THE NATIONAL GENOME RESEARCH NETWORKSCIENCE INSIDE

3

FOREWORD

a scientific network that is unique in the world. Scientistshere have joined together to utilize the new chances of inter-disciplinary research and to create a visible added value forsociety. The approaches pursued for this are quite diverse.They span clinical research, systematic approaches and indi-vidual Explorative Projects, whose significance will first be-come apparent in the future.

Since 2001 the German Federal Ministry for Education andResearch (BMBF) has been funding the NGFN to promote theelucidation of the genetic causes of common diseases. Sofar, this research work has been very successful. In the pastfour years the disease genes for allergies, chronic inflamma-tory bowel diseases, alcohol addiction, epilepsy and Parkin-son's disease have been discovered. The development ofDNA chips which can help detect disease-relevant gene mu-tations in kidney and breast cancer, leukemia or congenitalheart diseases has attracted international attention.

The NGFN has given disease-oriented human genome re-search a new dimension in Germany, and with this brochurewe would like to share this with you. Reports about selectedprojects will give you insight into the work of each individualresearch focus. Look over our shoulders! If we have awak-ened your interest, an enclosed CD-ROM with descriptions of all NGFN research projects will provide you with moreinformation.

We wish you fascinating reading!

"We used to think our fate was in the stars. Now we know, inlarge measure, our fate is in our genes." With his often quot-ed comment, the biologist and Nobel Prize laureate JamesWatson describes what we now know about the significanceof genes for health. Genetic factors play a crucial role in thedevelopment of most – if not all – human diseases. They cancause disease onset, increase disease risk, influence thecourse of a disease and modify the effectiveness of drugs.

Like hardly any other scientific field, human genome researchhas changed our understanding of the causes of human dis-eases and has rightly taken its place in medical research.Cutting-edge technologies and systematic research approach-es have led to a dynamic increase in knowledge. Findingsfrom human genome research have become increasingly sig-nificant for medical practice. Already today, the diagnosis,therapy and prevention of many diseases are based on in-sights gained about the molecular causes of disease. Indi-vidualized medicine seems to be within our reach.

Early on in the research to understand the genetic causes ofdisease, it became evident that one scientific discipline alonesoon comes up against its limits. Intensive interdisciplinarycooperation is necessary between genome researchers, clinicians, statisticians, engineers and other specialists, espe-cially because many diseases are elicited by a complex inter-action of various genes and by an interaction of genes withthe environment. In Germany, the National Genome ResearchNetwork (NGFN) is advancing interdisciplinary cooperation in

Professor Annemarie Poustka• Spokesperson of the Project Committee

of the NGFN• Director of the Department of Molecular

Genome Analysis at the German CancerResearch Center in Heidelberg

Professor Stefan Schreiber• Spokesperson of the Project Committee

of the NGFN• Department head at the Clinic for General and

Internal Medicine and Director of the Institutefor Clinical Molecular Biology at the Universityof Kiel

FOR THE NATIONAL GENOME RESEARCH NETWORK

4

SCIENCE INSIDE – THE NATIONAL GENOME RESEARCH NETWORK

3

8

10

12

14

16

18

20

22

24

26

FOREWORD

THE NATIONAL GENOME RESEARCH NETWORK

DISEASE-ORIENTED GENOME NETWORKS

CANCERGETTING THE EDGE ON CANCER – PREDICTING RESISTANCE PROFILES Network: "Combating Cancer through Integrated Functional Genomic Research"

CANCER IN CHILDREN – GENETIC ANALYSES ALLOW PERSONALIZED THERAPY Network: "Systems Biology of Embryonal Tumors – Neuroblastoma as a Model"

WHICH GENES CONTROL BRAIN TUMORS? Network: "Identification of Novel Diagnostic and Therapeutic Targets in Cranial Malignancies by Integrated Tumor Profiling, Brain Tumor Network (BTN)"

CARDIOVASCULAR DISEASESIMPACT OF GENES ON THE MOTOR OF OUR LIVESNetwork: "Genetic and Molecular Mechanisms of Common Cardiovascular Disorders: From Genes to Patients"

DISEASES OF THE NERVOUS SYSTEMRESEARCHERS AS CHANNEL WORKERS – SEARCHING FOR THE GENETIC CAUSES OF EPILEPSYNetwork: "Systematic Gene Identification and Functional Analyses in Common CNS Disorders"

OBESITY – ARE GENES TO BLAME?Network: "Obesity and Related Disorders"

STROKE – HELPING THE BRAIN TO REPAIR ITSELFNetwork: "Genomic Mechanisms of Functional Recovery from Stroke"

INFECTION AND INFLAMMATIONINFECTION AND INFLAMMATION – WHEN THE BODY CATCHES FIRENetwork: "Infection and Inflammation: From Pathogen-induced Signatures to Therapeutic Target Genes"

DISEASES DUE TO ENVIRONMENTAL FACTORSDANGEROUS INTERACTION OF GENES AND LIFESTYLENetwork: "Diseases due to Environmental Factors"

CONTENTS

5

CONTENTS

28

30

32

34

36

38

40

42

44

46

48

50

52

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

SEQUENCES IN ALL VARIATIONSSMP DNA

THE INTERPRETATION OF GENETIC INFORMATIONSMP Epigenetics

MOLECULAR SIGNATURES – CANCER LEAVES TRACES IN GENE ACTIVITYSMP RNA

OLD MOLECULE BACK IN FASHION – NGFN SCIENTISTS OPTIMIZE RNAi TECHNOLOGYSMP RNAi

FROM MASS TO CLASS – RESEARCHERS TRACK DOWN DISEASE-RELEVANT PROTEINSSMP Cell

PROTEINS – NETWORKERS OF THE CELLSMP Protein

OUR BRAIN – A NETWORK OF PROTEINSSMP Proteomics

A FACTORY FOR ANTIBODIESSMP Antibody Factory

A CLINIC JUST FOR MICESMP Mammalian Models

MOUSE GENES CAUGHT IN THE TRAPSMP Mammalian Models

BIOINFORMATICS – WHAT THE DATA REALLY MEANSMP Bioinformatics

ON THE WAY TOWARDS A GENETIC EPIDEMIOLOGYSMP Genetic Epidemiological Methods

RZPD IN BERLIN – 35 MILLION DEEP-FROZEN CLONESSMP Service and Resources

6

SCIENCE INSIDE – THE NATIONAL GENOME RESEARCH NETWORK

54

55

56

57

58

59

60

61

62

63

64

EXPLORATIVE PROJECTS (EP)

"SLEEPING BEAUTY" FINDS ITS WAY INTO GENOME RESEARCHEP "Determining the Genes Underlying the Williams-Beuren Syndrome byGenerating Allelic Series of Mutations Using Novel Transposon Approaches"

STEM CELLS – MUCH DEBATED, LITTLE UNDERSTOODEP "Cellular and Molecular Signatures of Human Pluripotent Stem Cells"

GASTROINTESTINAL DISEASES – WHAT ROLE DOES A SEROTONINRECEPTOR PLAY?EP "The Role of Serotonin Receptor Genes in the Etiology of Gastrointestinal Disorders"

SO TEST THEREFORE, WHO BIND FOREVER …EP "Near Infrared Nanocrystal-based Quantitative Protein Arrays"

THE ARCHITECTURE OF GENESEP "Exploration of the 3D Genome Organization and Its Impact on Gene Expression during Tumor Development and Aging with Novel High-Throughput Multicolor Imaging Technique"

FIGHTING HEART DISEASES WITH COMBINED STRENGTHSEP "Microarray Validation of Cardiovascular Risk Factors"

METHUSELAH'S HEIRS – WHERE IS THE GENETIC SECRET?EP "Genetic Etiology of Human Longevity"

HEALTHY DESPITE HIVEP "Genomic Variability of Host Factors in the AIDS Macaque Model – Role in Resistance to and Disease Course of Viral Infections in Humans and Non Human Primates"

ALTERNATIVE mRNA SPLICING – MORE INFORMATION FROM FEW GENESEP "Analyzing Global Regulators of Alternative Splicing in the Human System:A Combined RNAi and Microarray Approach"

CANCER VACCINATION – ARE WE MAKING PROGRESS?EP "A Novel Link between Genomics and Functional Proteomics with anImpact on Cancer Immunotherapy of the Future"

DEPRESSION – A WIDESPREAD ILLNESSEP "Identification of Autoantigens as Candidate Markers for Affective Disorders"

CONTENTS

7

65

66

67

68

69

70

71

72

74

76

NEWS FROM MASS SPECTROMETRYEP "Development of Infrared Scanning Microprobe MALDI MassSpectrometry for Analytical Imaging of Tissue and Molecular Phenotyping of Hypertension ("IR-SMALDI")"

INFERENCE MODELS FOR GENE REGULATIONEP "Inferring Genetic Interactions from Gene Expression Data"

IS OUR GENOME WASTEFUL?EP "Exploring the World of Non-Messenger RNAs: RNomics Meets Proteomics"

A "HOSPITAL BUG" BECOMES PREDICTABLEEP "Whole Cell Model for the Infection by the Bacterial PathogenPseudomonas aeruginosa – Experiment and Bioinformatics"

FROM GENOME TO ORGANISM – THE WHERE AND WHEN OF GENE EXPRESSIONEP "Synthetic Antibodies for ChIP-chip in Human Cells"

INNOVATIVE APPROACHES FOR TREATING RENAL CANCEREP "Combining Proteome-based Technologies with Epigenetics Using RenalCell Carcinoma (RCC) as a Model: Identification of RCC Specific CancerAntigens for Vaccine and Antibody Therapies"

LOOKING INTO THE CENTER – A NEW TECHNOLOGY ENABLES IN VIVO STUDIES IN THE BRAINEP "Brain Endoscopy for the Functional Analysis of Neuronal Networks in vivo"

A DRAGNET OPERATION TO FIND BINDING PARTNERSEP "Peptide Libraries on a Chip"

QUALITY MANAGEMENT IN THE NGFN

DISCOVERING AND MARKETING KNOWLEDGE – TECHNOLOGY TRANSFER COORDINATION IN THE NGFN

CONTENTS

8

NATIONAL GENOME RESEARCH NETWORK (NGFN)

THE NATIONAL GENOMERESEARCH NETWORK

Diseases of theNervous System*

CardiovascularDiseases

Diseases due toEnvironmental Factors

Infection andInflammation

Cancer*

ExplorativeProjects

Systematic-MethodologicalPlatforms

* three genome networks

The research activity of the National Genome ResearchNetwork (NGFN) focuses on investigating genetic causes ofcommon diseases. To accomplish this, a unique organiza-tional structure has been created so that leading expertsfrom both systematic genome research and clinical researchcan collaborate closely. Together, they seek to understandthe complex control mechanisms of the human body at theDNA, RNA and protein level in order to find keys to new treat-ments for currently incurable diseases.

The NGFN is continuing the successful work of the GermanHuman Genome Project (DHGP), which between 1995 and2004 conducted basic genetic analyses to decode the humangenome and continued this research with functional analy-ses. Thus far, the achievements of the NGFN scientists have

been impressive: Already in the initial grant period from 2001to 2004, the scientists' research efforts led to significantresults. In 2003 the work of the NGFN was evaluated by aninternational panel of experts. Based on the findings of thisevaluation, the German Federal Ministry for Education andResearch (BMBF) has awarded a second grant to the NGFNextending through 2007.

DISEASE-ORIENTED AND SYSTEMATIC GENOME RESEARCHWorking within nine Disease-oriented Genome Networks,NGFN scientists are investigating the processes underlyingdiseases of the nervous system, cardiovascular diseases andcancer. They are also studying diseases caused by infection,inflammation or environmental factors. NGFN's patient-

9

oriented research work goes hand in hand with researchinvolving systematic genome, transcriptome and proteomeanalysis. The know-how for such large-scale research ini-tiatives is provided by twelve Systematic-MethodologicalPlatforms (SMP). Highly specialized experts work in theseSMP, utilizing and continually developing powerful tech-nologies for modern high-throughput research. In addition,they provide all NGFN scientists with efficient data pro-cessing systems. Since 2004, nineteen Explorative Projects(EP) have been created in which NGFN scientists can testand implement their innovative research ideas. The aim ofthese EP is to generate new technologies and to open upadditional fields of application for disease-related humangenome research.

ESTABLISHING STANDARDS AND APPLYING RESULTSA comprehensive quality management system enables allNGFN scientists to comply with the same quality standards,which are in accordance with international guidelines.

The QM system ensures the consistent high quality of allmaterials and results produced by the NGFN as well as theoptimal exchange of data within the network. The NGFN hasalso established a central coordination office for technologytransfer. Its task is to identify research results generated in the NGFN that might prove economically valuable and tospeedily facilitate their commercialization by making themavailable to industry. This directly benefits patients by in-creasing their opportunities to access the new therapies and diagnostic methods developed by the NGFN.

ADVISING, COORDINATING AND ORGANIZINGThe research focus and the scientific strategy of the NGFNare primarily determined by the steering board, which func-tions as an independent, external body. It guides and super-vises program implementation, thus supporting the NGFN'sdevelopment. Membership on the board is an honorary posi-tion, and the eight members are prominent representativesof the academic and industrial research communities. Inter-nally, the NGFN is managed by a project committee, which is comprised of representatives of the Systematic-Methodo-logical Platforms, the Disease-oriented Genome Networks aswell as one member representing the Explorative Projects.The members of the Project Committee monitor the course of the scientific projects and coordinate research and publicrelations activities. The Project Management provides opera-tional support to the Project Committee. It regularly reportson the current situation and the development of the entireNGFN and also organizes scientific conferences. Through itspublic relations work, the project management seeks to pro-mote public acceptance of medical genome research.

Isolated mouse embryos

Protein separation by gel electrophoresis

10

DISEASE-ORIENTED GENOME NETWORKS

CANCER

GETTING THE EDGE ON CANCER –PREDICTING RESISTANCE PROFILES

When combating cancer diseases, doctors sometimes feelthey are fighting a clever opponent and not a disease: Justwhen a medication is found that is successful, the cancercells fight back with mutations, making it ineffective. Thesame is true for chronic myeloid leukemia (CML). Since2000, Professor Justus Duyster and his team in the Network"Combating Cancer through Integrated Functional GenomicResearch" of the NGFN have been focusing specifically onpreventing these drug-resistant relapses. In CML a reciprocaltranslocation between chromosome 9 and the long arm ofchromosome 22 leads to the development of a dangerousfusion protein: tyrosine kinase Bcr-Abl. It is present in nearlyall leukemia cells. This kinase permanently activates the Ras-Raf-MEK kinase signaling pathway, thus enabling the cells toproliferate in an uncontrolled way. Presumably, CML beginsin the bone marrow in a single pluripotent stem cell, a mye-loblast, from which granulocytes arise. In patients with CML,an abnormally high number of granulocytes and immatureprecursors of these cells are present both in the bone mar-row and in the blood. In the first years the disease respondswell to treatment, in fact many patients do not exhibit anysymptoms. This so-called chronic phase can last for up to sixyears. But over time the leukemia becomes more aggressive.In the subsequent advanced stages, a transitional phase withan accelerated growth of the leukemic clone is distinct fromthe final blast phase, in which an excessive number of imma-ture blood cells (blasts) are present in the bone marrow andin the blood. The blast phase ends fatally for all patients afterthree to six months.

CLEAR POINT OF ATTACKAt the end of the 1980s scientists already had a molecularunderstanding of CML, and they envisioned a very promisingnew therapy approach: inhibiting tyrosine kinase Bcr-Abl. In2001 the kinase inhibitor Imatinib Mesylate finally came onthe market. It targets the ATP-binding region of the enzyme.Thus, ATP cannot bind and Bcr-Abl is inactive. The cancercells no longer divide and finally die due to apoptosis. Theobjective of the therapy, the disappearance of all cells har-boring the chromosomal translocation (cytogenetic remissi-on), can be attained with Imatinib in over 90 percent of theCML patients who are in the chronic phase. "That was a tre-mendous success, because with previous treatment methods only a very small group of patients could be helped," JustusDuyster recalls. But unfortunately there is a down side as well.In the accelerated phase of CML relapses occur already aftersix to nine months. The cancer cells become resistant toImatinib and proliferate as they did before. This has spurredseveral research teams to begin searching for the molecularcauses of this resistance. In addition to amplifications of theBcr-Abl gene, overexpression of the Bcr-Abl kinase and elevat-ed expression of the multidrug resistance protein MDR1, pointmutations were found in the kinase domain of the protein. Dr. Nikolas von Bubnoff, group leader in Justus Duyster's lab,studied seven patients who had a relapse while being treatedwith Imatinib: "We found five distinct point mutations that ei-ther affect the ATP-binding domain or the activation domain,"says Nikolas von Bubnoff. All five mutations change the struc-ture of the kinase in such a way that Imatinib can no longer

Nikolas von Bubnoff

11

bind. However, the ATP binding remains unimpaired, so thatthe kinase is active. "In more than 70 percent of all CML pa-tients with a relapse studied up till now, such point mutationswere discovered. They are thus the primary cause for theImatinib resistance," he adds, summarizing the findings of thedifferent research groups.

A GOOD ALTERNATIVEIn the face of the occurring Imatinib resistance the scientistsbegan to look for alternative kinase inhibitors. They pursuedresearch utilizing the chemical class of pyrido-pyrimidines(PD). Crystal structure analyses showed that pyrido-pyrimi-dines also bind to the ATP-binding site of the kinase, but in a different way and apparently better than Imatinib. TheDuyster research group studied 13 different pyrido-pyrimi-dines. All tested substances inhibited Bcr-Abl and did so withconsiderably lower concentrations than Imatinib. And mostimportant of all, they were also able to inactivate clinicallyrelevant Imatinib-resistant forms of the kinase. Conversely,pyrido-pyrimidines evoke in part mutations that can be inhib-ited with Imatinib. In future, the physicians therefore plan tocombine different inhibitors in the treatment of CML from thevery beginning or to administer them sequentially. Two of thealternative kinase inhibitors are already in clinical phase I andII trials. "The first results from a study with 107 patients arevery promising. Perhaps alternative kinase inhibitors will beavailable within two or three years for CML therapy," JustusDuyster hopes.

PREDICTING RESISTANCE PROFILESFor the new therapy schemes, knowing which substance triggers which mutation prior to use on patients would be of great value. Lengthy, complicated clinical studies would then become unnecessary. Based on these considerations,Duyster's team has developed a cell-based method withwhich resistance mutations to kinase inhibitors can be predicted. "Of course, for this it was crucial that the occur-ring mutations in this cell culture system are also clinicallyrelevant," Justus Duyster explains. To accomplish this, theMunich team produced a cell line that expresses Bcr-Abl and can only survive when this kinase is constitutively active.First, they studied the effect of administering Imatinib to the cells in the concentrations which CML patients receive."Initially, the cell culture dishes remained empty," says JanaSänger, technical assistant in the Duyster lab. Then after sixor seven weeks the first resistant colonies grew. Sequencingof the Bcr-Abl domain of these clones revealed mutationsthat are also present in CML patients. Currently, by means ofits cell culture system, the Munich group is specifically inves-

tigating such substances that are used in therapy studies.For instance, tests with the pyridopyrimidine PD166326 showed that cells treated with this substance were less oftenresistant than with Imatinib.

Moreover, quite different Bcr-Abl mutations appeared. Inpart, these could be suppressed quite easily by increasingthe PD166326 concentration. Three more mutations could besuppressed with Imatinib. "The advantage of the cell culturesystem is that it is possible to test in advance which muta-tions occur with which substances and which inhibitor can beused as an alternative," Nikolas von Bubnoff explains. "Thus,therapy strategies can be derived directly from our system."The Munich scientists are also applying their method to otherforms of cancer which have similar mechanisms for the for-

mation and development of resistance. He adds, "I am firmlyconvinced that targeted treatment strategies can be directlyderived from our cell culture findings, which will improve thechances of healing cancer in the future."

Referencesvon Bubnoff N et al. (2002); BCR-ABL gene mutations in rela-tion to clinical resistance of Philadelphia-chromosome-positiveleukemia to STI571: a prospective study. Lancet 359: 487–491

von Bubnoff N et al. (2003); Inhibition of Wild-Type and MutantBcr-Abl by Pyrido-Pyrimidine-Type Small Molecule KinaseInhibitors. Cancer Research 63: 6395–6404

von Bubnoff N et al. (2005); A cell-based screen for resistanceof Bcr-Abl-positive leukemia identifies the mutation pattern forPD166326, an alternative Abl kinase inhibitor. Blood 105 (4):1652–1659

COORDINATORProfessor Walter BirchmeierMax Delbrück Center forMolecular Medicine, Berlinwbirch@mdc-berlin.de

Petra Seipel (Duyster group)

12

DISEASE-ORIENTED GENOME NETWORKS

Terrible. Shocking. Unfair. When children get cancer, feelingsof sadness are overwhelming. The emotional impact of eachchild's illness is immense. We have an acute need to find anexplanation: Why do the young and innocent have to sufferso much? But perhaps cancer specifically in children holdsthe key to this still baffling disease. "When cancer developsso early in life, the suspicion is warranted that genetic fac-

tors are mainly responsible for the cells' rampant growth.Negative environmental influences or an unhealthy lifestylehave not had enough time to affect the young organism,"says Professor Manfred Schwab of the NGFN. "We believethat by investigating tumors in children, we will be able tobetter understand the genetic causes of cancer overall."Manfred Schwab and Professor Angelika Eggert of the Uni-versity Hospital Essen have gathered together a team ofscientists intent on combating neuroblastoma, the most fre-quent solid tumor in children. Of 100,000 girls and boys, be-tween one and three will suffer from this insidious malignan-cy of the sympathetic nervous system. The NGFN resear-chers are working together in the Cancer Network "SystemsBiology of Embryonal Tumors – Neuroblastoma as a Model".

SOMETIMES THE CANCER GOES AWAY BY ITSELFAll in all, neuroblastoma has a bad prognosis. Only aboutevery second child survives the disease. However, the courseof the disease varies extremely. In about ten percent of thecases spontaneous healing is observed. That means thetumor goes away by itself – just as mysteriously as it

appears. Scientists are convinced that the rather benign andthe extremely malignant tumors are genetically distinct. "Ifwe could say, based on gene analysis, whether a sick child issuffering from an aggressive neuroblastoma or whether aspontaneous remission can be anticipated, it would be easierto choose the appropriate therapy from the very beginning,"explains Manfred Schwab. "We would know which childrendefinitely need an intensive treatment despite the seriousside effects, and whom we could spare this therapy."Precedence for this approach is the MYCN gene copy num-ber, which defines a subgroup of high-risk patients and isused worldwide as a robust stratification marker in neuro-blastoma. However, a considerable number of aggressiveneuroblastomas lack amplified MYCN, indicating that otherunfavorable molecular events are involved in neuroblastomatumor progression. In addition, the elucidation of the geneticcauses of neuroblastoma should provide approaches for newtherapies.

A SMALL PIECE ON CHROMOSOME 1 IS MISSINGIn the research on neuroblastoma in the Cancer Network, anumber of genetic factors have already been identified whichallow assertions to be made about the aggressiveness of thetumor and thus about the prognosis of the young patients.One of these genetic factors is called CAMTA1, which scien-tists of the German Cancer Research Center (DKFZ) inHeidelberg discovered. They happened upon CAMTA1 whileinvestigating which genes are located on a small fragment ofthe short arm of chromosome1. In previous studies theyfound that this fragment is often missing in neuroblastoma."Thus, the obvious conclusion was that the missing fragmentmust contain genes that have something to do with thedisease," explains Dr. Kai-Oliver Henrich, who together withDr. Frank Westermann co-supervised this project. UsingcDNA microarrays and real time RT PCR, Kai-Oliver Henrichand his colleagues correlated the expression of the gene tothe different tumor characteristics. Result: if CAMTA1 is onlyexpressed to a small extent, the neuroblastoma appears tobe particularly aggressive. "We think that CAMTA1 can contrib-ute to making the prognoses more accurate than they havebeen previously," he concludes. "Next, we want to clarifywhat exact function the gene has and how it is regulated. Bythis means we will be able to understand the biology of theneuroblastoma a little bit better and hopefully develop newtherapies." Until now, it was only known that CAMTA1 as atranscription factor might participate in the regulation of thecell cycle and also that it is expressed to a diminished degreeeven in other kinds of cancer, e.g. in certain brain tumorsand in a form of skin cancer.

CANCER

CANCER IN CHILDREN – GENETIC ANALYSES ALLOW PERSONALIZED THERAPY

Frank Westermann

13

GENE EXPRESSION-BASED TUMOR CLASSIFICATIONA joint project within the Neuroblastoma Cancer Network to predict the prognosis of neuroblastoma by using geneexpression analysis is being pursued by Dr. Matthias Fischer(University of Cologne) and Dr. Frank Westermann (DKFZ).Their hypothesis: The more genetic factors that are knownthrough which neuroblastomas can be characterized, themore exactly the aggressiveness of the disease can be deter-mined by using them. In collaboration with the partners with-in the Neuroblastoma Cancer Network the scientists at theDKFZ and the University of Cologne first devised a microar-ray with about 10,000 genes which are suspected of beingsomehow related to neuroblastoma. Matthias Fischer and histeam have demonstrated the validity of this hypothesis.Based on this array, they prepared gene expression profilesof the very malignant neuroblastomas on the one hand andof relatively benign tumors on the other hand. Depending onthe aggressiveness of the tumor, the profiles showed distinctdifferences. These differences enabled the identification of144 genes which explicitly influence the prognosis of thedisease. The expression pattern of the 144 genes was theninvestigated in 83 patients with a known prognosis, who suf-fered from neuroblastomas of varying aggressiveness. Thereit was shown that the gene expression profiles mirrored theprognosis of the disease very well – better than the usualclassification systems used until now, which are oriented on the age of the patients, the tumor stage and histologicalcharacteristics. Dr. André Oberthuer, a member of MatthiasFischer's team, explains the next steps: "To ultimately clarifywhether our systemallows reliable state-ments about the fur-ther course of thedisease, we now wantto apply it prognos-tically. This meansthat we want to in-vestigate the tumor tissue right at the be-ginning of the diseaseand then see whetherour predictions cometrue. Until now wehave only had to dealwith tumors whosebehavior was alreadyknown." If these newstudies also havepositive results, gene

expression analysis could soon be routine for tumor classifi-cation. Matthias Fischer: "Then we can treat the affectedchildren in a much more targeted way."

The project of Professor Angelika Eggert's team at theUniversity of Essen is at a similar stage. Using a commercial-ly available biochip, she and her colleagues have identified39 genes which likewise seem to characterize the malignancyof the neuroblastoma. By means of gene expression theycould track the course of the disease in 68 of the afflictedchildren with an accuracy of 80 percent. "Interestingly, how-ever, there was hardly any overlap between our results andthe results of our colleagues from Cologne. This means thatonly a few genes were considered to be prognostically rele-vant by both research groups. Angelika Eggert is thereforeawaiting with great interest the findings of the planned pros-pective studies from Cologne, and she is planning to supple-ment the gene expression analyses with studies on the pro-tein level.

ReferencesSchwab M et al. (2003); Neuroblastoma: biology and molecularand chromosomal pathology. The Lancet Oncology 4: 472–480

Henrich O et al. (2006); Reduced Expression of CAMTA1Correlates with Adverse Outcome in Neuroblastoma Patients.Clin Cancer Res 12(1): 131–138

Oberthuer A et al.; Gene-expression based classification ofneuroblastoma patients is superior to current risk stratifica-tion. (publication planned)

Schramm A et al. (2005); Prediction of clinical outcome andbiological characterization of neuroblastoma by expressionprofiling. Oncogene 24: 7902–7912

COORDINATORProfessor Manfred Schwab German Cancer Research Center,Heidelbergm.schwab@dkfz.de

Malgorzala Sawinska (Schwab group)

Steffen Bannert (Schwab group)

14

DISEASE-ORIENTED GENOME NETWORKS

On average, a patient has less than twelve months left to livewhen he is diagnosed with glioblastoma – despite intensiveradiation – and chemotherapy. Glioblastomas are the mostcommon and malignant form of brain tumors with one of thelowest survival rates of all types of cancer. "At present, wedo not yet have any curative therapy for glioblastoma," saysProfessor Otmar D. Wiestler. He coordinates the Disease-oriented Genome Network Brain Tumor Network (BTN) in the

NGFN. "That is why we are focusing on glioblastomas in ourresearch network." Glioblastomas originate like several otherbrain tumors from so-called glial cells. These support thenerve cells and ensure that an optimal milieu is maintainedfor them.

