analysed in one 2Dgel. This part is for pattern control only;the bulk of the three samples will be keptfor later tracer-controlled preparative

51 52 53

0.18 :to.Ol 0.33 :0.01 Q.49 ±O.O

0.09 ±O.OS 0.55 :1:0.0

0.09 ±O.O 0.57 :1:0.0

0.21 ±0.01 0.25 :1:0.0

0.56 :1:0.0 0.31 :1:0.0

2D gels for protein identification. Theradioactive labelling enables a very reli-able and exact quantification and differ-ential detection of even subtle changes ofprotein expression patterns (see also Fig.4). The quantification moreover is essen-tial for the understanding of the sequenceof molecular events in the given timeframe of experiments and thus for thebiological interpretation (Schrattenholzand Cahill, 2003; Cahill et aI., 2003; Vogtet aI., 2003).Identification of differential protein

biomarkers is performed by mass spec-trometry, for very scarce proteins aftertracer-controlled enrichment procedures.Bioinformatics at ProteoSys, has estab-lished effective data-mining procedures,screening real molecular data of identi-fied proteins and their respective modifi-cations (e.g. phosphorylations) againstcomplementary data from literature filesor additional experiments (e.g. fromnucleic acid arrays). Eventually, refinedmolecular signatures for human embry-onic neurotoxicity are generated, provid-ing the content for second generationhigh-throughput screening devices (e.g.antibody arrays against specific phos-phorylated protein fragments). As shownin Figure 6, a rigorous design of correlat-ed experimental levels enables the recog-nition of relevant contents of functional

45

KLEMM AND SCHRATTENHOLZ m....-------------~-

Establishment of humanembryonic stem cells

Differentiation intoneuronal cells

characterisation of postmitotic neurons: morphological,immunohistochemical and functional analysis (Ca-Imaging)

2dose-response curves for classified chemicals of theinternational validation study on in vitro embryotoxicity tests

proteomic analysis - fractions, multicolor 2D gels,ultrasensitiv detection, quantification and identification of to x-marker, bioinformation

3

II

4 human specific Neurotox-signatures

Fig. 3: System biology research approach. Project part 1 relates to in vitro cultureand differentiation of hES-celis into mature neurons, and a corresponding correlationalfunctional analysis; project part 2 covers the differential molecular analysis by proteomictechniques of dose-dependent consequences of exposure of cells to 20 selectedtoxic compounds in the human and murine in vitro ESe-based screening systems; finalobjective is the definition of human-specific embryonic neurotoxicity biomarkers.

extra- and intracellularchemical-induced changes

p::re:c::is:e::IY:c:o:n:tr:o:"e:d:i!nplu.t•••••• _l~. - _- . .'~Insult ~

Defined and relevantcell material

classifiedtest substances

LC/MS

Real timeCa-imaging

Neurotoxicity of testsubstances

(dose-response curve)

Fig. 4: Neurotoxicity of test substances. Functional correlational analysis, definition ofphysiological and/or pharmacological/toxic end points. The relevant cell status foreffective neurotoxic concentrations of the test substances was functionally defined by theoutcome of dose-response curves in calcium-signalling experiments.

proteinaceous information of innovativediscovery and screening methods inthe post-genome era, in the context ofsystem biology.The first step, the exact functional

tuning of the respective model system, is

also the most crucial step, because sub-sequent analytical methods are extremelypowerful and do not show up previousmistakes. Well-controlled functionalendpoints (and only they will allow, aftermolecular details become available, a

46

correct interpretation of data) are submit-ted first to quantitative differential pro-tein analysis, and then to tracer-con-trolled preparative experiments foridentification or complete sequencing.Quantification of differential proteins isthe basis for integrating related resultsfrom independent experiments and litera-ture, generating focused hypotheses. In asecond level of bioinformatics, iterativealgorithms verify or falsify in silica leadsusing the actual data set (organising pep-tide information from mass spectrometryand positional, e.g. pI and molecularmass information form 2D gels).

6 Validation of the human invitro ES-cellscreening system forhuman-specific embryo- andneurotoxicity

In general, the selection of test sub-stances is one of the critical points forvalidation studies. A substance can causetoxic effects in animals in high-dose,long-term studies, but this does not nec-essarily mean that it is a risk to humanhealth. To reach our project goal, namelydescribing human-specific relevant sig-natures for a developmental neurotoxici-ty biomarker, we decided to select testchemicals from the ECVAM Internation-al Validation Study with good quality"segment II"- type in vivo data and/orhuman data as in vivo reference data anda murine based ESC-screening systemfor animal reference data. In this design,a distinction between human-specificembryotoxicity and general toxicity, andlikewise discrimination between non-embryotoxic, weakly embryotoxic andstrongly embryotoxic substances can bemade. By means of this study criticalissues concerning the interpretation ofanimal study data for human develop-mental toxicity (NIH stem cell report,2001; Rosen, 2002), like specific ADME(Absorption, Distribution, Metabolismand Excretion of the substance) effectsand combinatorial effects of biological,physical und chemical agents (lead,PCBs, pesticides, phthalates, endocrinedisruptors) can be analysed. Comparativeprofiling of human and murine neuronaldifferentiated cell material, a key issuefor human differential display toxicopro-

