1. Developmental Biology
  2. Genetics and Genomics
Download icon

Pancreatic progenitor epigenome maps prioritize type 2 diabetes risk genes with roles in development

  1. Ryan J Geusz
  2. Allen Wang
  3. Joshua Chiou
  4. Joseph J Lancman
  5. Nichole Wetton
  6. Samy Kefalopoulou
  7. Jinzhao Wang
  8. Yunjiang Qiu
  9. Jian Yan
  10. Anthony Aylward
  11. Bing Ren
  12. P Duc Si Dong
  13. Kyle J Gaulton  Is a corresponding author
  14. Maike Sander  Is a corresponding author
  1. Department of Pediatrics, Pediatric Diabetes Research Center, University of California, San Diego, United States
  2. Department of Cellular & Molecular Medicine, University of California, San Diego, United States
  3. Sanford Consortium for Regenerative Medicine, United States
  4. Biomedical Graduate Studies Program, University of California, San Diego, United States
  5. Human Genetics Program, Sanford Burnham Prebys Medical Discovery Institute, United States
  6. Graduate School of Biomedical Sciences, Sanford Burnham Prebys Medical Discovery Institute, United States
  7. Ludwig Institute for Cancer Research, United States
Research Article
  • Cited 0
  • Views 853
  • Annotations
Cite this article as: eLife 2021;10:e59067 doi: 10.7554/eLife.59067

Abstract

Genetic variants associated with type 2 diabetes (T2D) risk affect gene regulation in metabolically relevant tissues, such as pancreatic islets. Here, we investigated contributions of regulatory programs active during pancreatic development to T2D risk. Generation of chromatin maps from developmental precursors throughout pancreatic differentiation of human embryonic stem cells (hESCs) identifies enrichment of T2D variants in pancreatic progenitor-specific stretch enhancers that are not active in islets. Genes associated with progenitor-specific stretch enhancers are predicted to regulate developmental processes, most notably tissue morphogenesis. Through gene editing in hESCs, we demonstrate that progenitor-specific enhancers harboring T2D-associated variants regulate cell polarity genes LAMA1 and CRB2. Knockdown of lama1 or crb2 in zebrafish embryos causes a defect in pancreas morphogenesis and impairs islet cell development. Together, our findings reveal that a subset of T2D risk variants specifically affects pancreatic developmental programs, suggesting that dysregulation of developmental processes can predispose to T2D.

Introduction

Type 2 diabetes (T2D) is a multifactorial metabolic disorder characterized by insulin insensitivity and insufficient insulin secretion by pancreatic beta cells (Halban et al., 2014). Genetic association studies have identified hundreds of loci influencing risk of T2D (Mahajan et al., 2018). However, disease-relevant target genes of T2D risk variants, the mechanisms by which these genes cause disease, and the tissues in which the genes mediate their effects remain poorly understood.

The majority of T2D risk variants map to non-coding sequence, suggesting that genetic risk of T2D is largely mediated through variants affecting transcriptional regulatory activity. Intersection of T2D risk variants with epigenomic data has uncovered enrichment of T2D risk variants in regulatory sites active in specific cell types, predominantly in pancreatic beta cells, including risk variants that affect regulatory activity directly (Chiou et al., 2019; Fuchsberger et al., 2016; Gaulton et al., 2015; Gaulton et al., 2010; Greenwald et al., 2019; Mahajan et al., 2018; Parker et al., 2013; Pasquali et al., 2014; Thurner et al., 2018; Varshney et al., 2017). T2D risk-associated variants are further enriched within large, contiguous regions of islet active chromatin, referred to as stretch or super-enhancers (Parker et al., 2013). These regions of active chromatin preferentially bind islet-cell-restricted transcription factors and drive islet-specific gene expression (Parker et al., 2013; Pasquali et al., 2014).

Many genes associated with T2D risk in islets are not uniquely expressed in differentiated islet endocrine cells, but also in pancreatic progenitor cells during embryonic development. For example, T2D risk variants map to HNF1A, HNF1B, HNF4A, MNX1, NEUROG3, PAX4, and PDX1 (Flannick et al., 2019; Mahajan et al., 2018; Steinthorsdottir et al., 2014), which are all transcription factors also expressed in pancreatic developmental precursors. Studies in model organisms and hESC-based models of pancreatic endocrine cell differentiation have shown that inactivation of these transcription factors causes defects in endocrine cell development, resulting in reduced beta cell numbers (Gaertner et al., 2019). Furthermore, heterozygous mutations for HNF1A, HNF1B, HNF4A, PAX4, and PDX1 are associated with maturity onset diabetes of the young (MODY), which is an autosomal dominant form of diabetes with features similar to T2D (Urakami, 2019). Thus, there is evidence that reduced activity of developmentally expressed transcription factors can cause diabetes later in life.

