Four projects were awarded during cycle 1 of the opportunity pool.
- A consortium-wide common tool box for transcript perturbation in multiple diabetes-relevant cell types
PIs: Anna Gloyn, Karen Mohlke
Abstract: A central goal of the AMP-CMD consortium is to deliver a robust set of genes or “effector transcripts” responsible for mediating T2D-GWAS associations. Perturbing candidate effectors is critical to provide evidence for causality. Many studies of function are limited to a single cell type and context, leading to incomplete or inaccurate conclusions about a gene’s contribution to disease mechanisms, especially genes with pleiotropic effects. Studies of transcript effects on multiple diabetes-related cell processes in multiple cell types/contexts will provide robust data to define risk genes and inform further studies of function. Aim 1: Prioritize genes for collaborative studies of knockdown effects on diabetes- relevant cell functions. We will collect genes identified in the initial AMP-CMD RFA, updated AMP-T2D analyses, and the Effector Transcript Working Group. Genes will be prioritized based on evidence for pleiotropy across diabetes-relevant cell types. In parallel we will generate a cell-type-specific pooled library focused on both genes and regulatory elements predicted to work through two major liver cell types. This will serve as a paradigm for how the library can be scaled and expanded to serve additional investigators. Aim 2: Generate lentiviral guideRNA libraries for CRISPR inhibition across cell types. We will clone gRNAs into plasmid vectors and generate lentivirus for ~180 gene promoters using 4 gRNAs/promoter in arrayed format and 10 gRNAs/promoter in pooled format. This cost-effective resource is scalable and future libraries could include gRNAs for enhancers. For the NAFLD library we will clone 4K gRNAs targeted at both genes and enhancers. Libraries will be infected in AAVS1-integrated doxycycline-inducible dCas9-KRAB inducible cell lines, which allows to generate homogenous, targeted cell populations in a reversible, tunable and efficient manner. Future Aim 3: Determine the consequences of gene knockdown. Multiple groups will use these libraries in at least islet beta cells, adipocytes, hepatocytes, and myocytes and measure, depending on cell type, insulin secretion, insulin-stimulated glucose uptake, lipid accumulation, mitochondrial function, and other assays.
- Comprehensive functional characterization of metabolic disease-associated variants in relevant cell types
PIs: Michael Stitzel and Ryan Tewhey
Abstract: Our understanding of the genetics of Type 2 diabetes (T2D) and other common metabolic diseases (CMD) has contributed significantly towards a fundamental goal of identifying common and rare genetic markers (e.g., single nucleotide polymorphisms, rare variants, structural variation etc.) associated with complex phenotypic traits including risk of disease. Variant catalogues have rapidly expanded, propelled by larger genome-wide association studies (GWAS) and biobanks. However, two characteristics of most GWAS variants impede efforts to translate genetic associations into molecular and pathophysiologic effects: (i) they reside in non-coding regions of the genome and modulate gene expression; and (ii) linkage disequilibrium and lagging efforts to leverage recombination across world populations imposes constraints on our ability to localize individual causal allele(s) within large haplotype blocks. Comprehensive first-principle, molecular-level understanding of CMD-associated non-coding variant function is essential to identify candidate causal gene regulatory variants, define their allelic effects on transcriptional activity, and determine the cell type(s) and context(s) in which they act. To identify functional T2D variants with expression-modulating activity in human (pre)adipocytes, we will test a 60k oligo T2D Module (T2DM) MPRA library containing ~8,800 fine-mapped, high priority T2D variants from DIAMANTE, UK Biobank, and Biobank Japan as well as a panel of positive and negative control sequences. Preliminary T2DM MPRA analyses nominated 176 T2D variants as functional expression modulating variants (emVars) in one of five ENCODE cell types (K562, A549, GM12878, SK-N-SH, and HepG2) and/or human EndoC-ßH3 ß-cells. We will extend this T2DM MPRA library to identify T2D emVars in SGBS (pre)adipocytes. Importantly, we will share the list of emVars and provide the T2DM library to AMP-CMD investigators to enable discovery of T2D variant effects in additional cell types and contexts. We will also contribute our expertise with MPRA to consortium efforts to create a new MPRA library representing new T2D/CMD signals, disease variants, and design parameters. We will test and validate cell- and context-specific effects of variants in this library using human EndoC-ßH3 ß-cells to identify and nominate functional emVars within these new association signals.