MISSING PIECES OF CHROMOSOMEIn the Brain Tumor Network the research groups of the uni-versity hospitals of Berlin, Bonn, Düsseldorf, and Tübingenhave joined together with colleagues of the German CancerResearch Center (DKFZ) in Heidelberg. The BTN wants to elu-cidate what happens on the molecular level when braintumors develop. The aim is to more accurately prognosticatetumors, thus enabling the development of new therapeuticapproaches. For glioblastoma the scientists have alreadydetected a number of chromosomal and genetic changes – inparticular, potential tumor suppressor genes – in the mutatedcells. Such genes regulate the cell cycle and prevent thedevelopment of cells that proliferate in an uncontrolled fash-ion. If tumor suppressor genes go missing or if they are de-fective, tumors can easily develop. A conspicuous observa-

tion led researchers to conclude that tumor suppressorgenes are possibly lacking in glioblastoma. The same seg-ments of chromosome 10 and/or chromosome 22 werealways absent in the tumor cells of most of the patients.

WHICH GENE PRODUCTS ARE AFFECTED?BTN researchers led by Professor Guido Reifenberger and Dr. Marietta Wolter of the University of Düsseldorf have iden-

tified four genes which are situated in the region in question of chromosome 10 and which are expressed to a lesserdegree in glioblastoma. Professor Andreas von Deimling andDr. Christian Hartmann of the Berlin Charité have similarlysucceeded with astrocytoma, a frequent precursor tumor ofglioblastoma. Both of them made their discoveries on themissing piece of chromosome 22. The potential tumor sup-pressor genes, which the scientists in Düsseldorf and Berlindiscovered, are now being investigated more closely at theDKFZ in Heidelberg. "We want to elucidate the disease-spe-cific regulation processes and the function of these genes,"says Dr. Meinhard Hahn, project leader at the DKFZ. TheHeidelberg BTN research groups are specialized in moleculargenetic high-throughput analyses. Among other achieve-ments, Meinhard Hahn and his colleagues have devised aspecial microarray with which they can create comprehensiveRNA expression profiles. The research group of his colleagueDr. Bernhard Radlwimmer is analyzing mutations of the ge-nomic DNA using array-based comparative genomic hybrid-ization (CGH). "It is especially interesting to correlate the

CANCER

WHICH GENES CONTROLBRAIN TUMORS?

Björn Tews, Meinhard Hahn and Sebastian Barbus

15

changes on the level of DNA and RNA with each other,"Meinhard Hahn says. "To accomplish this, we are developingspecial data analysis methods and computer programs." Thebasis of the genetic analyses is the so-called Glioma CoreCollection. This is a collection of 100 very well documentedtumor samples representing the different stages and formsof glioma. The tissue database was built up in a collaborationof the Düsseldorf and Berlin working groups.

NEW PROGNOSTIC MARKER FOR BRAIN TUMORS IN CHILDRENAnother tumor the BTN is studying is medulloblastoma. It isone of the most common malignant brain tumors in children.For about half of the patients this tumor is fatal. The BTNresearch group of Bernhard Radlwimmer and Frank Mendrzykat the DKFZ has now identified a gene that could possiblyenable prognoses for the afflicted girls and boys. Accordingto their research findings, an overproduction of the proteinCDK6 seems to accompany an unfavorable course of thedisease. The gene for CDK6 is situated on chromosome 7and apparently has a key function for the growth and differ-entiation of medulloblastoma cells. "CDK6 seems to allowmore accurate prognoses for patients with medulloblastomasthan already known markers," says Bernhard Radlwimmer."The cause of the overproduction of CDK6 is that the respon-sible gene occurs too frequently in the tumor cells and there-

fore is expressed more abundantly." The basis for this discov-ery was a tissue database with samples of 200 medulloblas-tomas. They originate from patients whose medical history isknown and well documented. "In the future, by means ofCDK6, the prognosis of the patients can be assessed morereliably and the therapy adapted accordingly," he hopes."Moreover, the protein is well suited for immunological detec-tion in routine diagnostics."

ReferencesTews B et al. (2006); Identification of novel oligodendroglioma-associated candidate tumor suppressor genes in 1p36 and19q13 using microarray-based expression profiling. Int JCancer (in press)

Mendrzyk F et al. (2005); Genomic and Protein ExpressionProfiling Identifies CDK6 as Novel Independent PrognosticMarker in Medulloblastoma. J Clin Oncol 23 (34): 8853–8862

Mendrzyk F et al. (2006); The Isochromosome Breakpoints on17p in Medulloblastoma are Flanked by Different Classes ofDNA Sequence Repeats. Genes Chromosomes Cancer (in press)

COORDINATORProfessor Otmar D. WiestlerGerman Cancer Research Center,Heidelbergo.wiestler@dkfz.de

Ute Schmidt (Lichter group)

16

DISEASE-ORIENTED GENOME NETWORKS

Today it is common knowledge that heart attack patientsrequire immediate medical attention and thorough aftercare.But what about their relatives? We should get into the habitof scrutinizing not only the patient, but also the patient'sfamily," says Dr. Christian Hengstenberg, cardiologist at theUniversity Clinic of Regensburg. He and his colleagues havebeen researching the genetic basis of heart diseases for anumber of years – and with success. Only recently, the scien-tists of the Disease-oriented Genome Network CardiovascularDiseases were able to show that genetic factors determinenot only the onset of coronary heart disease (CHD), but thecourse and severity of the disease as well. To a great extent,the presence of calcium deposits and the distribution of vas-cular stenoses are determined by genes. In particular, steno-ses of the aorta and the proximal segments of the coronaryarteries are hereditary. "If a patient with coronary heartdisease shows such an affliction pattern, we should performa medical check-up on the patient's relatives as to their riskof developing CHD. Simple screening tests are sufficient to enable us to react soon enough," Christian Hengstenbergexplains. And that can save lives, because stenoses of themain coronary vessels and the proximal segments of thecoronary arteries are especially dangerous. They often leadto heart rhythm disturbances and myocardial infarction – complications that are fatal for many afflicted people.

FAMILY-BASED PREVENTIONUnder the direction of Professor Heribert Schunkert from theUniversity Clinic of Regensburg, Christian Hengstenberg andhis colleagues studied the cardiac catheter films of 882 sib-lings from 401 families. The selection criteria for the familieswere that the index patient had suffered a heart attack priorto his or her 60th birthday and that at least one sibling suf-fered from a severe coronary heart disease. "As far as we

CARDIOVASCULAR DISEASES

IMPACT OF GENES ON THEMOTOR OF OUR LIVES

Calsarcin-deficient mouse heart (right), displaying dilated cardiomyopathy inresponse to aortic banding. A wild-type heart (left) subjected to banding servedas control.

currently know, our project is the first to distinguish betweenthe individual forms of coronary heart disease and to assignthe significance of hereditary factors to the respective disease picture (clinical phenotype)," Heribert Schunkertsays. The results are especially important so that preventivecare can be initiated for relatives of heart attack patientswho have no symptoms yet. Christian Hengstenberg refers to this as "family-based prevention". Next, the scientists wantto identify the genes responsible for the different forms ofcoronary heart disease.

ONGOING HEART RESEARCHThe project of Schunkert, Hengstenberg and colleagues isonly one of many in the genome network CardiovascularDiseases of the NGFN. Since 2001 investigations have beenongoing there to determine the genetic causes of coronaryheart disease and heart attacks. Other groups are success-fully researching hypertension, cardiac insufficiency and car-diomyopathy. For instance, researchers on the team of theHeidelberg scientist Dr. Wolfgang Rottbauer discovered oneof the genetic causes for cardiac insufficiency. Their study onzebrafish showed that mutations in the gene dead-beat re-duced the pumping capacity of the heart. Dr. Norbert Freyand his colleagues at the Medical University Clinic Heidelbergmoved a decisive step forward in the field of cardiomyopathy:They found a genetic mechanism which could play a decisiverole in the onset and development of cardiac insufficiency.For this achievement they were awarded the Franz MaximilianGroedel Research Prize of the German Cardiological Society.In the animal model the researchers detected calsarcin-1, aprotein that is primarily present in the heart muscle cells andis coded by the gene Myoz2.

Preparation for DNA gel electrophoresis

17

ADDITIONAL STRESS TRIGGERS DISEASETo see what function calsarcin-1 has, the scientists purpose-ly switched it off in mice. "Our mice mutants showed no disease symptoms whatsoever and had a normal life expec-tancy up to the point in time we exposed them to an additio-nal stress factor," Norbert Frey explains. Then, for instance,the increase in blood pressure had fatal consequences: Theanimals developed a pathological enlargement of the heart(dilated cardiomyopathy). Calsarcin-1 does not seem to playany role for the pre- and post-natal development of the heart. But for mice without calsarcin-1 a specific gene program wasactive from birth on which could be responsible for the factthat the animals reacted to an increase in blood pressurewith an exaggerated, pathological enlargement of the heart.

The results indicate that calsarcin-1 plays a key role in thefunction and modulation of heart muscle cells. As responseto biomechanical stress, e.g. elevation in blood pressure, itseems to trigger a specific biological reaction that leads toheart enlargement. The enzyme is accordingly an importantcomponent of the molecular machinery of the heart musclecells. It enables the heart to adjust to increased strain.

HOPE FOR NEW THERAPY OPTIONS"Calsarcin-1 is also present in human hearts. We thereforesurmise that it plays an important role in human heartmuscle diseases," Norbert Frey says. He and his colleaguesare now investigating whether a gene mutation exists in pa-tients with heart muscle diseases which has a functionlesscalsarcin-1 as consequence. He adds, "Possibly calsarcin-1could even be the starting point for a new treatment optionto improve the adjustment of the heart to high blood pres-sure."

ReferencesFischer M et al. (2005); Distinct heritable patterns of angio-graphic coronary artery disease in families with myocardial infarction. Circulation 111: 855–862

Frey N et al. (2004); Mice lacking calsarcin-1 are sensitized tocalcineurin signaling and show accelerated cardiomyopathy inresponse to pathological biomechanical stress. Nature Medicine10: 1336–1343

Rottbauer W et al. (2005); VEGF-PLCg1 pathway controls car-diac contractility in the embryonic heart. Genes & Development 19: 1624–1634

COORDINATORProfessor Hugo KatusUniversity of Heidelberghugo.katus@med.uni-heidelberg.de

Norbert Frey

18

Hardly any other disease has been given as many differentnames in the course of history as epilepsy. "Holy disease","lunacy", "demonic disease" or even the "scourge of Christ"are only a few examples. Today we know much more aboutthe actual causes of epilepsy. Nevertheless, we can stilldescribe it as a disease with a thousand names. In fact, theterm epilepsy covers a great number of different diseaseforms. What they all have in common is the repeated occur-rence of seizures. In the Network "Systematic Gene Identi-fication and Functional Analyses in Common CNS Disorders"of the NGFN several groups are studying epilepsy. In theirsearch for the genetic causes of this seizure affliction theyhave already achieved some striking successes, as for exam-ple in Bonn, where the research group of Dr. Armin Heils isconcerned with idiopathic generalized epilepsies (IGE).

MALFUNCTIONING CHLORIDE CHANNELSIdiopathic epilepsies comprise about 40 percent of all epilep-sies. In this form of the disease no clear cause in the braincan be found that triggers the seizures. Although ten geneshave already been found that cause very rare forms of idio-pathic epilepsies, the molecular background for the frequentsubforms of IGE has not yet been elucidated. In a study in-vestigating the complete genome of 130 families in which acase of IGE occurred, a segment on the chromosome 3q26attracted the researchers' attention. It seemed to contain

sequences which predispose IGE. On the corresponding chro-mosome segment the CLCN2 gene is also situated, whichcodes the voltage-dependent chloride ion channel ClC-2. Thischannel is very widespread in the brain, especially in the neu-rons, which are inhibited by the neurotransmitter GammaAmino Butyric Acid (GABA). In the search for relevant muta-tions, Dr. Heils' team investigated the CLCN2 gene in 47 families with IGE. The scientists analyzed the 24 exons of theCLCN2 gene and the neighboring splicing regions. "In threedifferent families we found three different mutations whichwere inherited together with the epilepsy disease. The muta-tions either lead to an exchange of amino acids, a too pre-mature stop codon or to atypical splicing," Armin Heilsexplains. To check whether the mutations in fact impair thefunction of the chloride channel, expression experimentswere performed at the universities of Ulm and Aachen in theresearch groups of Assistant Professor Holger Lerche andProfessor Christoph Fahlke. They discovered that the channelbecomes functionless due to two of the mutations. In con-trast, the third mutation influences the voltage-dependentopening behavior of the channel. All three mutations proba-bly lead to the fact that GABA can no longer effectively in-hibit the affected neurons and/or that the nerve cells can bemore easily stimulated. "Only recently, three other researchgroups found mutations in the GABA receptor in several fami-lies with idiopathic epilepsies. The mutations in the CLCN2

DISEASES OF THE NERVOUS SYSTEM

DISEASE-ORIENTED GENOME NETWORKS

RESEARCHERS AS CHANNEL WORKERS –SEARCHING FOR THE GENETIC CAUSES OF EPILEPSY

Armin Heils

19

gene which we identified are further evidence that the inhibi-tion transmitted by the GABA plays a critical role in epilepsy,"Dr. Heils explains.

MOUSE MODEL FOR EPILEPSYThe group of Dr. Dirk Isbrandt at the Hamburg Center forMolecular Neurobiology also belongs to this NeuroNet of theNGFN. Its research interest is a certain group of voltage-dependent potassium channels in the nerve cells of thebrain, the M-channels. These channels mediate a potassiumcurrent which regulates neuronal stimulation and responsepatterns. Mutations in the genes coding for subunits of thispotassium channel, KCNQ2 and KCNQ3, are the cause ofbenign familial neonatal convulsions in humans (BFNC). Thiscomplicated term denotes a very rare form of epilepsy, whichappears shortly after birth. Usually the seizures stop on theirown after weeks or months. The Hamburg scientists wantedto investigate in more detail the function of the M-channel in

the brain and its rolein the developmentof epilepsy. ChristianPeters, a doctoralstudent in Isbrandt'slaboratory, thereforeproduced mousemutants which carrya dominant negativemutation in theKCNQ2 gene. Themutation changesthe pore of thepotassium channel."Since a knockout inthe KCNQ2 subunitis lethal, we havebred mice that inaddition to the wild-

type genes also carry the mutated gene," Christian Petersexplains. "We wanted to impair the function of the M-channelby assembling the mutated subunits together with the wild-type subunits." Measurements of the M-current showed that this strategy in fact functioned: The amplitude of the M-current was significantly lower in mice with the mutatedpotassium channel subunit.

HYPERACTIVE AND FORGETFULThe scientists used an inducible system to be able to controlthe timing of the expression of the mutated gene. "We usedthe Tet-Off system. As long as the animals received doxycy-

cline, a derivative of tetracycline, in their drinking water, themutated gene was not expressed." Mice that expressed theKCNQ2 mutant from birth on suffered as adults under fre-quent epileptic seizures. Moreover, they were extremelyhyperactive. In brain slices of these mutants the Hamburgscientists could detect pathological changes in the hippo-campus. As comparison, Christian Peters bred mice that alsocarried the mutated transgene, but in these mice the expres-sion was suppressed during the first weeks. These animalsclearly exhibited fewer and less serious seizures, had mor-phologically normal brains and were not unduly active. "M-currents seem to play an important role in brain develop-ment in the first few weeks after birth," the scientist con-cludes. But M-currents are also important at an adult age:"When we induced the expression of the mutated KCNQ2subunits first in adult mice, so that their brain was able todevelop normally, we detected an overexcitability of the hip-pocampal neurons in these animals, too, along with reducedmemory performance." With their work the Hamburg teamproved directly for the first time that M-currents play a cru-cial role in the stimulation of cellular and neuronal networks,the development of the brain and in cognitive processes.Alongside basic insights, they expect from the mouse modelthat it will also facilitate the search for new therapies for epi-lepsy. Dirk Isbrandt says in summary: "Until now there haveonly been a few animal models suitable for researching themolecular causes of epilepsy. We hope that our mouse modelwill be a useful tool in the development of novel antiepilepticcompounds and therapies."

ReferencesHaug K et al. (2003); Mutations in CLCN2 encoding a voltage-gated chloride channel are associated with idiopathic general-ized epilepsies. Nat. Genet. 33: 527–532

Peters HC et al. (2005); Conditional transgenic suppression ofM channels in mouse brain reveals functions in neuronal ex-citability, resonance and behavior. Nat. Neurosci. 6: 51–60

COORDINATORProfessor Peter ProppingUniversity of Bonnpropping@uni-bonn.de

Joana Cobilanschi, Susanne Beyer and Armin Heils

Birgit Rau (Heils group)

20

Slender, physically fit people smile from the billboards; themagazines advocate diets, healthy nutrition and more exer-cise. Who doesn't obsess about a few unwanted pounds? But these "problem pounds" seem ridiculously lightweightwhen compared to people who are obese. Obesity is definedas having a body mass index (BMI) of 30 kg/m2 or higher. Individuals normally become obese if they have a specific genetic makeup that predisposes them for this condition. In the second funding phase of the NGFN, the network "Obesity and Related Disorders" has received a grant of more thanfour million euros. Within this network, the research group of Professor Johannes Hebebrand and Dr. Anke Hinney and17 other groups throughout Germany are studying the gene-tic causes of obesity. Their aim is to identify genes and genevariants relevant to the disorder to enable subsequent clini-cal, epidemiological and functional studies. Furthermore, the scientists want to find out if genes involved in obesity are also implicated in comorbidities such as type 2 diabetes,stroke or hypertension.

MUTATION WITH "WEIGHTY" CONSEQUENCESMutations in the gene encoding the melanocortin-4 receptor(MC4R) have been shown to contribute to obesity. The MC4Ris expressed in the hypothalamus and regulates energyhomeostasis and body weight. Both appetite-stimulating(orexigenic) and hunger-suppressing (anorexigenic) factorsbind to the MC4R. "Normally, appetite-suppressing factorsare likely to predominate at the MC4 receptor," saysJohannes Hebebrand, coordinator of the NGFN network."Knock-out mice lacking this receptor eat more and exerciseless than their wild-type counterparts." His own research iscurrently focused on the MC4R gene. At present more than70 gene variants of the receptor are known. Most of theseimpair receptor function or even cause a total loss of func-tion. Johannes Hebebrand and colleagues analyzed the actualquantitative effect of these MC4R mutations on BMI. "This

impact can easilybe overestimated,since in many stud-ies only extremelyoverweight personswere analyzed tofind out whetherthey carry an MC4Rmutation or not," heexplained. Finally, itis also known thatan individual's bodyweight is influenced

by several factors, both genetic and environmental. To studythe relevance of MC4R mutations for body weight, JohannesHebebrand and his colleagues analyzed extended family trees(pedigrees) of 22 individuals who harbor functionally relevantmutations. The families were ascertained out of a largerstudy group of 808 extremely obese children and adoles-cents. Within the 22 families the researchers determined thegenotype and phenotype of 207 persons including 26 obesepatients, 43 parents, 22 siblings and 116 more distant rela-tives. "Our results confirm that mutations in MC4R pose ahigh risk for people to become extremely overweight. Indivi-duals who carry functionally relevant mutations had a signi-ficantly higher BMI than their relatives who had no muta-tions," explains the mathematician Astrid Dempfle. Interest-ingly, this effect was twice as high in female mutation car-riers compared to men. This means that a man 180 cm tallwith a functionally relevant MC4R mutation weighs approxi-mately 13 kg more than a relative of the same sex who doesnot carry a mutation; a woman 170 cm tall weighs even 27 kg more than a non-carrier female relative. The scientistsnoted, however, that even in relatives without the mutationthe incidence of being overweight was noticeably high – a strong indication that yet other genetic and social factorscontribute to the obesity in mutation carriers. "In the up-coming years more gene variants that influence body weightwill certainly be discovered," Johannes Hebebrand adds,"including many that may not have such a pronounced effectas seen for mutations in MC4R but that occur more frequent-ly in the population."

GENETIC PROTECTION AGAINST OVERWEIGHT AND OBESITYHowever, MC4R variants do not always make people obese.In a meta-analysis, statistician Frank Geller and his col-leagues were able to show that the variant V103I, which hadpreviously not been considered functionally relevant, has aprotective, i.e. "weight reducing" effect. V103I is termed apolymorphism because its incidence in the population is

DISEASES OF THE NERVOUS SYSTEMDISEASE-ORIENTED GENOME NETWORKS

OBESITY – AREGENES TO BLAME?

Jitka Andrä (Hinney group)

Anke Hinney

21

greater than one percent. Thepolymorphism affects aminoacid position 103 of the recep-tor, whereby the amino acidvaline (V) is replaced with iso-leucine (I). In pharmacologicalstudies no difference betweenthe function of the I103 variantand the V103 wild-type receptorcould be detected. Hebebrand'sgroup only noticed the effect ofthe I103 variant because in astudy involving 1,040 parentsand their obese children, of 35parents who were heterozygousfor the polymorphism only 10passed on the I103 variant to their obese children. Based onthis result, Frank Geller calculated that the I103 allele isnegatively associated with obesity. Subsequently, the V103Ipolymorphism was analyzed in two large research groups inAugsburg and Essen. Frank Geller merged these data with allpublished data on this V103I polymorphism in obese andlean individuals. Altogether, more than 7,500 individuals wereanalyzed. Only due to these large numbers could it be shownthat carriers of the I103 variant have a reduced risk to be-come obese. Frank Geller calculated that an adult man withthe I103 variant weighs approximately 1.5 kilograms lessthan someone who does not harbor this variant. Dr. Iris MariaHeid and her colleagues confirmed this finding in a cross-sectional study of almost 8,000 individuals from the Augs-burg region. They likewise showed that the rare I103 allele isnegatively associated with overweight and obesity.

CATASTROPHIC ENVIRONMENTAL CONDITIONSIn 1997 the World Health Organization declared obesity anepidemic. More than 50 percent of the German population isconsidered to be overweight (BMI ≥ 25 kg/m2). Especiallyalarming is the increasing number of obese children and ado-lescents. Environmental factors such as changes in lifestyleare responsible for this increase. Most humans have genevariants which predispose them to develop obesity. Duringevolution these variants were beneficial. They first became aproblem in modern industrialized countries with their neverending supply of easily available, highly palatable and high-caloric food. Also, lack of exercise contributes to this devel-opment: cars, elevators and escalators are convenient, andchildren spend their time in front of the TV or at the com-

puter instead of outside. Even in our working environmentsedentary activity predominates. "For people with a geneticpredisposition for obesity these environmental conditions aredisastrous," Johannes Hebebrand says. Researchers like himhave for a long time recognized that the frequent failures inlosing weight cannot be traced back solely to a lack of will-power or motivation. But this is highly controversial in light ofthe general perception that fat people could become thinnerif they would only eat less. Obese people who don't succeedin maintaining the weight that they have lost with a hugeeffort often feel guilty and withdraw socially. "Obesity mustbe recognized as a disease and destigmatized," he demands."In this context, the insight into obesity we have gained atthe molecular level will make a significant contribution."

ReferencesDempfle A et al. (2004); Large quantitative effect of melano-cortin-4 receptor gene mutations on body mass index. J. Med. Genet. 41: 795–800

Geller F et al. (2004); Melanocortin-4 receptor gene variantI103 is negatively associated with obesity. A, J. Hum. Genet. 74: 572–581

Hebebrand J. et al. (2003); Perspectives: molecular geneticresearch in human obesity. Obes Rev. 4: 139–146

Heid IM et al. (2005); Association of the 103I MC4R allele withdecreased body mass in 7937 participants of two populationbased surveys. J Med Genet. 42(4): e21

COORDINATORProfessor Johannes HebebrandUniversity of Duisburg-Essenjohannes.hebebrand@uni-duisburg-essen.de

Jophia Carri and Anke Hinney

22

A stroke comes like a bolt from the blue. Suddenly one halfof your body is paralyzed, you can no longer speak or onlysee half of what you would normally see. What causes this is a circulatory disturbance in the brain, either because anartery is stopped up or because there is bleeding in the braintissue. The consequence is impaired circulation: Nerve cellsdie and can no longer fulfill their function. Cell death takesonly a few minutes. "However, the brain is also able to re-generate itself," says Professor Peter H. Seeburg of the MaxPlanck Institute for Medical Research (MPIMF) in Heidelberg."After a while patients can partially recover their lost capabil-ities. Due to the rewiring of the neuronal connections, thesurviving nerve cells can partially take over the function ofthe dead brain tissue. Professor Seeburg and his colleaguesin the NeuroNet "Genomic Mechanisms of Functional Re-covery from Stroke" are fascinated by these compensationmechanisms. The NGFN researchers want to support therewiring of the nerve cells and thus keep functional deficitsfollowing a stroke to a minimum. But to accomplish this theyfirst have to understand the network. How does our brain goabout repairing itself? Which genes are activated while doingthis? Which proteins are produced? Which cells step into thebreach for their dead neighbors? In the NeuroNet scientistsare trying to find out the answers to these questions.

TISSUE DAMAGE REDUCED BY 40 PERCENTA team led by Dr. Armin Schneider of Axaron Bioscience AGin Heidelberg is studying the granulocyte-colony stimulatingfactor (G-CSF). G-CSF plays an important role in the hemato-poiesis of the myeloid cell line – but apparently not onlythere. As the researchers proved, nerve cells in the brain

express both the G-CSF receptor and also the factor itself. A whole series of indications point to the fact that G-CSF sig-nificantly contributes to the elimination of the consequencesof stroke. Armin Schneider and his team triggered an artifi-cial stroke in mice by interrupting the blood flow in an impor-tant brain artery. Just two hours later the concentration of G-CSF in the brain tissue had risen by more than 100 times,and the number of receptors had also increased. Most ap-parent were these changes in the brain regions which were adjacent to the area affected by the stroke. Further experi-ments showed that G-CSF passes the intact blood-brain bar-rier. This fact gave the researchers the idea to use this factortherapeutically. They injected G-CSF into the mice intrave-nously following an induced cerebral ischemia. "The successwas impressive," says Armin Schneider, who heads the pro-ject. "In comparison to untreated mice the size of the infarc-tion area was reduced by more than 40 percent. Moreover,mice who were administered G-CSF developed fewer graveneurological deficits than the other experimental animals."Since this discovery the scientists have also figured out howG-CSF protects brain tissue. On the one hand, the substanceprevents the apoptosis of cells in the infarct area. On theother hand, it influences the neuronal stem cells of the brain.These also carry the G-CSF receptor on their surface. Underthe influence of G-CSF the stem cells develop further intonerve cells. These can then repair the damage that the strokehas incurred. Armin Schneider is therefore optimistic aboutbeing able to help afflicted patients by means of G-CSF. "It could turn out to be the ideal substance to treat strokesand neurodegenerative diseases."

DISEASES OF THE NERVOUS SYSTEM

DISEASE-ORIENTED GENOME NETWORKS

STROKE – HELPING THEBRAIN TO REPAIR ITSELF

Christel Bonnas (Priller group)

23

SUCCESS VIA GENE EXPRESSION ANALYSISResearchers at Berlin's Charité University Medical Center arealso searching for proteins that are involved in the formationof new circuits in the brain. To achieve this, the researchteam of Professor Ulrich Dirnagl and Professor Josef Priller isanalyzing which genes are more strongly expressed in braintissue after a stroke. They interrupted the blood flow of arather large area in the brain tissue of mice for two hours.

