German Network for Systems Genetics
The Helmholtz virtual institute GeNeSys (German Network for Systems Genetics), funded by the Presidential Fund of the Helmholtz-Association, has been established at the HZI. The institute comprises of 10 partners from universities and research centres in Germany, and the Vrije University in Amsterdam. The partners share a common mouse genetic reference population to perform different phenotypic screens related to human diseases. The phenotypes include infection responses to S. aureus, M. tuberculosis, Y. enterocolitica, L. donovani, septic shock, infection-induced liver fibrosis, diabetes, and regeneration of the nervous system. All data are collected in a single database allowing the identification of quantitative trait loci (QTL) that are influencing the various phenotypic traits. In this way, it will be possible to associate multiple phenotypes with multiple phenotypes (systems genetics).
The data generated by the different groups of the virtual center will be collected in a copy of the GeneNetwork database. GeneNetwork consists of a set of linked resources for systems genetics. It has been designed for multiscale integration of networks of genes, transcripts, and traits such as toxicity, cancer susceptibility, and behavior. This open resource combines more than 25 years of legacy data generated by hundreds of scientists with full genome sequence and deep transcriptome data sets. WebQTL is the leading GeneNetwork module, and has been optimized for online analysis of traits that are controlled by combinations of allelic variants and environmental factors. WebQTL exploits several permanent genetic reference populations (GRP) of mouse (BXD, LXS, etc.), rat (HXB), and Arabidopsis (BayXSha). Each GRP is accompanied by dense genetic maps used to locate modifiers that cause downstream differences in expression and higher-order phenotypes, including disease susceptibility.
Users can also enter their own private data directly into WebQTL to exploit the full range of analytic tools and to map upstream modulators in a powerful environment. Numerous statistical tools are combined with a database consisting of three million mouse SNPs. This combination allows relatively efficient analysis of possible relations between sequence variants and sets of functional variants.
GeneNetwork pages are extensively connected to external resources. Numerous links to the UCSC and Ensembl Genome Browsers, PubMed, Entrez Gene, GNF Expression Atlas, ABI Panther, and WebGestalt provide users with rapid interpretive information about genomic regions, published phenotypes and genes highlighted in WebQTL.
Helmholtz Centre for Infection Research, Braunschweig and University of Veterinary Medicine Hannover
Helmholtz Centre for Infection Research, Braunschweig
Center for Regenerative Therapies, Dresden
Reinhard Hoffmann, Roland Lang, Jörg Mages,
Technical University of Munich
University of Lübeck
Humboldt-University of Berlin
Research Center Borstel
Saarland University Medical Center, Homburg
Vrije University of Amsterdam
The associated partners of the virtual institute GeNeSys will not be funded directly through the Presidential Fund of the Helmholtz-Association but will be participants at meetings and strategic discussions.
Robert W. Williams, Ken Manly
University of Tennessee Health Science Center, Memphis, USA
Max Delbrück Center for Molecular Medicine, Berlin
Wellcome Trust Centre for Human Genetics, Oxford, UK
Tel-Aviv University, Israel
WP 1: Establishing the RI strain resource
August B Smit (University of Amsterdam)
The uniqueness of GeNeSys is to combine several research groups working on the same Genetic reference Population (GRP). None of the partners alone would be able to maintain the RI strains at their facility at reasonable costs just for the purpose of their own project, leave alone the enormous efforts to rederive the strains and introduce them into their facility. Therefore, the best solution was to set up a colony in Europe specifically for the purpose of this virtual institute.
A colony of BXD RI strains has already been established at a central facility at Harlan Netherlands B.V. in Amsterdam.
Our group will maintain the colony and deliver the requested mice to the partners of the consortium. No single laboratory in Germany and only very few institutions in the world will have a similar resource available.
WP 2: Susceptibility to Staphylococcus aureus infection
Eva Medina, Andreas Lengeling, Klaus Schughart (Helmholtz Centre for Infection Research and University of Veterinary Medicine Hannover)
Staphylococcus aureus is one of the most important life-threatening bacterial infection in the developed world. In the last decades, the incidence of S. aureus infections has increased dramatically which is in most part due to the extraordinary ability of these microorganisms to acquire antibiotic-resistance. The most common life-threatening manifestation of S. aureus infection is bacteremia, sepsis and septic shock. Septic shock due to S. aureus and other gram-positive bacteria is the most common cause of death in medical and surgical intensive care units in industrialized countries with a mortality rate of 40%-60% .