- Stem cell models for interrogating T2D effector gene function and benchmarking T2D relevant phenotypes
PIs: Wenli Yang, Paul Titchenell, and Klaus Kaestner
Abstract: A major goal of the AMP-CMD consortium is the functional interrogation of T2D-associated genes in relevant cell types. Increasing evidence suggests that many of the predicted candidate genes have pleiotropic effects across multiple cell types and may even function during tissue development. Critical for the success of AMP-CMD is the establishment, validation, and sharing of models and metabolic assays of insulin action and of endocrine hormone production. To this end, a set of well characterized human induced pluripotent stem cell (hiPSC) lines that serve as a renewable resource for generating cellular models for assaying gene function is required. Furthermore, PSC cell models of impaired insulin signaling, or defective pancreatic endocrine function are necessary as a ‘gold standard’ to validate any assay employed by AMP-CMD to evaluate candidate T2D genes. Establishing these cell lines and models will be extremely valuable to the AMP-CMD consortium for benchmarking T2D relevant phenotypes for gene validation. Constitutive CRISPR-mediated gene manipulation in PSC lines is not suitable for studying gene function in differentiated cells if the gene product affects cell differentiation. The inducible CRISPR interference (CRISPRi) and CRISPR activation (CRISPRa) gene perturbation systems are ideal alternatives that allow reversible and differentiation-stage specific gene suppression (“knockdown”) and activation, respectively. Aim 1: To generate and characterize inducible and reversible CRISPRi and CRISPRa human induced pluripotent human stem cell lines 1a. We will target the CLYBL safe harbor locus of a well characterized line to generate dox-inducible CRISPRi and CRISPRa cell lines expressing dCas9-KRAB and dCas9-VPR, respectively. 1b. We will extensively characterize the resulting lines for differentiation into hepatocytes, adipocytes, myocytes and pancreatic beta cells and tunability of the system. Aim 2: To benchmark assays of insulin action and insulin secretion in hiPSC-derived cells using CRISPRi and CRISPRa perturbation of validated model target genes 1a. We will suppress expression of AKT2 using CRISPRi and differentiate these cells into hepatocytes, myocytes and adipocytes and perform insulin action and metabolic physiological measurements in the AKT2 knockdown cells. 1b. We will suppress and activate expression of GCK and differentiate these cells into beta cells and perform glucose stimulation of insulin secretion assays.
- A consortium-wide modular MPRA resource to assess cellular contexts of T2D variants
PIs: Jacob Kitzman, Steve Parker, and Adelaide Tovar
Abstract: The AMP-CMD Genetics Working Group (GWG) recently released an updated trans-ancestry list of 667 T2D signals. Hence, there is urgent demand for a common resource to functionally comb through densely linked sets of candidate variants and define the cellular and environmental contexts of their regulatory activity. Massively parallel reporter assays (MPRAs) enable high-throughput regulatory element classification. Recent studies have employed MPRAs to test panels of putative regulatory sequences and to identify functional variants associated with a wide variety of traits [1–4], including in the context of T2D [5]. However, existing and ongoing T2D MPRAs were generated with outdated variant lists derived from 403 European signals reported by the DIAMANTE consortium in 2018 [6]. Thus, current MPRA libraries will miss an additional 264 T2D GWAS signals that have emerged in the past three years and all TOPMed-identified rare variants. To immediately address this urgent gap, we will create a modular, portable MPRA as a shared resource across AMPCMD to screen variants from the current T2D GWAS signals and TOPMed rare variants. Earlier T2D MPRA screens have focused on pancreatic beta cells; however, T2D etiology is complex and regions across several tissue types are enriched for heritability [7–11]. Additionally, genetic risk may be modified by environmental factors, meaning variant effects may only be detected after specific perturbations. Existing MPRAs are ill-suited to capture these effects because they generally (1) use generic housekeeping gene promoters, which do not reflect cell-specific regulatory effects, and (2) are conducted under basal conditions. To address this gap, we will use our state-of-the-art MPRA system to represent all current variants at T2D loci in the context of multiple cell-type-specific promoters. Aim 1: Construct an MPRA resource to classify regulatory activity at all known T2D-associated variants. We will use the ChromBPNet deep learning framework to optimally design oligos that span variants of interest, selected from the curated AMP-CMD GWG signal list and TOPMed rare variants. Then, we will clone these oligos into a multi-promoter MPRA vector. Using UM1-appropriated funds, we will screen human adipocytes, beta cells, hepatocytes, and myocytes to systematically assess enhancer activity for each element, in basal and stimulated states. We will make the library and screening results broadly available across AMPCMD, including data deposition in the Common Metabolic Diseases Genome Atlas (CMDGA) and Knowledge Portal (CMDKP). Aim 2: Design a comprehensive synthetic MPRA library to measure allelic effects at active T2Dassociated regions. We hypothesize that the primary mechanism through which T2D-associated common variants act is alteration of regulatory activity. Informed by the results of the activity screen in Aim 1, we will generate allelic versions of all oligos with enhancer activity. Additionally, we will target a subset of loci for (1) saturation mutagenesis and (2) driver nucleotide analysis, where densely tiled inserts spanning the nucleotide context of the variant(s) of interest will be synthesized. We will share broadly across AMP-CMD.