Christopher L. Plaisier

Systems Genetics Analysis of Gene-by-Environment Interactions in Human Cells

Gene by environment (GxE) interactions are clearly important in many human diseases, but they have proven to be difficult to study on a molecular level. We report genetic analysis of thousands of transcript abundance traits in human primary endothelial cell (EC) lines in response to proinflammatory oxidized phospholipids implicated in cardiovascular disease. Of the 59 most regulated transcripts, approximately one-third showed evidence of GxE interactions. The interactions resulted primarily from effects of distal-, trans-acting loci, but a striking example of a local-GxE interaction was also observed for FGD6. Some of the distal interactions were validated by siRNA knockdown experiments, including a locus involved in the regulation of multiple transcripts involved in the ER stress pathway. Our findings add to the understanding of the overall architecture of complex human traits and are consistent with the possibility that GxE interactions are responsible, in part, for the failure of association studies to more fully explain common disease variation.

Common Hepatic Nuclear Factor-4α Variants Are Associated With High Serum Lipid Levels and the Metabolic Syndrome

Hepatic nuclear factor-4α (HNF-4α), a transcription factor involved in the regulation of serum lipid and glucose levels, has recently been associated with type 2 diabetes. The HNF-4α gene (HNF4A) resides on chromosome 20q12-q13.1, which, in addition to type 2 diabetes, has also previously been linked to high triglycerides in Finnish familial combined hyperlipidemia (FCHL) families. FCHL, characterized by elevated levels of serum total cholesterol, triglycerides, or both, is a common dyslipidemia observed in up to 20% of patients with premature coronary heart disease. Considering the clear phenotypic overlap between type 2 diabetes and FCHL, both predisposing to high serum triglycerides and glucose intolerance, we tested this gene for association in dyslipidemic families originating from two distinct populations, Finnish and Mexican, and comprising 1,447 subjects. Our data show that common HNF4A variants and haplotypes are associated with elevated serum lipid levels and the metabolic syndrome (P = 0.008–0.04), as well as with elevated glucose parameters (P = 0.008–0.03), using family-based association analysis. Importantly, both Finnish and Mexican families shared two common lipid-associated HNF4A haplotypes (P = 0.005 for total cholesterol and 0.006 for triglycerides). In conclusion, we show for the first time that common HNF4A variants are associated with high serum lipid levels and the metabolic syndrome.

Association of Stearoyl-CoA Desaturase 1 Activity With Familial Combined Hyperlipidemia

Objective—Stearoyl-CoA desaturase 1 (SCD1) is the rate-limiting enzyme involved in the synthesis of monounsaturated fatty acids, and in mice SCD1 activity is associated with plasma triglyceride levels. We used the fatty acid desaturation index (the plasma ratio of 18:1/18:0) as a marker of SCD1 activity to investigate the relationship of SCD1 to familial combined hyperlipidemia (FCHL). Methods and Results—The fatty acid desaturation index was measured in 400 individuals from 18 extended FCHL pedigrees. FCHL-affected individuals exhibited increased SCD1 activity when compared to unrelated controls (P<0.0001). The fatty acid desaturation index was found to be highly heritable (h2=0.48, P=2.2×10−11) in this study sample. QTL analysis in 346 sibling pairs from 18 FCHL families revealed suggestive linkage of the desaturation index to chromosomes 3p26.1 to 3p13 (z=2.7, P=0.003), containing the peroxisome proliferator-activated receptor gamma (PPAR&ggr;) gene, and 20p11.21 to 20q13.32 (z=1.7, P=0.04), containing the hepatocyte nuclear factor 4, alpha (HNF4&agr;) gene. A specific haplotype of HNF4&agr; was found to be associated with the desaturation index in these FCHL families (P=0.002). Conclusion—Our results demonstrate that the fatty acid desaturation index is a highly heritable trait that is associated with the dyslipidemia observed in FCHL.