The role of these transcription factors in T2D and MODY could be explained by their functions in regulating gene expression in mature islet cells. However, it is also possible that their function during endocrine cell development could predispose to diabetes instead of, or in addition to, endocrine cell gene regulation. One conceivable mechanism is that individuals with reduced activity of these transcription factors are born with either fewer beta cells or beta cells more prone to fail under conditions of increased insulin demand. Observations showing that disturbed intrauterine metabolic conditions, such as maternal malnutrition, can lead to reduced beta cell mass and T2D predisposition in the offspring (Lumey et al., 2015; Nielsen et al., 2014; Portha et al., 2011) support the concept that compromised beta cell development could predispose to T2D. However, whether there is T2D genetic risk relevant to the regulation of endocrine cell development independent of gene regulation in mature islet cells has not been explored.

In this study, we investigated the contribution of gene regulatory programs specifically active during pancreatic development to T2D risk. First, we employed a hESC-based differentiation system to generate chromatin maps of hESCs during their stepwise differentiation into pancreatic progenitor cells. We then identified T2D-associated variants localized in active enhancers in developmental precursors but not in mature islets, used genome editing in hESCs to define target genes of pancreatic progenitor-specific enhancers harboring T2D variants, and employed zebrafish genetic models to study the role of two target genes in pancreatic and endocrine cell development.

Results

Pancreatic progenitor stretch enhancers are enriched for T2D risk variants

To determine whether there is a development-specific genetic contribution to T2D risk, we generated genome-wide chromatin maps of hESCs during their stepwise differentiation into pancreatic progenitors through four distinct developmental stages: definitive endoderm (DE), gut tube (GT), early pancreatic progenitors (PP1), and late pancreatic progenitors (PP2) (Figure 1A). We then used ChromHMM (Ernst and Kellis, 2012) to annotate chromatin states, such as active promoters and enhancers, at all stages of hESC differentiation as well as in primary islets (Figure 1—figure supplement 1A,B).

Figure 1 with 1 supplement see all
T2D-associated risk variants are enriched in stretch enhancers of pancreatic progenitors independent of islet stretch enhancers.

(A) Schematic illustrating the stepwise differentiation of human embryonic stem cells (hES) into pancreatic progenitors (solid arrows) and lineage relationship to islets (dotted arrow). Developmental intermediates include definitive endoderm (DE), gut tube (GT), early pancreatic progenitor (PP1), and late pancreatic progenitor (PP2) cells. (B) Box plots depicting length of typical enhancers (TE) and stretch enhancers (SE) at each developmental stage and in primary human islets. Plots are centered on median, with box encompassing 25–75th percentile and whiskers extending up to 1.5 interquartile range. Total numbers of enhancers are shown above each box plot. (C) Examples of stretch enhancers (denoted with red boxes) near the genes encoding the pancreatic lineage-determining transcription factors NKX6.1 and PDX1, respectively. Chromatin states are based on ChromHMM classifications: TssA, active promoter; TssFlnk, flanking transcription start site; TssBiv, bivalent promoter; Repr, repressed; EnhA, active enhancer; EnhP, poised enhancer. (D) Percentage of TE and SE overlapping with at least one ATAC-seq peak at PP2 or in islets. Enrichment analysis comparing observed and expected overlap based on random genomic regions of the same size and located on the same chromosome averaged over 10,000 iterations (***p<1 × 10−4; permutation test). ATAC-seq peaks were merged from two independent differentiations for PP2 stage cells and four donors for primary islets. (E) Genome-wide enrichment of T2D-associated variants (minor allele frequency >0.0025) in stretch enhancers, ATAC-seq peaks, and ATAC-seq peaks within stretch enhancers for all developmental stages when modeling each annotation separately. Points and lines represent log-scaled enrichment estimates and 95% confidence intervals from functional genome wide association analysis (fgwas), respectively. ATAC-seq peaks were merged from two independent differentiations for ES, DE, GT, PP1, and PP2 stage cells and from four donors for primary islets. (F) Genome-wide enrichment of T2D-associated variants (minor allele frequency >0.0025) in ATAC-seq peaks within stretch enhancers for all developmental stages and coding exons when considering all annotations in a joint model. Points and lines represent log-scaled enrichment estimates and 95% confidence intervals from fgwas, respectively. ATAC-seq peaks were merged from two independent differentiations for ES, DE, GT, PP1, and PP2 stage cells and from four donors for primary islets. See also Figure 1—figure supplement 1.

