SURVIVAL PREDICTION FOR BRAIN TUMOR PATIENTS USING GENE EXPRESSION DATA

Brain tumor is one of the most aggressive types of cancer in humans, with an estimated median survival time of 12 months and only 4% of the patients surviving more than 5 years after disease diagnosis. Until recently, brain tumor prognosis has been based only on clinical information such as tumor grade and patient age, but there are reports indicating that molecular profiling of gliomas can reveal subgroups of patients with distinct survival rates. We hypothesize that coupling molecular profiling of brain tumors with clinical information might improve predictions of patient survival time and, consequently, better guide future treatment decisions. In order to evaluate this hypothesis, the general goal of this research is to build models for survival prediction of glioma patients using DNA molecular profiles (U133 Affymetrix gene expression microarrays) along with clinical information. First, a predictive Random Forest model is built for binary outcomes (i.e. short vs. long-term survival) and a small subset of genes whose expression values can be used to predict survival time is selected. Following, a new statistical methodology is developed for predicting time-to-death outcomes using Bayesian ensemble trees. Due to a large heterogeneity observed within prognostic classes obtained by the Random Forest model, prediction can be improved by relating time-to-death with gene expression profile directly. We propose a Bayesian ensemble model for survival prediction which is appropriate for high-dimensional data such as gene expression data. Our approach is based on the ensemble "sum-of-trees" model which is flexible to incorporate additive and interaction effects between genes. We specify a fully Bayesian hierarchical approach and illustrate our methodology for the CPH, Weibull, and AFT survival models. We overcome the lack of conjugacy using a latent variable formulation to model the covariate effects which decreases computation time for model fitting. Also, our proposed models provides a model-free way to select important predictive prognostic markers based on controlling false discovery rates. We compare the performance of our methods with baseline reference survival methods and apply our methodology to an unpublished data set of brain tumor survival times and gene expression data, selecting genes potentially related to the development of the disease under study. A closing discussion compares results obtained by Random Forest and Bayesian ensemble methods under the biological/clinical perspectives and highlights the statistical advantages and disadvantages of the new methodology in the context of DNA microarray data analysis.

Host-pathogen interactions of secreted and surface Staphylococcus aureus factors

Staphylococcus aureus is an opportunistic bacterial pathogen that can infect humans and other species. It utilizes an arsenal of virulence factors to cause disease, including secreted and cell wall anchored factors. Secreted toxins attack host cells, and pore-forming toxins destroy target cells by causing cell lysis. S. aureus uses cell-surface adhesins to attach to host molecules thereby facilitating host colonization. The Microbial Surface Components Recognizing Adhesive Matrix Molecules (MSCRAMMs) are a family of cell-wall anchored proteins that target molecules like fibronectin and fibrinogen. The Serine-aspartate repeat (Sdr) proteins are a subset of staphylococcal MSCRAMMs that share similar domain organization. Interestingly, the amino-terminus, is composed of three immunoglobulin-folded subdomains (N1, N2, and N3) that contain ligand-binding activity. Clumping factors A and B (ClfA and ClfB) and SdrG are Sdr proteins that bind to fibrinogen (Fg), a large, plasma glycoprotein that is activated during the clotting cascade to form fibrin. In addition to recognizing fibrinogen, ClfA and ClfB can bind to other host ligands. Analysis of S. aureus strains that cause osteomyelitis led to the discovery of the bone-sialoprotein-binding protein (Bbp), an Sdr protein. Because several MSCRAMMs target more than one molecule, I hypothesized that Bbp may recognize other host proteins. A ligand screen revealed that the recombinant construct BbpN2N3 specifically recognizes human Fg. Surface plasmon resonance was used to determine the affinity of BbpN2N3 for Fg, and a dissociation constant of 540 nM was determined. Binding experiments performed with recombinant Fg chains were used to map the binding of BbpN2N3 to the Fg Aalpha chain. Additionally, Bbp expressed on the surface of Lactococcus lactis and S. aureus Newman bald mediated attachment of these bacteria to Fg Aalpha. To further characterize the interaction between the two proteins, isothermal titration calorimetry and inhibition assays were conducted with synthetic Fg Aalpha peptides. To determine the physiological implications of Bbp binding to Fg, the effect of Bbp on fibrinogen clotting was studied. Results show that Bbp binding to Fg inhibits the formation of fibrin. The consequences of this interaction are currently under investigation. Together, these data demonstrate that human Fg is a novel ligand for Bbp. This study indicates that the MSCRAMM Bbp may aid in staphylococcal attachment by targeting both an extracellular matrix and a blood plasma protein. The implications of these novel findings are discussed.