After reperfusion they performed an analysis of gene expres-sion in the brain tissue. In total, they found 83 transcriptswhich were strongly expressed during the ischemia phaseand 94 transcripts which were less strongly expressed thanin the healthy brain tissue. Most conspicuous was an in-creased concentration of the proteins metallothionein II (MT-II) and – less pronounced – its "sibling" MT-I. The rise ofMT-I and -II mRNA expression was relatively rapid and couldbe detected only two hours after the ischemia phase. In fur-ther experiments it became clear that mice in which thegenes for MT-I and MT-II had been silenced showed far great-er damage to the brain after a stroke than "normal" mice.The animals also suffered from more severe neurologicaldeficits. "Metallothioneins seem to have a protective effect inthe case of ischemic brain damage," Ulrich Dirnagl says. Bymeans of serial gene expression analysis he hopes to soonbe able to identify additional factors which could be possibletarget structures in stroke therapy.

BONE MARROW STEM CELLS BECOME NERVE CELLSScientists in the Dirnagl-Priller working group, however, arepursuing yet another approach. They are investigatingwhether bone marrow stem cells can develop into nerve cellsin the brain. If this should be the case, by means of a trans-plantation of these cells the damage to brain tissue followinga stroke could potentially be limited. The research team'sfindings so far appear to challenge current hypotheses about

the predetermination of thevarious stem cell lines. Fortheir experiments the scientistsmarked the hematopoieticstem cells from the bone marrow with green fluorescentprotein and administered thecells to experimental animalsintravenously. Ten to fifteenmonths later two kinds ofnerve cells marked with theprotein could be detected inthe brain of the animals."Despite the blood-brain barri-er, stem cells find their wayfrom the bone marrow into thebrain," Josef Priller concludes."There they apparently can for-get their actual hematopoieticdetermination and transformthemselves into nerve cells. It seems that the biological

barriers are less impervious than assumed." For stroke pa-tients these permeable barriers mean new hope.

ReferencesSchneider A et al. (2005); The hematopoietic factor G-CSF is aneuronal ligand that counteracts cell death and drives neuro-genesis. J. Clin. Invest. 115: 2083–2098

Trendelenburg G et al. (2002); Serial analysis of gene expression identifies metallothionein-II as major neuro-protectice gene in mouse focal cerebral ischemia. The Journal of Neuroscience 22 (14): 5879–5888

Priller J (2003); Grenzgänger: adult bone marrow cells populatethe brain. Histochem Cell Biol 120: 85–91

COORDINATORProfessor Peter SeeburgMax Planck Institute for MedicalResearch, Heidelbergseeburg@mpimf-heidelberg.mpg.de

Josef Priller and Daniela Bermpohl

24

The annual wave of flu, SARS, increasing resistance againstbacteria – despite immense medical progress, infectiousdiseases are not yet under control. Worldwide more than 15 million people die of infection every year. Usually the bodyresponds to disease-causing viruses and bacteria with aninflammation, a defense reaction to protect the organism.Inflammation, however, can also occur independently of aninfection, as for instance in rheumatoid arthritis. Both as-pects of inflammation – acute and chronic – are the subjectof investigation by scientists of the Disease-oriented Genome

Network Infection and Inflammation. Thus patient-orientedresearch is being carried out in the areas of sepsis andchronic-inflammatory rheumatic diseases, tuberculosis andparasite infections such as malaria as well as viral infectionsleading to hepatitis and immunodeficiency (AIDS). The objec-tive is to catalog and categorize disease-specific gene ex-pression profiles and to identify susceptibility genes in orderto facilitate the earliest diagnosis possible, to predict thecourse of the disease and to introduce an effective therapyas rapidly as this can be achieved. In addition, the networksupports investigations into the underlying mechanisms ofdisease using cell- and tissue-cultures, ex vivo organ- and ani-mal models, focusing mainly on the innate immune systemand in particular the family of Toll-like receptors (TLRs). Viathese receptors structural features of the pathogens are rec-ognized and defense reactions are triggered. In order to copewith the plethora of clinical and experimental data derivingfrom these studies, the areas of data management and bioin-formatics, microarray analysis and the bioinformatics coreunit as well as a cell-typing and sorting unit are affiliated withthe network as so-called "bridging projects" and play a cen-tral role. The main objectives within the framework of datamanagement are the linking of clinical findings with datacoming from functional genome analysis. An important focus

of research is the development of a central warehouse data-base which merges and interprets all of the data collected inthe network.

"Our network combines clinical research and basic researchinvolving the most common infectious diseases and chronicinflammatory rheumatic diseases of society. It provides uswith novel insights about the underlying disease mechanismsand helps us to transform these into new diagnosis andtherapy approaches. An essential focus of our research isthus investigating natural resistance against infectious dis-eases. This should give us information as to why not everyperson gets an equally severe case of the disease and whynot every person who has contact with a pathogen gets sick,"explains Professor Trinad Chakraborty, coordinator for thenetwork Infection and Inflammation.

SEPSIS DIAGNOSIS AS EARLY AS POSSIBLE "Severe sepsis and septic shock are the main causes ofdeath in non-cardiological intensive care units the world over.Despite the best possible intensive therapy, the mortality rateof severe sepsis has remained unacceptably high, at about40 percent, for decades," says Dr. Thilo Menges of the Uni-versity of Giessen, who is head of the subproject "ClinicalSepsis Phenotypes". His team is analyzing transcription pat-terns that are associated with the degree of severity of thedisease. To accomplish this, the scientists take blood sam-ples of the patients during their stay in the hospital to per-form microarray analyses. In their research the scientistsfocus on three large patient groups: polytraumatized pa-tients, patients with severe pneumonia of the lung and pre-mature babies who were born before the 32nd week of preg-nancy. All clinical data, findings and lab parameters of thesepatients are stored online and can be correlated with theexpression data from the microarray analyses. "In the poly-traumatized patients the gene expression analyses showclear differences between septic and non-septic patientsalready at the time of admittance into the intensive careunit," says Thilo Menges. "The gene expression profiles deter-mined in the premature babies show clear differences be-tween artificially respirated infants with congenital sepsisand infants without an infection. Our data from both of thesestudies suggest predictive diagnosis for sepsis soon afterinjury or insult. Perhaps they could enable diagnosis of afetal inflammatory response (FIRS) right at the time of birth."

JOINTS UNDER ATTACKA further focus of the genome network Infection andInflammation is on chronic inflammatory rheumatic diseases.

INFECTION AND INFLAMMATIONDISEASE-ORIENTED GENOME NETWORKS

INFECTION AND INFLAMMATION –WHEN THE BODY CATCHES FIRE

High-throughput DNA analysis

25

These include rheumatoid arthritis and spondyloarthropathiessuch as ankylosing spondylitis (AS). In these diseases theimmune system attacks the body's own tissue and destroysit, seemingly without cause. The working group of Dr. ThomasHäupl of the Charité in Berlin is investigating gene expressionpatterns in blood cells, joint liquid and synovial tissue of pa-tients with rheumatoid arthritis and AS. In cooperation withclinics, the scientists built up a database for this containingthe blood, tissue and DNA samples of 500 patients already inthe first NGFN funding period. "Over the long term we wantto describe gene expression patterns by means of whichrheumatic diseases can be diagnosed at a very early stageand the further course of the disease can be prognosticated,"Thomas Häupl says. "Up to now the diagnosis has often beenmade very late after irreversible damage has already oc-curred." He and his colleagues are therefore currently usinggene expression patterns to attempt to find the specific fac-tors that play a crucial role in the early stage of the inflam-mation.

Already in the NGFN's first funding period the Berlin re-searchers identified molecular patterns and markers in bloodmonocytes and synovial tissue which enable a diagnostic dif-ferentiation between rheumatoid arthritis and arthrosis. Inaddition, in both diseases they determined clear differencesto healthy tissue in the gene expression. Based on these find-ings, Dr. Bruno Stuhlmüller, who heads a further subprojectof the genome network at the Charité, developed a micro-array with the cDNA of 311 genes relevant to rheumatoidarthritis. By means of this array the scientists can classifythe degree of inflammation in patients with rheumatoidarthritis and predict how well a treatment would work. BrunoStuhlmüller now wants to increase the number of genes onthe chip and thus further improve the prognostic and predic-tion possibilities prior to anti-TNF (tumor necrosis factor)therapy. To identify new candidate genes, he is performingpreliminary experiments in monocytes, tissue macrophagesand T cells. The potentially significant genes detected in thecell cultures are then checked once again by comparing themwith gene expression patterns identified by the Häupl team incells of healthy blood donors and rheumatoid arthritis patients.

PREDICTING TREATMENT SUCCESSIn patients with rheumatoid arthritis, treatment with a mono-clonal antibody against the mediator TNF alpha has proved tobe very successful. TNF alpha has a central position in patho-genesis because it activates a whole cascade of additionalinflammatory cytokines. The therapeutic antibody neutralizesTNF alpha, whereby fewer inflammatory cells penetrate into

the joints. However, 20 to 40 percent of the patients do notrespond positively to therapy with TNF alpha blockers. "Wedo not know why this is so. Particularly on the molecularlevel we have no clear idea of what is actually going on here,"says Dr. Andreas Grützkau of the German Rheuma ResearchCenter in Berlin. To elucidate the molecular mechanisms oftherapy with TNF alpha blockers, he and Bruno Stuhlmüllerare studying the gene expression patterns of patients withchronic inflammatory rheumatic diseases before and afterthe beginning of treatment. For their expression analyses thescientists utilize highly purified cell populations (monocytesand CD4 positive lymphocytes), which they extract from theblood of patients. "From our earlier studies we know thathomogeneous, clearly defined cell populations are consider-ably better suited for data analysis than heterogeneous popu-lations with an unknown composition," Andreas Grützkauexplains. The methods utilized here for cell sorting werealready optimized. Soon it will be known why TNF alphablockers do not work or do not work well for some patients."Perhaps – even before starting a treatment – we will soonbe able to predict the optimal therapy for a patient, based onthe activity of his or her genes," the scientists conclude.

ReferencesBerghofer B et al. (2005); Common human Toll-like receptor 9polymorphisms and haplotypes: association with atopy andfunctional relevance. Clin Exp Allergy. 35 (9): 1147–1154

Haeupl T (2004); Perspectives and limitations of gene expres-sion profiling in rheumatology: new molecular strategies.Arthritis Res Ther. 6 (4): 140–146

Siegmund K (2005); Migration matters: regulatory T-cell com-partmentalization determines suppressive activity in vivo.Blood 106 (9): 3097–3104

COORDINATORProfessor Trinad Chakraborty University of Giessentrinad.chakraborty@mikrobio.med.uni-giessen.de

Thomas Häupl

26

DISEASE-ORIENTED GENOME NETWORKS

DISEASES DUE TO ENVIRONMENTAL FACTORS

DANGEROUS INTERACTION OF GENES AND LIFESTYLE

When hearing the term "body surface", most people think ofthe skin. However, the mucous membranes of the intestine,the respiratory passages and the lungs are also barrier sur-faces to the outer world and with 300 m2 are considerablylarger than that of the skin (1.4 m2). These are the barriersurfaces where the body has to deal with many hostile envi-

ronmental factors. In particular, the mucous membranes ofthe intestine and the respiratory passages provide an entryportal for pathogens. That is why these tissues have a num-ber of defense mechanisms for building up a barrier againstthe outside world. These include specialized proteins whichfend off invading bacteria, viruses or fungi and also "guards",like the intestinal flora. Our intestine's dense population ofbacteria prevents pathogens from being able to settle there.The mucous membranes also have a highly complex and dif-ferentiated immune system, the so-called mucous associateddefense system.

DRAGNET SEARCH FOR THE RESPONSIBLE GENESIn the genome network Diseases due to EnvironmentalFactors all research focuses on chronic inflammatory dis-eases in organs which exercise a barrier function against ourenvironment. Among these are diseases such as allergies,asthma, chronic inflammatory intestinal diseases (Crohn'sdisease, ulcerative colitis), atopic dermatitis, the lung diseasesarcoidosis, psoriasis and chronic obstructive respiratorydiseases. To be sure, they affect different organs, but they all

have a hereditary aspect to them. Moreover, they all have incommon that they can be triggered by lifestyle factors inindustrialized societies. In recent decades there has been agreat increase in the incidence of asthma and atopic dermati-tis in children as well as of chronic inflammatory intestinaldiseases – a clear indication that alongside disease genes,

environmental factors play asignificant role as well. "Inour network we are system-atically searching for dis-ease genes that togetherwith external influences cantrigger disease. We evenbelieve that the seeminglyso different environmentaldiseases are based in parton the same genetic fac-tors," says Professor StefanSchreiber, coordinator ofthe genome network.Scientists at the three loca-tions of the network – Kiel,Berlin and Munich – arecollaborating closely withthe Systematic-Methodo-logical Platforms DNA andGenetic EpidemiologicalMethods (GEMs). "Alone in

the first NGFN funding period of our network over four mil-lion single nucleotide polymorphisms (SNPs) were typed,"Stefan Schreiber says. These include substantial scientificbreakthroughs, such as the identification of the first and sec-ond disease genes for Crohn's disease or the discovery ofthe first disease gene for sarcoidosis.

SARCOIDOSIS – OFTEN NOT RECOGNIZED Sarcoidosis is an inflammatory disease. Due to the formationof inflammatory lesions, so-called granulomas, it destroys thelungs. Immunologically, there is a hyperactivity of macro-phages and CD4 T helper cells. Granulomas can occur every-where in the body and interfere with the function of the re-spective organs. Often, however, the lungs are affected. There-fore, scientists suspect that substances from the inhaled airactivate the immune system of genetically disposed persons.In Germany an estimated 30,000 people suffer from sarcoid-osis. "But most likely there is a high number of unreportedcases," says Dr. Jochen Hampe of the University of Kiel, who as young scientist together with his colleague Dr. RutaValentonyte discovered the first gene for sarcoidosis. "Due

Lena Bossen (Schreiber group)

27

COORDINATORProfessor Stefan SchreiberUniversity Hospital Schleswig-Holstein, Campus Kiels.schreiber@mucosa.de

to unspecific symptoms like coughing, joint pain and fever,the disease frequently goes unrecognized."

DISEASE GENE DISCOVEREDScientists have had indications that sarcoidosis is hereditaryfor quite some time. There are families in which several fami-ly members suffer from the disease, while in other families italmost never occurs. In 2001 Dr. Manfred Schürmann of theUniversity of Lübeck came a step closer to illuminating thegenetic causes of the disease when he studied 63 familieswith known sarcoidosis. He discovered that genetic changeson a segment of chromosome 6 play a role in the genesis ofthe disease. Together with the Institute for Human Geneticsin Lübeck, the Research Center Borstel, the Institute forMolecular Biotechnology in Jena and the Department ofPneumology of the University Hospital Freiburg, the Kielscientists have now tracked down the culprit: The diseasegene's name is butyrophilin-like 2 (BTNL2). Through an ex-change of nucleotides a premature stop codon develops in its mRNA. The protein resulting from this lacks important do-mains for anchoring in the cell membrane. "This loss couldlead to the fact that BTNL2 as co-activator stimulates the T cells too strongly and that thus an autoimmune disease istriggered," Jochen Hampe speculates.

If an allele of the BTNL2 gene exhibits this nucleotide ex-change, the risk of sarcoidosis increases by 60 percent. Ifboth alleles show this variant, the affected persons fall illwith sarcoidosis three times more often than people with"healthy" BTNL2 variants. As next step the scientists want toelucidate the exact role of BTNL2, in order to create a solid

basis for new therapy approaches for sarcoidosis. "The dis-covery of the BTNL2 gene is a breakthrough for clinical research on sarcoidosis," says Professor Joachim Müller-Quernheim, lung specialist at the University Hospital Freiburg. "Our long-term objective, to predict the course ofthe disease and the success of therapy, is thus closer tobeing realized."

NOT EVERY DISEASE HAS ITS OWN GENEAlready during the first experiments to circle in on diseasegenes for chronic barrier inflammation, the scientists noticedthat diseases which attack different organs show consider-able overlap. Now the discovery of the disease genes pro-vides certainty: The Munich working group of Dr. MichaelKabesch found out that the gene CARD15 (NOD2) and CARD4(NOD1) not only are significant for Crohn's disease, thechronic inflammation of the intestine, but also for asthma,periodontitis and joint inflammation in psoriasis. The Berlinresearchers in the working group of Professor Young-Ae Leereport a gene localization for atopic dermatitis which over-laps with a localization for psoriasis. "This will result in com-pletely new possibilities for therapy," Stefan Schreiber says."The medical implementation of these results means that wemust take a holistic view to solve the problem of 'inflamma-tion' across all organ systems."

ReferencesHampe J et al. (2002); Association of NOD2 (CARD 15) geno-type with clinical course of Crohn's disease: a cohort study.Lancet 359: 1661–1665

Stoll M et al. (2004); Genetic variation in DLG5 is associatedwith inflammatory bowel disease. Nat. Genet. 36 (5): 1–5

Valentonyte R et al. (2005); Sarcoidosis is associated with atruncating splice site mutation in BTNL2. Nat. Genet. 37 (4): 1–8Jochen Hampe

Robotics in genome research

28

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

Systematic sequencing and comparative sequence analysis –those are two main tasks of the Systematic-MethodologicalPlatform (SMP) DNA. "The sequencing data enable us to ana-lyze phenotypes, genes and promoter elements includingtheir modifications in detail," explains Professor HansLehrach, coordinator of the SMP DNA. "Moreover, we cangain insight into evolution by comparing the human genomewith other genomes, for instance with the genome of chim-panzees." Besides this, the scientists of the SMP DNA arefocusing on understanding the correlations between geno-type and phenotype, by means of which they want to identifydisease genes and elucidate disease processes.

THE LITTLE DIFFERENCEOn the DNA sequence level the variation among people isonly about 0.2 percent. However, in this seemingly so negli-gible variability lies the key to understanding complex dis-eases such as cancer, obesity or hypertension. Scientistsworldwide are therefore working on recording the geneticvariations within the human genome. Some scientists evendescribe the search for the variability of the human genomebetween individuals as the second act of the human genomeproject. To trace the differences, researchers are pursuingdifferent approaches. In genotyping, for instance, the DNAsequences are systematically scanned for single nucleotidepolymorphisms (SNPs). These deviations in single bases arethe most common differences occurring in the human ge-nome. Researchers estimate that there are about ten millionSNPs in humans. Although many SNPs do not have anyeffect, some of them increase the risk for getting a specificdisease or influence the effectiveness of a drug. Within theframework of the SMP DNA, six genotyping centers have

joined together to form the National Genotyping Platform(NGP): the GSF – National Research Center for Health andEnvironment in Munich under the direction of ProfessorThomas Meitinger, the University of Kiel (Professor StefanSchreiber), the Berlin research groups at the Max DelbrückCenter (Professor Norbert Hübner) in Berlin-Buch and at theMax Planck Institute for Molecular Genetics in Berlin (Pro-fessor Hans Lehrach, Dr. Sascha Sauer) and the universitiesof Cologne and Bonn under the coordination of ProfessorPeter Nürnberg. Together they can provide the NGFN scien-tists with the entire spectrum of presently known genotypingmethods.

TRACKING DOWN DISEASE GENES"Today, due to modern high-throughput procedures, we cananalyze 500,000 SNPs at one time," says Thomas Meitingerof the GSF in Munich. With the aid of genotyping, the scien-tists in the NGFN have already succeeded in identifying anumber of disease genes. For instance, they found sequencevariants in the gene for the protein leucine-rich repeat kinase 2(LRRK2), which are associated with different forms ofParkinson's Disease. Furthermore, they discovered thatseveral SNPs in the coding sequence for the protein FKBP5are linked with a fast response to antidepressant drugs and afrequent recurrence of depressive episodes. However, genet-ic variants that predispose for a specific disease are notevenly distributed throughout the world. On the contrary,there are differences evident depending on the population.For instance, a population genetic study in Iceland showedthat there are variants both in the gene for the 5-lipoxygen-ase activating protein (ALOX5AP) and in the gene for phos-phodiesterase 4D (PDE4D), which make people with thesegenetic variants more predisposed to suffer a stroke. A studyof the Munich genotyping platform on stroke patients fromcentral Europe showed that in them, too, variants of the genefor ALOX5AP are associated with an increased stroke risk,whereas variants of the gene for PDE4D had no significantinfluence on stroke risk.

NOTHING LEFT OUTDr. Richard Reinhardt of the Max Planck Institute forMolecular Genetics (MPIMG) in Berlin runs the resequencingplatform in the NGFN, consisting of the two groups in Berlinand Jena. "In contrast to genotyping we do not focus on indi-vidual SNPs when investigating a gene, but sequence thewhole gene, or at least all exons," Richard Reinhardt explains.The advantage over all other methods is its unsurpassedaccuracy. Resequencing captures all genetic variations – allSNPs but also deletions and insertions. The Berlin research-

SMP DNA

SEQUENCES IN ALL VARIATIONS

Marie-Laure Yaspo

29

ers already have several achievements to their credit, such asthe identification of the molecular causes of kidney diseasesand cilia defects that play a role in chronic respiratory dis-eases. The resequencing platform is continually being expand-ed and developed to reduce the great outlay of effort and thehigher costs involved in comparison to genotyping. "If wedevelop this here further, one day it will be possible to deter-mine the entire genome of a patient within a short time andwith a reasonable amount of effort, so that the costs will alsobe justifiable," says Richard Reinhardt, describing the futureof the resequencing platform.

CHIMP GENOME AS COMPARISONDr. Marie-Laure Yaspo is pursuing quite a different approachto decode the functions of the human genome. Within theframework of an international project (the SMP DNA cooper-ates with the groups from Berlin, Brauschweig and Jena), sheis working on the high-resolution sequence of the chimpan-zee genome. "Although the chimpanzee is the closest livingrelative of human beings, there are nevertheless immensedifferences: for instance with regard to linguistic or intellec-tual capabilities or the susceptibility for various diseases,"the biologist explains, who also conducts research at theMPIMG in Berlin. Within the framework of a German-Asianconsortium she was involved in the deciphering of the firstchimpanzee chromosome, the chromosome 22. "We assumethe comparative analysis of the chimpanzee genome and thehuman DNA sequence will contribute to discovering disease-relevant genes. But in particular, we will also learn a lot aboutthe evolution of the human genome," she adds. For this thescientists need an unbroken, top-quality sequence of thechimpanzee genome, because it differs from the human ge-nome by only 1.7 percent. Marie-Laure Yaspo is working onthe generation of a high-quality sequence of 42 megabases.Most of the segments she has investigated lie on the X chro-mosome. Numerous hereditary diseases are located on thischromosome, including many forms of mental disability. Thisis why decoding it is particularly significant. In addition, ge-nomic segments which have been identified as medicallyinteresting by the Disease-oriented Genome Networks in the

NGFN are being analyzed in the chimpanzee genome onceagain by Dr. Yaspo and her team.

INSIGHT INTO THE TRANSCRIPTION MACHINERYHowever, within the SMP DNA there are also groups that arenot focusing immediately on the whole genome, but rather onits specific segments. For instance, the group of Dr. MichalJanitz at the MPIMG: He and his staff are comparing the ac-tivity of promoter sequences in different cell types in order to understand the mechanisms of gene expression and reg-ulation. "We use cell arrays to be able to analyze the pro-moter regions on a large scale. Our objective is to investigate3,000 human promoter regions," Michal Janitz explains. To accomplish this task, plasmids containing a reporter genedriven by a promoter of interest are printed in an array for-mat and then transfected into a number of human cell lines.After the transfection the activities of the promoters can bemeasured by detecting the reporter gene expression level.Michal Janitz plans to evaluate his data together with thefindings from other projects, for example the promoter anal-ysis in mice, in order to trace the global regulation patternsin mammalian cells.

ReferencesBinder EB et al. (2004); Polymorphisms in FKBP5 are asso-ciated with increased recurrence of depressive episodes andrapid response to antidepressant treatment. Nat Genet. 36 (12): 1319–1325

Lohmussaar E et al. (2005); ALOX5AP gene and the PDE4Dgene in a central European population of stroke patients.Stroke 36 (4): 731–736

Watanabe H et al. (2004); DNA sequence and comparativeanalysis of chimpanzee chromosome 22. Nature 27 429 (6990): 382–388

Zimprich A et al. (2004); Mutations in LRRK2 cause auto-somal-dominant parkinsonism with pleomorphic pathology.Neuron 18 44 (4): 601–607

COORDINATORProfessor Hans LehrachMax Planck Institute forMolecular Genetics, Berlinlehrach@molgen.de

Yuhui Hu (Janitz group)

30

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

We inherit our genetic make-up from our parents in the formof the DNA sequence. But not all information regarding whichgenes are to be transcribed and which are not is passed onfrom one generation to the next. Epigenetics is concernedwith such mechanisms "outside" of the DNA sequence.Epigenetic information is acquired in the course of a lifetime.A central epigenetic mechanism is DNA methylation. Thisprocess usually occurs on a cytosine (C) which is immediate-ly followed by a nucleotide with the base guanine (G).Although the sequence CG does not occur very often in thehuman genome, there are certain regions where there ismore frequent incidence. They are defined as CpG islands. Ifsuch islands are methylated in the promoter region of a gene,it can no longer be transcribed. The CpG islands thus worklike a switch. "The DNA methylation pattern represents theepigenetic program of the genome and determines the inter-pretation of genetic information. Differences in methylationgenerate differences in the expression pattern," explains Dr. Jörg Hoheisel, coordinator of the Systematic-Methodo-logical Platform (SMP) Epigenetics of the German CancerResearch Center (DKFZ).

DECIPHERING THE METHYLATION CODECertain methylation patterns are associated with the appear-ance of different illnesses. With cancer, for example, a strongmethylation of tumor suppressor genes can frequently be

observed. "We know that changes in the DNA methylationpatterns belong to the earliest and most frequent events inthe development of cancer. Nevertheless, still very little isknown how these changes arise and exactly what role theyplay in the degeneration of a cell," says Professor Hermann-Josef Gröne, who oversees the tumor database at the DKFZin Heidelberg and who also works in the SMP Epigenetics.Furthermore, often there is not any information about howthe methylation pattern differs in healthy and diseased tis-sue. "And this knowledge could be useful for the diagnosisand prognosis of diseases or for the development of new therapies," he says. One reason for the knowledge gap is the lacking methodology: Until now the scientists have nothad any tools at their disposal for a genome-wide epigeneticanalysis. The researchers at the SMP Epigenetics intend tochange this.

MICROARRAYS FOR METHYLATION PATTERNSJörg Hoheisel explains the selected approach: "Using a trick,the methylation status of a nucleotide can be represented assingle nucleotide polymorphism (SNP). That is why we plan,analogous to the microarrays that search the genome forSNPs, to use microarrays also for a genome-wide and gene-specific analysis of the methylation pattern." The genomeresearchers then want to correlate the gained information toclinical data and the results of the transcription analyses.

SMP EPIGENETICS

THE INTERPRETATION OFGENETIC INFORMATION

Cora Mund and Verena Beier

31

Using this method, fundamental insights into the role of DNAmethylation during tumor development should be possible. Inaddition, Jörg Hoheisel and his colleagues want to identifymethylation patterns that allow diseases to be diagnosed andtheir prognosis to be assessed. Likewise, the scientists hopethat the methylation pattern will also give an indication howeffective a medication is.