Large and contiguous regions of active enhancer chromatin, which have been termed stretch- or super-enhancers (Parker et al., 2013; Whyte et al., 2013), are highly enriched for T2D risk variants in islets (Parker et al., 2013; Pasquali et al., 2014). We therefore partitioned active enhancers from each hESC developmental stage and islets into stretch enhancers (SE) and traditional (non-stretch) enhancers (TE) (Figure 1B). Consistent with prior observations of SE features (Parker et al., 2013; Whyte et al., 2013), SE comprised a small subset of all active enhancers (7.7%, 7.8%, 8.8%, 8.1%, 8.1%, and 10.4% of active enhancers in ES, DE, GT, PP1, PP2, and islets, respectively; Figure 1B and Figure 1—figure supplement 1C) and genes proximal to SE were more highly expressed than genes proximal to TE (p=4.68 × 10−7, 4.64 × 10−11, 1.31 × 10−5, 8.85 × 10−9, 5.34 × 10−6, and <2.2 × 10−16 for expression of genes near TE vs SE in ES, DE, GT, PP1, PP2, and islets, respectively; Figure 1—figure supplement 1D). Genes near SE in pancreatic progenitors included transcription factors involved in the regulation of pancreatic cell identity, such as NKX6.1 and PDX1 (Figure 1C). Since disease-associated variants are preferentially enriched in narrow peaks of accessible chromatin within broader regions of active chromatin (Greenwald et al., 2019; Thurner et al., 2018; Varshney et al., 2017), we next used ATAC-seq to generate genome-wide maps of chromatin accessibility across all time points of differentiation. Nearly all identified SE contained at least one ATAC-seq peak (Figure 1D and Figure 1—figure supplement 1E,F). At the PP2 stage, 62.3% of SE harbored one, 32.2% two or three, and 0.7% four or more ATAC-seq peaks (Figure 1—figure supplement 1F). Similar percentages were observed in earlier developmental precursors and islets.

Having annotated accessible chromatin sites within SE, we next tested for enrichment of T2D-associated variants in SE active in mature islets and in pancreatic developmental stages. We observed strongest enrichment of T2D-associated variants in islet SE (log enrichment = 2.18, 95% CI = 1.80, 2.54) and late pancreatic progenitor SE (log enrichment = 2.17, 95% CI = 1.40, 2.74), which was more pronounced when only considering variants in accessible chromatin sites within these elements (islet log enrichment = 3.20, 95% CI = 2.74, 3.60; PP2 log enrichment = 3.18, 95% CI = 2.35, 3.79; Figure 1E). Given that a subset of pancreatic progenitor SE is also active in islets, we next determined whether pancreatic progenitor SE contribute to T2D risk independently of islet SE. Variants in accessible chromatin sites of late pancreatic progenitor SE were enriched for T2D association in a joint model including islet SE (islet log enrichment = 2.94, 95% CI = 2.47, 3.35; PP2 log enrichment = 1.27, 95% CI = 0.24, 2.00; Figure 1F). We also observed enrichment of variants in accessible chromatin sites of pancreatic progenitor SE after conditioning on islet SE (log enrichment = 0.60, 95% CI = −0.87, 1.48), as well as when excluding pancreatic progenitor SE active in islets (log enrichment = 1.62, 95% CI = <-20, 3.14). Examples of known T2D loci with T2D-associated variants in SE active in pancreatic progenitors but not in islets included LAMA1 and PROX1. These results suggest that a subset of T2D variants may affect disease risk by altering regulatory programs specifically active in pancreatic progenitors.