THE CONSEQUENCES OF DISRUPTING THE MDM2-P53 BALANCE IN HEMATOPOIESIS

The bone marrow accommodates hematopoietic stem cells and progenitors. These cells provide an indispensible resource for replenishing the blood constituents throughout an organism’s life. A tissue with such a high turn-over rate mandates intact cycling checkpoint and apoptotic pathways to avoid inappropriate cell proliferation and ultimately the development of leukemias. p53, a major tumor suppressor, is a transcription factor that regulates cell cycle, and induces apoptosis and senescence. Mice inheriting a hypomorphic p53 allele in the absence of Mdm2, a p53 inhibitor, have elevated p53 cell cycle activity and die by postnatal day 13 due to hematopoietic failure. Hematopoiesis progresses normally during embryogenesis until it moves to the bone marrow in late development. Increased oxidative stress in the bone marrow compartment postnatally is the impediment for normal hematopoiesis via activation of p53. p53 in turn stimulates the generation of more reactive oxygen species and depletes bone marrow cellularity. Also, p53 exerts various defects on the hematopoietic niche by increasing mesenchymal lineage populations and their differentiation. Hematopoietic defects are rescued with antioxidants or when cells are cultured at low oxygen levels. Deletion of p16 partially rescues bone marrow cellularity and progenitors via a p53-independent pathway. Thus, although p53 is required to inhibit tumorigenesis, Mdm2 is required to control ROS-induced p53 levels for sustainable hematopoiesis and survival during homeostasis.

Mutations in the cofilin partner Aip1/Wdr1 cause autoinflammatory disease and macrothrombocytopenia.

A pivotal mediator of actin dynamics is the protein cofilin, which promotes filament severing and depolymerization, facilitating the breakdown of existing filaments, and the enhancement of filament growth from newly created barbed ends. It does so in concert with actin interacting protein 1 (Aip1), which serves to accelerate cofilin's activity. While progress has been made in understanding its biochemical functions, the physiologic processes the cofilin/Aip1 complex regulates, particularly in higher organisms, are yet to be determined. We have generated an allelic series for WD40 repeat protein 1 (Wdr1), the mammalian homolog of Aip1, and report that reductions in Wdr1 function produce a dramatic phenotype gradient. While severe loss of function at the Wdr1 locus causes embryonic lethality, macrothrombocytopenia and autoinflammatory disease develop in mice carrying hypomorphic alleles. Macrothrombocytopenia is the result of megakaryocyte maturation defects, which lead to a failure of normal platelet shedding. Autoinflammatory disease, which is bone marrow-derived yet nonlymphoid in origin, is characterized by a massive infiltration of neutrophils into inflammatory lesions. Cytoskeletal responses are impaired in Wdr1 mutant neutrophils. These studies establish an essential requirement for Wdr1 in megakaryocytes and neutrophils, indicating that cofilin-mediated actin dynamics are critically important to the development and function of both cell types.

Estimation of the incidence of a rare genetic disease through a two-tier mutation survey.

Recent attempts to detect mutations involving single base changes or small deletions that are specific to genetic diseases provide an opportunity to develop a two-tier mutation-screening program through which incidence of rare genetic disorders and gene carriers may be precisely estimated. A two-tier survey consists of mutation screening in a sample of patients with specific genetic disorders and in a second sample of newborns from the same population in which mutation frequency is evaluated. We provide the statistical basis for evaluating the incidence of affected and gene carriers in such two-tier mutation-screening surveys, from which the precision of the estimates is derived. Sample-size requirements of such two-tier mutation-screening surveys are evaluated. Considering examples of cystic fibrosis (CF) and medium-chain acyl-CoA dehydrogenase deficiency (MCAD), the two most frequent autosomal recessive disease in Caucasian populations and the two most frequent mutations (delta F508 and G985) that occur on these disease allele-bearing chromosomes, we show that, with 50-100 patients and a 20-fold larger sample of newborns screened for these mutations, the incidence of such diseases and their gene carriers in a population may be quite reliably estimated. The theory developed here is also applicable to rare autosomal dominant diseases for which disease-specific mutations are found.