Certain patients appear to be predisposed to severe infection, whereas others are protected, indicating important host genetic factors. We thus propose to identify genetic loci influencing the host response to S. aureus by studying the susceptibility of RI strains to experimental infection with S. aureus. This information will be very important for assessing the risk of individual S. aureus infected patients as well as for designing new strategies to enhance host innate resistance.
The goal of the proposed project is to identify multigenic host factors and their interactions controlling resistance or susceptibility to S. aureus. In an effort to map loci influencing the severity of staphylococcal infection, a mouse infection model will be used that is well established in our laboratory. In preliminary studies with this mouse infection model, we have found that inbred strains differ markedly in their susceptibility to S. aureus.
We will identify complex genetic traits controlling resistance or susceptibility of mice to S. aureus infection using the BXD RI strains. To determine the phenotype of individual strains, mice will be intravenously infected with S. aureus line SH1000 and survival time and bacteremia (CFU/ml of blood) will be determined. The strains presenting a clear resistant or susceptible phenotype will be selected for a more detailed analysis.
WP 3: Susceptibility to Mycobacterium tuberculosis infection
Stefan Ehlers (Research Center Borstel)
Tuberculosis is a major global health problem with about 8 million new cases and 2 million deaths occuring every year. A number of informative studies have led to correlations between gene polymorphisms and susceptibility to disease suggesting that susceptibility/resistance to tuberculosis is multigenic with marked interregional differences.
Much of our knowledge about the immunology and immunopathology of tuberculosis in humans has been gained from infection studies in resistant and susceptible, as well as transgenic and knock-out mouse lines. A highly reproducible aerosol infection model exists that mimics the natural route of low-dose infection in humans and allows the study of early, innate mechanisms of host defences as well as the elucidation of host andpathogen-derived factors that determine the persistence and recrudescence of infection.
In our laboratory, ample expertise exists in using a chronic mouse model of tuberculosis. Aerosol infections with Mycobacterium tuberculosis mimicking the natural, low-dose exposure are performed at biosafety level III conditions at the Research Center Borstel, and the course of infection has been characterized in terms of bacterial load, cellular recruitment into infected organs, gene expression in target organs, and survival in several knock-out mouse lines.
Funded by NGFN-2, comprehensive microarray analysis of gene expression in two resistant (C57BL/6, BALB/c) and two susceptible (DBA/2, CBA) mouse strains was recently performed and has guided optimization of the assay system. Differential gene regulation in these mouse strains pointed to an involvement of granulocytes in susceptibility to tuberculosis, and depletion of neutrophils resulted in prolonged survival of susceptible, but not resistant mouse strains.
WP 4: Susceptibility to Yersinia infection
Reinhard Hoffmann (Technical University München)
With about 6,000 cases annually, Yersinia is the third most common bacterial cause of acute gastroenteritis in Germany. After oral ingestion of contaminated food, Y. enterocolitica or Y. pseudotuberculosis invade the intestinal mucosa of the terminal ileum through lymphoid follicle-associated M-cells and subsequently penetrates the lymphoid tissue of Peyer’s patches. Yersinia is endowed with a unique capacity to withstand the host attack by injecting anti-host effector proteins (Yersinia outer proteins, Yops) into professional phagocytes via a type III protein secretion/translocation system (TTSS). Six effector proteins interfere with distinct signaling pathways resulting in paralysis of phagocyte function.
In contrast to these well-characterized bacterial virulence factors, much less is known about host factors determining resistance or susceptibility against bacterial infections. Different inbred mouse strains differ considerably in susceptibility against Yersinia, with LD50 doses of resistant strains (e.g. C57BL/6) being 100-1000 fold higher after intravenous (iv) infection compared to susceptible strains (e.g. DBA/2). This difference in susceptibility seems to be under multigenic control and is not linked to the MHC locus H-2.