COMPETENT PARTNERSThe SMP Epigenetics includes both research institutes andprivate companies. All participants can already look back onseveral years' experience in the field of epigenetics. Oneexample for the cooperation of research institutions and in-dustry is located at the University of Bonn. Here Dr. AndreasWaha has developed a microarray that contains 7,680 CpGislands of the human genome and which can be used tostudy epigenetic changes of the genomic DNA. One objectiveof his project is to extend the microarray so that all CpGislands of the human genome can be analyzed. In particular,he plans to identify all tumor-relevant CpG islands. To ac-complish this, he is supported by the Berlin company Epige-nomics, which is the leader worldwide in the commercial useof epigenetics. Researchers of the DKFZ, together with thecompany febit biotech are developing microarrays which are suitable for routine analyses of the methylation pattern. They, too, are focusing on sites that are typical for cancerdiseases. DNA methylation analysis has several characteris-tics which make it an ideal candidate for routine applications.DNA is a relatively stable molecule and thus easier to handlethan for example RNA, which is analyzed in transcription stud-ies. Moreover, the regulation mechanisms act much more slowly with reference to the methylation status than regula-tion mechanisms on the transcription level. The moment thesample is taken is therefore less critical. Another advantageis that the methylation signals "on" or "off" can be digitalizedfrom the very beginning and thus are excellently suited forevaluation at the computer. Within the framework of the SMP, scientists not only undertake a targeted search for significant patterns, but also gather fundamental informationabout the methylation status of a chromosome, quasi asstandard for later studies. This takes place in a coordinatedapproach between Professor Jörn Walter (University ofSaarbrücken), Professor Albert Jeltsch (International Uni-versity Bremen), Dr. Richard Reinhard (Max Planck Institutefor Molecular Genetics, Berlin) and Dr. Matthias Platzer(Leibniz Institute for Age Research – Fritz Lipmann Institute,Jena).

FIRST SUCCESSESJörg Hoheisel's team has already taken a first step in estab-lishing genome-wide microarrays in collaboration with theEpigenetics Division of the DKFZ headed by Dr. Frank Lyko. In the scope of a project carried out to prepare their jointwork in the Systematic-Methodological Platform, the scien-tists designed an oligonucleotide microarray with which theycan also determine parallel the methylation status of 53 cyto-sines in the promoter region of a gene. Project team member

Cora Mund is delighted: "While investigating the methylationstatus it is possible for the first time to combine a high-throughput with a high resolution." With the previousmethods, only the methylation status of merely a few cyto-sines of a gene could be determined. Next, the Heidelbergscientists want to expand their method: Currently they areworking on a chip to ascertain the methylation status of 250genes which play a role in prostate cancer. "We are con-vinced that with epigenetic analyses we can provide reliableresults that are suitable for routine applications," JörgHoheisel explains. "We expect a lot from epigenetics espe-cially for cancer research."

ReferencesMund C et al. (2005); Array-based analysis of genomic DNAmethylation patterns of the tumour suppressor gene p16 pro-moter in colon carcinoma cell lines. Nucleic Acids Res. 33: e73

Beier V et al.; Characterisation of genomic methylation pat-terns in tumours. Exocrine Pancreas Cancer (Gress TM,Neoptolemos JP, Lemoine NR & Real FX), Solvay, 254–260

COORDINATORDr. Jörg HoheiselGerman Cancer Research Center,Heidelbergj.hoheisel@dkfz.de

Cora Mund

32

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

"To combat cancer in a targeted way and with few side ef-fects, we first have to understand the complex dysfunctionsin the cells and combinations of cells which underlie this dis-ease," says Professor Annemarie Poustka, coordinator of theSystematic-Methodological Platform (SMP) RNA. In thesearch for new points of attack for targeted therapies, theanalysis of gene expression patterns has proved to be ofgreat value. For this, scientists use microarrays, which enablea fast and exact genome-wide investigation of which geneschange their activity in diseased tissue in comparison tohealthy tissue. In recent years systematic data on gene ex-pression in tumors and in healthy tissue have been gainedthis way. Mostly, however, these comparisons have not illumi-nated the role of these genes in tumor genesis. The subproj-ect "Systematic Analysis of Transcription Networks" of theSMP RNA is therefore aiming to uncover the significance ofthese genes in the development of tumors. To do this, thescientists are searching for the interaction partners of thegenes that are conspicuous in the microarray. Using RNAinterference (RNAi), they first suppress the expression ofgenes of interest in human cells. In the next step they isolatethe entire RNA out of these cells and submit in turn this RNAto a microarray analysis. Thus they can recognize whichgenes modify their activity when the expression of the initialgene is changed through RNAi. "In the ideal case we will

identify new connections to already known molecular pro-cesses in the cell," hopes Dr. Holger Sültmann, head of thesubproject at the German Cancer Research Center (DKFZ) inHeidelberg.

VALUABLE INFORMATIONIn the first funding period of the NGFN, Holger Sültmann'sworking group focused on collecting microarray data on geneexpression in tumors of the kidney, mammary glands and thebrain. "We worked very closely with clinicians. They contrib-uted tissue samples and their medical expertise to the proj-ects," says Dr. Ruprecht Kuner, post-doctoral fellow in thegroup. In fact, the scientists were able to assign specific sig-natures to certain kinds of cancer. How valuable this infor-mation can be has been shown with clear-cell renal cell carci-nomas: On the basis of the gene expression pattern the re-searchers can early recognize whether metastases will formin the patient. The physicians thus have the chance to adaptthe therapy protocol at an early stage in order to deceleratethe formation of metastases as much as possible.

VIRTUAL EXPERIMENTSIn the coming years, 150 of the differentially expressedgenes which were identified in the first funding period are toundergo a more detailed analysis, in order to identify further

SMP RNA

MOLECULAR SIGNATURES – CANCER LEAVES TRACES IN GENE ACTIVITY

Ruprecht Kuner, Armin Tresch and Silke Argo

33

tumor-associated genes. Here the focus is on genes whoseexpression shows differences either between tumor andhealthy tissue or changes between various tumor subtypes or stages. SMP RNA scientists are collaborating closely withscientists from the SMP Cell in order to characterize thegenes functionally. The work in the laboratory is thus dove-tailed with the data analysis of the microarray experimentsand the development of new analysis tools. "The computer isthe hub of the whole project; here all of the results are inte-grated so that we can simulate models of the biological net-works," says the computational scientist Dr. Achim Tresch.The in silico modeling is more than just an attempt to pro-cess and manage the flood of data. The computer-calculatedmodels of gene regulation enable the simulation of as manyinteraction variants as desired. Such experiments, which lastonly a few seconds virtually, would mean months of work inthe laboratory. "The analysis and modeling of the results bycomputer simulation will enable us to identify networks ofgene expression which facilitate the understanding of thepathological mechanisms of human diseases," HolgerSültmann explains.

POSITION DETERMINATION IN THE MOUSE EMBRYOProfessor Bernhard Herrmann of the Max Planck Institute forMolecular Genetics (MPIMG) in Berlin has taken a differentapproach to finding out how the signal cascades in our bodyare networked and which dysfunctions cause diseases todevelop. "If we want to understand the signal mechanismsthat steer the development of organs but that are also in-volved in the genesis of diseases, we have to identify thegenes that control these processes," explains the Berlinscientist. In his subproject he is therefore studying where inthe body a gene is active by using whole-mount in situ hy-bridizations (WISH) on mouse embryos. To accomplish this,whole mouse embryos are isolated, processed according tocertain techniques and incubated with the cDNA of the candi-date gene. "The cDNA only binds in regions where the geneis expressed. Only here does it find the suitable mRNA towhich it can bind according to the principle of complemen-tary base pairing," explains Manuela Scholze, who as tech-nical assistant has already performed countless numbers ofthese in situ hybridizations. Professor Herrmann's researchgroup has already investigated over 10,000 cDNAs in theframework of the German Human Genome Project (DHGP).To accomplish this, 9.5-day-old mouse embryos were hybrid-ized. "We have developed WISH into a high-throughputmethod in order to be able to study gene expression inmouse embryos on a large scale," Bernhard Herrmann says.The Berlin researchers place their data at the disposal of

their scientific partners. In this way, they obtain valuable ad-ditional information about a disease gene that has just beenidentified. The gene set which has been analyzed shall nowbe enhanced. The scientists are concentrating on geneswhich are involved in signal processes during tissue and organ development and on genes whose expression is dis-turbed due to human diseases. "9.5- to 11.5-day-old embryosare particularly well suited for these investigations. In thisdevelopment stage 40 percent of the genes which show aspecific expression pattern are involved either in gene regula-tion or in signal mechanisms," explains Bernhard Herrmann.Moreover, Herrmann's group is focusing on genes that wereidentified as being of interest in microarrays in the SMP RNA:"We plan to study a total of up to 3,000 genes in 9.5- and11.5-day-old mouse embryos," Bernhard Herrmann says. Theyevaluate their data together with the results from the SMPCell and SMP RNAi and enter them into the internationaldatabase of the Molecular Anatomy of the Mouse Project(MAMEP). In this way, the Berlin scientist expects to identifyadditional disease genes, especially for cancer diseases, verysoon.

ReferencesSültmann H (2005); Gene expression in kidney cancer is asso-ciated with cytogenetic abnormalities, metastasis formation,and patient survival. Clinical Cancer Research 11(2): 646–655

Huber W et al. (2002); Variance stabilization applied to micro-array data calibration and to the quantification of differentialexpression. Bioinformatics 18: 96–104

Group 1: Gitton Y, Dahmane N, Baik S, Ruiz A, Altaba I; group2: Neidhardt L, Scholze M, Herrmann BG; group 3: Kahlem P,Kahla AB, Schrinner S, Yildirimman R, Herwig R, Lehrach H,Yaspo ML (2002); A gene expression map of human chromo-some 21 orthologs in the mouse. Nature 420: 586–590

COORDINATORProfessor Annemarie PoustkaGerman Cancer Research Center,Heidelberga.poustka@dkfz.de

Manuela Scholze and Ralf Spörle

34

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

They had already lost interest in it. Thought they knew every-thing there was to know about it. But all at once it appearedagain on researchers' radar screens: RNA. The reason wasthe discovery of RNA interference (RNAi). Already back at thebeginning of the nineties, scientists noticed that in plants,short RNA molecules could inhibit the translation of genesinto proteins. They at first suspected the known antisensephenomenon to account for this "gene silencing". Theresingle-stranded RNA molecules bind to complementarymRNA molecules and thus block transcription. But in experi-ments with the nematode Caenorhabditis elegans the re-searchers discovered to their surprise that "gene silencing"was considerably intensified when double-stranded RNAmolecules were used. This puzzling observation awakenedgreat interest, because the double RNA strands inhibitedgene expression very efficiently. Researchers throughout theworld worked feverishly trying to understand this so-calledRNA interference (RNAi) mechanism. They were successful:Within only a few years scientists were able to explain howthis mechanism works (see box). RNAi is a natural processwhich takes place in many organisms ranging from plants tohuman beings. Researchers surmise that it is a mechanismto protect the cell against foreign genes, such as parasite or

virus genes. The newly discovered and greatly promisingRNAi technology will now be further developed by the Sys-tematic-Methodological Platform (SMP) RNAi into a routinemethod. To accomplish this, the two coordinators ProfessorWolfgang Wurst and Professor Tony Hyman have broughttogether leading German groups in the fields of cell biology,molecular biology, genetics and bioinformatics. "Our goal isto optimize RNAi technology for gene function assays in vitroand in the whole organism and to make it available for medi-cal research," Wolfgang Wurst says. The SMP RNAi com-prises ten subprojects forming a pipeline for RNAi gene func-tion analysis. This pipeline starts with sophisticated cell cul-ture tests, followed by analysis of mouse embryos and finallyculminates in studies of adult knock-down mice. In contrastto the classic knock-out, in which a gene of the genome ismade incapable of functioning (null mutation), the knock-down effect achieved with RNAi takes place on the posttrans-criptional level and leaves the respective gene undamaged."RNAi technology is considerably less expensive and time-consuming than the classical knock-out methods. Moreover,with RNAi technology we can even inactivate several genes at the same time," Tony Hyman explains. Furthermore, thistechnology can be used for gene function analysis in all spe-

SMP RNAi

OLD MOLECULE BACK IN FASHION –NGFN SCIENTISTS OPTIMIZE RNAi TECHNOLOGY

Ralf Kühn (Wurst group)

35

cies (e.g. rats, productive livestock), for which traditionalknock-out strategies are not available due to the lack ofembryonic stem cells.

RNAi TECHNOLOGY AS ROUTINE METHOD In the SMP RNAi three subprojects are optimizing RNAi tech-nology for use in cell culture, whereas two additional subproj-ects are using RNAi strategies to analyze early stages ofmouse embryogenesis. The objective of still other subproj-ects is to generate adult RNAi knock-down mice as models ofhuman disease. Moreover, the researchers want to produceknock-down mice for candidate genes identified in otherNGFN projects. Some of these mice are systematically inves-tigated for relevant signs of disease by the German MouseClinic in the SMP Mammalian Models. "We want to develop

new standard operating procedures (SOPs) and make themavailable to all NGFN scientists and the entire scientific com-munity," Wolfgang Wurst explains. In addition, a strong bioin-formatics group is included in the SMP RNAi which supportsthe design of experiments, the evaluation of data and theirpresentation to the scientific community. Cenix BioScienceGmbH, which has many years of expertise in the field of RNAi technology, is also a member of the SMP RNAi: It isactive in RNAi technology development and service. In anoth-er subproject RiNA GmbH offers workshops and laboratorycourses to present the knowledge and insight gained by theSMP RNAi researchers to interested NGFN participants. "TheSMP RNAi includes all currently available key technologies forthe design and construction of interfering RNAs. Moreover, it will make RNAi technology utilizable for high-throughputscreening in human cells and for in vivo functional studies inmice," Wolfgang Wurst adds. "We are therefore convincedthat we can develop RNAi technology into a routine proce-dure for in vitro and in vivo gene function analysis."

ReferencesSonnichsen B et al. (2005); Full-genome RNAi profiling of earlyembryogenesis in Caenorhabditis elegans. Nature 434: 462–469

Kittler R et al. (2004); An endoribonuclease-prepared siRNAscreen in human cells identifies genes essential for cell divi-sion. Nature 432: 1036–1040

Elbashir SM (2001); Duplexes of 21-nucleotide RNAs mediateRNA interference in cultured mammalian cells. Nature 411: 494–498

Fire A (1998); Potent and specific genetic interference bydouble-stranded RNA in Caenorhabditis elegans. Nature 391: 806–811

COORDINATORProfessor Wolfgang WurstGSF – National Research Center forEnvironment and Health, Neuherbergwurst@gsf.de

RNA INTERFERENCE (RNAi) RNA interference (RNAi) is triggered by the pres-ence of double-stranded RNA molecules (dsRNA).These are interpreted by the host cell as "foreign"and therefore as "not wanted". First the endonucle-ase Dicer cleaves these trigger dsRNAs into 21–23base pair fragments called "short interfering RNAs"(siRNAs). These siRNAs are then integrated into anRNA-Induced Silencing Complex (RISC). The RISCbinds to those mRNAs which are complementary tothe siRNAs and catalyzes their degradation. If RNAiis used as an experimental tool, the natural RNAinterference mechanism can be used to specificallyinhibit the translation of desired genes. To ac-complish this, the scientists introduce specific trig-ger RNA into cells or organisms. In invertebratessuch as Caenorhabditis elegans and Drosophila melanogaster this objective is best achieved byusing relatively long dsRNA molecules (200–1,500bp), which contain coding sequences of the targetmRNA. However, in mammals the situation is morecomplicated because here a second mechanismexists which can be influenced by dsRNA: the inter-feron response. It leads to downregulation of proteinsynthesis and ends with programmed cell death(apoptosis). The scientists avoid this problem byusing siRNA molecules instead. In most cases theseevoke strong, reproducible and specific RNAiresponses without activating the interferon signalcascade.

Frozen cell lysates

36

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

SMP CELL

The work of the Systematic-Methodological Platform (SMP)Cell is reminiscent of a filter: At the top the cDNAs go in – at the bottom disease-relevant human genes come out. In between a lot of work takes place, which in the SMP Cell is divided among several groups (www.smp-cell.org). Dr. Stefan Wiemann, at the German Cancer Research Center(DKFZ) and coordinator of the SMP Cell, describes the work-ing groups as modules. Their most important aim is to iden-tify genes and proteins which are involved in the genesis ofcancer. "However, our assays also yield results which are rel-evant for infectious and inflammatory diseases and for theresearch of neurodegenerative processes." The work of SMPCell is based on the experience and results of the GermancDNA Consortium, the first module of the platform. Since1996 its researchers have been systematically sequencingand analyzing cDNAs and have identified many previouslyunknown genes. The cDNA Consortium is therefore signifi-cantly involved in an international project to describe allhuman genes. Furthermore, the scientists of the consortiumare building up a cDNA-library. Full-length cDNAs for eachgene and each splicing variant from this library are providedto other research projects. "In the meantime the cDNAswhich have been sequenced up to now cover 50 megabasesof sequence. That is more than the length of the human chro-mosome 21," Stefan Wiemann explains. The protein-codingregions of cDNAs are cloned in transport vehicles, so-calledvectors, out of which they in turn can quickly and convenient-ly be cloned into other expression vectors. In the meantimeabout 1,200 cDNAs are available as such "entry clones".Nomen est omen – these clones are the entry point for allfurther analyses of the SMP Cell.

SNAPSHOTS FROM THE CELLThe module "Protein Localization" already has a long history:"Ever since 1998, in cooperation with the DKFZ, we havebeen investigating where proteins are localized in the cell,"explains Dr. Rainer Pepperkok from the European MolecularBiology Laboratory (EMBL). For this, the scientists fall backon the cDNA clones of the cDNA Consortium. In their investi-gations they fuse the cDNAs with the green fluorescent pro-tein (GFP) and transfect mammalian cells with it. These thenexpress the fusion proteins generated. "Via fluorescentmicroscopy we thus obtain snapshots of the protein distribu-tion in a cell," Rainer Pepperkok explains. The scientists in-vestigate each protein in two work steps. In doing so, theytag the GFP onto the C-(3') terminus and also onto the N-(5')

terminus. The reason is that at both ends of proteins thereare often signal sequences that steer the localization of aparticular protein in the cell. The GFP could mask these sig-nals, making the protein lose its orientation in the cell andlocalize to the "wrong site". Of 567 proteins studied, this wastrue in 219 proteins. In all cases, the scientists have discov-ered which of the two ends is responsible for the correctlocalization by making bioinformatic comparisons with otherproteins, among other techniques. With their methods theresearchers have localized about 1,000 previously unknownproteins already at the subcellular level. Moreover, StefanWiemann and his colleagues observed that protein localiza-tion is not a stable factor. It changes, for instance, dependingon the cell density. "Cells are dynamic systems," RainerPepperkok explains. "To be able to document and analyzethese dynamics we will use a new method in the future, ‘livecell imaging'." Here a computer-automated camera takes pictures of the cells for several hours, and then from thesephotos a time-lapse video is produced at the computervisualizing the dynamics of the proteins within the cells.

THE CORE OF SMP CELLIn the next module "Functional Assays", the actual core partof the SMP Cell, high-throughput methods are developed andapplied to perform cell-based assays in microtiter plate for-mat. With these the DKFZ researchers investigate e.g. thesignificance of proteins in cell proliferation, apoptosis or signaling. The results should aid in more exactly analyzingtumorgenesis and inflammation processes. Furthermore, theyhave established assays with which mitosis, anchorage inde-pendent growth and cell invasion can be investigated. Thus,the scientists want to find proteins which play a role in theformation of metastases. "We receive 30 to 3,000 potentialdisease genes from our partners in the NGFN, especiallyfrom the SMP RNA," says Dr. Dorit Arlt, project director inSMP Cell at the DKFZ in Heidelberg. In order to handle themass of candidate genes that are to be characterized, thescientists first sort the proteins using screens into functionalcategories. "In doing so, we check whether the promisingcandidate genes are in fact disease relevant," Dr. Arlt goeson to say. The number of proteins to be more closely investi-gated then usually shrinks. "But for us there still remains alot to do," the biologist says, laughing. A whole series of dif-ferent kinds of assays is needed to find out what role theproteins coded by the cDNAs play in the course of a disease."Only then will we be able to formulate reliable hypotheses

FROM MASS TO CLASS –RESEARCHERS TRACK DOWN DISEASE-RELEVANT PROTEINS

37

COORDINATORDr. Stefan WiemannGerman Cancer Research Center,Heidelbergs.wiemann@dkfz.de

regarding functionand disease rele-vance," Dr. Arlt says."We suspected, forexample, that a pro-tein is involved in thesignaling pathway ofthe MAP kinase. The reason is that a small interfering (si)RNA, specific for the gene of this protein, has dramaticallyreduced the activation of p42/44 MAP kinases." To pursuethis hypothesis further, the gene in question was further in-vestigated in DNA replication assays. With success: The pro-tein did in fact stimulate cell division when it was overex-pressed. RNA expression profiles ultimately showed that thegene is increasingly expressed in kidney tumors and in me-tastasizing tumors. It is therefore now a hot candidate fordetailed functional and disease-relevant analyses.

BIOINFORMATICS BRINGS ORDER TO THE SYSTEMA further important module is "Bioinformatics". It managesand interprets the data of the experimental projects. "Whiledeveloping the high-throughput functional assays, it was clearto us very early on that the data could only be correctly inter-preted using statistical methods," says Heiko Rosenfelder, abioinformatician in the SMP Cell at the DKFZ. "Cells are com-plex biological systems, and they act and react accordingly.There one can only dream of tidy linear dependencies. Onlythrough the evaluation of high cell numbers does one obtainresults with statistic significance." Therefore, after the labo-ratory work in the SMP Cell, the analysis and management ofthe experiments take place at the computer. For this, thecomputer scientists, together with the laboratory research-ers, are developing laboratory information managementsystems (LIMS) that can be quickly and easily adapted to therequirements of the particular experiment. That includes sta-tistical models or meta-analyses for the comparison of multi-ple experiments. "Due to LIMS our experiments are objectiveand robust. We try to run as many analyses as we can via thecomputer to exclude subjective errors as far as possible,"Heiko Rosenfelder explains. Moreover, there is an automatic"annotation system" and a half-automatic protein sequenceanalysis with which protein functions can be predicted. Alldata of the SMP Cell are stored in a central server, which isaccessible to the public. Furthermore, records from otherdatabases are integrated there, in order to directly evaluatethe results gained in the SMP Cell, in context with externalinformation.

FURTHER ANALYSISThe identification of a protein's interaction partners alsosheds light on the cellular signal pathways the particular pro-tein is involved in. With the aid of the yeast two-hybridsystem (Y2H) and immune precipitation with downstreammass spectroscopy the scientists are therefore investigatingprotein-protein interactions. In addition, the SMP Cell is colla-borating especially closely with the SMP RNA in the NGFN.There scientists are investigating which genes are differen-tially expressed in normal as compared to diseased tissues.Genes that have become "conspicuous" due to their modifiedexpression in tumors will then be more closely inspected bySMP Cell researchers. "By systematically linking gene identi-fication and high-throughput function analysis we have al-ready traced a number of potential disease-relevant pro-teins," Stefan Wiemann says in summation. "Now we mustanalyze in detail what role these proteins play in diseases inorder to validate new therapeutic targets."

ReferencesImanishi T et al. (2004); Integrative annotation of 21,037human genes validated by full-length cDNA clones. PLoS Biol 2: 856–875

Wiemann S et al. (2004); From ORFeome to biology: a functional genomics pipeline. Genome Res 14: 2136–2144

Arlt DH et al. (2005); Functional profiling: from microarrays viacell-based assays to novel tumor relevant modulators of thecell cycle. Cancer Res 65: 7733–7742

KEY QUESTIONSTo understand the biological activity of a gene product,certain key questions need to be answered. At whatpoint in time during development and growth is a geneexpressed? "This is one of the central questions whenone wants to identify disease-related genes," StefanWiemann declares. "To cite an example, for this, ex-pression profiles of diseased and healthy tissue can becompared." In which tissues and cell types is a geneexpressed and where in the cell is the gene productactive? What biological activity does the gene producthave and how does the cell react to elevated or lowerquantities of this protein? How is the activity of this pro-tein regulated in the cell? In what kind of biological con-text is the protein active? With which partners does itinteract, and what are its substrates? If one answersthese questions, one obtains important clues about thebiochemical processes in which the protein participates.Taken together, all of these questions and answers are akey to an approach for discovering new drugs.

Dorit Arlt

38

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

SMP PROTEIN

PROTEINS – NETWORKERSOF THE CELLInside a large cabinet made of tinted glass, an arm-like robot-ic device moves up and down, forward and backward. Auto-matically, it lifts lids off agar plates, dips numerous pointedtips into a microtiter plate and dispenses tiny spots of liquidonto the gelatinous surface of the agar. This is the roboticsunit of the Max Delbrück Center for Molecular Medicine(MDC) in Berlin-Buch. "Here, thousands of protein-protein-interactions have been tested in an automated screeningapproach," Professor Erich Wanker explains. He is coordina-tor of the Systematic-Methodological Platform (SMP) Protein."We are employing a yeast two-hybrid system that we haveespecially adapted to the screening of human proteins. Usingit, we can identify protein interactions on a very large scale."

Since the robotics unit was established four years ago, 25 million protein pairs have been tested, and two large pro-tein interaction networks have been generated. One contains180 interactions that are all connected to the protein thatcauses Huntington's disease, a severe neurological disorder;the other is a global network of the human proteome con-sisting of 3,186 interactions between 1,705 proteins. Throughthese networks, the NGFN scientists hope to gain new in-sight into the elementary processes in the cell, particularlyinto their pathological changes. "Knowledge about the inter-actions helps us find out more about the function of the pro-teins," says Dr. Ulrich Stelzl, the scientist in Erich Wanker's

group responsible for the interaction studies. "By identifyinga protein's interactors – its neighbors, as it were – we getclues as to what this protein does. If the interacting proteinsare involved in a particular signaling cascade, the odds arehigh that the protein in question plays a part in that cascadeas well."

PROTEIN INTERACTION NETWORKS: A NEW RESOURCEThe ultimate goal of interactomics – the systematic study ofall protein-protein interactions – is the generation of a reli-able map showing how the whole proteome of a cell con-nects. This ambitious plan, which can only be realized withthe automated procedures established in the SMP Protein,serves several purposes: First, it enables the researchers todiscover which cellular proteins interact and share function.Yeast two-hybrid interaction studies try to model what hap-pens between cellular proteins in an experimental set-up andtherefore provide many clues about the real situation in thecell. Second, the generated networks are needed and usedas resources. By analyzing these networks, it can be foundout which proteins form clusters, which proteins act as hubsthat interconnect many others and which proteins are fairlyisolated. Third, interaction data are collated with data fromstudies on signaling pathways and disease mechanisms. "If aprotein interaction that appeared in a study can be mappedto a pathway that is relevant, for instance, in tumorgenesis,

Jan Timm and Ulrich Stelzl

39

COORDINATORProfessor Erich WankerMax Delbrück Center forMolecular Medicine, Berlinewanker@mdc-berlin.de

we have confirmed the proteins' importance for the disease,"Erich Wanker points out. "In this way, interactomics contrib-utes significantly to disease research."

INTEGRATED RESEARCH ACTIVITIES TO FACILITATE THERAPYThe yeast two-hybrid interaction studies, however, are onlyone part of a larger concerted proteomics strategy intendedto yield concrete results to facilitate therapy development forcomplex human disorders. "The ultimate goal of the SMPProtein is to gain insight about disease-relevant proteinswhich would enable the development of more specific drugsin a shorter time," Erich Wanker says. To achieve this, re-search requires clear organization. This is why a unique infra-structure was established that guarantees exchange of dataand know-how among groups within the SMP Protein, groupsfrom other SMP, from the Disease-oriented Genome Net-works and Explorative Projects. A central office headed byDr. Patrick Umbach at the MDC makes sure that this ex-change functions smoothly. It serves both as an internal con-trol unit for project progress and as a link between the indi-vidual laboratories. Regular meetings within the platformensure close contact among the scientists involved.

ANALYZING PROTEINS, COLLECTING DATA, PROVIDING ACCESSThe production line starts with the preparation of the cDNAclones at the RZPD, the German Resource Center for Ge-nome Research (Dr. Uwe Radelof). From these clones, humanproteins are expressed and purified at the Max Planck Insti-tute for Molecular Genetics in Berlin-Dahlem (MPIMG, Dr. Konrad Büssow). At the Protein Structure Factory (PSF)proteins are crystallized and undergo X-ray structure analysis

(Dr. Yvette Roske). Next, various approaches are employed toanalyze the manifold relationships proteins can form. Protein-protein interactions are studied at the MDC (Dr. Ulrich Stelzl); protein-DNA interactions are analyzed at the MPIMG(Dr. Claus Hultschig). At the GBF, the German ResearchCenter for Biotechnology in Braunschweig, protein-drug inter-actions are detected with the immediate aim of identifyinglead structures (Dr. Ronald Frank). Protein complexes are iso-lated at the MPIMG from cell extracts and characterized bymass spectrometry (Dr. Bodo Lange and Dr. Johan Gobom).In this way, different methods and techniques permit lookingat protein function from a multitude of angles.

While these studies are being carried out in the eight experi-mental labs of the SMP Protein, two further project units are busily gathering and processing the generated data.Professor Olaf Wolkenhauer and his group at the Universityof Rostock specialize in computer modeling of signalingpathways, while Dr. Ralf Herwig from the MPIMG is responsi-ble for setting up the databases needed for the integration ofthe results. On his computer screen, Dr. Stelzl zooms into adetail of the interaction map that consists of many hundredsof differently colored dots representing the proteins: "At theend of the funding period, we want all our scientific partnersto be able to click on any dot on the map and receive thecomplete information for that specific protein."

ReferencesStelzl U et al. (2005); A human protein-protein interaction net-work: a resource for annotating the proteome. Cell 122 (6): 957–968

Goehler H et al. (2004); A protein interaction network linksGIT1, an enhancer of huntingtin aggregation, to Huntington'sdisease. Mol Cell. 15 (6): 853–865

Sauer S et al. (2005); Miniaturization in functional genomicsand proteomics. Nat Rev Genet. 6 (6): 465–476Ulrich Stelzl

Automated two-hybrid screen

40

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

SMP PROTEOMICS

Proteome – a term first coined in 1995 – is the key to under-standing complex cellular processes in our body. It can beroughly defined as the protein equivalent of the genome, asthe complete set of proteins in a biological system. Protein-protein interaction is involved in all biological processes. Theproteome is larger than the genome and more complex: Forexample, the genomes of the caterpillar and the butterfly areidentical – but interaction among proteins is what makes theinsect first crawl and then fly.

THE HUMAN BRAIN PROTEOME PROJECT (HBPP)Scientists of the Systematic-Methodological Platform (SMP)Proteomics in the NGFN are studying the proteome of thebrain. In their Human Brain Proteome Project (HBPP) they aresystematically analyzing protein expression in the brains ofhuman and mouse. In addition, they are investigating pro-teins in body fluids such as blood plasma and the cerebrospi-nal fluid (CSF), which are relevant for the nervous system.The aim of their research is to understand the complex inter-action of proteins. Furthermore, the scientists are working onways to improve already established methods of proteomeresearch. The emphasis of the clinically oriented projects ison neurodegenerative diseases, especially on Alzheimer'sand Parkinson's diseases (AD and PD): "We want to contrib-ute to uncovering the mechanisms of neurodegenerative dis-eases by systematically identifying and characterizing theproteins of the nervous system," explains Professor HelmutE. Meyer, coordinator of the SMP Proteomics at the Ruhr-

University in Bochum. "Our work can lead to an enhancedunderstanding of these diseases. It is even conceivable thaton the basis of molecular patterns which characterize thedisease, a new classification system may evolve. Besideshistological and clinical criteria, this classification would bebased on the molecular phenotype, that is, on the mRNA andthe protein expression patterns." Ultimately, the new know-ledge should improve both the diagnosis and the treatmentof neurodegenerative diseases. To accomplish this enormoustask, ten academic partners have joined together with thebioinformatics company MicroDiscovery GmbH. In addition,the proteome researchers are collaborating closely with theDisease-oriented Genome Network "Systematic Gene Identi-fication and Functional Analyses in Common CNS Disorders"and the SMP Bioinformatics, RNAi, and Mammalian Models.

CELLULAR INTERDEPENDENCIES – A CELL WIDE WEBIn a subproject of the SMP, Professor Joachim Klose is work-ing to detect the rules of protein networking. "Our workinghypothesis is that proteins not only interact within a processor a complex, but that all proteins of a cell have at least indi-rect contact with each other," he explains. "All of a cell's pro-teins compete for the same, but limited resources - freespace, free water, energy and amino acids." This could func-tion on the basis of a regulatory network encompassing all ofthe proteins in a cell. Pathological changes in the concentra-tion of a single protein would induce changes in the concen-tration of the other proteins in a cell, in order to preserve the fine balance between protein quantity and resources.Starting from this hypothesis, the scientists want to showthat genetic disturbances and drugs can change the concen-tration of quite a number of proteins. But not every modifiedprotein may be significant for the disease. Investigations arebeing carried out on mouse brains with neurodegenerativediseases. "Looking at the proteome of a cell as a whole of-fers a completely new angle," says Joachim Klose. "With ourstrategy we hope to learn more about the pathogenesis ofneurodegenerative diseases and show starting points foreffective treatments." To analyze the functional implications,the scientists are using a high-resolution two-dimensional gelelectrophoresis with which they can separate and differenti-ate 10,000 proteins per gel. "With our experiments we willnot only discover disease-specific protein mutations, but alsodifferences which occur in general in several diseases or withaging. Especially proteins which play a central role in the net-work, e.g. alpha-B-crystalline, which is responsible for theassembly of the proteins, will react to many influences andthus have a nonspecific effect with regard to the disease,"Joachim Klose explains. That is why the scientists divide the

OUR BRAIN – A NETWORK OF PROTEINS

Katrin Marcus

41

COORDINATORProfessor Helmut E. MeyerRuhr-University, Bochumhelmut.e.meyer@rub.de

proteins into three groups prior to characterizing them fur-ther: specific for a particular neurodegenerative disease, spe-cific for all investigated diseases or nonspecific.

RESEARCH ADVANCES IN PARKINSON'S ANDHUNTINGTON'S DISEASESResults in mice modeled for Parkinson's disease show thatKlose's working group is headed in the right direction: Themice lack the protein parkin. Mutations in this protein causein humans a hereditary and early-onset form of Parkinson.The scientists discovered that the expression of 13 proteinsin the brains of diseased mice is significantly reduced. Fourof the proteins are directly involved in the oxidative phos-phorylation in the mitochondria, whereby this reaction is lim-ited in mice without parkin. As a consequence, the scientistsfound an increase of harmful reactive oxidative products(ROS) in the brain of the diseased mice. "It is known thatROS can inactivate parkin. It may be that a cycle could devel-op that leads to the onset of Parkinson," Joachim Klose sur-mises. "Less parkin reduces the activity of the mitochondria.This generates more ROS which in turn inactivate parkin." Histeam can also report research advances for Huntington'sdisease: In the mouse model they found a number of proteinswhose quantity diminishes in the course of the disease. Re-searchers were able to confirm this result by examining post-mortem brain tissue from Huntington patients.

REFINED TECHNIQUESThe research group of Professor Helmut E. Meyer and JuniorProfessor Katrin Marcus is working on the establishment ofan alternative method for 2-D gel electrophoresis. They aredeveloping new multidimensional separation methods for thequantitative and qualitative analysis of complex protein com-binations. "At present 2-D gel electrophoresis is the methodof choice to carry out comparative proteome studies. But thistechnology has its limits. It is unable to reliably detect verylow quantities of proteins," Katrin Marcus explains. The scien-tists want to remedy just this disadvantage with a new liquidchromatography based multidimensional separation method.Furthermore, the Bochum team is optimizing techniques todetect posttranslational modifications such as phosphoryla-tion and ubiquitinylation. Such modifications appear to play asignificant role in Alzheimer's and Parkinson's disease. "De-tecting phosphorylations is a real challenge since the crux ofthe problem is to detect transient changes that go hand inhand with dephosphorylations," she adds. But already back in the first NGFN funding period the Bochum team, togetherwith the companies Bruker BioSpin GmbH and Protagen AG,succeeded in developing a method to determine the phos-

phorylation status of proteins. "Now we want to refine andenhance this technology," Helmut E. Meyer says. "Using thetechnology we have optimized, we hope to identify proteinpatterns characteristic for Alzheimer and Parkinson in thebrain as well as in the blood and the cerebrospinal fluid(CSF)." To achieve this, his team is working with other scien-tific groups who provide material from mouse brains modeledfor Alzheimer's and Parkinson's diseases. Furthermore, theyare investigating the CSF and blood plasma of Alzheimer pa-tients and post-mortem brain tissue from Alzheimer's andParkinson's patients. He stresses, "Since these diseases cannot be traced back to a genetic defect alone, protein ex-pression analysis is essential for understanding the patho-mechanisms of Alzheimer's and Parkinson's diseases."

ReferencesKlose J et al. (2002); Genetic analysis of the mouse brain pro-teome. Nat Gen 30: 385–393

Tribl F et al. (2005); "Subcellular proteomics" of neuromelaningranules isolated from the human brain. Mol Cell Proteomics 4 (7): 945–957

Schäfer H et al. (2004); A peptide preconcentration approachfor nano-high-performance liquid chromatography to diminishmemory effects. Proteomics 4 (9): 2541–2544

Katrin Marcus, Romano Hebeler and Cornelia Joppich

42

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

SMP ANTIBODY FACTORY

Antibodies – what molecular biologist or biochemist has notyet worked with these precision weapons of the immunesystem? They are indispensable even for the analysis ofgenes which have not yet been decoded. The human genomeproject (HGP) has brought forth a huge number of genes andgene products whose function is still to be determined.Professor Stefan Dübel from the Technical University (TU)Braunschweig estimates that up to 100,000 proteins awaittheir analysis. "Our goal is to establish a pipeline within threeyears that will produce antibodies for proteomics research ina high-throughput process," Stefan Dübel explains. "This isthe only way we can meet the need for antibodies for thefunctional analysis of genes." The conventional production ofantibodies in mice, rats or rabbits often reaches its limitsbecause not every protein evokes an immune response andthus the formation of antibodies. Moreover, this method can-not be completely conducted with high-throughput tech-niques. Alternative possibilities are provided by in vitro selec-tion methods. With their aid the whole process can be auto-mated – from the gene identification number to the finishedantibody. Antibodies against non-immunogenic or even toxicproteins can be produced in this way as well. "During the invitro selection process we can control the biochemical condi-tions," says Stefan Dübel. "Thus, antibodies can also be pro-duced that recognize posttranslational modifications or con-formation mutants, e.g. after the binding of a co-factor." Theselection method of choice is antibody phage display (seebox). Until now, most in vitro antibodies have been gained inthis way. "Hitherto, the methods and the form of the anti-bodies have been optimized for the use of human antibodiesin therapy," Stefan Dübel explains. "Therefore, antibodiesfrom phage display libraries have not been used for basicresearch in larger numbers." (see table)

CUSTOM ANTIBODIESTo improve the work with the antibodies, the Institute forBiochemistry and Biotechnology of the TU Braunschweig, theSociety for Biotechnological Research, the Max Planck Insti-

tute for Molecular Genetics and the German Resource Centerfor Genome Research in the Systematic-Methodological Plat-form (SMP) Antibody Factory have pooled their resourcesand have formed a collaboration. In four closely interlockingsubprojects scientists in this antibody factory are working to make in vitro selection useful for proteomics research. Toaccomplish this, they are, for example, concerned with theform of antibodies in their research. The bacterium Esche-richia (E.) coli used for phage display is not able to producecomplete IgG molecules. Small antibody fragments such assingle chain variable fragments (scFv fragments) or Fab frag-ments, on the contrary, can be produced by microorganismswithout any problems. Until now, the scientists have gener-ally fallen back on scFv fragments, because the yield of cor-rectly folded and thus functional molecules is greater thanwith Fab fragments. But the use of scFv fragments can causeproblems. On the one hand, they are often not stable enough.On the other hand, they are frequently not compatible withthe detection systems of existing assays, which are based onthe detection of IgG molecules. In the production of thera-peutic antibodies these problems are circumvented by clon-ing the antigen-binding sequence into other vectors for theproduction of complete IgG molecules. Proteomics research,however cannot afford to lose time and money with thesedetours. "That is why we are working on optimizing the use of Fab fragments," explains Dr. Michael Hust, a member onDübel's research team. "They are more stable and are alsousually recognized by secondary antibodies already used inexisting assays." He and his colleagues plan to build up dif-ferent Fab antibody libraries and to compare them with eachother. Furthermore, they are testing different methods forpanning, i.e. the in vitro purification of specific antibodies. Toaccomplish this, the researchers use ELISA plates, magneticbeads and peptide arrays which are still in development.

SELECTION PROCESSAn additional project is devoted to the preparation of pro-teins for which the compatible antibody is to be found in the

A FACTORY FOR ANTIBODIES

Thomas Schirrmann, Martina Kirsch and Michael Hust

43

COORDINATORProfessor Stefan DübelTechnical University Braunschweigs.duebel@tu-bs.de

phage antibody library. The reason is that researchers oftenonly have the DNA sequence but not the protein. The recom-binant expression of cDNAs sometimes delivers the desiredprotein, but frequently the production in E. coli is not easy.That is why scientists are seeking an alternative to proteinsto obtain antibodies: synthetic peptides, which in the selec-tion can be used as antigens. "Of course, not all antibodiesthat are specific to a peptide will recognize the correspon-ding native protein," says Stefan Dübel. "But the high-through-put method allows parallel screening of several peptides perantigen so that the chances of making a hit are greater.Besides, even an antibody recognizing only denatured anti-gen is still a useful agent, e.g. for immunoblots or immuno-histochemistry." The most recent developments in bioinfor-matics are contributing to the improvement of epitope pre-diction. Based on computer prognoses, scientists are tryingto produce better targeted peptides for antibody selection.An additional option for protein synthesis is in vitro transcrip-tion/translation from cDNA libraries. Scientists can usethese methods whenever they are not successful in fishingthe compatible antibody out of the library using syntheticpeptides. "We are trying to integrate these individual compo-nents of antibody production into one single pipeline and atthe same time to take into account researchers' require-ments. In this way the current bottleneck in generating anti-bodies for proteomics research could soon be cleared, en-abling the production of antibodies against problematic anti-gens," Stefan Dübel emphasizes.

ReferencesHust M and Dübel S (2004); Mating antibody phage displaywith proteomics. Trends in Biotechnol. 22: 8–14

Kirsch M et al. (2005); Parameters affecting the display of anti-bodies on phage. J. Immunol. Methods. 301: 173–185

Konthur Z et al. (2005); Perspectives for systematic in vitroantibody generation. Gene 364: 19–29

PRINCIPLE OF ANTIBODY PHAGE DISPLAYBacteriophages (phages for short) are viruses that adhereto the surfaces of bacteria and transfect their genome intothe bacterium. For antibody phage display the genes whichcode for antibody fragments are fused with a gene for a surface protein of the phages. Subsequently, infectedEscherichia coli bacteria thus produce viruses which carryantibody fusion proteins on their surface. A mixture ofphage clones, all of which carry different antibody genesand therefore express different fusion proteins, is called aphage display library. Using the protein for which one wouldlike to have an antibody, this library is searched. To do this,the protein is coated onto a firm surface (such as magneticbeads or microtiter plates) and incubated with the phage.After several washing steps only phage with the specificantibody fragment remain bound to the immobilized protein.They can be eluted in the next step. This technique of invitro selection via the binding activity is termed "panning".Normally, two to five such panning rounds are performedbefore individual phage clones can be isolated.

THERAPEUTICS

• few• known and available• well characterized• production cost not important

• arbitrary number of selection rounds • cost of the actual selection relatively

unimportant • optimized for maximal success rate • re-engineering of the antibody format

essential

• optimized for minimal immunogenity• optimized for good pharmacokinetics

PROTEOMICS

• large number• mostly unknown and not available • unknown characteristics• production costs limiting

• number of selection rounds should be minimal

• optimized for low selection costs • maximal success rate not primary goal• avoidance of re-engineering of every

individual antibody clone

• optimized to be robust• optimized for maximal compatibility

to established assays

APPLICATION

ANTIGENS

SELECTION PROCESS

PRODUCT (ANTIBODIES)

COMPARISON OF THE REQUIREMENTS OF IN VITRO SELECTION SYSTEMS FOR ANTIBODIES IN THERAPY OR PROTEOME RESEARCH

44

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

At the age of five weeks the process begins: In the GermanMouse Clinic at the GSF - National Research Center for Envi-ronment and Health in Neuherberg the mouse mutants aregiven thorough medical check-ups. First the clinic staff weighs and measures the mice and looks for abnormalities.Then all of the results and the birth date are entered into adatabase. But that is not all: The mice will progress through14 stations in the next four months. The German MouseClinic (GMC) is a diagnostic clinic in which genetically modi-

fied mice are systematically phenotyped. Behavior and lungfunction are included in the examination modules along withneurology, cardiology and allergies – in the screen no impor-tant area is left out. "Hardly any patient is examined as thor-oughly as our mice," says Professor Martin Hrabé de Angelis,director of the GMC and coordinator of the Systematic-Meth-odological Platform (SMP) Mammalian Models, to which theGerman Mouse Clinic belongs. "Our goal is to find mamma-lian models for genetically determined human diseases inorder to understand them better." For that reason, the SMPMammalian Models covers the entire spectrum of mousegenetics, ranging from the creation of genetically modifiedmouse lines, phenotyping, to the archiving of mouse lines."Creating mouse mutants is one aspect," Martin Hrabé deAngelis explains. "But most important is the subsequentthorough diagnosis. Only when we have understood every-thing down to the last detail about what happens in themutants can we elucidate gene functions with medical and

biological relevance." The German Mouse Clinic is thereforean essential element of his scientific concept.

SCHEDULED DIAGNOSISDr. Valérie Gailus-Durner and Dr. Helmut Fuchs, the coordina-tors of the GMC, are responsible for ensuring that everythingin the mouse clinic runs smoothly. They organize a detailedschedule according to which the individual mouse linesadvance through the different investigation modules in a

weekly rhythm. "We have taken spe-cial care that the tests which are per-formed on an animal do not interferewith each other. In this way we pre-vent false results," Valérie Gailus-Durner explains. As a worldwide unique offer, the mouse clinic estab-lished this workflow to be able toaccept mouse lines of other scientificworking groups for investigation.Every two weeks these "foreign"mouse mutants, which are only a fewweeks old, are delivered. First of all,the new arrivals have two weeks toget used to their new environment –the food, the smells, the sounds andtheir keepers – before they areexamined. "All mice should have thesame situation at the beginning, sothat we can compare them with eachother," explains Helmut Fuchs.

240 PARAMETERS PER MOUSEThe mouse clinic cooperates with numerous specialists inperforming the multifaceted, comprehensive examinations."Without the close collaboration of clinicians and genomeresearchers our project would not be conceivable," ValérieGailus-Durner says. Thus, for instance, neurologists of theLudwig Maximilian University in Munich test reflexes andmuscle tension. At the Technical University Munich bloodspecimens of mice are screened for infectious and auto-immune diseases. A working group of the NGFN Cardiovas-cular Diseases Network is presently building up the cardiol-ogy department. To make a diagnosis, the scientists use de-vices like in a normal clinic: electrocardiogram, X-ray andultrasound devices and blood analysis machines. The only dif-ference is in the size of the machines. For example, a micro-MRI machine was developed especially for mice. After com-pleting the primary screening, the scientists can refer to 240different parameters. With this data they can judge whether

SMP MAMMALIAN MODELS

A CLINIC JUST FOR MICE

Valérie Gailus-Durner and Helmut Fuchs

45

the particular mouse line represents a phenotype that is simi-lar to a human disease. Since the GMC was established, theNeuherberg scientists have already examined forty genetical-ly modified mouse lines. "In almost all of them we foundnumerous deviations in comparison to the genetically un-modified mice," Helmut Fuchs explains. For example, theCra1 mouse mutant, a model for hereditary degenerativediseases of the motor neurons, is one of the success stories.The mouse line was discovered with a simple neurologicaltest: When the scientists held up the mouse, the back legs cramped. With increasing age the motor abilities then dete-riorated in this mouse line. A single amino acid exchange inthe intracellular transport protein dynein causes the motorneurons in the medulla to degenerate. But even mouse linesthat have existed for years are put through a thorough medi-cal check-up once again. Only recently this effort was re-warded: In an important cytoskeletal component in a ten-year-old mouse line the scientists detected a pronounced immu-nological phenotype. Until then, attempts to find out theeffects of this mutation had been in vain. Another example isa mouse mutant that represents a model system for Down'ssyndrome. In these mice, which are difficult to breed, GMCscientists not only detected immunological and hematologicaleffects that are known in humans as well, but also changesin the eye and in behavior.

THE ART OF FREEZINGBut how can a scientifically relevant mouse line be pre-served? It would involve too much effort to always keep onbreeding the lines, and it would also be too risky. Thereforethe scientists utilize cryopreservation: Sperm or embryos ofmutant mouse lines are frozen in liquid nitrogen. Then, asneeded, they can be thawed, revitalized and investigated further. However, not all research institutes master this intri-cate technique, which requires a high degree of technical

know-how. That is why the European Mouse Mutant Archive(EMMA) was founded, which offers all scientists the preser-vation of their mouse lines free of charge. Hrabé de Angelisis the director of EMMA and coordinates this unique archive,which is comprised of seven institutes in six European coun-tries. All work processes – freezing, medical examinations,handling, and the transport of both the living mice and thefrozen material are meticulously regulated in standard oper-ating procedures (SOP). Consequently, the work of EMMAmeets the highest quality standards. At the GSF in Neuher-berg mainly sperms are frozen; the partner institutes inEngland, France, Italy, Portugal and Sweden have specializedin embryo freezing.

With its comprehensive concept the GMC has set a new stan-dard in mouse phenotyping. Both in the US and in Asia scien-tists are already building up mouse clinics according to thesame concept. In Neuherberg work is already going on toexpand the GMC. Tests are planned with changed environ-mental conditions: The mice will be exposed e.g. to preciselydefined pathogens or allergens for a certain time. Six newexamination units and an MRI tomograph for mice are planned for this project.

ReferencesGailus-Durner V et al. (2005); Introducing the German MouseClinic: open access platform for standardized phenotyping. Nat Methods 2 (6): 403–404

Reinhard C et al. (2005); Genomewide linkage analysis identi-fies novel genetic Loci for lung function in mice. Am J Respir Crit Care Med. 171 (8): 880–888

Pasche B et al (2005); Sex dependent susceptibility pattern toListeria monocytogenes infection is mediated by differential IL-10 production. Infect. Immun. 73 (9): 5952–5960

COORDINATORProfessor Martin Hrabé de AngelisGSF – National Research Center forEnvironment and Health, Neuherberghrabe@gsf.de

Mouse X-ray

46

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

An important mutagenesis project within the SMP Mam-malian Models is the German Gene Trap Consortium (GGTC).The researchers of this project apply different genetic toolsto decode the function of genes. Using gene trap technology,they are able to randomly mutagenize genes of the mousegenome very efficiently. The bases of this approach aremouse embryonic stem (ES) cells which are transfected withespecially designed gene trap vectors. "Gene trap technologyis at present the most efficient validated mutagenesisapproach for functional gene analysis", explains ProfessorWolfgang Wurst, coordinator of the GGTC. "We perform ourfunctional analyses in mice because they are very similar to

humans with respect to embryonic development, physiologyand behavior. In addition, 99 percent of all human geneshave orthologs in the mouse genome. Moreover, many re-search findings in mouse disease models are relevant to theelucidation of the molecular cause of human disease condi-tions."

TRAPPERS AT WORKIn gene trap mutagenesis, a vector integrates into a gene,thereby disrupting its sequence. As a consequence, the pro-tein coded by the respective gene can no longer be formed,and this can have serious effects in cells of the organism.Initially, researchers assumed that gene trap vector integra-

tions occur at random throughout the genome. However, itbecame apparent that each vector type has its favorite ge-nome integration sites. Thus, to achieve saturation muta-genesis, a whole repertoire of retroviral and plasmid vectorsneeds to be used. First-generation gene trap vectors have incommon that they contain a promoterless reporter/selectorgene cassette. As soon as the vector inserts into an activegene, this reporter/selector gene is expressed via the controlelements of the trapped endogenous gene, which can beeasily visualized in the entire organism. In this way the geneitself announces its disruption and place of activity duringembryogenesis and in the adult.

In some cases, however, gene trap integration cannot bedetected. For instance, when the gene trap vector integratesinto a gene that codes for a secretory protein, the reporter/selector protein will – just like the original protein – be ex-creted from the cell, thereby preventing its detection by invitro selection. For these classes of genes, the GGTC hasdeveloped a special secretory gene trap vector which allowsthe detection and identification of a single peptide containinggenes: U3Ceo. This vector contains a transmembrane pep-tide which causes the reporter protein to be retained on theinner side of the cell membrane, thus remaining selectableand detectable. With the aid of this secretory gene trap vec-tor, mutations in ligands and transmembrane proteins can be

SMP MAMMALIAN MODELS

MOUSE GENES CAUGHTIN THE TRAP

Mouse under investigation

47

detected approximately five times more efficiently than with"classical gene trap vectors". Moreover, the GGTC research-ers have constructed new, conditional gene trap vectorswhich allow the mutation of genes in a time- and space-con-trolled manner. This conditional mutagenesis approach isvery versatile because it can model human disease condi-tions more precisely – for example an adult-onset form ofParkinson's disease which affects specific neuronal popula-tions in the adult brain.

AN EFFICIENT METHODSince its foundation the GGTC has already produced 22,828mutated mouse embryonic stem cell lines. "For 16,862 ofthese clones we could determine the gene trap vector inte-gration site in the genome and identified the disrupted gene.This reflects how successful these strategies are," saysThomas Floss, who is a member of the Wurst research group.In this process, typically some genes were hit several times,but nevertheless 3,779 different genes have been mutated,

among them 279 genes that are associated with humandiseases. "Besides, we detected 173 integrations in genesthat had not been annotated before. Fifty-two integrationstook place in sequence segments that apparently code forregulatory RNAs", he adds. Altogether, the GGTC scientistshave already mutated about 15 percent of all mouse genes:"With our project we have created the largest publicly acces-sible library of mutated mouse embryonic stem cells world-wide. Already several hundred mouse models have beenestablished from our mutant ES cells to investigate genefunctions in vivo", Wolfgang Wurst says. Among them, manyinteresting mouse mutants with disease-relevant phenotypes

were created: for instance, a mouse line with a mutation inthe Nephrin (Nphs1) gene. Nphs1 codes for a structural pro-tein that plays a significant role in urine filtration in the kid-ney. Nephrin (trap/trap) mice excrete an increased numberof proteins with their urine and die soon after birth. This so-called proteinuria is an inherited kidney disease which is fre-quently found in the Finnish population. Another example aremice that bear a mutation in the gene coding for the latenttransforming growth factor beta (TGF-beta) binding protein 4.This protein is a component of connective tissue microfibrilsand regulates the deposition and activation of TGF-beta. Micein which both alleles of this gene are disrupted develop se-vere heart muscle disease, colorectal tumors and lung em-physema. Furthermore, mutants have been established in DJ-1 which is a candidate gene for Parkinson's disease. Someof these established mouse mutants are being studied in theGerman Mouse Clinic, which is also part of the SMP Mam-malian Models. "Thus, we are in a position to generate largenumbers of mutations and analyze the role of these partic-ular genes in development and adulthood and their potentialcontribution to disease phenotypes," Wolfgang Wurst ex-plains.

INTERNATIONALLY RECOGNIZED CONCEPTThe GGTC is currently composed of four collaborating researchgroups: the groups led by Professor Wolfgang Wurst and Dr. Jens Hansen at the GSF - National Research Center for En-vironment and Health in Neuherberg, the research group ofProfessor Harald von Melchner at the University of Frankfurt/Main and the team of Professor Patricia Ruiz at the Charité in Berlin. The GGTC was significantly involved in the foundingof the International Gene Trap Consortium (IGTC) whose mem-bers include BayGenomics, Berkeley, CA, USA; the SangerInstitute, Hinxton, UK; the Fred Hutchinson Cancer ResearchCenter, Seattle, WA, USA; the University of Manitoba, Winni-peg, MB, Canada and the Mount Sinai Hospital, Toronto, ON,Canada. This international consortium is involved in a world-wide effort striving for complete mutagenesis of the mousegenome and endeavoring to provide animal models for everyhuman disease.

ReferencesHansen J (2003); A large scale, gene-driven mutagenesisapproach for the functional analysis of the mouse genome.Proc. Natl. Acad. Sci. USA 100: 9918–9922

Schnutgen F (2005); Genomewide production of multipurposealleles for the functional analysis of the mouse genome. Proc. Natl. Acad. Sci. USA 102: 7221–7726

COORDINATORProfessor Martin Hrabé de AngelisGSF - National Research Center forEnvironment and Health, Neuherberghrabe@gsf.de

The German Gene Trap Consortium publishes all se-quence data of its gene traps at www.genetrap.de, onthe websites of the International Gene Trap Consortium(www.igtc.org.uk/cgi-bin/advanced_search, select‘GGTC' as cell line source prior to clicking on search)and of the National Center for Biotechnology Information(NCBI; www.ncbi.nlm.nih.gov, in 'nucleotides' searchfor 'gene trap GGTC' to obtain a list of the GGTC genes)as well as on the Ensembl website (www.ensembl.org/Mus_musculus, enter gene of interest and then theDAS source: select gene trap). Gene traps of the GGTCand of other origins can be found at ftp://ftp.ncbi.nih.gov/genomes/M_musculus/GeneTrap.

48

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

Munich, Spring 2005. In seminar room 5 in the GrosshadernClinic, 26 young scientists are sitting engrossed in their workat computer terminals. In four days the participants are tolearn as much as possible about the statistical analysis ofDNA microarray data. On this Wednesday morning the topicsare molecular diagnosis, classification by nearest shrunkencentroids and support vector machines, and model assess-ment and selection.

"Just a few years ago everything revolved around questions ofbasic research," says Dr. Rainer Spang of the Max PlanckInstitute for Molecular Genetics (MPIMG), in his introductionto the third course day. "Today the main concern is utilizingthe molecular data for the diagnosis and therapy of patients."The data analysis often makes decisions necessary thatdirectly affect the patient and must therefore be made verycarefully. However, it is still uncommon for a medical doctoror biologist to master algorithms, plot analyses and performhierarchical clustering. "That is why we decided to offer semi-nars on selected topics even back in the first NGFN fundingphase," says Dr. Florian Markowetz of the MPIMG in Berlin.The courses in microarray data analysis take place four timesa year, twice in Munich and twice in Heidelberg. All NGFNmembers are eligible to register – and if there are any open-ings left – scientists from other institutions can register aswell. "Ideally the participants should be familiar with at leastone programming language. However, it is especially impor-tant for them to refresh their knowledge of statistics prior tothe course, because without statistics you cannot evaluatedata," Florian Markowetz recommends. "People interested in

participating can find a number of links to reference litera-ture and software to help get acquainted with the topic atwww.compdiag.molgen.mpg.de/ngfn/.

THREE BIG "I's" – INTEGRATE, INTERPRET AND INFORMThe courses on the analysis of microarray data are part ofthe module "Service, Training, Consultation and QualityManagement" of the Systematic-Methodological Platform

(SMP) Bioinformatics. They are only one of many activities which have beeninitiated to transfer know-how to thepartners of the NGFN. The scientistsalso develop software programs, forexample, and provide consulting servic-es for intricate problems. Apart fromthat, they work closely with numerousprojects from the Disease-orientedGenome Networks and help with statis-tical and mathematical problems. "Inthe previous NGFN funding phase wewere involved in more than 30 projectsfrom all of the clinical areas," saysProfessor Roland Eils of the GermanCancer Research Center (DKFZ) inHeidelberg, who coordinates the SMPBioinformatics. However, the achieve-ments of the SMP Bioinformaticsextend far beyond specific support for

individual projects. Besides the module "Service, Training,Consultation and Quality Management" the bioinformaticiansare concerned with data management, the standardization ofdata and data analysis for the entire NGFN. "We speak of the"3-I" approach: integrate, interpret and inform," Professor Eilsexplains.

iCHIP"Integrate" is a keyword for data management, an area inwhich bioinformaticians conceive, develop and install data-bases and standards for the various NGFN findings. "A com-plex research network like the NGFN can only work efficient-ly when all of its members use the same vocabulary and canalso access the results of the other groups," Professor Eilsexplains. An example is the database iCHIP. It was conceivedand introduced specifically for clinical research, already backin the last NGFN phase. iCHIP helps scientists manage theirfindings from investigations with DNA chips, Affymetrix chips and molecular cytogenetic experiments and to linkthem to clinical data. iCHIP is already firmly established inseveral Disease-oriented Genome Networks. Next, Roland

SMP BIOINFORMATICS

BIOINFORMATICS – WHATTHE DATA REALLY MEAN

Elena Reiffel, Stefanie Scheid, Florian Markowetz and Stefan Bentink

49

Eils and his colleagues want to extend iCHIP so that findingsfrom cellular assays can also be combined with data fromRNAi experiments, proteomics research, matrix-CGH and tissue microarrays (TMA). "Clinical research is changing con-stantly. New methods are being established and thus newdata is generated," Roland Eils explains. "That is why general-ly valid standards must continually be developed further." Toalways meet the current quality demands, the SMP Bioinfor-matics works closely with international committees, for in-stance with the European Bioinformatics Institute.

INTERPRETING DATA CORRECTLYDr. Jörg Rahnenführer of the Max Planck Institute for Infor-matics in Saarbrücken is active in the field of data analysis.Together with his colleagues, he works out the bases for the statistical analysis of microarray data. In addition, as amathematician he is involved in clinically oriented projects.Thus, in collaboration with medical doctors from theUniversity of Düsseldorf, the Charité in Berlin and theSaarbrücken University Clinic, Jörg Rahnenführer has devel-oped a prognostic marker with which the survival time of

tumor patients and the period until a relapse can be estimat-ed. Such time predictions are very important to classify atumor and to select the suitable therapy," says Dr. BerndWullich of the Clinic for Urology and Children's Urology inSaarbrücken, who is significantly involved in the project. Upto now, physicians have primarily relied on clinical and histo-logical parameters in assessing a tumor. But interest in ge-netic markers is increasing. With their help, scientists arehoping to predict the course of cancer in a patient muchmore accurately. "A considerable number of research find-ings already exist that are based on individual genetic mu-tations in order to assess a tumor," Jörg Rahnenführer says."But actually we need more complex analyses that take allgenetic changes in the course of a cancer disease into con-

sideration. Moreover, these analyses must be capable of eval-uating the probabilities for possible disease courses."Together with clinicians, he developed a method to calculatethis, the so-called genetic progression score (GPS). Geneticchanges that have already been recorded are also included inthe calculation of this score. "GPS allows us to estimate howthe tumor, with the greatest probability, will develop in thepatient," Bernd Wullich explains. GPS has already proved tobe a suitable prognostic marker for both glioblastomas andprostate cancer to estimate the further course of the dis-ease. Since these are two tumor types with quite differentgenetic backgrounds, the scientists are convinced that theirapproach can be applied to other kinds of cancer.

Back to Munich. At 5:00 p.m., after seven hours crammedfull of exercises, lectures and discussions, the scientists turnoff their computers. Many a head is spinning, but most of theparticipants seem satisfied as they leave the GrosshadernClinic. Anja Weigmann from Hannover is also happy that shetook the time for this course: "For me, today was very re-warding. This afternoon we evaluated and discussed my data,which I brought with me from Hannover because we simplyhad not got any further. Now, at last, I have reliable results."

ReferencesRalser M et al. (2005); An integrative approach to gain insightsinto the cellular function of human ataxin-2. J Mol Biol 346 (1): 203-214

Brors B (2005); Microarray annotation and biological informa-tion on function. Methods Inf Med 44 (3): 468-472

Lottaz C, Spang R (2005); Molecular decomposition of com-plex clinical phenotypes using biologically structured analysisof microarray data. Bioinformatics 21 (9): 1971-1978

COORDINATORProfessor Roland EilsGerman Cancer Research Center,Heidelbergr.eils@dkfz-heidelberg.de

"The primary purpose of our SMP is tosimplify the utilization of data generat-ed in the NGFN initiative. We want to

facilitate the transformation of data intobiomedical knowledge. Our platform

has the infrastructure and know-how forthis 'data to knowledge' transfer."

Professor Roland Eils

Florian Markowetz and Martin Lange

50

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

The mechanisms of major common diseases remain elusive –even genetic researchers have not yet really been able to findtheir causes. And no wonder: In complex diseases like can-cer, cardiovascular diseases and allergies hundreds of differ-ent genes interact with each other. In addition, environmentand lifestyle also play a role. To fully understand this interac-tion it is not enough to study a few affected persons – largesample surveys and the most advanced high-throughput anal-ysis are necessary. Usually, thousands of people must takepart in the relevant studies. It has just recently become pos-sible to carry out genetic analyses on this large scale and tocorrelate the results to environment and lifestyle variables.These recent advances have greatly increased the demandfor expertise in the etiology, distribution and control of hu-man diseases in groups of relatives and – with the inheritedpredisposition to diseases – in populations. In the NGFN thisgenetic epidemiological expertise is provided by eight re-search groups located at universities or research institutesfrom all over Germany (Berlin, Bonn, Göttingen, Heidelberg,Kiel, Lübeck, Marburg and Munich). Coordinated by ProfessorMax P. Baur (University of Bonn), they form the Systematic-Methodological Platform (SMP) Genetic Epidemiological Meth-ods (GEMs) and closely cooperate with the Disease-orientedGenome Networks and the high-throughput genotyping plat-form to further promote research on complex diseases. Butalso new methods are needed. To accomplish this, four spe-cialized research modules within the SMP GEMs work on qual-ity improvement and data standardization (Module A), thedevelopment of new statistical methodologies to analyze anever increasing complexity of genetic data (Module B), appro-priate training programs in genetic epidemiology (Module C)– or the two large bio-database projects (Module D).

BIO-DATABASE IN GERMANY'S FAR NORTHThe first project is coordinated by the gastroenterologistProfessor Stefan Schreiber and the bioinformatician and stat-istician Michael Krawczak at the University Clinic Kiel. Thename of the research project: PopGen (Population GeneticRecruiting of Patients and Control Groups). The goals of theKiel researchers are ambitious. Approximately 25,000 pa-tients in the northern part of the state of Schleswig-Holsteinwho suffer from one of twelve chronic diseases are to becontacted, and data is to be gathered on them. StefanSchreiber, who is at the same time one of the spokespersonsfor the NGFN, says: "Individual genes that play a role in thesediseases are already known from smaller studies. Within thescope of PopGen and using a larger patient collective, wewant to elucidate how great their influence on the course ofthe disease actually is. Moreover, we hope to gain informa-tion about how often these disease-causing gene variantsactually occur. This objective can only be achieved through aclose cooperation of clinicians and methodological scientists,just like we have in PopGen." To enable a comparison withthe "normal population", the data of 25,000 randomly select-ed healthy inhabitants of Schleswig-Holstein will be collected.A sample of 30 ml of blood will be taken from every partici-pant of PopGen. Diseased patients will also answer questionsabout their lifestyle and possible risk factors. It is planned torepeatedly question half of them briefly about the course oftheir disease every six months.

Other project groups in the NGFN can turn to PopGen withspecial requests. Stefan Schreiber explains, "For instance,when colleagues want to know how many people bear a cer-tain variant of the gene X that plays a role in asthma, wecarry out the relevant analyses in Kiel and pass the resultson to the partners." To prevent an abuse of the project, espe-cially through a linking of genetic and personal data by thirdparties, PopGen is subject to especially strict data protectionguidelines. A key aspect of these guidelines is that the bio-logical data and personal data are managed in separate data-bases.

SMP GENETIC EPIDEMIOLOGICAL METHODS

ON THE WAY TOWARDS A GENETIC EPIDEMIOLOGY

51

GENOMIC DNA FROM TWO-AND-A-HALF DECADESThe second large population-based project of the SMP GEMsis located at the other end of Germany. The KORA-gen proj-ect is based on a data and sample pool which has beencreated and maintained since the middle of the eighties inthe framework of the so-called MONICA and KORA studies(MONICA stands for Monitoring of Trends and Determinantsin Cardiovascular disease, KORA for Cooperative HealthResearch in the Region of Augsburg). In charge of the projectare scientists of the GSF – National Research Center forEnvironment and Health. While in MONICA and KORA thefocus at first was "only" on identifying risk factors for cardio-vascular diseases, the research spectrum has meanwhilebeen expanded to include diseases such as cancer and dia-betes. To date, the researchers have gathered data in fourphases from more than 20,000 men and women between theages of 25 to 74 years. In contrast to PopGen, MONICA andKORA do not specifically collect data from people with cer-tain diseases, but rather from a cross-section of the popula-tion that is as representative as possible.

All MONICA/KORA participants undergo medical check-ups,provide information about their living habits and provide theresearchers with urine and blood specimens. KORA spokes-man Professor Erich Wichmann: "Thus data and bio-samplesrepresentative for the population are available to researchthe most varied problems; these include the genomic DNAfrom about 18,000 individuals." This bio-database can serveas the basis for estimating the incidence of different geno-types in the general population and enables a comparisonwith certain patient groups. KORA-gen was established tofacilitate the use of the extensive data and sample collectionfor interested researchers. When they send a specific re-quest, KORA-gen processes the needed information in codedform and carries out the relevant genotypings. As additionalservice KORA-gen provides expert advice and aid in ques-tions concerning study design and statistical evaluation. LikePopGen, KORA-gen is liable to strict regulations with regardto data protection.

Max P. Baur, coordinator of the SMP GEMs, describes a vi-sion of the impact of genetic-epidemiological studies: "In acooperative effort of clinical scientists, molecular geneticists,and genetic epidemiologists we hope to find genetic compo-nents which are causally involved in the etiology of complexdiseases." The verification of these risk factors on the popu-lation level will then be the first step towards defining indivi-dual genetic risks for these diseases. The possible conse-quences for prevention (lifestyle), diagnostics and treatmentare manifold, but it would be unrealistic to expect quick suc-cess. Genetic medicine for complex diseases will to a largeextent consist of individualized prevention.

More information on the genetic-epidemiological projects in the NGFN is available at: www.ngfn.de

www.popgen.de www.gsf.de/kora-gen

COORDINATORProfessor Max P. BaurUniversity of Bonnmax.baur@ukb.uni-bonn.de

Christian Gieger (KORA-gen)

52

SYSTEMATIC-METHODOLOGICAL PLATFORMS (SMP)

To do their work, gene researchers need genetic referencematerial – optimally in easily accessible and well-processedform. So that each single NGFN research group need notmaintain its own biobank, the German Resource Center forGenome Research (RZPD) was founded in 1995 – at that timeunder the auspices of the German Human Genome Project(DHGP). The RZPD is the holder of two records: It is the larg-est service center for functional genome research in Europeand has the most comprehensive public clone collection inthe world. Located in Berlin, the RZPD houses more than1,200 genomic and cDNA libraries stored in deep freezers at-80° Celsius. Altogether there are more than 35 million clonesets from 32 different species, in particular from humans andrelevant model organisms. NGFN scientists and also otherresearch groups can access the services of the Berlin not-for-profit company. At present more than 9,000 users draw onRZPD's products and services.

The materials of the RZPD can be ordered via the Internet(www.rzpd.de). Using the search engine GenomeCube®, inter-ested researchers can run a search to find out which materi-als relevant to their research are available and can be or-dered. All materials are linked to internationally standardizedkeywords and accession numbers, so that a systematicsearch can be made with different search entries. Genes andgene products are the most important and largest compo-nents of GenomeCube®. For each gene of a specific speciesthe RZPD can provide different clones, DNAs or antibodiesagainst the respective gene product. Every year the RZPDdistributes approximately 60,000 clones in this way.

RESEARCHERS CAN FOCUS ON THEIR RESEARCH"Our main task is to provide research groups with standard-ized biological reference materials of high quality. In this waywe support the scientific progress of the projects and con-tribute to the most efficient use possible of available person-nel, technical and financial resources for genome research,"says Dr. Johannes Maurer, scientific managing director of the

SMP SERVICE AND RESOURCES

RZPD IN BERLIN – 35 MILLIONDEEP-FROZEN CLONES

Uwe Radelof, Ronny Kalis and Tom-Oliver Weiss (scientists at the RZPD)

53

RZPD. In other words: Gene researchers can concentrate ontheir actual biomedical research objectives and leave a lot ofthe tedious preliminary work to their RZPD colleagues. At thesame time, due to the consistent high quality of RZPD's ma-terials, the data in different experiments can be comparedvery well.

Usage rights to the results gained with RZPD materials orinformation remain with the researchers: "The RZPD raisesno claims to the results and pursues no commercial interestof any kind with them," Maurer adds. "In addition, we consol-idate the information gained into a database, which is freelyaccessible to the scientific community, and link it to addition-al information from public databases."

TECHNOLOGICAL SUPPORTIn addition to genes, DNA, etc., the Berlin resource centeroffers technological support to scientists in industry andresearch. Included are e.g. customized high-throughput tech-nologies and automation solutions as well as extensive ser-vices for analyzing gene activity and gene regulation alongwith their bioinformatics analysis. For all of its services theRZPD guarantees consistent high quality. All products complywith the strict guidelines of the international standard DIN EN ISO 9001:2000 for quality management systems.Furthermore, close collaboration with German and internatio-nal academic and industrial partners ensures the highestscientific and technical standard.

The RZPD is managed by a scientific director and an adminis-trative director. In addition, an internationally renownedscientific advisory board oversees all activities. Shareholdersof the RZPD are the Max Planck Society (50 percent), theMax Delbrück Center for Molecular Medicine (25 percent)and the German Cancer Research Center (25 percent).

ReferencesStelzl U et al. (2005); A human protein-protein interaction net-work: A resource for annotating the proteome. Cell 122 (6): 957–968

Kittler R et al. (2004); An endoribonuclease-prepared siRNAscreen in human cells identifies genes essential for cell divi-sion. Nature 432 (7020): 1036–1040

Weaver T et al. (2004); Sharing genomes: an integratedapproach to funding, managing and distributing genomic cloneresources. Nat Rev Genet. 5 (11): 861–866

COORDINATORDr. Johannes MaurerGerman Resource Center forGenome Research, Berlinj.maurer@rzpd.de

DNA chip

54

EXPLORATIVE PROJECTS (EP) PROJECT LEADERDr. Zoltan IvicsMax Delbrück Center forMolecular Medicine, Berlinzivics@mdc-berlin.de

"SLEEPING BEAUTY" FINDS ITSWAY INTO GENOME RESEARCH

In the middle of the last century Barbara McClintock hypoth-esized for the first time the existence of "jumping genes" thatchange their positions in the genome. She surmised thatthese mobile genetic elements (transposons) are responsiblefor the changed color patterns that maize cobs develop in thecourse of generations. This theory, however, was questionedin scientific circles for a long time. It was not until about

twenty years later that the existence of these transposonswas definitely proven through improved molecular biologicalmethods – and that Barbara McClintock was awarded theNobel Prize for her discovery. Today, transposons are an im-portant tool for genome researchers. In many model organ-isms such as Drosophila melanogaster and Caenorhabditiselegans, they are used to systematically decode the functionof genes in large mutagenesis screens. In vertebrates, how-ever, this technology could not be used up to now, since only"defective" transposons are known for these organisms. Thesequence of these transposons is so severely mutated thatthey have lost their ability to "jump".

Dr. Zoltán Ivics and his team from the Max Delbrück Centerfor Molecular Medicine (MDC) in Berlin-Buch are now en-abling the use of transposon technology in vertebrates aswell. Through a comparative phylogenetic approach they havereconstructed the sequence of an active transposon from thefragments of inactive elements in the genomes of fish. "Likewith the Prince's kiss in the fairy tale, we have wakened this

transposon after probably20 million years of sleep byreversing the mutations thathad switched off the trans-poson. That is also the reason why we named it'Sleeping Beauty'," he jokes.

However, transposons donot only find applications in basic research. In collabo-ration with Dr. Thomas Flossat the GSF in Neuherberg,Dr. Ivics' laboratory team isusing it to trace a geneticdisorder: the Williams-Beuren syndrome. This raredisease generally leads tomental retardation, but isoften associated with partic-ular abilities, e.g. with pro-nounced musical ability. Thegenetic causes of the dis-ease are largely unknown,but a microdeletion in the

gene for elastin could be detected in 95 percent of the pa-tients. Elastin is the main component of the elastic fibers in the extracellular matrix. The expression of elastin is influ-enced by about 30 additional genes (so-called modifier genes)that are located in a region of 500 bp around the elastingene. Since the deletions in the elastin gene in the mousemodel do not lead to the phenotype of the disease, it is as-sumed that the patients have additional mutations in the modifier genes. The objective of this Explorative Project is to generate a series of mutations in the Williams-Beuren syndrome region of the mouse genome using the "SleepingBeauty" transposon, with the hope that these mutations willrecapitulate the phenotypes observed in humans. Analysis ofthese mutations and the affected genes will help researchersto uncover the genetic basis of this disease.

DNA gel electrophoresis

55

EXPLORATIVE PROJECTS (EP)

functions to the two fractions. The cells of the FDF are com-mitted progenitor cells which differentiate and mature intodifferent specialized cell types, whereas cells of the SDF, on the other hand, are destined for the purpose of self-re-newal. For their experiments the scientists used stem cellsderived from human bone marrow as a model. "Our aim wasto pinpoint the molecular and genotypic differences betweendaughter cells that demonstrate the self-renewing potentialversus those that are destined to differentiate and becomespecialists. Using a human transcriptome microarray, whichrepresents about 95 percent of human gene sequences, wehave compared the expression profiles of the two cell formswith each other and have been successful in identifyingnumerous genes which are up-regulated in the SDF as com-pared to those in the FDF," Anthony Ho explains. Further-more, the scientists also found morphological and functionaldifferences between the SDF and FDF: Primitive and self-renewing stem cells, found in the SDF, form numerous mem-brane protrusions, so-called lamellipodia, and express strong-ly the protein CD133 at the tip of these protrusions.

The interaction between stem cells and their surroundings –in particular with the cellular microenvironment – play a cru-cial role in determining the division behavior and subsequent-ly the long term fate of the stem cells and their progenies.Stem cells communicate both with cells of their own type aswell as with other cells that constitute the cellular microenvi-ronment. The results of the Heidelberg researchers suggestthat this communication takes place via junctions and junc-tion complexes. Concurrently, the scientists of the group arestudying how the cellular microenvironment influences thegenetic profiles of stem cells. Anthony Ho emphasizes,"Decoding which factors govern the decision process self-renewal versus differentiation will not only lead to a betterunderstanding of stem cell control and stem cell functions,but will provide a sound foundation for identifying moleculesand signal transducers that play a key role in the various dif-ferentiation process. This knowledge will subsequently leadto a breakthrough for reproducibly controlling stem cell fatesin vitro and in vivo."

PROJECT LEADERProfessor Anthony D. HoUniversity of Heidelberganthony_dick.ho@urz.uni-heidelberg.de

Many areas of biomedical research place great hope in stemcells, but the molecular secret of these "miracle weapons" is still far from being understood. One reason is that manyexperiments are neither comparable nor reproducible. "Tobetter compare and utilize findings from global research ef-forts on stem cells, we need common definitions and uni-form standards. We urgently need better and more precisemethods for characterizing stem cells that are rapid and re-producible," says Professor Anthony Ho, visualizing the goalsof his Explorative Project "Cellular and Molecular Signaturesof Human Pluripotent Stem Cells".

Stem cells possess a unique characteristic distinguishingthem from other cells: They have the dual abilities to self-renew and to differentiate into different cell types. To do thisthey must undergo asymmetric divisions. Thus, one motherstem cell gives rise to two functionally unequal daughtercells. In previous years of research using hematopoietic stemcells as models, the research team has shown that stem cellscan be separated into two groups according to the kinetics ofcell division: a slow-dividing fraction (SDF) and a fast-dividingfraction (FDF). The Heidelberg researchers could match cell

STEM CELLS – MUCH DEBATED,LITTLE UNDERSTOOD

Morphology of slow- and fast-dividing hematopoietic progenitor cells. The CD34+/CD38- fraction is enriched in primitive hematopoietic stem cells.Using the fluorescent membrane dye PKH26, these cells were divided in a slow-dividing fraction (SDF; A and B) that keeps the membrane dye and a fast-dividingfraction (FDF; C and D) that dilutes the dye to their progeny. Cells in the SDF arehighly polarized and demonstrate higher migratory activity and podia formation.

A B

DC

56

EXPLORATIVE PROJECTS (EP)

Irritable bowel syndrome (IBS) and "nervous stomach" ornon-ulcer dyspepsia (NUD) are widespread in the population,affecting approximately one in ten people. Every year thesedisorders lead to immense costs for the healthcare system.The patients suffer from muscle spasms in the stomach orbowel, persistent pain as well as constipation or diarrhea.

Symptoms may last for a long time and recur again andagain, which is especially stressful for the affected person.Despite the high incidence, little has been known until nowabout the etiology of the diseases. Scientists assume thatthe stimulation process is disturbed between the entericnervous system (ENS) of the gastrointestinal system and thebrain (central nervous system, CNS). The serotoninergic

system is also very likely involved in the etiology of the dis-eases. Serotonin is a messenger substance which contributesto the regulation of movements and excitability both in thebrain as well as in the gastrointestinal tract. For this reasonresearch is specifically focusing on combating gastrointesti-nal diseases with the serotoninergic system.

The team from the Institute for Human Genetics in Heidelbergis searching for the cause of functional gastrointestinal dis-eases in the serotonin type 3 receptor system (HTR3), one ofthe seven known serotonin receptor types. "By conducting ahomology search in databases with sequences of the humangenome, we found two novel HTR3 genes: HTR3D and HTR3E.Both genes reside in close proximity and on the same chro-mosome," explains Dr. Beate Niesler, head of the ExplorativeProject "The Role of Serotonin Receptor Genes in the Etiologyof Gastrointestinal Disorders". Through expression analysis,the researchers were able to determine where the two recep-tor genes may play a role: HTR3D is found in the kidneys,colon and liver, while the expression of HTR3E is limited tothe digestive tract. "All further HTR3 genes are expressed innumerous tissues – including the brain. The limited activity ofthe HTR3E gene in the gastrointestinal tract may be an indi-cation that it is relevant to the pathomechanism of gastro-intestinal diseases. We will now look for mutations in thisgene, e.g. in irritable bowel syndrome patients. These muta-tions could be the cause of the gastrointestinal diseases,"explains Dr. Beate Niesler. The genetic investigation shouldbe a step forward in clarifying the role of the receptor in theetiology of the diseases. To help patients as soon as possible,the results will be collected in a public database and madeavailable to medical and scientific institutions.

GASTROINTESTINAL DISEASES: WHAT ROLEDOES A SEROTONIN RECEPTOR PLAY?