The recombinant adeno-associated virus vector (rAAV2)-mediated apolipoprotein B mRNA-specific hammerhead ribozyme: a self-complementary AAV2 vector improves the gene expression.

BACKGROUND: In humans, overproduction of apolipoprotein B (apoB) is positively associated with premature coronary artery diseases. To reduce the levels of apoB mRNA, we have designed an apoB mRNA-specific hammerhead ribozyme targeted at nucleotide sequences GUA6679 (RB15) mediated by adenovirus, which efficiently cleaves and decreases apoB mRNA by 80% in mouse liver and attenuates the hyperlipidemic condition. In the current study, we used an adeno-associated virus vector, serotype 2 (AAV2) and a self-complementary AAV2 vector (scAAV2) to demonstrate the effect of long-term tissue-specific gene expression of RB15 on the regulation apoB mRNA in vivo. METHODS: We constructed a hammerhead ribozyme RB15 driven by a liver-specific transthyretin (TTR) promoter using an AAV2 vector (rAAV2-TTR-RB15). HepG2 cells and hyperlipidemic mice deficient in both the low density lipoprotein receptor and the apoB mRNA editing enzyme genes (LDLR-/-Apobec1-/-; LDb) were transduced with rAAV2-TTR-RB15 and a control vector rAAV-TTR-RB15-mutant (inactive ribozyme). The effects of ribozyme RB15 on apoB metabolism and atherosclerosis development were determined in LDb mice at 5-month after transduction. A self-complementary AAV2 vector expressing ribozyme RB15 (scAAV2-TTR-RB15) was also engineered and used to transduce HepG2 cells. Studies were designed to compare the gene expression efficiency between rAAV2-TTR-RB15 and scAAV2-TTR-RB15. RESULTS: The effect of ribozyme RB15 RNA on reducing apoB mRNA levels in HepG2 cells was observed only on day-7 after rAAV2-TTR-RB15 transduction. And, at 5-month after rAAV2-TTR-RB15 treatment, the apoB mRNA levels in LDb mice were significantly decreased by 43%, compared to LDb mice treated with control vector rAAV2-TTR-RB15-mutant. Moreover, both the rAAV2-TTR-RB15 viral DNA and ribozyme RB15 RNA were still detectable in mice livers at 5-month after treatment. However, this rAAV2-TTR-RB15 vector mediated a prolonged but low level of ribozyme RB15 gene expression in the mice livers, which did not produce the therapeutic effects on alteration the lipid levels or the inhibition of atherosclerosis development. In contrast, the ribozyme RB15 RNA mediated by scAAV2-TTR-RB15 vector was expressed immediately at day-1 after transduction in HepG2 cells. The apoB mRNA levels were decreased 47% (p = 0.001), compared to the control vector scAAV2-TTR-RB15-mutant. CONCLUSION: This study provided evidence that the rAAV2 single-strand vector mediated a prolonged but not efficient transduction in mouse liver. However, the scAAV2 double-strand vector mediated a rapid and efficient gene expression in liver cells. This strategy using scAAV2 vectors represents a better approach to express small molecules such as ribozyme.

Genetics and epidemiology of gallbladder disease in New World native peoples.

Native peoples of the New World, including Amerindians and admixed Latin Americans such as Mexican-Americans, are highly susceptible to diseases of the gallbladder. These include cholesterol cholelithiasis (gallstones) and its complications, as well as cancer of the gallbladder. Although there is clearly some necessary dietary or other environmental risk factor involved, the pattern of disease prevalence is geographically associated with the distribution of genes of aboriginal Amerindian origin, and levels of risk generally correspond to the degree of Amerindian admixture. This pattern differs from that generally associated with Westernization, which suggests a gene-environment interaction, and that within an admixed population there is a subset whose risk is underestimated when admixture is ignored. The risk that an individual of a susceptible New World genotype will undergo a cholecystectomy by age 85 can approach 40% in Mexican-American females, and their risk of gallbladder cancer can reach several percent. These are heretofore unrecognized levels of risk, especially of the latter, because previous studies have not accounted for admixture or for the loss of at-risk individuals due to cholecystectomy. A genetic susceptibility may, thus, be as "carcinogenic" in New World peoples as any known major environmental exposure; yet, while the risk has a genetic basis, its expression as gallbladder cancer is so delayed as to lead only very rarely to multiply-affected families. Estimates in this paper are derived in part from two studies of Mexican-Americans in Starr County and Laredo, Texas.

Phosphorylation of the proline-rich domain of Xp95 modulates Xp95 interaction with partner proteins.

The mammalian adaptor protein Alix [ALG-2 (apoptosis-linked-gene-2 product)-interacting protein X] belongs to a conserved family of proteins that have in common an N-terminal Bro1 domain and a C-terminal PRD (proline-rich domain), both of which mediate partner protein interactions. Following our previous finding that Xp95, the Xenopus orthologue of Alix, undergoes a phosphorylation-dependent gel mobility shift during progesteroneinduced oocyte meiotic maturation, we explored potential regulation of Xp95/Alix by protein phosphorylation in hormone-induced cell cycle re-entry or M-phase induction. By MALDI-TOF (matrix-assisted laser-desorption ionization-time-of-flight) MS analyses and gel mobility-shift assays, Xp95 is phosphorylated at multiple sites within the N-terminal half of the PRD during Xenopus oocyte maturation, and a similar region in Alix is phosphorylated in mitotically arrested but not serum-stimulated mammalian cells. By tandem MS, Thr745 within this region, which localizes in a conserved binding site to the adaptor protein SETA [SH3 (Src homology 3) domain-containing, expressed in tumorigenic astrocytes] CIN85 (a-cyano-4-hydroxycinnamate)/SH3KBP1 (SH3-domain kinase-binding protein 1), is one of the phosphorylation sites in Xp95. Results from GST (glutathione S-transferase)-pull down and peptide binding/competition assays further demonstrate that the Thr745 phosphorylation inhibits Xp95 interaction with the second SH3 domain of SETA. However, immunoprecipitates of Xp95 from extracts of M-phase-arrested mature oocytes contained additional partner proteins as compared with immunoprecipitates from extracts of G2-arrested immature oocytes. The deubiquitinase AMSH (associated molecule with the SH3 domain of signal transducing adaptor molecule) specifically interacts with phosphorylated Xp95 in M-phase cell lysates. These findings establish that Xp95/Alix is phosphorylated within the PRD during M-phase induction, and indicate that the phosphorylation may both positively and negatively modulate their interaction with partner proteins.

Population amalgamation and genetic variation: observations on artificially agglomerated tribal populations of Central and South America.

The interpretation of data on genetic variation with regard to the relative roles of different evolutionary factors that produce and maintain genetic variation depends critically on our assumptions concerning effective population size and the level of migration between neighboring populations. In humans, recent population growth and movements of specific ethnic groups across wide geographic areas mean that any theory based on assumptions of constant population size and absence of substructure is generally untenable. We examine the effects of population subdivision on the pattern of protein genetic variation in a total sample drawn from an artificial agglomerate of 12 tribal populations of Central and South America, analyzing the pooled sample as though it were a single population. Several striking findings emerge. (1) Mean heterozygosity is not sensitive to agglomeration, but the number of different alleles (allele count) is inflated, relative to neutral mutation/drift/equilibrium expectation. (2) The inflation is most serious for rare alleles, especially those which originally occurred as tribally restricted "private" polymorphisms. (3) The degree of inflation is an increasing function of both the number of populations encompassed by the sample and of the genetic divergence among them. (4) Treating an agglomerated population as though it were a panmictic unit of long standing can lead to serious biases in estimates of mutation rates, selection pressures, and effective population sizes. Current DNA studies indicate the presence of numerous genetic variants in human populations. The findings and conclusions of this paper are all fully applicable to the study of genetic variation at the DNA level as well.

Single-stranded DNA-binding proteins regulate the abundance of LIM domain and LIM domain-binding proteins.

The LIM domain-binding protein Ldb1 is an essential cofactor of LIM-homeodomain (LIM-HD) and LIM-only (LMO) proteins in development. The stoichiometry of Ldb1, LIM-HD, and LMO proteins is tightly controlled in the cell and is likely a critical determinant of their biological actions. Single-stranded DNA-binding proteins (SSBPs) were recently shown to interact with Ldb1 and are also important in developmental programs. We establish here that two mammalian SSBPs, SSBP2 and SSBP3, contribute to an erythroid DNA-binding complex that contains the transcription factors Tal1 and GATA-1, the LIM domain protein Lmo2, and Ldb1 and binds a bipartite E-box-GATA DNA sequence motif. In addition, SSBP2 was found to augment transcription of the Protein 4.2 (P4.2) gene, a direct target of the E-box-GATA-binding complex, in an Ldb1-dependent manner and to increase endogenous Ldb1 and Lmo2 protein levels, E-box-GATA DNA-binding activity, and P4.2 and beta-globin expression in erythroid progenitors. Finally, SSBP2 was demonstrated to inhibit Ldb1 and Lmo2 interaction with the E3 ubiquitin ligase RLIM, prevent RLIM-mediated Ldb1 ubiquitination, and protect Ldb1 and Lmo2 from proteasomal degradation. These results define a novel biochemical function for SSBPs in regulating the abundance of LIM domain and LIM domain-binding proteins.

Conversion of myoblasts to physiologically active neuronal phenotype.

Repressor element 1 (RE1)-silencing transcription factor (REST)/neuron-restrictive silencer factor (NRSF) can repress several terminal neuronal differentiation genes by binding to a specific DNA sequence (RE1/neuron-restrictive silencer element [NRSE]) present in their regulatory regions. REST-VP16 binds to the same RE1/NRSE, but activates these REST/NRSF target genes. However, it is unclear whether REST-VP16 expression is sufficient to cause formation of functional neurons either from neural stem cells or from heterologous stem cells. Here we show that the expression of REST-VP16 in myoblasts grown under muscle differentiation conditions blocked entry into the muscle differentiation pathway, countered endogenous REST/NRSF-dependent repression, activated the REST/NRSF target genes, and, surprisingly, activated other neuronal differentiation genes and converted the myoblasts to a physiologically active neuronal phenotype. Furthermore, in vitro differentiated neurons produced by REST-VP16-expressing myoblasts, when injected into mouse brain, survived, incorporated into the normal brain, and did not form tumors. This is the first instance in which myoblasts were converted to a neuronal phenotype. Our results suggest that direct activation of REST/NRSF target genes with a single transgene, REST-VP16, is sufficient to activate other terminal neuronal differentiation genes and to override the muscle differentiation pathways, and they suggest that this approach provides an efficient way of triggering neuronal differentiation in myoblasts and possibly other stem cells.

Three linked enzyme loci in fishes: implications in the evolution of vertebrate chromosomes.

A three-point linkage group comprised of loci coding for adenosine deaminase (ADA), glucose-6-phosphate dehydrogenase (G6PDH), and 6-phospho-gluconate dehydrogenase (6PGD) is described in fish of the genus Xiphophorus (Poeciliidae). The alleles at loci in this group were shown to assort independently from the alleles at three other loci--isocitrate dehydrogenase 1 and 2, and glyceraldehyde-3-phosphate dehydrogenase 1. Alleles at the latter three loci also assort independently from each other. Data were obtained by observing the segregation of electrophoretically variant alleles in reciprocal backcross hybrids derived from crosses between either X. helleri guentheri or X. h. strigatus and X. maculatus. The linkage component of chi2 was significant (less than 0.01) in all crosses, indicating that the linkage group is conserved in all populations of both species of Xiphophorus examined. While data from X. h. guentheri backcrosses indicate the linkage relationship ADA--6%--G6PDH--24%--6PGD, and ADA--29%--6PGD (30% when corrected for double crossovers), data from backcrosses involving strigatus, while supporting the same gene order, yielded significantly different recombination frequencies. The likelihood of the difference being due to an inversion could not be separated from the possibility of a sex effect on recombination in the present data. The linkage of 6PGD and G6PDH has been shown to exist in species of at least three classes of vertebrates, indicating the possibility of evolutionary conservation of this linkage.

THE ROLE OF MUCOSAL EPITHELIAL CELLS IN HIV-1 INFECTION

The predominant route of human immunodeficiency virus type 1 (HIV-1) transmission is infection across the vaginal mucosa. Epithelial cells, which form the primary barrier of protection against pathogens, are the first cell type at these mucosal tissues to encounter the virus but their role in HIV infection has not been clearly elucidated. Although mucosal epithelial cells express only low levels of the receptors required for successful HIV infection, productive infection does occur at these sites. The present work provides evidence to show that HIV exposure, without the need for productive infection, induces human cervical epithelial cells to produce Thymic Stromal Lymphopoietin (TSLP), an IL7-like cytokine, which potently activated human myeloid dendritic cells (mDC) to cause the homeostatic proliferation of autologous CD4+ T cells that serve as targets for HIV infection. Rhesus macaques inoculated with simian immunodeficiency virus (SIV) or with the simian-human immunodeficiency virus (SHIV) by the vaginal, oral or rectal route exhibited dramatic increases in: TSLP expression, DC and CD4+ T cell numbers, and viral replication, in the vaginal, oral, and rectal tissues, respectively within the first 2 weeks after virus exposure. Evidence obtained showed that HIV-mediated TSLP production by cervical cells is dependent upon the expression of the cell surface salivary agglutinin (SAG) protein gp340. Epithelial cells expressing gp340 exhibited HIV endocytosis and TSLP expression and genetic knockdown of gp340 or use of a gp340-blocking antibody inhibited TSLP expression by HIV. On the other hand, gp340-null epithelial cells failed to endocytose HIV and produce TSLP, but transfection of gp340 resulted in HIV-induced TSLP expression. Finally, HIV-induced TSLP expression was found to be mediated by TLR7/8 signaling and NF-kB activity because silencing these pathways or use of specific inhibitors abrogated TSLP expression in gp340-postive but not in gp340-null epithelial cells. Overall these studies identify TSLP as a key player in the acute phase of HIV-1 infection in permitting HIV to successfully maneuver the hostile vaginal mucosal microenvironment by creating a conducive environment for sustaining the small amount of virus that initially crosses the mucosal barrier allowing it to successfully cause infection and spread to distal compartments of the body

Influence of Anchoring on Miscarriage Risk Perception Associated with Amniocentesis

INFLUENCE OF ANCHORING ON MISCARRIAGE RISK PERCEPTION ASSOCIATED WITH AMNIOCENTESIS Publication No. ___________ Regina Nuccio, BS Supervisory Professor: Claire N. Singletary, MS, CGC Amniocentesis is the most common invasive procedure performed during pregnancy (Eddleman, et al., 2006). One important factor that women consider when making a decision about amniocentesis is the risk of miscarriage associated with the procedure. People use heuristics such as anchoring, the action of using a prior belief regarding the magnitude of risk as a frame of reference for new information to be synthesized, to better understand risks that they encounter in their lives. This study aimed to determine a woman’s perception of miscarriage risk associated with amniocentesis before and after a genetic counseling session and to determine what factors are most likely to anchor a woman’s perception of miscarriage risk associated with amniocentesis. Most women perceived the risk as low or average pre-counseling and were likely to indicate the numeric risk of amniocentesis as <1% risk. A higher percentage of patients correctly identified the numeric risk as <1% post-counseling when compared to pre-counseling. However, the majority of patients’ feeling about the risk perception did not change after the genetic counseling session (60%), regardless of how they perceived the risk before discussing amniocentesis with a genetic counselor. Those whose risk perception did change after discussing amniocentesis with a genetic counselor showed a decreased risk perception (p<0.0001). Of the multitude of factors studied, only two showed significance: having a friend or relative with a personal or family history of a genetic disorder was associated with a lower risk perception (p=0.001) and having a child already was associated with a lower risk perception (p=0.038). The lack of significant factors may reflect the uniqueness of each patient’s heuristic framework and reinforces the importance of genetic counseling to elucidate individual concerns.