ALTEX 21, Supp\. Linz 0312004

m..... KLEMM AND SCHRATTENHOLZ--~--------------------------

teomics and a better protection of healthand safety of unborn human lives, can beaccomplished.We have already established a propri-

etary murine ESC-screening systembased on neuronal differentiation andperformed during in vitro cultivation,differentiation to functional neurons andfunctional-physiological tests. Based onthe functional analysis of specific testsubstances, protein pellets are generatedfrom treated and untreated neuronaldifferentiated murine stem cells at appro-priate end point conditions, and theprotein material is submitted to a quanti-tative and differential protein patternanalysis. The results clearly indicate theexistence of murine molecular toxicitysignatures (Sommer et al., 2003a and2003b). Even with detection sensitivitiesof bulk methods for protein detection,neuro-specific proteins (stress-relatedproteins and protein isoforms) represent-ing toxicity signatures were identified.To jointly analyse human and murine cellmaterial in comparative proteomic analy-sis, the whole procedure must be devel-oped and repeated with hESC-lines.Now, for the first time, it seems possi-

ble to establish an innovative screeningmethod for human-specific embryo- andneurotoxicity and to identify toxicitybiomarkers without animal-based in vitroor in vivo systems by exploiting the out-standing features of human embryonicstem cells. Moreover, the various possi-ble "organotypic" derivatives of embry-onic stem cells (neural, cardiomyocyte,adipocyte, various muscle cell types)available today provide genetically ho-mogeneous models for future, morecomprehensive experiments, perfectlysuited for innovative validation methodslike RNAi.

ReferencesCahill, M. A., Wozny, w., Schwall, G. etal. (2003). Analysis of relative isotopo-logue abundances for quantitative pro-filing of complex protein mixtures la-belled with the acrylamide/D3-acrylamide alkylation tag system.Rapid-Commun.- Mass-Spectrom. 17(12), 1283-90.

Commission of European Communities,(2003). Commission staff working pa-

ALTEX 21, Supp!. Linz 03/2004

Neurotox-signature

Cell lysate+ test substance

fractionation

Cell lysate- test substance

Maldi-TOF MS, ESI-identification

quantitative and qualitativedifferential display

Fig. 5: Functional protein pattern comparison - "differential display". Samples fromfunctionally verified stem cell cultures are prepared on a preparative scale. Subsequentlythey are differentially labelled by radioisotopic techniques, separated by 2-dimensional gelelectrophoresis with immobilised pH gradients. Eventually they are fragmented andidentified by automated peptide mass fingerprints via MALDI- TOF mass spectrometry.

Tab. 6: Sequence of steps in a system biology/proteomic study. ProteoSys hasdeveloped proprietary platforms for each of the stages (www.proteosys.com).

Sample Defined state of differentiated hESC

Functional Control Ca- or pH imaging, cytokine- or neurotransmitter release,morphological parameters

Sample preparation Denaturation, inactivation of all protein activity, labelling withradioactive or stable isotope tracers

Pattern analysis Two-dimensional gel electrophoresis, differential andquantitative protein detection

Protein identification Mass spectrometry (MALDI-TOF, ESI-MS, MS/MS

Bioinformatics 1 Integration of external data from literature or independentexperiments with actual quantitative protein expression;integrative project data base

Bioinformatics 2 Data Mining: hypothesis-driven structuring of project database, focussed experimental verification

lytical, pharmacotoxicological andclinical standards and protocols in re-spect of the testing of medicinal prod-ucts.

EC Directive 20011201EC and of theCouncil of 4 April 2001 on the approx-imation of the laws, regulations andadministrative provisions of the Mem-ber States relating to the implementa-tion of good clinical practice in theconduct of clinical trials on medicinalproducts for human use.

EC Directive 2001/83IEC of the Euro-pean parliament and of the council of 6

47

per: Report on human embryonic stemcell research, Brussels.

EC Council Directive 89/341IEEC of 3May 1989 amending Directives65/65IEEC, 75/318IEEC and 75/319/EEC on the approximation of provi-sions laid down by law, regulation oradministrative action relating to pro-prietary medicinal products.

EC Commission Directive 1999/83IECof 8 September 1999 amending theAnnex to Council Directive 75/318/EEC on the approximation of the lawsof the Member States relating to ana-

KLEMM AND SCHRATTENHOLZ m.....-------------~-

November 2001 on the Communitycode relating to medicinal products forhuman use.

Evans, M. J. and Kaufman, M. H. (1981).Establishment in culture of pluripotentcells from mouse embryos. Nature292, 154-156.

Heywood, R. (1990). In Animal ToxicityStudies: Their Relevance for Man. InC. E. Lumley und S. R. Walker (Hrsg.),Quay Publishing.