Pancreatic progenitor-specific stretch enhancers are near genes that regulate tissue morphogenesis

Having observed enrichment of T2D risk variants in pancreatic progenitor SE independent of islet SE, we next sought to further characterize the regulatory programs of SE with specific function in pancreatic progenitors. We therefore defined a set of pancreatic progenitor-specific stretch enhancers (PSSE) based on the following criteria: (i) annotation as a SE at the PP2 stage, (ii) no classification as a SE at the ES, DE, and GT stages, and (iii) no classification as a TE or SE in islets. Applying these criteria, we identified a total of 492 PSSE genome-wide (Figure 2A and Figure 2—source data 1).

Figure 2 with 1 supplement see all
Candidate target genes of pancreatic progenitor-specific stretch enhancers regulate developmental processes.

(A) Schematic illustrating identification of pancreatic progenitor-specific stretch enhancers (PSSE). (B) Heatmap showing density of H3K27ac ChIP-seq and ATAC-seq reads at PSSE, centered on overlapping H3K27ac and ATAC-seq peaks, respectively, and spanning 5 kb in ES, DE, GT, PP1, PP2, and islets. PSSE coordinates in Figure 2—source data 1. (C) Percentage of PSSE exhibiting indicated chromatin states at defined developmental stages and in islets. (D) Percentage of PSSE overlapping with at least one ChIP-seq peak at PP2 for the indicated transcription factors. Enrichment analysis comparing observed and expected overlap based on random genomic regions of the same size and located on the same chromosome averaged over 10,000 iterations (***p<1×10−4; permutation test). (E) Gene ontology analysis for nearest expressed genes (fragments per kilobase per million fragments mapped (FPKM) ≥1 at PP2) to the 492 PSSE. See also Figure 2—source data 2. (F) Enrichment (LD score regression coefficient z-scores) of T2D, developmental, and metabolic GWAS trait-associated variants at accessible chromatin sites in PSSE as compared with PP2 and islet stretch enhancers. Significant enrichment was identified within accessible chromatin at PP2 stretch enhancers for lean type 2 diabetes (Z = 2.06, *p=3.94 × 10−2), at PP2 stretch enhancers for type 2 diabetes (Z = 3.57, ***p=3.52 × 10−4), at islet stretch enhancers for type 2 diabetes (Z = 2.78, **p=5.46 × 10−3), at islet stretch enhancers for fasting proinsulin levels (Z = 2.83, **p=4.61 × 10−3), at islet stretch enhancers for HOMA-B (Z = 2.58, **p=9.85 × 10−3), at PP2 stretch enhancers for disposition index (Z = 2.18, *p=2.94 × 10−2), at islet stretch enhancers for acute insulin response (Z = 2.24, *p=2.51 × 10−2), at islet stretch enhancers for HbA1c (Z = 1.98, *p=4.72 × 10−2), and at islet stretch enhancers for fasting glucose levels (Z = 2.64, **p=8.31 × 10−3). See also Figure 2—source data 3 and Figure 2—figure supplement 1.

Figure 2—source data 1

Chromosomal coordinates of pancreatic progenitor-specific stretch enhancers (PSSE).

https://cdn.elifesciences.org/articles/59067/elife-59067-fig2-data1-v2.xlsx
Figure 2—source data 2

Enriched gene ontology terms for PSSE-associated genes.

https://cdn.elifesciences.org/articles/59067/elife-59067-fig2-data2-v2.xlsx
Figure 2—source data 3

Proportion of variants nominally associated with beta cell functional traits.

https://cdn.elifesciences.org/articles/59067/elife-59067-fig2-data3-v2.xlsx
Figure 2—source data 4

Tissue identity of downloaded data from ROADMAP consortium.

https://cdn.elifesciences.org/articles/59067/elife-59067-fig2-data4-v2.xlsx

As expected based on their chromatin state classification, PSSE acquired broad deposition of the active enhancer mark H3K27ac at the PP1 and PP2 stages (Figure 2B,C). Coincident with an increase in H3K27ac signal, chromatin accessibility at PSSE also increased (Figure 2B), and 93.5% of PSSE contained at least one accessible chromatin site at the PP2 stage (Figure 2—figure supplement 1A,B). Further investigation of PSSE chromatin state dynamics at earlier stages of pancreatic differentiation revealed that PSSE were often poised (defined by H3K4me1 in the absence of H3K27ac) prior to activation (42%, 48%, 63%, and 17% of PSSE in ES, DE, GT, and PP1, respectively; Figure 2C), consistent with earlier observations that a poised enhancer state frequently precedes enhancer activation during development (Rada-Iglesias et al., 2011; Wang et al., 2015). Intriguingly, a subset of PSSE was classified as TE earlier in development (13%, 23%, 29%, and 46% of PSSE in ES, DE, GT, and PP1, respectively; Figure 2C), suggesting that SE emerge from smaller regions of active chromatin seeded at prior stages of development. During differentiation into mature islet cells, PSSE lost H3K27ac but largely retained H3K4me1 signal (62% of PSSE) (Figure 2C), persisting in a poised state in terminally differentiated islet cells.

To gain insight into the transcription factors that regulate PSSE, we conducted motif enrichment analysis of accessible chromatin sites within PSSE (Figure 2—figure supplement 1C). Consistent with the activation of PSSE upon pancreas induction, motifs associated with transcription factors known to regulate pancreatic development (Conrad et al., 2014; Masui et al., 2007) were enriched, including FOXA (p=1 × 10−34), PDX1 (p=1 × 10−30), GATA (p=1 × 10−25), ONECUT (p=1 × 10−17), and RBPJ (p=1 × 10−14), suggesting that pancreatic lineage-determining transcription factors activate PSSE. Analysis of the extent of PSSE overlap with ChIP-seq binding sites for FOXA1, FOXA2, GATA4, GATA6, PDX1, HNF6, and SOX9 at the PP2 stage substantiated this prediction (p<1 × 10−4 for all transcription factors; permutation test; Figure 2D).

Annotation of biological functions of predicted target genes for PSSE (nearest gene with FPKM ≥1 at PP2 stage) revealed gene ontology terms related to developmental processes, such as tissue morphogenesis (p=1 × 10−7) and vascular development (p=1 × 10−8), as well as developmental signaling pathways, including BMP (p=1 × 10−5), NOTCH (p=1 × 10−4), and canonical Wnt signaling (p=1 × 10−4; Figure 2E and Figure 2—source data 2), which have demonstrated roles in pancreas morphogenesis and cell lineage allocation (Ahnfelt-Rønne et al., 2010; Li et al., 2015; Murtaugh, 2008; Sharon et al., 2019; Sui et al., 2013). Consistent with the temporal pattern of H3K27ac deposition at PSSE, transcript levels of PSSE-associated genes increased upon pancreatic lineage induction and peaked at the PP2 stage (p=1.8 × 10−8; Figure 2—figure supplement 1D). Notably, expression of these genes sharply decreased in islets (p<2.2 × 10−16), underscoring the likely role of these genes in regulating pancreatic development but not mature islet function.

Pancreatic progenitor-specific stretch enhancers are highly specific across T2D-relevant tissues and cell types

We next sought to understand the phenotypic consequences of PSSE activity in the context of T2D pathophysiology. Variants in accessible chromatin sites of PSSE genome-wide were enriched for T2D association (log enrichment = 2.85, 95% CI = <-20, 4.09). We determined enrichment of genetic variants for T2D-related quantitative endophenotypes within accessible chromatin sites of PSSE, as well as all pancreatic progenitor SE (not just progenitor-specific) and islet SE, using LD score regression (Bulik-Sullivan et al., 2015; Finucane et al., 2015). As expected based on prior observations (Parker et al., 2013; Pasquali et al., 2014), we observed enrichment (Z > 1.96) of variants associated with quantitative traits related to insulin secretion and beta cell function within islet SE, exemplified by fasting proinsulin levels, HOMA-B, and acute insulin response (Z = 2.8, Z = 2.6, and Z = 2.2, respectively; Figure 2F). Conversely, PSSE showed a trend toward depletion for these traits, although the estimates were not significant. We further tested for enrichment in the proportion of variants in PSSE and islet SE nominally associated (p<0.05) with beta cell function traits compared to background variants. There was significant enrichment of beta cell trait association among islet SE variants (χ2 test; p<0.05 for all beta cell functional traits except for insulin secretion rate), but no corresponding enrichment for PSSE (Figure 2—source data 3).