PROJECT LEADERDr. Beate NieslerUniversity of Heidelbergbeate_niesler@med.uni-heidelberg.de

Mapping of HTR3CDE on chromosome 3q27

57

EXPLORATIVE PROJECTS (EP)

Nanotechnology meets protein chemistry: The project "NearInfared Nanocrystal-based Quantitative Protein Arrays" aims to detect proteins by using antibodies labeled with nano-crystals. These crystals are characterized by a wide absorp-tion profile and a narrow emission spectrum. Due to the dif-ferent emission spectra of the individual crystals, the emittedsignals can be perceived clearly separate from each other.Thus, they enable simultaneous detection of many differentproteins in one solution. A further advantage – the crystalsshow excellent photostability, enabling longer measurementtimes.

Proteins are often characterized by their binding to specificantibodies. High-throughput techniques for this purpose usemicroarray slides on which many different capture antibodiesare immobilized to facilitate the simultaneous detection ofmultiple proteins. However, in complex mixtures like serum orcellular lysates, errors can occur: The high protein concentra-tion in these solutions frequently causes the antibodies tobind to unrelated proteins as well. These unspecific bindingsseverely limit the sensitivity of the high-throughput proce-dures. Scientists at the German Cancer Research Center(DKFZ) in Heidelberg go for more precision: To increase thesensitivity of the procedure, they monitor the antibody-pro-tein interactions. To this purpose, Dr. Ursula Klingmüller andher team use two antibodies: Both of them must recognizethe protein and bind to it for the binding to be viewed as spe-cific. After this a signal is emitted. But how can the bindingof both antibodies be linked to signal emission? The idea ofthe Heidelberg scientists is based on the principle of "fluo-rescence quenching" – a special case of "fluorescence reso-nance energy transfer" (FRET). Of the two antibodies used,one is coupled to a nanocrystal, the other to a so-called"quencher". Light of a specific wave-length, which stimulatesthe nanocrystal, is used to detect the antibody binding. Assoon as the stimulated electrons of the nanocrystal revertback to their original state, they emit energy of a differentwave-length. This emitted light corresponds to the absorptionspectrum of the quencher. If only one antibody binds to the

protein, this energy is detected as fluorescence. However, ifboth antibodies have bound to the same protein and thus arein close proximity to each other, the quencher absorbs thelight which the nanocrystal releases. Thus, if both antibodiesbind to the protein, the emitted fluorescence is reduced orquenched. "Many diseases are based on subtle changes inprotein structure or quantity. High-throughput techniquesmake it possible to recognize these changes rapidly and reli-ably. Our technique is a big step in advancing high-through-put technology. Through our use of nanocrystals and anti-bodies we will increase the sensitivity of detection enor-mously," Ursula Klingmüller explains.

SO TEST THEREFORE,WHO BIND FOREVER …

PROJECT LEADERDr. Ursula KlingmüllerGerman Cancer Research Center,Heidelbergu.klingmueller@dkfz.de

Specific detection in the protein array by nanocrystals (quantum dots) in thenear-infrared range. The protein of interest (red) is immobilized by a capturingantibody and detected by detection antibodies coupled to nanocrystals (QD;blue) or quencher (QE; orange).

QD QE

58

EXPLORATIVE PROJECTS (EP)

A method for detecting precursor forms of individual cancercells among thousands of healthy cells – for oncologiststoday that still seems a long way off. But when will it becomea reality? "Right now we are still far away from large-scalescreening at the single-cell level, but we are working on it.Moreover, we are searching for additional diagnostic markersto characterize single cells as potential cancer cells," ex-plains Dr. Michael Speicher, head of the Explorative Project"Exploration of the 3D Genome Organization and Its Impacton Gene Expression during Tumor Development and Agingwith Novel High-Throughput Multicolor Imaging Technique".Until now, a change in the number of copies in certain ge-nome segments has been the primary indication of a cancercell. However, it is not yet clear to what extent chromatin orposition rearrangements take place prior to that. If these dotake place, they could be used as markers for very earlystages of tumor evolution.

The aim of the Munich scientists is clear: They want to eluci-date the relationship between chromatin structure and geneexpression and thus identify disease- or age-related changesin the architecture of genes. To accomplish this, 3D maps ofgenomes from different cell systems are produced and corre-lated with the corresponding gene expression data. By ana-lyzing defined chromosome regions, scientists can discoverwhat influence the position in the genome has on the expres-sion of a gene. High-resolution 3D maps are a prerequisitefor these analyses: Their purpose is to describe the spatialarrangement of certain DNA segments down to the positionof individual genes. Next, using this information, the re-searchers are focusing their search on disease-relatedchanges which can be used as early diagnostic markers, e.g. for cancer. But even structural changes of the chromatin,which are caused by aging processes, could be studied inthis way. In old age gene activity changes – a process aboutwhich little has been known until now. "Nevertheless, thisinformation is important. It is the only way to distinguishbetween age-related and disease-related changes," explainsMichael Speicher.

An important prerequisite for three-dimensional positionalmapping is the simultaneous analysis of several genome seg-ments in numerous cell nuclei. However, these comparativestudies are also necessary for the later application: In orderto definitely diagnose a disease, many different cells must bescreened for the diagnostic markers. At present this is still avery complicated procedure that has to be repeated for eachcell nucleus. It is not suited to high-throughput methods.Apart from the analysis of the 3D organization in various cellsystems, a further goal of the Munich scientists is thereforeto carry out these analyses in high-throughput procedures. To achieve this, they want to employ a new procedure whichthey have developed, based on FISH methodology (fluores-cent in situ hybridization). For this method, probes are usedwhich hybridize to specific regions in the genome. Using fluo-rochrome-labeled antibodies which bind to the probes, theseregions can be visualized. Michael Speicher's team is nowdeveloping a microscope which will enable three-dimensionalimaging with high spatial resolution and analysis with a highdegree of automation.

THE ARCHITECTUREOF GENES

PROJECT LEADERDr. Michael SpeicherGSF - National Research Center forEnvironment and Health, Neuherbergspeicher@humangenetik.med.tu-muenchen.de

Chromosomes in three-dimensional space: Using new techniques DNA can be dyed in its natural position and environment.

59

EXPLORATIVE PROJECTS (EP)

FIGHTING HEART DISEASES WITHCOMBINED STRENGTHS

PROJECT LEADERProfessor Gerd WalzUniversity of Freiburgwalz@mm41.ukl-uni-freiburg.de

"It's the combination that counts!" jokes Professor GerdWalz, head of the research project "Microarray Validation ofCardiovascular Risk Factors". His "winning combination" ishis project team, which is made up of medical doctors, biolo-gists and physicists. Together they are searching for factorsleading to arteriosclerosis in terminal renal failure. Merely ashort time after startingthe treatment, dialysispatients often sufferfrom coronary heartdisease, which is other-wise more likely to befound in older people.

The search begins in theDepartment of InternalMedicine of the Uni-versity Clinic in Freiburg.By means of microar-rays, Professor GerdWalz and his colleaguesare narrowing down thenumber of potential culprit genes. They com-pare cDNA samples thatwere taken from patientsbefore and during a dial-ysis session. In addition,samples of healthy peo-ple of the same age are included in the evaluation. The re-searchers are especially interested in the genes that showdifferent activity in the three sample groups. Using this meth-od, the medical scientists hope to be able to narrow thenumber of potentially relevant genes down to between 1,000and 750. At this point Professor Jens Timmer, a physicist,takes over. As data specialist he statistically determines themost promising candidates. Out of the 750 to 1,000 genes,probably about 150 candidates still remain on the short-list.

The last station in the search is the Department for Bioinfor-matics and Molecular Genetics of the Institute for Biology inFreiburg. Here biologists on Professor Ralf Baumeister's teamtest the remaining candidate genes on the animal model.They switch off the genes in the nematode Caenorhabditiselegans – and thus obtain information about the functions of

the gene products. "Our threefold approach enables us toclose the information gaps between the high-throughputscreenings, the genetic backgrounds and the diseases – amodel that can set a precedent," Gerd Walz says delightedly.

Robotics in microarray validation

60

EXPLORATIVE PROJECTS (EP)

Today more and more people attain a "biblical" age whileoften still leading healthy and active lives. Epidemiologicalstudies indicate that the phenomenon of healthy or success-ful aging in humans is inherited to about 30 percent. The Ex-plorative Project (EP) "Genetic Etiology of Human Longevity"aims to characterize the genetic factors which contribute to a long and healthy life as well as to identify molecular path-

ways that are associated with the physiology of aging and/orage-related diseases. This EP is a multicenter collaboration inwhich the following institutes participate: Institute for ClinicalMolecular Biology, University Hospital Schleswig-Holstein,Kiel (Professor Stefan Schreiber); Department of Dermatol-ogy, Charité University Medical Center Berlin (ProfessorChristos C. Zouboulis); Max Planck Institute for MolecularGenetics, Berlin (Professor Hans Lehrach).

Already during the NGFN's first funding phase, scientists atthe Institute for Clinical Molecular Biology in Kiel collectedDNA samples from more than 2,500 long-lived (95 years andolder) and younger control individuals (60–75 years) fromacross Germany. Of these, 430 samples are from centena-rians. The DNA collection in Kiel is thus one of the largestbiodatabases of its kind worldwide. The first clues aboutgenes that can extend life span were gained from studies

on various model organisms such as flies, nematodes andmice. The scientists in Kiel are currently conducting a large-scale study to investigate to what extent these "longevitygenes" also play a role in humans. In their search the resear-chers make use of single nucleotide polymorphisms (SNPs),the most common type of genetic variation in the human ge-nome. Thousands of SNPs are analyzed in several hundred

functional candidates, which in modelorganisms are known to be relevant tolongevity and aging (e.g. energy meta-bolism, anti-oxidation or apoptosis). TheSNP data of the long-lived individuals arethen compared to those of the youngercontrol group. The idea is to find outwhich of the tested genetic variantsoccurs more or less frequently in the eld-erly than in the younger population. Theso-identified SNPs indicate genes thatmay be involved in longevity or aging inhumans.

In addition, the researchers of the EP arealso interested in the metabolic pathwaysof aging. Using DNA arrays, the teams in Berlin are comparing the expressionpatterns of skin cells of different origins:

skin cells from human cell cultures, from human tissue sam-ples of twenty-year-olds and sixty-year-olds, and from tissuesamples of long-lived mice. Differences in the expression pro-files will allow conclusions about the molecular pathways ofaging. The scientists are searching for genes that are more or less highly expressed in tissues of younger people than inthose of elderly individuals. Subsequently, using the technol-ogy of siRNA, the Berlin scientists plan to identify the func-tional role of these genes on cell life and behavior and com-pare their findings with the Kiel data.

METHUSELAH'S HEIRS – WHERE IS THE GENETIC SECRET?

PROJECT LEADERProfessor Stefan SchreiberUniversity Hospital Schleswig-Holstein,Campus Kiels.schreiber@mucosa.de

High-throughput genotyping

61

EXPLORATIVE PROJECTS (EP)

In some people an effective protection against AIDS(Acquired Immune Deficiency Syndrome) is encoded in theirgenes. Even more than ten years after the infection with HIV(Human Immunodeficiency Virus), the disease does not breakout in these patients. Only extremely small quantities of thevirus can be found in their blood – a sign that it hardly repli-cates at all. "This knowledge should make us hope. We mustfind out what factors are holding the virus in check in thesepeople. Perhaps that is an important step in the struggleagainst AIDS", says Dr. Matthias Platzer, head of the Explo-rative Project "Genomic Variability of Host Factors in theAIDS Macaque Model – Role in Resistance to and DiseaseCourse of Viral Infections in Humans and Non HumanPrimates" at the Leibniz Institute for Age Research – FritzLipmann Institute (FLI) in Jena.

An interdisciplinary team is now searching for the crucialgenes in the rhesus monkey. The team's members includethe immunologist and virologist Dr. Ulrike Sauermann andProfessor Gerhard Hunsmann of the German Primate Centerin Göttingen, the Kiel mathematician Professor MichaelKrawczak, the genetics professor Peter Nürnberg of theUniversity of Cologne and the microbiologist Dr. Roman A.Siddiqui, who also works at the FLI in Jena. In Rhesus mon-keys the SI-Virus (Simian Immunodeficiency Virus) elicits animmune deficiency which is very similar to the human HIVinfection. Since some of these animals do not develop anAIDS-like disease after infection, they are an ideal model forresearching resistance mechanisms. "The monkey modelgives us the advantage of having a great deal more dataabout the infection, for instance the kind and time of thetransmission, the quantity and type of the infectious virusand the complete course of the disease. Above all, we knowthe personal and familial medical history of the infected pri-mates. Such data are very valuable for systematic studiesand are never available for humans," Matthias Platzer adds.

The NGFN researchers are searching through the entire ge-nome of the infected primates. They are looking for geneticmarkers which only occur in resistant animals. Using thesemarkers they can limit the areas in the genome that makesurvival possible after such an infection. Sequencing theseDNA segments should then identify the particular factors inthe genome, which protect resistant monkeys from falling ill.The results of the search will be entered into a databasemade available by the genome networks Infection and Inflam-mation and Environmental Diseases. The cooperation withthe competence network HIV/AIDS in turn ensures that theresults gained in the animal model are quickly incorporatedinto human AIDS research.

HEALTHY DESPITE HIV

PROJECT LEADERDr. Matthias PlatzerLeibniz Institute for Age Research –Fritz Lipmann Institute, Jenamplatzer@imb-jena.de

DNA sequencing

62

EXPLORATIVE PROJECTS (EP)

ALTERNATIVE mRNA SPLICING: MORE INFORMATION FROM FEW GENES

PROJECT LEADERProfessor Albrecht BindereifUniversity of Giessenalbrecht.bindereif@chemie.bio.uni-giessen.de

also occur in the DNA. From a pre-mRNA always the samemRNA is produced. Due to alternative splicing the sequenceof the mRNA is varied: individual exons can be shortened orlengthened, entire exons can be skipped. Through these pro-cesses many mRNA variants and thus also many differentproteins are generated from one single gene. Often these

processes are regulated in a tissue-specific and de-velopment-specific manner.

The Explorative Project"Analyzing Global Regula-tors of Alternative Splicingin the Human System: ACombined RNAi and Micro-array Approach" deals withthe underlying mechanismsof alternative splicing. "Untilnow little has been knownabout which proteins regu-late these processes andwhat their target sequenceslook like in the RNA. Ourgoal is to systematicallyidentify the complete net-works of splicing regulatorproteins and their target

genes in the human genome," says project leader ProfessorAlbrecht Bindereif. A new methodical approach is to helpwith this: Via RNA interference putative splicing regulatorswill be targeted individually and switched off. By using micro-arrays it can then be analyzed what effects switching off theregulator has on the splicing pattern of certain target genes.For this purpose splicing-sensitive microarrays are to be usedthat simultaneously allow the tracking of the alternative splic-ing pattern of a great number of candidate genes.

How can a complex organism like the human being functionwith only about 30,000 genes? After all, even the simplenematode worm Caenorhabditis elegans has about 19,000genes. One answer is alternative mRNA splicing. Probablymore than half of all human proteins are produced via thismechanism. But what lies behind this?

Alternative splicing can best be explained by stating what itis not: constitutive splicing. In constitutive splicing the codingsegments (exons) – after removal of the non-coding se-quences (introns) – are joined in the sequence in which they

Sample preparation

63

EXPLORATIVE PROJECTS (EP)

The immune system is our best weapon in the fight againstdiverse pathogens. Utilizing this fact, scientists are attempt-ing to enlist the immune system to combat one of the great-est scourges of humanity: cancer. Vaccination against cancerusing T lymphocytes (T cells) to activate the immune defenseis a very promising research approach. T cells have highlyspecific receptors for the identification of their antigens. Thebinding to the receptor triggers an immune response which

destroys the "agitator" (see box). However, a successful vac-cination implies the presence of tumor-associated antigensthat the immune system can target. Characterizing theseantigens is therefore one of the central tasks of today's can-cer research. But where do we find these antigens? Goodcandidates are proteins frequently produced in tumor cellsbut rarely or never present in healthy cells. In general, cancerscientists search for these proteins by means of the cells'mRNA quantity. With high-throughput screening methods

such as DNA arrays, this is effective and cost-efficient. Ulti-mately, however, it does not play a decisive role how muchprotein is present in the cell. More important for a successfulimmune response is the frequency with which a protein ispresented on the surface, for here is where the contact withthe T cells occurs. Until recently, it was not possible to deter-mine the quantity of MHC-bound (see box) protein fragmentspresented on the cell surface. Therefore, it was not clearwhether proteins with increased mRNA quantity in tumorcells are also presented with higher frequency on the surfaceof the cell. That, however, is a prerequisite for proteins to be used as effective antigens in the fight against cancer. The Explorative Project headed by Professor Hans-GeorgRammensee has solved this problem: Using mass spectrom-etry, scientists can focus their investigations on how fre-quently an MHC-bound protein fragment appears on the cellsurface. "First we are checking the methods established upto now. If we can prove that proteins with an elevated mRNAquantity are also present in increased quantity on the cellsurface, we can go ahead and continue working with thesemethods. However, if this is not the case, this would castmuch doubt on these previously used approaches. Our mass spectrometry analyses would then be the only reliablepossibility to look for tumor-specific antigens," ProfessorRammensee explains.

CANCER VACCINATION: ARE WE MAKING PROGRESS?

PROJECT LEADERProfessor Hans-Georg RammenseeUniversity of Tübingenrammensee@uni-tuebingen.de

HOW DOES A T LYMPHOCYTE IDENTIFY "ITS" ANTIGEN? The cell places small fragments of all proteins whichare present in the interior of the cell in a "gallery"onto the cell surface. The cell's own mechanismensures that all proteins present in the cell interiorare continually presented on the surface. A constantcirculation of degradation and production of proteinstakes place in the cell. During the degradation pro-cess protein fragments are created, of which smallpieces are bound by specialized receptors (MHC =Major Histocompatibility Complex) and transported tothe surface. The T lymphocytes can differentiatewhether these fragments are either known and nor-mal, or foreign and abnormal, respectively. If theyidentify the protein as foreign or pathogenic, they killthe cell and stimulate the production of antibodiesagainst the antigen.

64

EXPLORATIVE PROJECTS (EP)

Depression and anxiety disorders are among the greatestmedical and healthcare challenges of our time. In the indus-trial countries twelve to 20 percent of the population will fall ill with a depression at least once in their lives. The basicmood of the afflicted persons is sad, their drive is dimin-ished, and they lose interest in their social surroundings.They feel helpless and are despairing. Forty to eighty percentof the affected persons even think of committing suicide.One major problem is that depression is often not recognizedand is diagnosed much too late.

The Explorative Project "Identification of Autoantigens asCandidate Markers for Affective Disorders", headed byProfessor Christoph W. Turck, wants to contribute to an im-proved diagnosis of affective disorders. His team is searchingfor characteristic markers which only occur in combinationwith affective disorders by taking advantage of the immenseprecision of the human immune system. The scientists aresearching for autoantibodies which target the proteins of thebrain tissue in the CSF of the patients. CSF is an abbreviationfor cerebrospinal fluid, which circulates through the ventri-

cles of the brain and the cavities of the spinal cord. "The CSFprovides a reflection of the healthy or sick brain," ChristophTurck explains, "which we were able to show in large-scaleanalyses of CSF proteins."

In their experiments the scientists were also able to showthat antibodies from the CSF recognize proteins of thehuman brain. In previous experiments they established animmunoassay that enabled this determination. They separat-ed the proteins of the human brain tissue electrophoreticallyand transferred them to a membrane. They incubated thismembrane with CSF samples of patients with affective dis-orders. Here they discovered first interactions between auto-antibodies and brain proteins. Currently, in a larger scaleexperiment, scientists are systematically screening the CSFof patients with affective disorders to detect other autoanti-bodies. The brain proteins identified by means of the auto-antibodies are the sought-after diagnostic markers for affec-tive disorders. Preventive measures can then be taken in theclinic as treatment.

Moreover, these proteins could also provide the basis for fur-ther experiments: Professor Turck and the members of hisworking group want to decode the functions of the proteinsusing various molecular biological and biochemical approach-es. Knowledge about the protein functions could then lead toconclusions about their significance for the pathogenesis andcourse of affective disorders. That could be an importantstarting point for pharmaceutical research to develop newmedications for affective disorders.

DEPRESSION – A WIDESPREADILLNESS

PROJECT LEADERProfessor Christoph W. TurckMax Planck Institute of Psychiatry, Munichturck@mpipsykl.mpg.de

CSF Western blot analysis:

Human brain proteins immobilized on a

membrane are probed with CSF antibo-

dies. If the CSF contains a brain-specific

antibody it will bind to its antigen on the

membrane and generate a signal that can

be detected by an X-ray film.

CSF antibodies

Membrane withhuman brainproteins

Cerebral spinalfluid (CSF)

sample of patientwith affective

disorder

65

EXPLORATIVE PROJECTS (EP)

Biological macromolecules are often large and very complex.This is why for a long time it was not possible to use massspectrometric analyses in biology, although in analyticalchemistry, elements and small molecules – or their frag-ments – were routinely determined with this method. It wasnot until the development of gentle ionization methods(MALDI, ESI) in the middle of the 1980s that mass spectrom-etry came into use in biological and biochemical laboratories.

This was a milestone: From then on biological macromole-cules could be transferred intact into the gas phase andcould be ionized for analysis. In 2002 the Nobel Prize inChemistry was awarded for these achievements, illustratingthe current significance of mass spectrometry. Since thenthe technology has been developed even further, especiallywith regard to the achievable sensitivity. At present even thesmallest quantities of less than one femtogram (10-15 gram)can be detected. Advancements in mass spectrometry arebeing pursued worldwide: Professor Bernhard Spengler andhis team from the University of Giessen are working on amass spectrometry method which can image the compositionand the spatial distribution of biochemical compounds. Theprojected methodology is an extension of the SMALDI tech-nology (scanning microprobe matrix-assisted laser desorptionionization) developed by Spengler and his team in recent

years, and it has unique analytical features. It involves a high-power laser that scans e.g. a cell surface; the laser-ablatedmolecules are subsequently analyzed using mass spectrom-etry. Finally, a software program evaluates the data with re-spect to composition as well as to the local distribution of theindividual substances. In this way an image is created thate.g. shows the concentration gradient of a substance on thecell surface. "In the future we plan to use an infrared laser

instead of a UV laser for this system. That would have severaladvantages: Infrared light penetrates up to one micrometerdeep into the specimens – while UV light reaches layers thatare maximally 100 nanometers deep. Infrared lasers wouldenable us, for instance, to study even deeper regions of a tis-sue," Bernhard Spengler explains. But this is not the onlyadvantage: For many biological substances, such as carbo-hydrates and lipids, infrared light is superior to ultravioletlight with respect to accessible mass range. Moreover, withan infrared laser the laborious task of sample preparationwould be simpler. While for an analysis with UV light the sam-ple has to be embedded in a special matrix, for an infraredlaser it is in principle sufficient to deep-freeze it. He adds,"Until this is accomplished we still have some work to do. Itis important to note that there is not yet any infrared laserthat can ensure such a high resolution. Likewise, a systemhas yet to be developed to evaluate these data."

NEWS FROM MASSSPECTROMETRY

PROJECT LEADERProfessor Bernhard SpenglerUniversity of Giessenbernhard.spengler@anorg.chemie.uni-giessen.de

Mass spectrometer

66

EXPLORATIVE PROJECTS (EP)

The goal of the bioinformatics team under the direction ofProfessor Andreas Zell is to develop inference models forgene regulation networks. "Based on the vast quantity ofdata available from microarrays, we reconstruct parts of theregulatory network of the genome," Andreas Zell explains."For example, if we have information on time-series geneexpression – which genes are active when – we can conclude

from that how they interact with each other." The team fromthe Center for Bioinformatics at the University of Tübingenhas developed a software program that enables these con-clusions: From the experimental data JCell directly infersregulatory relationships among the genes studied.

Until now microarrays were mainly used to compare theactivity of genes independently of each other, e.g. to analyzegene expression in healthy and diseased tissue. However,mutations in a single gene may often lead to changes inmany different cell functions. JCell can help scientists iden-tify regulatory relationships – and thus glean more informa-tion from the data. "With the current version of JCell, gene

researchers can reconstruct small geneticnetworks, at least if they have sufficientexperimental data. And we are continuallyoptimizing the system to improve the in-ference algorithm design of natural inter-actions," explains Christian Spieth, mem-ber of the Explorative Project "InferringGenetic Interactions from Gene Expres-sion Data". JCell also accesses publicdatabases. "This is how we guarantee that the most up-to-date expert knowledge isincluded in the analysis of gene regulationprocesses," he adds. In the future, it is planned that JCell will also simulate com-plex biological processes and their dis-ease-related variants. The program couldthen make a considerable contribution tothe understanding of complex diseasessuch as infections or cancer and open upnew therapy approaches.

JCell is currently one of the most effectiveanalysis tools for reconstructing and simulating the dynamicsystems of gene regulation. For this achievement the projectwon third prize in the 2005 doIT Software Award Competi-tion, a research initiative of the state of Baden-Württemberg.

INFERENCE MODELS FORGENE REGULATION

PROJECT LEADERProfessor Andreas ZellUniversity of Tübingenzell@informatik.uni-tuebingen.de

Screenshot of the software framework JCell

67

EXPLORATIVE PROJECTS (EP)

In recent years it has become ever more evident that ourgenome produces several thousand transcripts which do notcode for proteins. Does that mean that our genome is waste-ful? Or do these mainly small RNAs have tasks in the cellsthat are not yet known? "In my opinion many of these RNAsperform other functions, for example, they regulate geneactivity. The emphasis is on many, not all. Today we still donot know whether most of them have a job to do in the cellsor not," explains Professor Jürgen Brosius, head of theExplorative Project "Exploring the World of Non-MessengerRNAs: RNomics Meets Proteomics". The Münster scientistswant to find answers to these questions. With different high-throughput procedures they are searching specifically fornon-protein-coding RNAs in mammals. "When such RNAs arefound, we try to find out whether they take over functions inthe cells. And next, if this is the case, we want to find outwhat they do there," Jürgen Brosius says, describing theobjective of his project.

One thing is already considered certain. These new RNAsenable genetic diversity because they represent the rawmaterial for new sequence modules in the evolution of ge-nomes. These develop through reverse transcription of RNAsor through random integration into the genome. Dependingon where these modules land, the effect can be disadvanta-geous, neutral or even beneficial. For instance, this is hownew exons are formed out of non-coding sequences – an

important mechanism of evolution. Often, splice variants ofalready established genes are generated. The old and newforms of the gene then exist parallel in the genome. Thus, thenew form can be tested while the old form continues to be inuse. If the new splice variant does not offer the organism anyadvantage, it disappears in the course of time. The research-ers of the Institute for Experimental Pathology are also pur-suing this approach. They have investigated 153 exons in thegenome in whose development Alu elements were probablyinvolved. Alu elements belong to the family of short inter-spersed elements (SINE) and are specific to primates. Someof them are so similar to splice sites that they induce alterna-tive splicing. Other Alu elements can be transformed into asplice sequence through a simple nucleotide exchange. Theresearch group led by Jürgen Brosius compared four selectedexamples more exactly between the species of primates.Result: They were able to reconstruct the evolutionary stepsthat occurred in the various species and thus could decodethe significance of the elements for evolution. According tothese findings, after integration into the genome the Alu ele-ments either induced the alternative splicing directly or slum-bered unrecognized for millions of years until they becameactive.

IS OUR GENOMEWASTEFUL?

PROJECT LEADERProfessor Jürgen BrosiusUniversity of Münsterrna.world@uni-muenster.de

"Nor ought we to marvel if all the contrivances in nature be not,as far as we can judge, absolutely per-fect; and if some of them be abhorrent

to our ideas of fitness. We need notmarvel at … the astonishing waste of

pollen by our fir trees."