ICH International Conference on Har-monisation of Technical Requirementsfor Registration of Pharmaceuticals forHuman Use. ICH Harmonised Tripar-tite Guideline: Detection of Toxicity toHuman Reproduction for MedicinalProducts. Recommended for Adoptionat Step 4 of the ICH Process on 24 Ju-ly 1993 by the ICH Steering Commit-tee. ICH Secretariat, Geneva.

Itskovitz-Eldor, J. and Benvenisty, N.(2000). Differentiated human embry-oid cells and a method for producingthem. WO 00/70021.

Kitano, H. (Hrsg.) (2001). Foundationsof Systems Biology. Boston: MITPress.

Lowe, J. B. and Marth, J. D. (2003). AGenetic Approach to Mammalian Gly-can Function. Annu. Rev. Biochem. 72(epub ahead of print).

Martin, G. R. (1981). Isolation of apluripotent cell line from early mouseembryos cultured in medium condi-tioned by teratocarcinoma stem cells.Proc. Natl. Acad. Sci. USA. 78 (12),7634-7638.

Nichols, J., Evans, E. P. and Smith, A. G.(1990). Establishment of germ-linecompetent embryonic stem (ES)cells using differentiation inhibitingactivity. Development 110,1341-1348.

NIH stem cell report (2001). Internet:www.nih.gov/news/stemcell/scireport.htm

OECD (1982). OECD Guidelines for thetesting of chemicals. OECD Paris.

48

OECD (1996). Final Report of theOECD Workshop on Harmonization ofValidation and Acceptance Criteria forAlternative Toxicological Test Meth-ods. OECD Paris, July 1996; 60 pp.

Parke, D. V. (1994). Clinical pharma-cokinetics in drug safety evaluation.ATLA 22, 207-209.

Pera, M. F., Reubinoff, B. E. and Troun-son, A. (2000). Human embryonicstem cells. Journal of Cell Science113,5-10.

Richards, M., Fong, C-Y., Chan, W-K. etal. (2002). Human feeders supportprolonged undifferentiated growth ofhuman inner cell masses and embry-onic stem cells. Nature Biotech. 20 (9),933-936.

Robertson, E. (Hrsg.) (1987). Embryo-derived stem cell lines. In E. Robertson(ed), Teratocarcinomas and Embryon-ic Stem Cells: A Practical Approach(71-112). Oxford: IRL Press.

Rosen, J. D. (2002). Acrylamide in food:Is it a real threat to public health? Aposition paper of the American councilon Science and Health (acsh).http://www.acsh.org.

Scholer, H. R., Balling, R., Hatzopoulos,A. K. et al. (1989). A family of oc-tamer-specific proteins present duringmouse embryogenesis: evidence forgermline-specific expression of an Octfactor. EMBO 1. 8, 2543-2550.

Schrattenholz, A. (Hrsg.) (2001).Methoden der Proteomforschung,Molekulare Analyse der Proteinex-pression. Heidelberg, Berlin: Spek-trum Akademischer Verlag.

Schrattenholz, A. and Cahill, M. A.(2003). Potential of ComprehensiveToxico-Proteomics: Quantitative andDifferential Mining of FunctionalProteomes from Native Samples. AT-LA (in press).

Smith, A. G. and Hooper, M. L. (1987).Buffalo rat liver cells produce a

diffusible activity which inhibits thedifferentiation of murine embryonalcarcinoma and embryonic stem cells.Dev. Bioi. 121 (1),1-9.

Sommer, S., Klemm, M., Biefang-Amdt,K. et al. (2003a). Differential pro-teomics analysis of neuronal stress in-duced by homocysteic acid: phospho-rylated and methylated isoforms ofchaperones as molecular signature ofhomocysteic acid -induced excitotoxi-city. ENS abstract, accepted for publi-cation in European Journal of Neurol-ogy.

Sommer, S., Klemm, M., Biefang-Amdt,K. et al. (2003b). The molecular signa-ture of homocysteic acid-induced neu-ronal stress. Submitted to publication.

Solter, D. and Damjanov, 1. (1979).Teratocarcinoma and the expression ofoncodevelopmental genes. Methods inCancer Research 18, 277-332.

Thomson, J. A., Itskovitz-Eldor, J.,Shapiro, S. S. et al. (1998). Embryonicstem cell lines derived from humanblastocysts. Science 282 (5391), 1145-1147.

Vogt, J. A., Schroer, K., HOlzer, K. et al.(2003). Protein abundance quantifica-tion in embryonic stem cells usingincomplete metabolic labelling with15N amino acids, MALDI-TOFMS,and analysis of relative isotopologueabundances of peptides. Rapid-Com-mun.-Mass-Spectrom. 17 (12), 1273-82.

Correspondence toDr. Martina KlemmProteoSys AGCarl-Zeiss-Strasse 51D-55129 MainzGermanyphone: +49-(0)-6131 50192 34fax: +49-(0)-6131 50192 11e-mail: martina.k1emm@proteosys.com

ALTEX 21, Supp!. Linz 0312004