A prior study found that variants at the LAMA1 locus had stronger effects on T2D risk among lean relative to obese cases (Perry et al., 2012). Since we identified a PSSE at the LAMA1 locus, we postulated that variants in PSSE collectively might have differing impact on T2D risk in cases segregated by BMI. We therefore tested PSSE, as well as pancreatic progenitor SE and islet SE, for enrichment of T2D association using GWAS of lean and obese T2D (Perry et al., 2012), using LD score regression (Bulik-Sullivan et al., 2015; Finucane et al., 2015). We observed nominally significant enrichment of variants in pancreatic progenitor SE for T2D among lean cases (Z = 2.1). Variants in PSSE were mildly enriched for T2D among lean (Z = 1.1) and depleted among obese (Z = −0.70) cases, although neither estimate was significant. By comparison, islet SE showed positive enrichment for T2D among both lean (Z = 1.9) and obese cases (Z = 1.3; Figure 2F). Together, these results suggest that PSSE may affect T2D risk in a manner distinct from islet SE function.

Having observed little evidence for enrichment of PSSE variants for traits related to beta cell function, we asked whether the enrichment of PSSE for T2D-associated variants could be explained by PSSE activity in T2D-relevant tissues and cell types outside the pancreas. We assessed PSSE activity by measuring H3K27ac signal in 95 representative tissues and cell lines from the ENCODE and Epigenome Roadmap projects (Kundaje et al., 2015). Interestingly, there was group-wide specificity of PSSE to pancreatic progenitors relative to other cells and tissues including those relevant to T2D, such as adipose tissue, skeletal muscle, and liver (Figure 2—figure supplement 1E and Figure 2—source data 4). Since gene regulation in adipocyte precursors also contributes to T2D risk (Claussnitzer et al., 2014), we further examined PSSE specificity with respect to chromatin states during adipogenesis, using data from human adipose stromal cell differentiation stages (hASC1-4) (Mikkelsen et al., 2010; Varshney et al., 2017). PSSE exhibited virtually no active chromatin during adipogenesis (9, 8, 6, and 8 out of the 492 PSSE were active enhancers in hACS-1, hASC-2, hASC-3, and hASC-4, respectively; Figure 2—figure supplement 1F). These findings identify PSSE as highly pancreatic progenitor-specific across T2D-relevant tissues and cell types.

Identification of pancreatic progenitor-specific stretch enhancers harboring T2D-associated variants

Given the relative specificity of PSSE to pancreatic progenitors, we next sought to identify T2D-associated variants in PSSE at specific loci which may affect pancreatic development. We therefore identified variants in PSSE with evidence of T2D association (at p=4.7 × 10−6) after correcting for the total number of variants in PSSE genome-wide (n = 10,738). In total there were 49 variants in PSSE with T2D association exceeding this threshold mapping to 11 loci (Figure 3A). This included variants at nine loci with known genome-wide significant T2D association (PROX1, ST6GAL1, SMARCAD1, XKR6, INS-IGF2, HMGA2, SMEK1, HMG20A, and LAMA1), as well as at two previously unreported loci with sub-genome-wide significant association, CRB2 and PGM1. To identify candidate target genes of the T2D-associated PSSE in pancreatic progenitors, we analyzed the expression of all genes within the same topologically associated domain (TAD) as the PSSE in PP2 cells and in primary human embryonic pancreas tissue (Figure 3B and Figure 3—figure supplement 1A). These expressed genes are candidate effector transcripts of T2D-associated variants in pancreatic progenitors.

Figure 3 with 1 supplement see all
Identification of T2D risk variants associated with pancreatic progenitor-specific stretch enhancers.

(A) Manhattan plot showing T2D association p-values (from Mahajan et al., 2018) for 10,738 variants mapping within PSSE. The dotted line shows the threshold for Bonferroni correction (p=4.66 × 10−6). Novel loci identified with this threshold and mapping at least 500 kb away from a known locus are highlighted in blue. Chromosomal coordinates of T2D-associated PSSE are indicated. (B) mRNA levels (measured in fragments per kilobase per million fragments mapped [FPKM]) at PP2 (blue) and in human embryonic pancreas (54 and 58 days gestation, gold) of nearest expressed (FPMK ≥1) gene at PP2 for PSSE harboring T2D variants identified in A. (C) PP2 specificity of H3K27ac signal at PSSE harboring T2D variants identified in A. Z-score comparing H3K27ac signal at PP2 to H3K27ac signal in tissues and cell lines from the ENCODE and Epigenome Roadmap projects. See also Figure 3—figure supplement 1.

As many pancreatic progenitor SE remain poised in mature islets (Figure 2C), we considered whether T2D-associated variants in PSSE could have gene regulatory function in islets that is re-activated in the disease state. We therefore assessed overlap of PSSE variants with accessible chromatin of islets from T2D donors (Khetan et al., 2018). None of the strongly T2D-associated variants in PSSE (p=4.7 × 10−6) overlapped an islet accessible chromatin site in T2D islets, arguing against the relevance of PSSE in broadly regulating islet gene activity during T2D.

A pancreatic progenitor-specific stretch enhancer at LAMA1 harbors T2D risk variants and regulates LAMA1 expression selectively in pancreatic progenitors

Variants in a PSSE at the LAMA1 locus were associated with T2D at genome-wide significance (Figure 3A), and LAMA1 was highly expressed in the human embryonic pancreas (Figure 3B). Furthermore, the activity of the PSSE at the LAMA1 locus was almost exclusively restricted to pancreatic progenitors (Figure 3—figure supplement 1B,C), and was further among the most progenitor-specific across all PSSE harboring T2D risk variants (Figure 3C). In addition, reporter gene assays in zebrafish embryos have shown that this enhancer drives gene expression specific to pancreatic progenitors in vivo (Cebola et al., 2015). We therefore postulated that the activity of T2D-associated variants within the LAMA1 PSSE is relevant for gene regulation in pancreatic progenitors, and we sought to characterize the LAMA1 PSSE in greater depth.

Multiple T2D-associated variants mapped within the LAMA1 PSSE, and these variants were further in the 99% credible set in fine-mapping data from the DIAMANTE consortium (Mahajan et al., 2018; Figure 4A). No other variants in the 99% credible set mapped in an accessible chromatin site active in islets from either non-diabetic or T2D samples. The PSSE is intronic to the LAMA1 gene and contains regions of poised chromatin and TE at prior developmental stages (Figure 4A). Consistent with its stepwise genesis as a SE throughout development, regions of open chromatin within the LAMA1 PSSE were already present at the DE and GT stages. Furthermore, pancreatic lineage-determining transcription factors, such as FOXA1, FOXA2, GATA4, GATA6, HNF6, SOX9, and PDX1, were all bound to the PSSE at the PP2 stage (Figure 4B). Among credible set variants in the LAMA1 PSSE, rs10502347 overlapped an ATAC-seq peak as well as ChIP-seq sites for multiple pancreatic lineage-determining transcription factors. Additionally, rs10502347 directly coincided with a SOX9 footprint identified in ATAC-seq data from PP2 cells, and the T2D risk allele C is predicted to disrupt SOX9 binding (Figure 4B). Consistent with the collective endophenotype association patterns of PSSE (Figure 2F), rs10502347 showed no association with beta cell function (p=0.81, 0.23, 0.46 for fasting proinsulin levels, HOMA-B, and acute insulin response, respectively; Figure 4—figure supplement 1A). Thus, T2D variant rs10502347 is predicted to affect the binding of pancreatic transcription factors and does not appear to affect beta cell function.

Figure 4 with 2 supplements see all
A T2D risk-associated LAMA1 pancreatic progenitor-specific stretch enhancer regulates LAMA1 expression specifically in pancreatic progenitors.

(A) (Top) Locus plots showing T2D association p-values for variants in a 35 kb window (hg19 chr18:7,050,000–7,085,000) at the LAMA1 locus and LAMA1 PSSE (red box). Fine mapped variants within the 99% credible set for the LAMA1 locus are colored black. All other variants are colored light gray. (Bottom) Chromatin states and ATAC-seq signal in ES, DE, GT, PP1, and PP2. TssA, active promoter; TssFlnk, flanking transcription start site; TssBiv, bivalent promoter; Repr, repressed; EnhA, active enhancer; EnhP, poised enhancer. (B) FOXA1, FOXA2, GATA4, GATA6, HNF6, SOX9, and PDX1 ChIP-seq profiles at the LAMA1 PSSE in PP2. The variant rs10502347 (red) overlaps transcription factor binding sites and a predicted ATAC-seq footprint for the SOX9 sequence motif. Purple dotted lines indicate the core binding profile of the average SOX9 footprint genome-wide and the blue dotted line indicates the position of rs10502347 within the SOX9 motif. (C) LAMA1 mRNA expression at each developmental stage determined by RNA-seq, measured in fragments per kilobase per million fragments mapped (FPKM). Data shown as mean ± S.E.M. (n = 3 replicates from independent differentiations). Light blue and purple indicate classification of the LAMA1 PSSE as typical enhancer (TE) and stretch enhancer (SE), respectively. (D) LAMA1 mRNA expression at each developmental stage determined by qPCR in control and ∆LAMA1Enh cells. Data are shown as mean ± S.E.M. (n = 3 replicates from independent differentiations for control cells. ∆LAMA1Enh cells represent combined data from two clonal lines with three replicates for each line from independent differentiations. n = 3 technical replicates for each sample; p=0.319, 0.594, 0.945, 0.290, and <1 × 10−6 for comparisons in ES, DE, GT, PP1, and PP2, respectively; student’s t-test, two sided; ***p<0.001, n.s., not significant). Light blue and purple indicate classification of the LAMA1 PSSE as TE and SE, respectively. Each plotted point represents the average of technical replicates for each differentiation. (E) mRNA expression determined by RNA-seq at PP2 of genes expressed in either control or ∆LAMA1Enh cells (FPKM ≥ 1 at PP2) and located within the same topologically associated domain as LAMA1. Data are shown as mean FPKM ± S.E.M. (n = 2 replicates from independent differentiations for control cells. ∆LAMA1Enh cells represent combined data from two clonal lines with two replicates for each line from independent differentiations. p adj. = 0.389 and 8.11 × 10−3 for ARHGAP28 and LAMA1, respectively; DESeq2). See also Figure 4—figure supplements 1 and 2.

Figure 4—source data 1

Genes downregulated in ∆LAMA1Enh PP2 stage cells compared to control cells (p adj. <0.05).

https://cdn.elifesciences.org/articles/59067/elife-59067-fig4-data1-v2.xlsx
Figure 4—source data 2

Genes upregulated in ∆LAMA1Enh PP2 stage cells compared to control cells (p adj. <0.05).

https://cdn.elifesciences.org/articles/59067/elife-59067-fig4-data2-v2.xlsx

Enhancers can control gene expression over large genomic distances, and therefore their target genes cannot be predicted based on proximity alone. To directly assess the function of the LAMA1 PSSE in regulating gene activity, we utilized CRIPSR-Cas9-mediated genome editing to generate two independent clonal human hESC lines harboring homozygous deletions of the LAMA1 PSSE (hereafter referred to as ∆LAMA1Enh; Figure 4—figure supplement 1B). We examined LAMA1 expression in ∆LAMA1Enh compared to control cells throughout stages of pancreatic differentiation. Consistent with the broad expression of LAMA1 across developmental and mature tissues, control cells expressed LAMA1 at all stages (Figure 4C). LAMA1 was expressed at similar levels in ∆LAMA1Enh and control cells at early developmental stages, but was significantly reduced in PP2 cells derived from ∆LAMA1Enh clones (p=0.319, 0.594, 0.945, 0.290, and <1 × 10−6 for comparisons in ES, DE, GT, PP1, and PP2, respectively; Figure 4D). To next investigate whether the LAMA1 PSSE regulates other genes at this locus, we utilized Hi-C datasets from PP2 cells to identify topologically associated domains (TADs). We then examined expression of genes mapping in the same TAD as the LAMA1 PSSE. ARHGAP28 was the only other expressed gene within the TAD, and albeit not significantly different from controls (p.adj >0.05), showed a trend toward lower expression in ∆LAMA1Enh PP2 cells (Figure 4E), raising the possibility that ARHGAP28 is an additional target gene of the LAMA1 PSSE. Together, these results demonstrate that while LAMA1 itself is broadly expressed across developmental stages, the T2D-associated PSSE regulates LAMA1 expression specifically in pancreatic progenitors.

To determine whether deletion of the LAMA1 PSSE affects pancreatic development, we generated PP2 stage cells from ∆LAMA1Enh and control hESC lines and analyzed pancreatic cell fate commitment by flow cytometry and immunofluorescence staining for PDX1 and NKX6.1 (Figure 4—figure supplement 1C,D). At the PP2 stage, ∆LAMA1Enh and control cultures contained similar percentages of PDX1- and NKX6.1-positive cells. Furthermore, mRNA expression of PDX1, NKX6.1, PROX1, PTF1A, and SOX9 was either unaffected or only minimally reduced (p adj. = 3.56 × 10−2, 0.224, 0.829, 8.14 × 10−2, and 0.142, for comparisons of PDX1, NKX6.1, PROX1, PTF1A, and SOX9 expression, respectively; Figure 4—figure supplement 1E), and the overall gene expression profiles as determined by RNA-seq were similar in ