Bayesian variable selection in structured high-dimensional covariate spaces with applications in genomics

We consider the problem of variable selection in regression modeling in high-dimensional spaces where there is known structure among the covariates. This is an unconventional variable selection problem for two reasons: (1) The dimension of the covariate space is comparable, and often much larger, than the number of subjects in the study, and (2) the covariate space is highly structured, and in some cases it is desirable to incorporate this structural information in to the model building process. We approach this problem through the Bayesian variable selection framework, where we assume that the covariates lie on an undirected graph and formulate an Ising prior on the model space for incorporating structural information. Certain computational and statistical problems arise that are unique to such high-dimensional, structured settings, the most interesting being the phenomenon of phase transitions. We propose theoretical and computational schemes to mitigate these problems. We illustrate our methods on two different graph structures: the linear chain and the regular graph of degree k. Finally, we use our methods to study a specific application in genomics: the modeling of transcription factor binding sites in DNA sequences. © 2010 American Statistical Association.

C. elegans germline-deficient mutants respond to pathogen infection using shared and distinct mechanisms.

Reproduction extracts a cost in resources that organisms are then unable to utilize to deal with a multitude of environmental stressors. In the nematode C. elegans, development of the germline shortens the lifespan of the animal and increases its susceptibility to microbial pathogens. Prior studies have demonstrated germline-deficient nematodes to have increased resistance to gram negative bacteria. We show that germline-deficient strains display increased resistance across a broad range of pathogens including gram positive and gram negative bacteria, and the fungal pathogen Cryptococcus neoformans. Furthermore, we show that the FOXO transcription factor DAF-16, which regulates longevity and immunity in C. elegans, appears to be crucial for maintaining longevity in both wild-type and germline-deficient backgrounds. Our studies indicate that germline-deficient mutants glp-1 and glp-4 respond to pathogen infection using common and different mechanisms that involve the activation of DAF-16.

Transcription factors MYOCD, SRF, Mesp1 and SMARCD3 enhance the cardio-inducing effect of GATA4, TBX5, and MEF2C during direct cellular reprogramming.

Transient overexpression of defined combinations of master regulator genes can effectively induce cellular reprogramming: the acquisition of an alternative predicted phenotype from a differentiated cell lineage. This can be of particular importance in cardiac regenerative medicine wherein the heart lacks the capacity to heal itself, but simultaneously contains a large pool of fibroblasts. In this study we determined the cardio-inducing capacity of ten transcription factors to actuate cellular reprogramming of mouse embryonic fibroblasts into cardiomyocyte-like cells. Overexpression of transcription factors MYOCD and SRF alone or in conjunction with Mesp1 and SMARCD3 enhanced the basal but necessary cardio-inducing effect of the previously reported GATA4, TBX5, and MEF2C. In particular, combinations of five or seven transcription factors enhanced the activation of cardiac reporter vectors, and induced an upregulation of cardiac-specific genes. Global gene expression analysis also demonstrated a significantly greater cardio-inducing effect when the transcription factors MYOCD and SRF were used. Detection of cross-striated cells was highly dependent on the cell culture conditions and was enhanced by the addition of valproic acid and JAK inhibitor. Although we detected Ca(2+) transient oscillations in the reprogrammed cells, we did not detect significant changes in resting membrane potential or spontaneously contracting cells. This study further elucidates the cardio-inducing effect of the transcriptional networks involved in cardiac cellular reprogramming, contributing to the ongoing rational design of a robust protocol required for cardiac regenerative therapies.

Rapid behavioral and genomic responses to social opportunity.

From primates to bees, social status regulates reproduction. In the cichlid fish Astatotilapia (Haplochromis) burtoni, subordinate males have reduced fertility and must become dominant to reproduce. This increase in sexual capacity is orchestrated by neurons in the preoptic area, which enlarge in response to dominance and increase expression of gonadotropin-releasing hormone 1 (GnRH1), a peptide critical for reproduction. Using a novel behavioral paradigm, we show for the first time that subordinate males can become dominant within minutes of an opportunity to do so, displaying dramatic changes in body coloration and behavior. We also found that social opportunity induced expression of the immediate-early gene egr-1 in the anterior preoptic area, peaking in regions with high densities of GnRH1 neurons, and not in brain regions that express the related peptides GnRH2 and GnRH3. This genomic response did not occur in stable subordinate or stable dominant males even though stable dominants, like ascending males, displayed dominance behaviors. Moreover, egr-1 in the optic tectum and the cerebellum was similarly induced in all experimental groups, showing that egr-1 induction in the anterior preoptic area of ascending males was specific to this brain region. Because egr-1 codes for a transcription factor important in neural plasticity, induction of egr-1 in the anterior preoptic area by social opportunity could be an early trigger in the molecular cascade that culminates in enhanced fertility and other long-term physiological changes associated with dominance.

Quantitative genetics of CTCF binding reveal local sequence effects and different modes of X-chromosome association.

Associating genetic variation with quantitative measures of gene regulation offers a way to bridge the gap between genotype and complex phenotypes. In order to identify quantitative trait loci (QTLs) that influence the binding of a transcription factor in humans, we measured binding of the multifunctional transcription and chromatin factor CTCF in 51 HapMap cell lines. We identified thousands of QTLs in which genotype differences were associated with differences in CTCF binding strength, hundreds of them confirmed by directly observable allele-specific binding bias. The majority of QTLs were either within 1 kb of the CTCF binding motif, or in linkage disequilibrium with a variant within 1 kb of the motif. On the X chromosome we observed three classes of binding sites: a minority class bound only to the active copy of the X chromosome, the majority class bound to both the active and inactive X, and a small set of female-specific CTCF sites associated with two non-coding RNA genes. In sum, our data reveal extensive genetic effects on CTCF binding, both direct and indirect, and identify a diversity of patterns of CTCF binding on the X chromosome.

Chromatin accessibility mapping identifies mediators of basal transcription and retinoid-induced repression of OTX2 in medulloblastoma.

Despite an emerging understanding of the genetic alterations giving rise to various tumors, the mechanisms whereby most oncogenes are overexpressed remain unclear. Here we have utilized an integrated approach of genomewide regulatory element mapping via DNase-seq followed by conventional reporter assays and transcription factor binding site discovery to characterize the transcriptional regulation of the medulloblastoma oncogene Orthodenticle Homeobox 2 (OTX2). Through these studies we have revealed that OTX2 is differentially regulated in medulloblastoma at the level of chromatin accessibility, which is in part mediated by DNA methylation. In cell lines exhibiting chromatin accessibility of OTX2 regulatory regions, we found that autoregulation maintains OTX2 expression. Comparison of medulloblastoma regulatory elements with those of the developing brain reveals that these tumors engage a developmental regulatory program to drive OTX2 transcription. Finally, we have identified a transcriptional regulatory element mediating retinoid-induced OTX2 repression in these tumors. This work characterizes for the first time the mechanisms of OTX2 overexpression in medulloblastoma. Furthermore, this study establishes proof of principle for applying ENCODE datasets towards the characterization of upstream trans-acting factors mediating expression of individual genes.

Inactivation of the von Hippel-Lindau tumor suppressor leads to selective expression of a human endogenous retrovirus in kidney cancer.

A human endogenous retrovirus type E (HERV-E) was recently found to be selectively expressed in most renal cell carcinomas (RCCs). Importantly, antigens derived from this provirus are immunogenic, stimulating cytotoxic T cells that kill RCC cells in vitro and in vivo. Here, we show HERV-E expression is restricted to the clear cell subtype of RCC (ccRCC) characterized by an inactivation of the von Hippel-Lindau (VHL) tumor-suppressor gene with subsequent stabilization of hypoxia-inducible transcription factors (HIFs)-1α and -2α. HERV-E expression in ccRCC linearly correlated with HIF-2α levels and could be silenced in tumor cells by either transfection of normal VHL or small interfering RNA inhibition of HIF-2α. Using chromatin immunoprecipitation, we demonstrated that HIF-2α can serve as transcriptional factor for HERV-E by binding with HIF response element (HRE) localized in the proviral 5' long terminal repeat (LTR). Remarkably, the LTR was found to be hypomethylated only in HERV-E-expressing ccRCC while other tumors and normal tissues possessed a hypermethylated LTR preventing proviral expression. Taken altogether, these findings provide the first evidence that inactivation of a tumor suppressor gene can result in aberrant proviral expression in a human tumor and give insights needed for translational research aimed at boosting human immunity against antigenic components of this HERV-E.

Les facteurs de transcription du groupe PEA3: Régulateurs transcriptionnels dans le processus de cancérisation

Erm, Er81, and Pea3 are the three members of the PEA3 group which belong to the Ets transcription factors family. These proteins regulate transcription of multiple target genes, such as those encoding several matrix metalloproteinases (MMP), which are enzymes degrading the extracellular matrix during cancer metastasis. In fact, PEA3-group genes are often overexpressed in different types of human cancers that also over-express these MMP and display a disseminating phenotype. In experimental models, regulation of PEA3 group member expression has been shown to influence the metastatic process, thus suggesting that these factors play a key role in metastasis. © John Libbey Eurotext.

Regulation of the alpha-fetoprotein promoter: Ku binding and DNA spatial conformation

This work shows that the proximal promoter of the mouse Afp gene contains a Ku binding site and that Ku binding is associated with down-regulation of the transcriptional activity of the Afp promoter. The Ku binding site is located in a segment able to adopt a peculiar structured form, probably a hairpin structure. Interestingly, the structured form eliminates the binding sites of the positive transcription factor HNF1. Furthermore, a DNAse hypersensitive site is detected in footprinting experiments done with extracts of AFP non-expressing hepatoma cells. These observations suggest that the structured form is stabilised by Ku and is associated with extinction of the gene in AFP non-expressing hepatic cells.

Pubertal impairment in Nhlh2

Pubertal development is impaired in mice lacking the basic helix-loop-helix transcription factor Nhlh2. The mechanisms underlying changes in reproduction in Nhlh2-deficient mice (Nhlh2-/-) are unclear. Here we show that hypothalamic gonadotropin-releasing hormone-1 (GnRH-1) content is reduced in adult Nhlh2-/- mice as is the number of GnRH-1 neurons localized to mid- and caudal hypothalamic regions. This reduction was detected postnatally after normal migration of GnRH-1 neurons within nasal regions had occurred. Phenotype rescue experiments showed that female Nhlh2-/- mice were responsive to estrogen treatment. In contrast, puberty could not be primed in female Nhlh2-/- mice with a GnRH-1 regimen. The adenohypophysis of Nhlh2-/- mice was hypoplastic although it contained a full complement of the five anterior pituitary cell types. GnRH-1 receptors (GnRHRs) were reduced in Nhlh2-/- pituitary gonadotropes as compared to wild type. In vitro assays indicated that Nhlh2 expression is regulated in parallel with GnRHR expression. However, direct transcriptional activity of Nhlh2 on the GnRHR promoter was not found. These results indicate that Nhlh2 plays a role in the development and functional maintenance of the hypothalamic-pituitarygonadal axis at least at two levels: 1) in the hypothalamus by regulating the number and distribution of GnRH-1 neurons and, 2) in the developing and mature adenohypophysis.

Functional interaction of E1AF and Sp1 in glioma invasion.

Transcription factor E1AF is widely known to play critical roles in tumor metastasis via directly binding to the promoters of genes involved in tumor migration and invasion. Here, we report for the first time E1AF as a novel binding partner for ubiquitously expressed Sp1 transcription factor. E1AF forms a complex with Sp1, contributes to Sp1 phosphorylation and transcriptional activity, and functions as a mediator between epidermal growth factor and Sp1 phosphorylation and activity. Sp1 functions as a carrier bringing E1AF to the promoter region, thus activating transcription of glioma-related gene for beta1,4-galactosyltransferase V (GalT V; EC 2.4.1.38). Biologically, E1AF functions as a positive invasion regulator in glioma in cooperation with Sp1 partly via up-regulation of GalT V. This report describes a new mechanism of glioma invasion involving a cooperative effort between E1AF and Sp1 transcription factors.

Polyomavirus enhancer activator 3 protein promotes breast cancer metastatic progression through Snail-induced epithelial-mesenchymal transition.

Polyomavirus enhancer activator 3 protein (Pea3), also known as ETV4, is a member of the Ets-transcription factor family, which promotes metastatic progression in various types of solid cancer. Pea3-driven epithelial-mesenchymal transition (EMT) has been described in lung and ovarian cancers. The mechanisms of Pea3-induced EMT, however, are largely unknown. Here we show that Pea3 overexpression promotes EMT in human breast epithelial cells through transactivation of Snail (SNAI1), an activator of EMT. Pea3 binds to the human Snail promoter through the two proximal Pea3 binding sites and enhances Snail expression. In addition, knockdown of Pea3 in invasive breast cancer cells results in down-regulation of Snail, partial reversal of EMT, and reduced invasiveness in vitro. Moreover, knockdown of Snail partially rescues the phenotype induced by Pea3 overexpression, suggesting that Snail is one of the mediators bridging Pea3 and EMT, and thereby metastatic progression of the cancer cells. In four breast cancer patient cohorts whose microarray and survival data were obtained from the Gene Expression Omnibus database, Pea3 and Snail expression are significantly correlated with each other and with overall survival of breast cancer patients. We further demonstrate that nuclear localization of Pea3 is associated with Snail expression in breast cancer cell lines and is an independent predictor of overall survival in a Chinese breast cancer patient cohort. In conclusion, our results suggest that Pea3 may be an important prognostic marker and a therapeutic target for metastatic progression of human breast cancer. © 2011 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

A GREM1 gene variant associates with diabetic nephropathy

Gremlin, a cell growth and differentiation factor, promotes the development of diabetic nephropathy in animal models, but whether GREM1 gene variants associate with diabetic nephropathy is unknown. We comprehensively screened the 5' upstream region (including the predicted promoter), all exons, intron-exon boundaries, complete untranslated regions, and the 3' region downstream of the GREM1 gene. We identified 31 unique variants, including 24 with a minor allele frequency exceeding 5%, and 9 haplotype-tagging single nucleotide polymorphisms (htSNPs). We selected one additional variant that we predicted to alter transcription factor binding. We genotyped 709 individuals with type 1 diabetes of whom 267 had nephropathy (cases) and 442 had no evidence of kidney disease (controls). Three individual SNPs significantly associated with nephropathy at the 5% level, and two remained significant after adjustment for multiple testing. Subsequently, we genotyped a replicate population comprising 597 cases and 502 controls: this population supported an association with one of the SNPs (rs1129456; P = 0.0003). Combined analysis, adjusted for recruitment center (n = 8), suggested that the T allele conferred greater odds of nephropathy (OR 1.69; 95% CI 1.36 to 2.11). In summary, the GREM1 variant rs1129456 associates with diabetic nephropathy, perhaps explaining some of the genetic susceptibility to this condition.

Elevated expression of Runx2 as a key parameter in the etiology of osteosarcoma

To understand the molecular etiology of osteosarcoma, we isolated and characterized a human osteosarcoma cell line (OS1). OS1 cells have high osteogenic potential in differentiation induction media. Molecular analysis reveals OS1 cells express the pocket protein pRB and the runt-related transcription factor Runx2. Strikingly, Runx2 is expressed at higher levels in OS1 cells than in human fetal osteoblasts. Both pRB and Runx2 have growth suppressive potential in osteoblasts and are key factors controlling competency for osteoblast differentiation. The high levels of Runx2 clearly suggest osteosarcomas may form from committed osteoblasts that have bypassed growth restrictions normally imposed by Runx2. Interestingly, OS1 cells do not exhibit p53 expression and thus lack a functional p53/p21 DNA damage response pathway as has been observed for other osteosarcoma cell types. Absence of this pathway predicts genomic instability and/or vulnerability to secondary mutations that may counteract the anti-proliferative activity of Runx2 that is normally observed in osteoblasts. We conclude OS1 cells provide a valuable cell culture model to examine molecular events that are responsible for the pathologic conversion of phenotypically normal osteoblast precursors into osteosarcoma cells.