Charles Darwin

68

EXPLORATIVE PROJECTS (EP)

Contact with Pseudomonas aeruginosa cannot be avoided,because these resistant bacteria are found nearly every-where in our environment. In most cases, however, thehuman immune system prevents disease caused by this typi-cal moist-site, waterborne pathogen. By contrast, immuno-compromised patients often have no protection against infec-tion with Pseudomonas aeruginosa. In cystic fibrosis (CF)patients, for example, the predominant cause of death ispneumonia caused by Pseudomonas aeruginosa. Cancer pa-tients and burn victims are also particularly susceptible tothis pathogen, since Pseudomonas aeruginosa can triggerwound infections, sepsis and cardiovascular diseases.

In CF patients Pseudomonas aeruginosa frequently attacksthe respiratory tract and colonizes the thick mucus which istypical of the disease. There the bacteria form a biofilmwhich enables them to evade the body's own defense mecha-nism. Even many antibiotics cannot penetrate this protectiveenvelope and are therefore ineffective against the pathogen.Moreover, Pseudomonas also generates enzymes whichdirectly inactivate the antibiotic. Once the bacteria have colo-

nized the lung, they cause inflammation which leads to pro-gressive lung damage of the patients and is the major causefor morbidity and mortality in CF. Professor Dieter Jahn andhis team at the Institute of Microbiology at the TechnicalUniversity Braunschweig are attempting to elucidate the ex-act processes which lead to the formation of the biofilm andto resistance against antibiotics. To achieve this, they want tothoroughly analyze the metabolic products of Pseudomonasaeruginosa. "What we need is a combination of very ad-vanced high-throughput procedures and sophisticated bio-informatics to attain this objective. After all, we want to in-vestigate all metabolic products of this bacterium during an

infection of the lungs. But we first have todevelop the appropriate high-throughput pro-cedures," Dieter Jahn explains. The findingsfrom this approach will first be comparedwith existing data from proteome and ge-nome analyses in order to draw possible con-clusions about the development of metabo-lites. Two established databases are availablefor this: BRENDA (BRaunschweig ENzymeDAtabase) and PRODORIC (PROcariotIC Data-base Of gene Regulation). BRENDA is man-aged by Professor Dietmar Schomburg andhis colleagues at the Cologne Institute forBiochemistry. The database contains the larg-est collection of data anywhere in the worldon enzymes and metabolic pathways. PRO-DORIC, on the other hand, is managed by theBraunschweig scientists and is the largestdatabase worldwide on regulatory networksof prokaryotes. In particular, the PRODORICdatabase organizes information on patho-

gens. "Our objective is to develop a model for Pseudomonasaeruginosa – a model that will represent all gene-regulatoryand metabolic pathways of the pathogen," says Dieter Jahn,looking to the future.

A "HOSPITAL BUG"BECOMES PREDICTABLE

PROJECT LEADERProfessor Dieter JahnTechnical University Braunschweigd.jahn@tu-bs.de

Culture of Pseudomonas aeruginosa

explains Professor Michael Meisterernst, who heads theExplorative Project. Using synthetic antibodies, he and histeam aim to develop ChIP-chip technology further for thehuman system. These synthetic immunoglobulins bind highlyspecifically to certain regions in their target protein. In con-trast to their natural counterparts, synthetic antibodies canalso target short signal peptides which do not occur natural-ly. This enables researchers to express numerous differentproteins in the cells which all contain the same signal peptideand which therefore can be detected by the same antibody.Alternatively to this, the Munich scientists are developingantibodies that directly recognize a natural protein. Sincehighly specific antibodies which can be used for ChIP-chiptechnology are relatively rare, the genome researchers areplacing great hope in the synthetic substitute.

69

EXPLORATIVE PROJECTS (EP)

FROM GENOME TO ORGANISM – THEWHERE AND WHEN OF GENE EXPRESSION

PROJECT LEADERProfessor Michael MeisterernstGSF - National Research Center forEnvironment and Health, Neuherbergmeisterernst@gsf.de

A textbook "Molecular Biology of the Cell" (Bruce Alberts etal., 3rd edition) introduces the chapter about gene expressionas follows: "An organism's DNA sequences encode all of theRNA and protein molecules that are available to construct itscells. Yet a complete description of the DNA sequence of agenome – be it the few million nucleotides of a bacterium orthe three billion nucleotides of a human – would provide rela-tively little understanding of the organism itself. It has beensaid that the genome represents a complete 'dictionary' foran organism, containing all of the 'words' for its construction.But we can no more reconstruct the logic of an organismfrom such a dictionary than we can reconstruct a play by Shakespeare from a dictionary of English words." In both cases the problem is the same: Where and when is eachword used, or rather, which gene is expressed?

Gene expression in the cell is strictly controlled. Transcrip-tion plays a key role in this. Many regulatory mechanismsadhere to this first step of gene expression and thus enablethe different developmental and differentiation processes ofthe cells. Transcriptional control is enabled via protein-DNAinteractions. For scientists to understand these controlmechanisms, they must trace these interactions and analyzethem. This generally takes place via chromatin immunopreci-pitation (ChIP) technology: First the DNA of a cell is cross-linked with regulatory proteins. Via antibodies that specifical-ly bind to the cross-linked proteins, the scientists purify theDNA protein complexes. The thus-extracted DNA sequencesare copied using PCR and subsequently analyzed. In modelorganisms such as yeast, DNA microarrays representing theentire genome of the organism have been used in recentyears.

The Explorative Project "Synthetic Antibodies for ChIP-chip inHuman Cells" is searching for the target sequences of theregulatory proteins in humans. "In the human genome theuse of microarrays for ChIP technologies has not been possi-ble until just recently. The reason for this is that the humangenome is considerably larger and more complex than that ofyeast, for instance. The development of comparable methodspresents an enormous technical challenge for science,"

70

EXPLORATIVE PROJECTS (EP)

Renal cancer is the tenth most common cancer in developedcountries, accounting for approximately three percent of alldiagnosed malignancies with up to 95,000 deaths per yearworldwide. More than 75 percent of renal cancer cases areclassified as renal cell carcinomas (RCC). At presentationless than 40 percent of the patients have tumors confined to the kidney – they have developed metastatic disease andhave a five-year survival rate ranging from 16 to 32 percent.For localized disease, nephrectomy is the main treatment,

whereas the efficacy of other treatment modalities, such asradiotherapy and/or chemotherapy, is very limited. No infor-mation is available regarding valuable prognostic and diag-nostic as well as broadly applicable therapeutic markers forRCC. Currently, the performance status, tumor stage andgrade are the most useful clinical prognostic indicators inpatients with RCC. Although RCC is – analogous to mela-noma – a relatively immunogenic tumor and biological thera-pies appear to be successful, the number of RCC-specificantigens is still very limited and the subjective response rateis less than 20 percent. Therefore the identification of noveltumor-associated antigens (TAAs) as well as novel conceptsfor the treatment of this disease are urgently needed.

The Explorative Project "Combining Proteome-based Technol-ogies with Epigenetics Using Renal Cell Carcinoma (RCC) asa Model: Identification of RCC Specific Cancer Antigens forVaccine and Antibody Therapies" is taking on this task. Epi-genetic events, such as deregulated methylation of CpGdinucleotides and aberrant histone acetylation are frequentlyoccurring modifications at the DNA level. If such eventsoccur close to and/or within open reading frames (ORF), thiswill likely lead to a reduced or even fully silenced gene ex-pression and thus may have an impact on diverse proteinfunctions. Abnormalities in DNA methylation and histoneacetylation have been shown to play an important role inboth tumor development and progression, thereby impairingthe immunogenic potential of cancer. It has recently beenshown that DNA hypermethylation and/or histone deacety-lation contribute to the lack or deficient expression of com-ponents involved in tumor recognition like molecules of theantigen processing and presentation machinery and/or ofthe costimulatory pathways.

Thus, the scientists treat a panel of primary tumor-derivedcell lines, representing both malignant RCC subtypes such as (i) clear cell, (ii) chromophobic and chromophilic RCC aswell as the benign oncocytoma with demethylating or de-acetylating agents to overcome given tumor-related epige-netic blocks. In a systematic effort to identify such cancer-related epigenetic changes, the research team will employtwo-dimensional gel electrophoresis followed by mass spec-trometry and PROTEOMEX, a method which combines pro-teome analysis with immunoblotting using sera from healthydonors and RCC patients to analyze pattern changes in theprotein expression profiles under experimentally designedepigenetic conditions and thereby assess their impact on theimmunogenicity of the treated tumor cells. "In our previousexperiments we succeeded in defining differentially ex-pressed proteins between primary tumors and healthy renalepithelia," Professor Barbara Seliger explains. Now the scien-tists are attempting to characterize the mechanisms by whichtumor cells install epigenetic pattern changes. The analysis of the expression profiles of untreated versus treated tumorcells might not only allow to define novel biomarkers, butalso might lead to the development of novel immunothera-peutic relevant concepts linked to the pursued reversion ofepigenetic phenomena.

INNOVATIVE APPROACHES FORTREATING RENAL CANCER

PROJECT LEADERProfessor Barbara SeligerUniversity of Halle-Wittenbergbarabra.seliger@medizin.uni-halle.de

3 10pH

Molecular w

eight

Two-dimensional polyacrylamid gel: Differences in the protein expression pat-tern of treated versus untreated RCC cells. Exclusively expressed or significant-ly upregulated proteins in the treated variants are marked by red arrows.

71

EXPLORATIVE PROJECTS (EP)

The brain awakens the curiosity of scientists and non-scien-tists alike. For years, Professor Arthur Konnerth and his col-leagues at the Institute for Neuroscience of the TechnicalUniversity in Munich have been studying the function of thecentral nervous system. The researchers focus on the devel-opment and the plasticity of neural networks in the brains ofanimal models, mostly mice and rats. Their results show thatduring brain development, large groups of nerve cells areactive simultaneously. Characteristic activity patterns pro-duce prominent changes in the calcium concentration in vir-tually all neurons. These changes in concentration spread ina wave-like fashion through the cortex and have a key func-tion during the early phase of brain development. "However,in vivo – that is, in the intact brain – we were not able toprove the presence of these calcium waves because welacked the proper technology. Until recently we could onlyperform in vivo experiments on anesthetized test animals. We now know, however, that the anesthesia of the mice pre-vents the calcium waves," Professor Konnerth explains.

The NGFN scientists thus had no other choice – they had todevelop a new in vivo technology. In the framework of theExplorative Project "Brain Endoscopy for the Functional Anal-ysis of Neuronal Networks in vivo" a suitable method has nowbeen established. "To monitor Ca2+ levels in awake behavingmice, we have developed an optical fiber-based recording

system consisting of a chronically implanted fiber used bothfor excitation of the dye and for detection of the emitted fluo-rescence. This single-fiber endoscope allows recording ofCa2+ signals from any brain region in anesthetized as well asawake behaving animals. Therefore, it offers a unique oppor-tunity for studying brain function and for detecting the actionof drugs without the influence of anesthetics," Dr. HelmuthAdelsberger comments. As a first step, the desired brain re-gions have to be stained with calcium-sensitive fluorescentindicator dyes. Upon binding Ca2+ ions, such dyes change the spectral characteristics of the emitted light. The emittedlight is monitored with sensitive detectors and translated intointracellular alterations of the calcium concentration. "Fur-ther advantages of this system in contrast to other methodsused for these purposes are the possibility to implant theoptical fibers in deeper layers of the brain. By using severalfibers simultaneously, the interaction of different brain re-gions can be studied in real time," Dr. Helmuth Adelsbergeradds. "With this new system we can extend our studies onbrain development and address open questions. Utilizing thenew approach, we recently found that anesthesia blocks theintracellular calcium oscillations. In non-anesthetized micewe were able to clearly identify and characterize the calciumoscillations. We have now started to study altered calciumsignaling in mutant mouse disease models with the single-fiber endoscope," says Professor Konnerth.

LOOKING INTO THE CENTER – A NEW TECHNOLOGY ENABLES IN VIVO STUDIES IN THE BRAIN

PROJECT LEADERProfessor Arthur KonnerthTechnical University Munichkonnerth@lrz.uni-muenchen.de

ARGON LASER Scheme of the optical fiber-based calcium recordingsystem. (Em = emission; Ex = excitation; ND = neutral density; PMT = photomultiplier)

Ex. filter

Argon laser

PMT

Beamsplitter

Em. filterND filter

Shutter

Collimator

Optical fiber

Optical fiberSkullDental cement

Stainedregion

Illuminatedarea

CORTEX

72

EXPLORATIVE PROJECTS (EP)

A DRAGNET OPERATION TOFIND BINDING PARTNERS

While genome research today has highly complex DNA arraysat its disposal, the spot density of peptide arrays is compara-tively low: Between 10,000 and 50,000 peptides can be syn-thesized on a chip of 20 x 20 cm using current technology.One reason for this is the more complex binding chemistry ofthe peptides, which makes their production more difficult.But also the greater number of components in comparison to

DNA play a role. For the synthesis of peptides 20 differentamino acids have to be combined – instead of only fournucleotides in the production of DNA arrays. The ExplorativeProject "Peptide Libraries on a Chip" under the direction of Dr. Ralf Bischoff is undertaking a new approach to facilitatehighly complex peptide arrays. The individual amino acids arefirst embedded in a solid carrier material (e.g. diphenyl form-amide), which is subsequently processed into well-definedmicroparticles. These microparticles are then addressed todefined locations on the chip via electrostatic interactions.Not until all 20 amino acids have been distributed – and eachfield is addressed with an amino acid – are the amino acidsfirmly bound. For this the carrier material is melted and theamino acids are released, thus forming the peptide bonds.After several washing and drying steps, the next synthesiscycle can be performed and a further amino acid can beadded to the growing peptide chain. A great advantage ofthis method is the possibility to wait until all 20 differentamino acid particles are distributed to their designated loca-tion on the chip before the formation of the peptide bondshas to be carried out. Compared with various lithographicmethods (see box), significantly fewer individual reactions arerequired. Thus the probability of faulty synthesis is reduced.

The highly specific addressing of amino acids to the corre-sponding spots on the chip is ensured via electrostatic forces: Initially, the amino acid particles (consisting of e.g. the amino acid threonine and the carrier material sur-rounding it) are electrically charged in such a way that allparticles have the same charge. The spots on which thethreonine particles are to be affixed are provided with the

respective countercharge. The charged particles are thenbrought into contact with the chip via aerosol delivery andattach precisely at the selected spots (see figure). However,before the chip is brought into contact with the next aminoacid particle containing aerosol (e.g. isoleucine particles), the charge pattern of the chip is varied: Now the sites on thechip are provided with a countercharge where the amino acidisoleucine is to bind (see figure). In total, this cycle is repeat-ed 20 times – once for each of the 20 amino acids. The par-ticles are subsequently melted, and the peptide bonds areformed for all 20 amino acids simultaneously. The team atthe German Cancer Research Center (DKFZ) uses a semicon-ductor chip comprising an array of individually addressablepixel electrodes as support for the peptide arrays. By apply-ing voltages to the electrode array, any arbitrary charge pat-tern can be generated and varied on the chip. "One of ourgoals is to improve this chip. We want to design highly com-plex chips that consist of up to 40,000 pixel electrodes percm2. The complete proteome of a bacterium could find roomon such a chip," Ralf Bischoff explains.

Chip

73

PROJECT LEADERDr. F. Ralf BischoffGerman Cancer Research Center,Heidelbergr.bischoff@dkfz.de

Deposition of amino acidparticles on the chip using

electrostatic forces.Charge patterns on the

chip surface are generated and configured by the

microchip itself.

MET

+ - - + + -

THR ILEMETTHR ILE

+ + +

nextaminoacid

melting,coupling and

washing

next cycle Charge patterns onthe chip surface

Chip

THR

- --------

- --

ILE

---- - - - --

-- -

THR

- --------

- --

THR

- --------

- --

THR

- --------

- --THR

- --------

- --

THR

- --------

- --

THR

- --------

- --

THR

- --------

- --

ILE

---- - - - --

-- -

ILE

---- - - - --

-- -

ILE

---- - - - --

-- -

ILE

-- - - - --- -

ILE

---- - - - --

-- -

THR

- ----

- ---

PHOTOLITHOGRAPHIC METHODThe photolithographic method enables synthesizing oli-gonucleotides directly onto the surface of the chip. Forthis approach a support is used on which the individualspots are chemically protected by light-sensitive protec-tive groups. When these protective groups are radiatedwith light, they are destroyed, and the spot is deprotect-ed. The deprotection of the spot ensues through litho-graphic masks which only let the light pass at clearlydefined locations. In the next step the first nucleotidesare added (e.g. dATP) that in turn also have a protectivegroup. They bind selectively to the spots whose protec-tive groups have been destroyed. Before the secondnucleotide (e.g. dCTP) comes into contact with the chip,an additional exposure – via a new mask – is implement-ed. Thus, exactly those spots are deprotected to whichdCTP is now supposed to bind. This cycle is conductedfor all four nucleotides, so that after the fourth cycle allspots have a nucleotide bound to them.

The nucleotides that are used for synthesis have thesame light-sensitive protective groups as found on thesupport. This prevents the subsequent nucleotides fromcoupling to the wrong spots.

MODIFIED INK JET PRINTERSUsing modified ink jet printers, nucleotides can be trans-ferred directly to the synthesis spots. Instead of thecolors cyan, magenta, yellow and black, the nozzles dis-perse the nucleotides dATP, dTTP, dGTP and dCTP.Using this method, already synthesized DNA samples canalso be transferred to the chip. The disadvantage withthis application, however, is that all samples have to besynthesized separately beforehand.

MECHANICAL MICROSPOTTINGWith this method a robot transfers the sample directly tothe chip. Here, too, both nucleotides and finished DNAsamples can be transferred.

PRODUCTION OF DNA ARRAYS

74

QUALITY MANAGEMENTIN THE NGFN

NATIONAL GENOME RESEARCH NETWORK (NGFN)

Successful collaboration depends on common standards. Forscientists to share knowledge, they must be able to trust thatall researchers are equally conscientious and conform to thesame rules. This is especially true for a large network like theNGFN, which has around 350 projects from 80 scientific in-stitutes and companies. For this reason, since the end of2003 the NGFN has been building up a comprehensive qual-ity management (QM) system. Quality management coordi-nators have been named, and currently seven working groupsare concerned with QM in the following fields: genotyping;microarrays; production of open reading frame (ORF) re-sources; functional assays and cell cultures; proteomics andprotein analysis; clinic management, data management andanalysis; and animal models. Apart from assuring optimaldata exchange, the QM measures aim to ensure that all ma-terials and results produced within the NGFN maintain thesame high quality. The respective quality standards followinternational guidelines, if such exist.

STANDARDIZED TECHNIQUES AND QUALITY-CONTROLLED MATERIALSMany work steps have already been standardized and docu-mented for the entire NGFN network, including techniques toproduce samples and to process microarrays. These stan-

dards improve the quality of the results and ensure the trans-parency of experimental procedures. They are intended to beused by as many partners as possible in the NGFN on thebasis of voluntary self-commitment. Procedure descriptionsare subjected to the respective QM working group for ap-proval before they are made available on the NGFN intranet.

The NGFN also specifies strict quality criteria for raw materi-als such as DNA and RNA. "We ensure the good quality ofDNA and RNA specimens by performing complete checks.For example, we test the concentration and the purity of the material with spectrophotometric analyses," explains Dr. Holger Sültmann, spokesman of the working group"Microarrays". "The quality of each individual microarray is carefully checked and documented following a standardizedprocedure."

In the production of ORF resources every clone is checked to make sure that it really contains the complete and correctORF sequence. The National Genotyping Platform (NGP) per-forms a validation of the respective genotyping in the form of ring trials. The results of these ring trials are evaluatedparticularly with respect to reproducibility.

Standardized procedures for high quality data

75

data are supplemented by specific parameter sets for theindividual indication areas. Together with the users, we con-tinually develop these parameter sets further." To ensuresmooth data transfer, all participants must use a common vo-cabulary. "That is why NGFN ontology is rendered in a specialXML-based language, the Web Ontology Language (OWL),"Christian Lawerenz explains. Due to the semantic descriptioncapabilities of OWL, the ontology is presented exactly on thebasis of rules and terminologies and can serve as templatefor the data transfer. The NGFN ontology is also integratedinto the NGFN portal for high-throughput technologies (MIPSExpress). Important NGFN databases, such as the epidemio-logical metadatabases in Munich, Kiel and Bonn, the heart/cardiovascular database in Munich and iCHIP as local data-base platform for various Disease-oriented Genome Net-works, have already been integrated into the NGFN ontology.

However, all these measures do not suffice – quality manage-ment in the NGFN is a continuing process. To further opti-mize the quality and reproducibility of the research results,additional quality assurance strategies are being developedin regularly scheduled workshops. Publishing the QM resultsis also being considered. "In this way we want to stimulate abroader discussion on the subject," explains Professor StefanSchreiber, spokesman of the NGFN Project Committee. "Dueto the diversity and the large number of leading experts inthe field of disease-oriented genome research and in theplatforms, the NGFN could thus make a contribution to ele-vating quality standards in German genome research overall."

MANAGEMENT SYSTEMS FOR LAB INFORMATIONSeveral Systematic-Methodological-Platforms (SMP) haveintroduced lab data management systems. They ensure thatall experiments are accurately documented. From the regis-tration of sample entry to the scientific evaluation, all workprocesses must be documented without any gaps. For in-stance, within the SMP Cell, in the production of ORF re-sources, a management system assigns distinctive names for samples, thus creating a standardized nomenclature forthe entire process sequence. In the SMP Protein a manage-ment system records different physical and chemical condi-tions for automated protein crystallization experiments.Within the NGP each microtiter plate has its own identifiercode, corresponding to a linear bar code. This is linked viathe plate coordinates to the sample IDs. Increasingly, more-over, sample test tubes labeled with a 2D barcode are usedfor incoming DNA samples which can then be pipetted ontothe microtiter plates in an automated procedure.

DATABASE STRUCTURES AND INSTRUMENTS FOR DATA INTEGRATIONThe QM working group "Clinic Management, Data Manage-ment and Analysis" has developed standards for describingclinical data. Christian Lawerenz of the German Cancer Re-search Center (DKFZ), quality coordinator of the group, says,"Now we are able to transfer data within the NGFN and builddatabases in which patient records involving different dis-ease pictures are merged. The core parameter set includesattributes such as age, diagnosis, sample type and time tak-en that are relevant for all disease networks. These basic

Material check

76

NATIONAL GENOME RESEARCH NETWORK (NGFN)

DISCOVERING AND MARKETINGKNOWLEDGE – TECHNOLOGY TRANSFERCOORDINATION IN THE NGFN

"Know-how and inventions often slumber unused in the re-searchers' laboratory books," Dr. Isabel von Korff says."Worldwide only five to seven percent of the patents fromscientific research institutions are utilized commercially." Dr. von Korff, who is a natural scientist with several years ofexperience in the pharmaceutical industry, is director of theTechnology Transfer Coordination Office in the NGFN (KTT).This office coordinates the activities of the more than twentytechnology transfer offices that are involved in the NGFN andsees to that the results from NGFN-research are commercial-ized. "Intellectual assets have to be managed actively justlike any other assets to make a profit," she explains.

Since the German Federal Ministry of Education and Research(BMBF) started its commercialization initiative in 2001, muchhas changed in the German technology transfer landscape.Alongside technology transfer centers (German abbreviation:TTE) that manage the intellectual property of research institu-tions (e.g. Ascenion for life science centers in the Helmholtzand the Leibniz Associations, or Garching Innovation GmbHfor the Max Planck Society), a nation-wide network has sinceemerged comprising 22 patent and licensing agencies. Theseagencies (German abbreviation: PVA) are private companiescommissioned by universities which manage the patentingand commercialization of the research results from all disci-plines. The agencies are organized regionally and work oncommission for all institutions of a region. "For that reason, a PVA team must simultaneously represent the interests of amechanical motor, a substance used in making porcelain, anasthma drug, or a microchip," Dr. von Korff explains. In re-cent years a number of technology transfer agencies havebeen founded that specialize in a certain research area. Oneexample is Ascenion GmbH, which evolved out of an initiativeof several Helmholtz institutes and which concentrates onthe life sciences.

FROM ASSESSMENT TO COMMERCIALIZATIONThe new Technology Transfer Coordination Office in theNGFN links the many different groups involved. Already atthe start of the second NGFN funding phase, all scientistswere obligated to make a binding choice of the transfer facili-ty they wanted to work with in the future. The project mana-gers were free to choose whether to stay with the regionaltechnology transfer center or to accept the offer of the Tech-nology Transfer Coordination Office in the NGFN. "This wayclear competences were created from the beginning," Isabelvon Korff says. The legally binding collaboration always in-

cludes all of the steps from an initial assessment to commer-cialization. She continues, "We want to avoid that some of-fices only pick out the raisins in the cake and just shuffle offthe economically less interesting projects after assesment."

THE "GENOME MARKETPLACE"Since May 2005 the Internet platform www.genome-market-place.de provides potential industry partners with access tointeresting results from the NGFN which could be commer-cialized. "To accomplish this, we implement our principle ofpresenting one face to the customer," Dr. von Korff explains.On our website we offer patented inventions and technolo-gies, but also laboratory and research materials. Interestedcompanies can contact the relevant technology transfer of-fice through the NGFN Coordination Office. The genome mar-ketplace evolved out of the numerous coordination processeswithin the network. These ensure that all research results inthe second funding phase of the NGFN are analyzed to meetdefined high quality standards and are patent protected bysuitable intellectual property rights. Among other services to facilitate this, the Technology Transfer Coordination Officeprovides a central publication screen. Then the research re-sults are offered to industry via the genome marketplacewebsite. In addition, continuing education symposia are orga-nized for scientists in the field of Intellectual Property (IP)Asset Management. In individual cases, the NGFN Coordi-nation Office provides further services, particularly for com-mercial exploitation of IP to technology transfer centers, ifthis is expressly requested. For example, this could be thecase when special industry know-how or experience in found-ing and assistance with spin-offs is needed. "Our conceptenables a close-meshed network of research and industry,"Dr. von Korff adds. "That is the best precondition for an effi-cient and successful technology transfer."

In November 2005 Ascenion established a collaboration withBioDeutschland. The new biotechnology industry associationalready has more than 80 member companies. This will en-sure a close and efficient link to the German biotechnologyindustry, thus leading to more efficient and faster commer-cialization of NGFN technologies in Germany.

PUBLISHED BYNational Genome Research Network (NGFN)Project Management NGFN at PT-DLRHeinrich-Konen-Straße 153227 BonnGermanyPhone: +49 228 3821-331Fax: +49 228 3821-332E-mail: pm-ngfn@dlr.deInternet: www.ngfn.de

PROJECT MANAGEMENT NGFNDr. Markus Albertini, Dr. Uta Straßer

EDITED BYNational Genome Research NetworkMasterMedia GmbH, Hamburg

LAYOUT BYMasterMedia GmbH, Hamburg

PHOTOGRAPHS AND ILLUSTRATIONSBMBF, NGFN except: Dr. Norbert Frey: 16 Professor Anthony D. Ho: 55 Dr. Beate Niesler: 56 Dr. Ursula Klingmüller: 57 (adapted)Dr. Michael Speicher: 58Professor Gerd Walz: 59 DHGP: 61Professor Christoph W. Turck: 64 (adapted)Stiftung Tierärztliche Hochschule Hannover/Institut für Mikrobiologie: 68Professor Barbara Seliger: 70Professor Arthur Konnerth: 71 (adapted)Dr. F. Ralf Bischoff: 72 + 73 (adapted)

PRINTED BYmsp druck und medien gmbh, Mudersbach

2006

Project Management NGFNat PT-DLRHeinrich-Konen-Straße 153227 BonnGermany

Phone: +49 228 3821-331Fax: +49 228 3821-332E-mail: pm-ngfn@dlr.deInternet: www.ngfn.de