In vivo Bovine Brucellosis Pathogenesis: An in silico Interactome Model . Garry Adams 1, Carlos A. Rossetti 2, Sangeeta Khare 1, Sara D. Lawhon 1, Harris A. Lewin 3, Mary S. Lipton 4, Joshua E. Turse 4, Dennis C. Wylie 5, Yu Bai 5, and Kenneth L. Drake 5 1 Texas A&M University, College Station, TX, 2 IP – INTA, Bs. As. , Argentina 3 University of Illinois at Urbana-Champaign, IL, 4 Pacific Northwest National Laboratory, Richland, WA, and 5 Seralogix LLC, Austin, TX Abstract Interactome Modeling Results Systems biology is bringing a new, more robust approach to pathogenesis that is based upon understanding the host-pathogen interactions and its impact on the host’s molecular network of the immune system. A computational systems biology method was utilized to create an in silico interactome model (pathogen protein interactions with the host) of the bovine host responses to Brucella melitensis (BMEL). A bovine ligated ileal loop biological model was employed to capture the host gene expression response at multiple time points post infection concurrently with the pathogen gene expressions. Computational methods based on Dynamic Bayesian Network (DBN) machine learning were employed to conduct pathogenicity analysis and to create the in silico interactome model. For the bovine host, 219 signaling and metabolic pathways and 1620 Gene Ontology (GO) categories were profiled for perturbation against the control (uninfected condition) which defined the host’s biosignatures to the BMEL response. For the pathogen, 119 signaling and metabolic pathways and 781 GO categories were similarly profiled. Through this DBN computational approach, the method identified significantly perturbed pathways and GO category groups of genes that define the pathogenicity signatures of the host and pathogen response. Our preliminary results provide deeper understanding of the overall complexity of the bovine host innate immune response as well as the identification of host gene perturbations that defined a unique host temporal biosignature response to BMEL. The application of advanced computational methods for developing interactome models based on DBNs has proven to be instrumental in identifying novel host responses and improved functional biological insight into the host defensive mechanisms. Introduction The complexity of host-pathogen interactions requires a system level understanding of the entire hierarchy of biological interactions and dynamics. A systems biology approach can provide systematic insights into the dynamic/temporal difference in gene regulation, interaction, and function, and thereby deliver an improved understanding and more comprehensive hypotheses of the underlying mechanisms Moreover, the ability to consolidate complex data and knowledge into plausible interactome models is essential to promote the effective discovery of key points of interaction. Accordingly, a systems biology approach to study molecular pathway gene expression profiles of host cellular responses to microbial pathogens holds great promise as a methodology to identify, model and predict the overall dynamics of the host-pathogen interactome (1). Computational capabilities are emerging for creating host-pathogen interactome models. Such models, utilizing data at the genomic, proteomic, metabolomics, etc. levels can be used to learn and understand the underlying mechanisms and points of interaction governing the host innate and adaptive responses to pathogens. Such models are envisioned to play an increasingly integral part in the vaccine and immunotherapeutic development process, with incremental model improvements accruing as new biological knowledge is collected from translational in vivo and ex vivo efficacy and safety studies (nonclinical through clinical trials). The interactome model cannot just be a list of possible interaction prediction, but must be part of a dynamic model in which the relationships governing the host immune response can be captured, interpreted and refined. The interactome model becomes a tool that can be interrogated and employed in simulation to help guide vaccine development and/or immunotherapeutic drug candidate selections (2). Experimental verification will always be a necessary element, and as such experiments are conducted the resulting biological information should be retained and employed as new biological knowledge for creating the next refined interactome model. Fig. 1. MAPK Pathway trained and scored employing DBGGA. This model represents the state of gene expression at 120 min p. i. By interrogating the model, important mechanistic genes and their strength of relationships (down stream influence) can be ascertained to achieve improved understanding of the influence of B. melitensis on the bovine host response. Experimental Methods Gene expression data were collected at Texas A&M University College of Veterinary Medicine from a methodology developed for simultaneously collecting total RNA from both pathogen and host. The bovine ligated ileal loop model below was used to study the host response infected with Brucella melitensis 16 M. Under BSL 3 condition, non-survival surgeries were performed on 3 -week old male pathogen-free beef calves. Host RNA was collected and co-hybridized against bovine reference RNA & 13 K custom bovine arrays (UIUC) to allow for cross-comparison between experimental conditions (3). Non-infected ligated ileal loops where used as healthy state controls. Pathogen RNA was enriched & amplified from total RNA samples (4). The intracellular in vivo pathogen gene expression was compared to the gene expression in the inoculum (cultures of the pathogen at late-log growth phase) (5). Figure 3. A. Bovine HOST pathways perturbation scores, and B. B. melitensis Fig. 2. Heatmap of gene expression for the HOST. MAPK Pathway. RED is UP regulated and GREEN is DOWN regulated. GRAY is no or MINIMAL change between infection & control. Pathogen pathways perturbation scores. Selected pathways are from important immunological categories: Cell Communication, Cell Growth and Death, Cell Motility, Immune System, Membrane Transport, Signal Transduction, and Signaling Molecules and Interaction. Pathway descriptions and networks were obtained from publically available resources, i. e. , KEGG and Biocarta. Computational Methods Pathway Modeling and Analyses. A novel computational approach based on dynamic Bayesian network (DBN) machine learning was employed to conduct pathogenicity analysis and to create the in silico interactome model (6). The computational tools scored and ranked 219 known bovine metabolic and signaling pathways. The computational approach identified groups of interrelated genes that as a whole represents the activation (perturbation) or suppression of a pathway over the time-course of these experiments (15 min, 30 min, 1 hr, 2 hr, 4 hr p. i. ). This technique is termed “Dynamic Bayesian Gene Group Activation” (DBGGA). DBGGA enables the scoring and ranking of groups of genes across all time points in lieu of just individual genes in a single time point to determine differences between infected and control conditions. For example, the MAPK pathway has a known network structure which defines the static causal relationships for our DBN model (Fig 1). This model is trained with the control (non-infected) expression data which determines the DBN parameters (covariances, weights, means) through an Expectation Maximization algorithm. The trained model is used to test the “goodness of fit” of the experimental (infected) data that results in a log likelihood measure of divergence that is then transformed to a z-score test statistic (Bayesian Z-score). Interrogation of the model identifies important relationships (arcs between genes) and mechanistic genes (Fig 2) that are the source of pathway perturbation. The magnitude of perturbation of these pathways are ranked and visualized as a heatmap (Fig 3) that converges attention on those gene groups that are uniquely perturbed and selected for use as building blocks in the pathogen-host interactome model. Interactome Modeling Between Pathogen and Host. New computational methods for creating hostpathogen “interactome models” have been developed for predicting protein-protein interactive (PPI) relationships and networks from in vivo experimental studies where host-pathogen gene co-expressions are measured simultaneously (6). The method integrates existing prior biological knowledge with gene coexpression data to make PPI predictions combined with structure learning techniques. The gene expression data (host and pathogen) are employed to learn potential points of interaction between the two biological systems that produces a unique mapping of PPIs between the pathogen and host (Fig 4). Three predictive learning algorithms are combined in the PPI structure learning process: 1) Protein-domain Interaction Transference; 2) Sequence similarity Interaction Transference; and 3) Gene Ontology-based Functional Algorithm. Prior knowledge comes from public databases that include KEGG, Pfam, NCBI, and others. Fig. 4. The pathogen-host interactome model is composed of 526 gene nodes of which 117 are BMEL genes from the BMEL pathways (Fig 3 B) and the remaining host genes selected from 12 highly perturbed pathways (Fig 3 A). Interrogation of the model produced a list of novel PPIs (Table 1) which potentially interact with host immune response pathways. The PPI list represent highly positive and negative correlated relations learned from prior knowledge and the co-expressed gene data. Selected References 1. 2. 3. 4. 5. 6. Adams, L. G. , S. Khare, S. D. Lawhon, C. A. Rossetti, H. A. Lewin, M. S. Lipton, J. E. Turse, D. C. Wylie, Y. Bai, K. L. Drake. 2011. Multi-comparative systems biology analysis reveals time-course biosignatures of in vivo bovine pathway responses to B. melitensis, S. enterica Typhimurium and M. avium paratuberculosis. BMC Proceedings of the International Symposium for Animal Genomics for Animal Health (AGAH 201). Paris, France, May 30 -June 2, 2010. BMC Proceedings 2011, 5(Suppl 4): S 6 doi: 10. 1186/1753 -6561 -5 -S 4 -S 6. Adams, L. G. , S. Khare, S. D. Lawhon, C. A. Rossetti, H. A. Lewin, M. S. Lipton, J. E. Turse, D. C. Wylie, Y. Bai, K. L. Drake. 2011. Enhancing the Role of Veterinary Vaccines in Reducing Zoonotic Disease Of Humans: Linking Systems Biology with Vaccine Delivery. Vaccine. doi: 10. 1016/j. vaccine. 2011. 05. 080 PMC 21651944. Rossetti, C. A. , C. L. Galindo, R. E. Everts, H. A. Lewin, H. R. Garner, and Adams, L. G. 2011. Comparative analysis of the early transcriptome of Brucella abortus - infected monocytederived macrophages from cattle naturally resistant or susceptible to brucellosis. Research in Veterinary Science. 91: 40 -51. DOI: 10. 1016/j. rvsc. 2010. 09. 002. Rossetti CA, Galindo C. L. , Garner H. R. , Adams L. G. 2010. Selective amplification of Brucella melitensis m. RNA from a mixed host-pathogen total RNA. BMC Res Notes. 2010; 3: 244. PMCID: 2954846. Rossetti, C. A. , C. L. Galindo, H. R. Garner and L. G. Adams. 2011. Transcriptional profile of the intracellular pathogen Brucella melitensis following He. La cells invasion. Microbial Pathogenesis. 51: 338 -344. Bai, Y. , D. C. Wylie, L. G. Adams, S. Khare, S. D. Lawhon, C. A. Rossetti, and K. L. Drake. Bio. Signature Discovery System (Biosignature. DS): A next generation solution for OMICS analysis and modeling in drug discovery and personalized medicine. Submitted. Acknowledgements We thank Dr. Thomas A. Ficht for providing the B. melitensis 16 M strain. We are grateful to the NIH Western Regional Center of Excellence (WRCE) Pathogen Expression Core (Drs. Mitchell Mc. Gee, Rhonda Friedberg and Stephen A. Johnston, Arizona State Univ. ) for developing and printing the B. melitensis c. DNA microarrays. This study was supported by U. S. Department of Homeland Security – National Center of Excellence for Foreign Animal and Zoonotic Disease (FAZD) Defense grant ONR-N 00014 -04 -1 -0 and a NIH grant 2 U 54 AI 05715606. Conclusions • Computational methods were developed for predicting host-pathogen protein-protein interactions (PPIs) from combined host-pathogen gene expressions • Putative host-pathogen interaction mechanisms were identified by the interactome network • Systems biology analysis of molecular pathway gene expression profiles of host cellular responses to B. melitensis were modeled to identify and predict the overall dynamics of the in silico host-pathogen interactome • Current experiments are underway to validate the in silico interactome model using in vitro phenotyping of mutants of putative mechanistic genes of B. melitensis • Current experiments are underway to validate the in silico interactome model using in vitro phenotyping si. RNA knockdowned putative host mechanistic genes in RAW 267. 4 macrophages • Systems biology analysis of the in silico interctome may potentially guide the search for improved therapeutics, vaccines and diagnostics for brucellosis
Скачать презентацию In vivo Bovine Brucellosis Pathogenesis An in silico
4075750bf8ec93aa88e0930d4d85af08.ppt