Earlier studies from our group have compared Yersinia-elicited gene expression patterns in bone marrow derived macrophages from resistant (C57BL/6) and susceptible (BALB/c) mice. These studies identified a number of candidate genes which have known immunomodulatory functions, so that it is likely that they modify the host immune response. Therefore, we will perform phenotype analysis of RI strains of mice to identify genetic traits determining resistance or susceptibility to Yersinia infections and to identify the corresponding gene regulatory networks.
Array-based gene expression analysis will be performed on spleen RNA from C57BL/6 and DBA/2 parental strains. Spleens will be removed from mock- and Y. enterocolitica-, infected mice and studied with Affymetrix MOE430_2 arrays. At the dose defined above, groups of 10 mice from each of the 30 BXD RI strains will be infected iv, and survival recorded. These survival data will be entered into a common database (GeneNetwork), and a first analysis for complex genetic traits will be performed to identify loci associated with resistance or susceptibility to infection.
WP 5: Susceptibility to Leishmania infection
Werner Solbach, Tamás Laskay (University of Lübeck)
Leishmaniasis is caused by infection with protozoan parasites of the genus Leishmania. Leishmaniasis is not a single disease but constitutes a variety of syndromes ranging from local, self-healing skin ulcers (cutaneous leishmaniasis) to a severe sand life-threatening systemic disease (visceral leishmaniasis). Transmitted to vertebrates by phlebotomine sandflies, the intracellular parasite Leishmania infects professional phagocytes such as neutrophils, monocytes, macrophages and dendritic cells. Studies of mouse models of leishmaniasis have provided important insights into the host response to Leishmania, and revealed a multigenic basis of susceptibility. Several studies indicated that resistance or susceptibility to infection is determined by the activation of different classes of T helper cells.
After intravenous injection of L. donovani, parasites spread to the spleen and liver during the early stage of infection. In susceptible strains a hundred-fold multiplication of the parasites is observed whereas in resistant strains, the parasites multiply only for a limited extent. Mapping of acute susceptibility to L. donovani using RI strains led to the identification of the Lsh locus on chromosome 1 which was subsequently identified as Nramp1. Among others, Nramp1 has been shown to contribute to an increased risk to visceral leishma-niasis in humans demonstrating that genetic analysis in mice can reveal susceptibility/resistance genes that are of major relevance for the human disease.
Our group has been working for many years on projects related to susceptibility and resistance to Leishmania infection. We have aimed to investigate the early events that are decisive for the preferential development of Th1 or Th2 cells. A local containment of parasites, e.g. the lack of systemic dissemination of L. major, was found to be characteristic for resistant but not for susceptible mice. Activation of Natural Killer (NK)-cells and the early expression of NK-cell activating chemokines were shown to be associated with resistance to Leishmania. On the other hand, an early CD62L-dependent migration of lymphocytes into the draining lymph node was found to correlate with susceptibility for the infection. Moreover, our group showed that neutrophil granulocytes can serve as temporary host cells and transport vehicles forLeishmania in the early phase of infection.
The aim of the proposed project is to identify genetic factors that control host resistance or susceptibility to Leishmania. For this, the well established murine model for the early stage L. donovani infection will be used. After intravenous infection, C57BL/6 are highly susceptible in the early stage of infection, whereas in DBA/2 mice only a limited level of parasite growth can be observed in liver and spleen. The strategy proposed here will be based on the use of BXD RI strains. To determine the phenotype (resistant or susceptible) of the RI strains, mice will be infected intravenously L. donovani amastigotes. The early course of infection will be monitored over time and compared with those of the parental strains. Parasite load will be determined at day 7 and day 14 in liver and spleen using limiting dilution parasite culture and quantitative PCR.
WP 6: Sepsis
Roland Lang, Jörg Mages (Technical University of Munich)
Even with the advent of modern intensive care and effective antibiotic treatment, severe sepsis remains the third leading cause of death. Sepsis is caused by an over-shooting immune response. Innate immune cells sense infectious danger via Toll like receptors (TLR) and activate a transcriptional program that includes cytokines, chemokines and inflammatory mediators. While this response is required to control local infection, over-whelming and systemic innate immune activation can lead to hypotonic shock, multiple organ failure and death. The prognosis of sepsis in patients is influenced by factors such as age, co-morbidity and environmental factors, however, the genetic make-up is also a very important component. In animal models using different inbred mouse strains a clear genetic contribution to the septic response was found.
The cell wall component lipopolysaccharide (LPS) is the major factor activating the innate response to gram-negative bacteria and has been used for decades in a very well characterized model of severe sepsis. The host response in the absence of a replicating pathogenrepresents a unique difference of the LPS challenge model compared to the infection models analysed by the other partners in the consortium. Recently, two studies have made use of RI strains to map complex genetic traits controlling cytokine expression after LPS-challenge. Here, we propose to analyse the genetics of global gene expression in the LPS septic shock model using the BXD RI strains.
We have studied the systemic response to LPS and mycobacterial infection in transgenic mice overexpressing IL-10 in macrophages and employed genome-wide expression analysis to in-vestigate the underlying mechanisms. To dissect the functional importance of individual components of the IL-10-Stat3 pathway and of IL-10-induced genes (e.g. DUSP1, SOCS3), we have made use of the respective gene targeted mice and analysed the response to LPS and cytokines in macrophages and in vivo. From this work, we have ample experience with the LPS challenge. Furthermore, we have extensively used tissue RNA for microarray analysis and obtained gene sets significantly associated with specific pathways and conditions.
The goal is to map genetic traits that control the systemic inflammatory re-sponse and the development of severe sepsis in the LPS challenge model using BXD RI strains. In addition to the cytokine profiles, we will monitor changes in body temperature and clinical chemistry parameters as indicators of organ damage. Differences in the genome-wide transcriptional response to LPS challenge between C57BL/6 and DBA/2 parental strains will be determined.
WP 7: Obesity and activation of the immune system
Gudrun Brockmann, Ayca Dogan, Armin Schmitt, Christina Neuschl (Humboldt-University of Berlin)
Obesity is a major risk factor for severe health impairment. Epidemiological observations indicate an association between malnutrition and susceptibility to infection on the one hand, and obesity and autoimmune diseases on the other hand. In particular, the role of adipose tissue as an endocrine organ and mediator of inflammation has been described. Beside leptin as the master adipose tissue hormone, the fat tissue releases proinflammatory cytokines as well as anti-inflammatory molecules. Transcriptional profiling studies have revealed that inflammatory and stress response genes are among the most abundantly regulated genes in adipose tissues of obese animals. Leptin has been identified as satiety factor, but also links the proinflammatory T helper-1 immune response to the nutritional status and energy balance.
Body weight and obesity as well as immune response are influenced by many genes. The studies proposed here will contribute to recognize early inflammation processes as result of obesity and to develop new strategies to suppress these inflammation processes in order to prevent later diseases.
In the past, we have used long term-selected mouse strains (DU6, BFMI) as polygenic models for human obesity. Linkage analyses in intercross pedigrees and advanced intercross lines, differential gene expression analyses and mutation screens in candidate genes have been performed. A set of chromosome substitution strains has been generated for the high body weight-selected mouse strain DU6, which allows the analysis of cis- and trans-acting genes located on the substituted chromosome. Five genes have been selected as candidate genes, which are currently investigated for genetic association to obesity in human populations.
C57BL/6 and DBA/2 represent average strains with respect to body weight and fat percentage and do not differ much on standard breeding diet. However, the weight gain varies drastically after eight weeks on atherogenic diet: DBA/2 responds by extremely high increase of body weight in females and in males, while C57BL/6 females show an intermediate response and males do not respond at all. Therefore, the BXD RI strain mapping panel represents a well suited genetic resource for the proposed research project.
Based on our expertise in mouse genetics of obesity, we will use a set of 30 RI strains to map quantitative trait loci that respond to high fat diet by diet-induced obesity and increased immune response. Initially, 10 males and 10 females of each of 30 BXD strains, C57BL/6, DBA/2, B6D2F1 and D2B6F1 will be phenotyped and genetic traits affecting diet-induced obesity and activation of immune response will be mapped.
WP 8: Susceptibility to liver fibrogenesis
Frank Lammert, Sonja Hillebrandt, Rabea Hall (Saarland University Medical Center, Homburg)
Liver fibrogenesis, or scarring of the liver, is the common end-stage of chronic liver diseases, in particular after chronic viral infections, causing at least 10,000 deaths per year in Germany. In the past decade key molecular pathomechanisms of hepatic fibrogenesis due to chronic viral infections were identified: activated hepatic stellate cells (HSCs) drive the process of de novo deposition of abnormal extracellular matrix, which is modulated by complex interactions between cytokines, receptors, and matrix components.
Several studies demonstrated that the course and progression of the fibrogenic response to chronic liver injury display significant variability among individual patients. The marked variability of fibrosis progression has been attributed to aetiology, age, gender, and environmental factors. Host genetic factors are critical, but are yet to be identified systematically.
Recently, our group identified new genetic determinants of liver fibrogenesis by quantitative trait locus analysis in experimental crosses between fibrosis-susceptible and resistant mouse strains. A quantitative trait locus that confers susceptibility to hepatic fibrosis was refined by in silico mapping and, using congenic mice and transgenesis, the group demonstrated that the Hc gene (encoding complement factor C5) underlies this locus. Furthermore, common haplotype-tagging polymorphisms of the human gene C5 were associated with advanced fibrosis in chronic hepatitis C virus infection. Thus, the mouse quantitative trait gene led to the identification of an unknown gene underlying human susceptibility to liver fibrosis, supporting the idea that C5 has a causal role in chronic inflammatory disorders and organ fibrogenesis across species. This study shows that the combination of current genomic technologies in the mouse can identify single small-risk genes in human.
We will study liver fibrogenesis in GRPs as a model for chronic liver injury secondary to viral infections, and map complex genetic traits that modulate gene expression and determine gene networks during liver fibrogenesis in GRPs. The project will be based on BXD RI strains.
The following assays are planned:
Liver fibrosis studies. Phenotyping protocols will include standardized histology, morphometry, biochemical quantification of hepatic collagen contents, serum surrogate markers of fibrosis, immunohistochemistry and expression profiling of proinflammatory and profibrogenic genes by qRT-PCR and Affymetrix microarrays.
Characterisation of liver cells. Liver immune cell fractions will be isolated and sorted according to SOPs developed in our laboratory. In addition, in co-operation with the technology platforms of the HepatoSys network of excellence (http://www.systembiologie.de), we will characterize primary HSCs that play critical roles in liver fibrogenesis with respect to proinflammatory responses during chronic liver inflammation.
WP 9: Activity-dependent control of adult neurogenesis
Gerd Kempermann, Norbert Hübner (Center for Regenerative Therapies, Dresden and Max Delbrück Center for Molecular Medicine, Berlin)
The adult hippocampus lifelong generates new neurons and does so in an activity-dependent manner. Adult hippocampal neurogenesis is discussed as an important factor in the development of neuropsychiatric disorders such as Alzheimer Disease and Major Depression. The adult brain can be trained and activity has preventive effects against neu-rodegenerative disorders and age-related cognitive decline.
One of the key issues that remains to be addressed is that the genes determining the process of adult neurogenesis are not necessarily identical to those genes that govern its activity-dependent control. It is thought that genes controlling regulation are of great biomedical rele-vance because they would help to identify targets for the therapeutic enhancement of cellular plasticity. Our current working hypothesis is that the sensitivity or responsiveness to environmental stimuli and activity is strongly influenced by genetic determinants, e.g. that the plastic response is genetically controlled and that these determinants are not identical to the underlying mechanisms of plasticity.
Our laboratory has studied adult neurogenesis and hippocampus-dependent learning in 12 BXD RI strains and found a correlation between parameters describing the acquisition of the task with adult neurogenesis. Based on this study we studied adult hippocampal neurogenesis in a total of 52 BXD and AXB/BXA strains. We analyzed the genetic covariance of adult neurogenesis with gene expression data to identify candidate genes with presumably key positions within the regulatory gene networks. We have obtained gene expression data from the hippocampi of all RI strains and are currently analyzing our data on adult neurogenesis in rela-tion to these new transcription data. The project proposed here, builds upon these studies. Availability of the BXD RI strains in the context of the proposed virtual institute will for the first time allow to introduce a behavioral/environmental variable and directly study a genotype - activity/environment interaction.
We intend to analyze the 26 classical strains used in our previous study plus 10 – 20 new strains that are selected by the presence of strand breaks in the chromosomal regions of inter-est identified in our previous studies, in particular proximal chromosome 5. Animals from each genotype will be housed in standard cages or an enriched environment for 28 days. Neurogenesis data will be obtained from the additional strains that have not yet been characterized. All strains will be tested with the Intellicage automated behavior analysis system (New-Behavior, Zürich) to obtain information about hippocampal learning (performance).
This system allows to generate a large amount of behavioral parameters (behavioral pheno-types) in the absence of an experimenter. mRNA will be collected from microdissected hip-pocampus from all strains and transcription profiles will be obtained. Data will be analyzed from various perspectives: classical complex genetic traits, analysis of genetic covariance, expression genetics. Covariance networks will be build and models of regulation will be generated.
Of particular interest will be the comparison of results from this study to the results of the other network partners who are mainly investigating biological responses with respect to in-fection and the immune system. We expect that in some cases, more general gene regulatory networks, e.g. cell proliferation, stress response, etc. will be in common to all systems whereas infection-related responses will group separately from the neurogenesis-specific gene activation program. Therefore, we particularly chose adult neurogenesis as a phenotypic trait in this consortium.
WP 10: Expression analysis
Jörg Mages, Reinhard Hoffmann, Roland Lang (Technical University of Munich)
Genetics of gene expression (eQTL) based on the synergistic analysis of phenotypical quantitative trait loci (QTL) and the associated gene expression patterns is an emerging method for identifying loci (or genes) responsible for the appearance of a phenotype. Several model systems have been successfully examined with eQTL analysis in mice until now, including obesity, type2 diabetes or ethanol sensitivity.
Over the last years, numerous RNA samples from various experimental sources (e.g. cell culture, ex vivo sorted cells and whole organs from mice) have been analysed at the BMBF-aided Microarray- and Bioinformatics Core Unit using the Affymetrix Microarrays with well estab-lished Standard Operating Procedures (SOP). In collaboration with Dr. Stefan Ehlers, different knock-out mice were studied after infection with M. tuberculosis. Using a stringent statistic (2-way ANOVA with multiple testing correction), we were able to identify two gene sets of which the expression was dependent on the respective mutated gene locus. Analysis of the global expression pattern revealed an unexpectedly high consistency between biological replicates with correlation factors > 0.980. Similar results were obtained analysing RNA from lungs of mice infected with Chlamydia pneumoniae or RNA from spleen of LPS treated mice.
All the genome-wide expression analysis proposed in this network program will be performed at the Microarray- and Bioinformatics Core Unit which is operated at the Institute of Medical Microbiology (TU Munich). This will ensure comparability of data between projects. For the statistical analysis of strain specific effects on the infection induced changes in expression profiles a two-way ANOVA model with subsequent correction for multiple testing will be used.
All RNA samples for microarrays from all work packages will be quality checked, labelled and hybridized under SOP in our facility. The workflow includes pre-experimental counselling in the experimental design, state of the art quality control of the starting material by Lab-on-a-chip technology (Bio-Rad Experion system) and spectrophotometry (NanoDrop ND-1000), as well as extensive quality control of the primary data (image analysis on spatial artefacts, background noise, 5´3´ ratios and overall homogeneity of mating samples). After nor-malization, screening for outliers in the samples due to impure organ preparation, etc. is done by visualisation methods and manually by using biological background information (e.g. detection of abundant transcripts of pancreatic enzymes in spleen samples).
WP 11: Results database
Klaus Schughart (Helmholtz Centre for Infection Research and University of Veterinary Medicine Hannover)
The data generated by the different groups of the virtual center will be collected in a common data base: GeneNetwork.
GeneNetwork consists of a set of linked resources for quantitative trait mapping and systems genetics and has been designed for multi-scale integration of networks of genes, transcripts, and phenotypic traits. In the case of the BXD RI strains, a large research community has collectively generated hundreds of thousands of transcript phenotypes in different tissues and cells, as well as hundreds of protein, cellular, pharmacologic, anatomical, and behavioural data types. Interpretation of all data from GeNeSys into GeneNetwork will allow correlations between phenotypes and expression data from various genetic traits.
WP 12: Susceptibility to influenza infection
Klaus Schughart (Helmholtz Centre for Infection Research and University of Veterinary Medicine Hannover)
WP 13: Immunophenotyping
Thure Adler (National Research Center for Environment and Health, Neuherberg)