TCF7L2 is associated with high serum triacylglycerol and differentially expressed in adipose tissue in families with familial combined hyperlipidaemia

Aims/hypothesisCommon DNA variants of the transcription factor 7-like 2 gene (TCF7L2) are associated with type 2 diabetes. Familial combined hyperlipidaemia (FCHL) is characterised by hypertriacylglycerolaemia, hypercholesterolaemia, or both. Additionally, disturbances in glucose metabolism are commonly seen in FCHL. Therefore, we hypothesised that TCF7L2 may contribute to the genetic susceptibility for this common dyslipidaemia.MethodsWe investigated the effect of the TCF7L2 variants, rs7903146 and rs12255372, on FCHL and its component traits triacylglycerol (TG), total cholesterol (TC) and apolipoprotein B (ApoB) in 759 individuals from 55 Mexican families. As a replication sample, 719 individuals from 60 Finnish FCHL families were analysed. We also used quantitative RT-PCR to evaluate the transcript levels of TCF7L2 in 47 subcutaneous fat biopsies from unrelated Mexican FCHL and normolipidaemic participants.ResultsSignificant evidence for association was observed for high TG for the T alleles of rs7903146 and rs12255372 (p = 0.005 and p = 0.01) in Mexican FCHL families. No evidence for association was observed for FCHL, TC, ApoB or glucose in Mexicans. When testing rs7903146 and rs12255372 for replication in Finnish FCHL families, these single nucleotide polymorphisms were associated with TG (p = 0.01 and p = 0.007). Furthermore, we observed statistically significant decreases in the mRNA levels (p = 0.0002) of TCF7L2 in FCHL- and TG-affected individuals. TCF7L2 expression was not altered by the SNP genotypes.Conclusions/interpretationThese data show that rs7903146 and rs12255372 are significantly associated with high TG in FCHL families from two different populations. In addition, significantly decreased expression of TCF7L2 was observed in TG- and FCHL-affected individuals.

WW-domain-containing oxidoreductase is associated with low plasma HDL-C levels.

Low serum HDL-cholesterol (HDL-C) is a major risk factor for coronary artery disease. We performed targeted genotyping of a 12.4 Mb linked region on 16q to test for association with low HDL-C by using a regional-tag SNP strategy. We identified one SNP, rs2548861, in the WW-domain-containing oxidoreductase (WWOX) gene with region-wide significance for low HDL-C in dyslipidemic families of Mexican and European descent and in low-HDL-C cases and controls of European descent (p = 6.9 x 10(-7)). We extended our investigation to the population level by using two independent unascertained population-based Finnish cohorts, the cross-sectional METSIM cohort of 4,463 males and the prospective Young Finns cohort of 2,265 subjects. The combined analysis provided p = 4 x 10(-4) to 2 x 10(-5). Importantly, in the prospective cohort, we observed a significant longitudinal association of rs2548861 with HDL-C levels obtained at four different time points over 21 years (p = 0.003), and the T risk allele explained 1.5% of the variance in HDL-C levels. The rs2548861 resides in a highly conserved region in intron 8 of WWOX. Results from our in vitro reporter assay and electrophoretic mobility-shift assay demonstrate that this region functions as a cis-regulatory element whose associated rs2548861 SNP has a specific allelic effect and that the region forms an allele-specific DNA-nuclear-factor complex. In conclusion, analyses of 9,798 subjects show significant association between HDL-C and a WWOX variant with an allele-specific cis-regulatory function.

Genome-wide CRISPR-Cas9 Screens Reveal Loss of Redundancy between PKMYT1 and WEE1 in Glioblastoma Stem-like Cells.

To identify therapeutic targets for glioblastoma (GBM), we performed genome-wide CRISPR-Cas9 knockout (KO) screens in patient-derived GBM stem-like cells (GSCs) and human neural stem/progenitors (NSCs), non-neoplastic stem cell controls, for genes required for their in vitro growth. Surprisingly, the vast majority GSC-lethal hits were found outside of molecular networks commonly altered in GBM and GSCs (e.g., oncogenic drivers). In vitro and in vivo validation of GSC-specific targets revealed several strong hits, including the wee1-like kinase, PKMYT1/Myt1. Mechanistic studies demonstrated that PKMYT1 acts redundantly with WEE1 to inhibit cyclin B-CDK1 activity via CDK1-Y15 phosphorylation and to promote timely completion of mitosis in NSCs. However, in GSCs, this redundancy is lost, most likely as a result of oncogenic signaling, causing GBM-specific lethality.

A high-resolution network model for global gene regulation in Mycobacterium tuberculosis

The resilience of Mycobacterium tuberculosis (MTB) is largely due to its ability to effectively counteract and even take advantage of the hostile environments of a host. In order to accelerate the discovery and characterization of these adaptive mechanisms, we have mined a compendium of 2325 publicly available transcriptome profiles of MTB to decipher a predictive, systems-scale gene regulatory network model. The resulting modular organization of 98% of all MTB genes within this regulatory network was rigorously tested using two independently generated datasets: a genome-wide map of 7248 DNA-binding locations for 143 transcription factors (TFs) and global transcriptional consequences of overexpressing 206 TFs. This analysis has discovered specific TFs that mediate conditional co-regulation of genes within 240 modules across 14 distinct environmental contexts. In addition to recapitulating previously characterized regulons, we discovered 454 novel mechanisms for gene regulation during stress, cholesterol utilization and dormancy. Significantly, 183 of these mechanisms act uniquely under conditions experienced during the infection cycle to regulate diverse functions including 23 genes that are essential to host-pathogen interactions. These and other insights underscore the power of a rational, model-driven approach to unearth novel MTB biology that operates under some but not all phases of infection.

Niche adaptation by expansion and reprogramming of general transcription factors

Numerous lineage‐specific expansions of the transcription factor B (TFB) family in archaea suggests an important role for expanded TFBs in encoding environment‐specific gene regulatory programs. Given the characteristics of hypersaline lakes, the unusually large numbers of TFBs in halophilic archaea further suggests that they might be especially important in rapid adaptation to the challenges of a dynamically changing environment. Motivated by these observations, we have investigated the implications of TFB expansions by correlating sequence variations, regulation, and physical interactions of all seven TFBs in Halobacterium salinarum NRC‐1 to their fitness landscapes, functional hierarchies, and genetic interactions across 2488 experiments covering combinatorial variations in salt, pH, temperature, and Cu stress. This systems analysis has revealed an elegant scheme in which completely novel fitness landscapes are generated by gene conversion events that introduce subtle changes to the regulation or physical interactions of duplicated TFBs. Based on these insights, we have introduced a synthetically redesigned TFB and altered the regulation of existing TFBs to illustrate how archaea can rapidly generate novel phenotypes by simply reprogramming their TFB regulatory network.

A system-level model for the microbial regulatory genome

Microbes can tailor transcriptional responses to diverse environmental challenges despite having streamlined genomes and a limited number of regulators. Here, we present data‐driven models that capture the dynamic interplay of the environment and genome‐encoded regulatory programs of two types of prokaryotes: Escherichia coli (a bacterium) and Halobacterium salinarum (an archaeon). The models reveal how the genome‐wide distributions of cis‐acting gene regulatory elements and the conditional influences of transcription factors at each of those elements encode programs for eliciting a wide array of environment‐specific responses. We demonstrate how these programs partition transcriptional regulation of genes within regulons and operons to re‐organize gene–gene functional associations in each environment. The models capture fitness‐relevant co‐regulation by different transcriptional control mechanisms acting across the entire genome, to define a generalized, system‐level organizing principle for prokaryotic gene regulatory networks that goes well beyond existing paradigms of gene regulation. An online resource (http://egrin2.systemsbiology.net) has been developed to facilitate multiscale exploration of conditional gene regulation in the two prokaryotes.

miRvestigator: web application to identify miRNAs responsible for co-regulated gene expression patterns discovered through transcriptome profiling

Transcriptome profiling studies have produced staggering numbers of gene co-expression signatures for a variety of biological systems. A significant fraction of these signatures will be partially or fully explained by miRNA-mediated targeted transcript degradation. miRvestigator takes as input lists of co-expressed genes from Caenorhabditis elegans, Drosophila melanogaster, G. gallus, Homo sapiens, Mus musculus or Rattus norvegicus and identifies the specific miRNAs that are likely to bind to 3′ un-translated region (UTR) sequences to mediate the observed co-regulation. The novelty of our approach is the miRvestigator hidden Markov model (HMM) algorithm which systematically computes a similarity P-value for each unique miRNA seed sequence from the miRNA database miRBase to an overrepresented sequence motif identified within the 3′-UTR of the query genes. We have made this miRNA discovery tool accessible to the community by integrating our HMM algorithm with a proven algorithm for de novo discovery of miRNA seed sequences and wrapping these algorithms into a user-friendly interface. Additionally, the miRvestigator web server also produces a list of putative miRNA binding sites within 3′-UTRs of the query transcripts to facilitate the design of validation experiments. The miRvestigator is freely available at http://mirvestigator.systemsbiology.net.