Pathway Semantics: An Algebraic Data Driven Algorithm to Generate Hypotheses about Molecular Patterns Underlying Disease Progression

The overarching goal of the Pathway Semantics Algorithm (PSA) is to improve the in silico identification of clinically useful hypotheses about molecular patterns in disease progression. By framing biomedical questions within a variety of matrix representations, PSA has the flexibility to analyze combined quantitative and qualitative data over a wide range of stratifications. The resulting hypothetical answers can then move to in vitro and in vivo verification, research assay optimization, clinical validation, and commercialization. Herein PSA is shown to generate novel hypotheses about the significant biological pathways in two disease domains: shock / trauma and hemophilia A, and validated experimentally in the latter. The PSA matrix algebra approach identified differential molecular patterns in biological networks over time and outcome that would not be easily found through direct assays, literature or database searches. In this dissertation, Chapter 1 provides a broad overview of the background and motivation for the study, followed by Chapter 2 with a literature review of relevant computational methods. Chapters 3 and 4 describe PSA for node and edge analysis respectively, and apply the method to disease progression in shock / trauma. Chapter 5 demonstrates the application of PSA to hemophilia A and the validation with experimental results. The work is summarized in Chapter 6, followed by extensive references and an Appendix with additional material.

Sequencing of TGF-beta pathway genes in familial cases of intracranial aneurysm.

BACKGROUND AND PURPOSE: Familial aggregation of intracranial aneurysms (IA) strongly suggests a genetic contribution to pathogenesis. However, genetic risk factors have yet to be defined. For families affected by aortic aneurysms, specific gene variants have been identified, many affecting the receptors to transforming growth factor-beta (TGF-beta). In recent work, we found that aortic and intracranial aneurysms may share a common genetic basis in some families. We hypothesized, therefore, that mutations in TGF-beta receptors might also play a role in IA pathogenesis. METHODS: To identify genetic variants in TGF-beta and its receptors, TGFB1, TGFBR1, TGFBR2, ACVR1, TGFBR3, and ENG were directly sequenced in 44 unrelated patients with familial IA. Novel variants were confirmed by restriction digestion analyses, and allele frequencies were analyzed in cases versus individuals without known intracranial disease. Similarly, allele frequencies of a subset of known SNPs in each gene were also analyzed for association with IA. RESULTS: No mutations were found in TGFB1, TGFBR1, TGFBR2, or ACVR1. Novel variants identified in ENG (p.A60E) and TGFBR3 (p.W112R) were not detected in at least 892 reference chromosomes. ENG p.A60E showed significant association with familial IA in case-control studies (P=0.0080). No association with IA could be found for any of the known polymorphisms tested. CONCLUSIONS: Mutations in TGF-beta receptor genes are not a major cause of IA. However, we identified rare variants in ENG and TGFBR3 that may be important for IA pathogenesis in a subset of families.

ATM Signaling to TSC2: Mechanisms and Implications for Cancer Therapy

Ataxia telangiectasia mutated (ATM) is a critical component of the cellular response to DNA damage, where it acts as a damage sensor, and signals to a large network of proteins which execute the important tasks involved in responding to the damage, namely inducing cell cycle checkpoints, inducing DNA repair, modulating transcriptional responses, and regulating cell death pathways if the damage cannot be repaired faithfully. We have now discovered that an additional novel component of this ATM-dependent damage response involves induction of autophagy in response to oxidative stress. In contrast to DNA damage-induced ATM activation however, oxidative stress induced ATM, occurs in the cytoplasm, and does not require nuclear-to-cytoplasmic shuttling of ATM. Using several cell culture systems including MCF7 breast carcinoma cells, SKOV3 ovarian cancer cells, and various lineages of mouse embryonic fibroblasts, we showed that once activated by reactive oxygen species (ROS), ATM signals to mTORC1 to induce autophagy via the LKB1-AMPK-TSC2 pathway. Targeting dysregulation of mTORC1 in Atm-deficient mice, which succumb to lymphomagenesis within 3-4 months of age with daily administration of rapamycin, could significantly extend survival and cause regression of tumors, suggesting that pharmacologically targeting this pathway has therapeutic implications in cancer. We also identified a second contrasting pathway for DNA damage-induced mTORC1 repression which does not require AMPK activation, but does require ATM and TSC2. Several potential mechanisms including mTOR localization and p53-mediated pathways were ruled out however we identified that TSC2 may be an additional cytoplasmic direct ATM substrate that is engaged in response to DNA damage specifically. Lastly, a study was performed to examine whether autophagy induced by ovarian cancer therapeutics (focusing on cisplatin, since paclitaxel does not induce autophagy in the SKOV3 cell line model we used) plays a role in resistance to therapy since autophagy can play both pro-survival mechanisms or be a mechanism of cell death. Using a genetic approach to knock-down Atg5 expression with shRNA in SKOV3 ovarian carcinoma cells, we compared the cytotoxicity of cisplatin in vector or Atg5 knock-down cells, and demonstrated that autophagy does not play any significant role in the response to cisplatin in this cell line.

Identification of germ cell tumor modifier genes from the 129.MOLF-Chr19 consomic mouse strain

Naturally occurring genetic variants confer susceptibility to disease in the human population, including in testicular germ cell tumor development. Disease susceptibility loci for testicular germ cell tumors have been identified by genetic mapping in humans and mice. However, the identity of many of the susceptibility genes remains unclear. My study utilized a chromosome substitution strain, the 129.MOLF-Chr 19 (or M19 strain), to identify candidate testicular germ cell tumor susceptibility genes. Males of this strain have a high incidence of germ cell tumors in the testes. By forward genetic approaches, five susceptibility loci were fine-mapped and the genetic interactions were dissected. In addition, I identified three protein-coding genes and one micro-RNA as testicular tumor susceptibility genes by genomic screening. Using reverse genetic approaches, I verified one of the candidates, Splicing factor 1, as a modifier of testicular tumor. Deficiency of SF1 significantly reduces the incidence of testicular tumors in mice. This study highlights the advantage of the 129.MOLF-Chr 19 consomic strain in disease gene identification and validation. It also sets the stage to elucidate the molecular mechanisms of tumorigenesis in the testis. ^

Identification of diabetic retinopathy genes through a genome-wide association study among Mexican-Americans from Starr County, Texas

To identify genetic susceptibility loci for severe diabetic retinopathy, 286 Mexican-Americans with type 2 diabetes from Starr County, Texas completed detailed physical and ophthalmologic examinations including fundus photography for diabetic retinopathy grading. 103 individuals with moderate-to-severe non-proliferative diabetic retinopathy or proliferative diabetic retinopathy were defined as cases for this study. DNA samples extracted from study subjects were genotyped using the Affymetrix GeneChip® Human Mapping 100K Set, which includes 116,204 single nucleotide polymorphisms (SNPs) across the whole genome. Single-marker allelic tests and 2- to 8-SNP sliding-window Haplotype Trend Regression implemented in HelixTreeTM were first performed with these direct genotypes to identify genes/regions contributing to the risk of severe diabetic retinopathy. An additional 1,885,781 HapMap Phase II SNPs were imputed from the direct genotypes to expand the genomic coverage for a more detailed exploration of genetic susceptibility to diabetic retinopathy. The average estimated allelic dosage and imputed genotypes with the highest posterior probabilities were subsequently analyzed for associations using logistic regression and Fisher's Exact allelic tests, respectively. To move beyond these SNP-based approaches, 104,572 directly genotyped and 333,375 well-imputed SNPs were used to construct genetic distance matrices based on 262 retinopathy candidate genes and their 112 related biological pathways. Multivariate distance matrix regression was then used to test hypotheses with genes and pathways as the units of inference in the context of susceptibility to diabetic retinopathy. This study provides a framework for genome-wide association analyses, and implicated several genes involved in the regulation of oxidative stress, inflammatory processes, histidine metabolism, and pancreatic cancer pathways associated with severe diabetic retinopathy. Many of these loci have not previously been implicated in either diabetic retinopathy or diabetes. In summary, CDC73, IL12RB2, and SULF1 had the best evidence as candidates to influence diabetic retinopathy, possibly through novel biological mechanisms related to VEGF-mediated signaling pathway or inflammatory processes. While this study uncovered some genes for diabetic retinopathy, a comprehensive picture of the genetic architecture of diabetic retinopathy has not yet been achieved. Once fully understood, the genetics and biology of diabetic retinopathy will contribute to better strategies for diagnosis, treatment and prevention of this disease.^

Genetic variability and the hypothalamic control of energy balance

Background. Obesity is a major health problem throughout the industrialized world. Despite numerous attempts to curtail the rapid growth of obesity, its incidence continues to rise. Therefore, it is crucial to better understand the etiology of obesity beyond the concept of energy balance.^ Aims. The first aim of this study was to first investigate the relationship between eating behaviors and body size. The second goal was to identify genetic variation associated with eating behaviors. Thirdly, this study aimed to examine the joint relationships between eating behavior, body size and genetic variation.^ Methods. This study utilized baseline data ascertained in young adults from the Training Interventions and Genetics of Exercise (TIGER) Study. Variables assessed included eating behavior (Emotional Eating Scale, Eating Attitudes Test-26, and the Block98 Food Frequency Questionnaire), body size (body mass index, waist and hip circumference, waist/hip ratio, and percent body fat), genetic variation in genes implicated related to the hypothalamic control of energy balance, and appropriate covariates (age, gender, race/ethnicity, smoking status, and physical activity. For the genetic association analyses, genotypes were collapsed by minor allele frequency, and haplotypes were estimated for each gene. Additionally, Bayesian networks were constructed in order to determine the relationships between genetic variation, eating behavior and body size.^ Results. We report that the EAT-26 score, Caloric intake, percent fat, fiber intake, HEAT index, and daily servings of vegetables, meats, grains, and fats were significantly associated with at least one body size measure. Multiple SNPs in 17 genes and haplotypes from 12 genes were tested for their association with body size. Variation within both DRD4 and HTR2A was found to be associated with EAT-26 score. In addition, variation in the ghrelin gene (GHRL) was significantly associated with daily Caloric intake. A significant interaction between daily servings of grains and the HEAT index and variation within the leptin receptor gene (LEPR) was shown to influence body size.^ Conclusion. This study has shown that there is a substantial genetic component to eating behavior and that genetic variation interacts with eating behavior to influence body size.^

Genetic epidemiology of childhood brain tumors

A rare familial cancer syndrome involving childhood brain tumors (CBT), breast cancer, sarcomas and an array of other tumors has been described (Li and Fraumeni 1969, 1975, 1982, 1987). A survey of CBT identified through the Connnecticut Tumor Registry in 1984 revealed a high frequency of CBT, leukemia and other childhood cancer in siblings of CBT patients (Farwell and Flannery, 1984). Other syndromes such as neurofibromatosis and nevoid basal cell carcinoma syndrome have also been associated with CBT; however, no systematic family studies have been conducted to determine the extent to which cancer aggregates in family members of CBT patients. This family study was designed to determine the frequency of cancer aggregation overall or at specific sites, to determine the frequency of known or potentially hereditary syndromes in families of CBT patients, and to determine a genetic model to characterize familial cancer syndromes and to identify specific kindreds to which such a model(s) might apply. This study includes 244 confirmed CBT patients referred to the University of Texas M. D. Anderson Cancer Center between the years 1944 and 1983, diagnosed under the age of 15 years and resident in the U.S. or Canada. Family histories were obtained on the proband's first (parents, siblings and offspring) and second degree (proband's aunts, uncles and grandparents) relatives following sequential sampling scheme rules. To determine if cancer aggregates in families, we compared the cancer experience in the population to that expected in the general population using Connecticut Tumor Registry calendar year, age, race and sex-specific rates. The standardized incidence ratio (SIR) for cancer overall was 0.91 (41 observed (O) and 44.94 expected (E); 95% Confidence Interval (CI) = 0.65-1.24). We observed a significant excess of colon cancer among the proband's first degree relatives (O/E = 5/1.64; 95% CI = 1.01-7.65), in particular those under age 45 year. Segregation analysis showed evidence for multifactorial inheritance in the small percentage (N = 5) of the families. ^

HEMATOCRIT AND HEMOGLOBIN, ATP AND DPG CONCENTRATIONS IN ANDEAN MAN: VARIABILITY BY SEX, AGE, VILLAGE, ALTITUDE, WEIGHT, SMOKING HABITS AND GENETIC CONSTITUTION

The Departmento de Arica in northern Chile was chosen as the investigation site for a study of the role of certain hematologic and glycolytic variables in the physiological and genetic adaptation to hypoxia.^ The population studied comprised 876 individuals, residents of seven villages at three altitudes: coast (0-500m), sierra (2,500-3,500m) and altiplano (> 4,000m). There was an equal number of males and females ranging in ages from six to 90 years. Although predominantly Aymara, those of mixed or Spanish origin were also examined. The specimens were collected in heparinized vacutainers precipitated with cold trichloroacetic acid (TCA) and immediately frozen to -196(DEGREES)C. Six variables were measured. Three were hematologic: hemoglobin, hematocrit and mean cell hemoglobin concentration. The three others were glycolytic: erythrocyte 2,3-diphosphoglycerate (DPG), adenosine triphosphate (ATP) and the percentage of phosphates (DPG + ATP) in the form of DPG.^ Hemoglobin and hematocrit were measured on site. The DPG and ATP content was assayed in specimens which had been frozen at -196(DEGREES)C and transported to Houston. Structured interviews on site provided information as to lifestyle and family pedigrees.^ The following results were obtained: (1) The actual village, rather than the altitude, of examination accounted for the greatest proportion of the variance in all variables. In the coast, a large difference in levels of ionic lithium in the drinking water exists. The chemical environment of food and drink is postulated to account, in part, for the importance of geographic location in explaining the observed variance. (2) Measurements of individuals from the two extreme altitudes, coast and altiplano, did not exhibit the same relationship with age and body mass. The hematologic variables were significantly related to both age and body build in the coast. The glycolytic variables were significantly related to age and body mass in the altiplano. (3) The environment modified male values more than female values in all variables. The two sexes responded quite differently to age and changes in body mass as well. The question of differing adaptability of the two sexes is discussed. (4) Environmental factors explained a significantly higher proportion of total variability in the altiplano than in the coast for hemoglobin, hematocrit and DPG. Most of the ATP variability at both altitudes is explained by genetic factors. ^

Effects of thymus size and involution on the contribution of recent thymic emigrants to the peripheral T cell pool

The contribution of recent thymic emigrants (RTEs) to the peripheral naïve T cell population is necessary to maintain diversity of the T cell receptor (TCR) repertoire and produce immune responses against newly encountered antigens. The thymus involutes with age, after irradiation or chemotherapy, and due to severe viral infections. Thymus involution results in decreased thymopoiesis and RTE output leading to a reduced diversity of peripheral T cells. This increases susceptibility to disease and impairs immune responsiveness to vaccines. Therefore, studies aimed at maintaining or regenerating thymic function are integral for maintaining and restoring peripheral TCR diversity. Mice that express a K5.CyclinD1 transgene expression have a severely hyperplastic thymus that fails to undergo involution. Both thymocyte and TEC development appear normal in these mice. We have used the K5.CyclinD1 transgenic model to test the hypothesis that preventing thymus involution will sustain RTE output and incorporation into the peripheral T cell pool to prevent naïve T cell depletion with age. The K5.CyclinD1 transgene was crossed to the RAG2p-GFP transgenic model so that RTEs could be tracked by the intensity of the GFP signal. The frequency and number of RTEs in naïve CD4 splenic T cells was analyzed at monthly intervals to 5 months of age. Using this double transgenic approach, we determined that preventing thymus involution does maintain or enhance the number of RTEs in the peripheral T cell pool before and after thymus involution.

ATXA -dependent gene expression in Bacillus anthracis

The Bacillus anthracis toxin genes, cya, lef , and pag, can be viewed as a regulon, in which transcription of all three genes is activated in trans by the same regulatory gene, atxA, in response to the same signal, CO2. I determined that several phenotypes are associated with the atxA gene. In addition to being toxin-deficient, an atxA -null mutant grows poorly on minimal media and sporulates early compared to the parent strain. Furthermore, an atxA-null mutant has an altered 2-D gel protein profile. I used a genetic approach to find additional atxA-regulated genes. Random transcriptional lacZ fusions were generated in B. anthracis using transposon Tn 917-LTV3. Transposon-insertion libraries were screened for mutants expressing increased β-galactosidase activity in 5% CO2. Introduction of an atxA-null mutation in these mutants revealed that 79% of the CO2-regulated fusions were also atxA-dependent. DNA sequence analysis of transposon insertion sites in mutants carrying CO 2/atxA-regulated fusions revealed ten mutants harboring transposon insertions in loci distinct from the toxin genes. The majority of the tcr (toxin co-regulated) loci mapped within the pXO1 pathogenicity island. These results indicate a clear association of atxA with CO2-enhanced gene expression in B. anthracis and provide evidence that atxA regulates genes other than the structural genes for the anthrax toxin proteins. ^ Characterization of one tcr locus revealed a new regulatory gene, pagR. The pagR gene (300 nt) is located downstream of pag. pagR is cotranscribed with pag and is responsible for autogenous control of the operon. pagR also represses expression of cya and lef. Repression of toxin gene expression by pagR may be mediated by atxA. The steady state level of atxA mRNA is increased in a pagR mutant. Recombinant PagR protein purified from Escherichia coli did not specifically bind the promoter regions of pagA or atxA. An unidentified factor in B. anthracis crude extracts, however, was able to bind the atxA promoter in the absence of PagR or AtxA. These investigations increase our knowledge of virulence regulation in B. anthracis and ultimately will lead to a better understanding of anthrax disease. ^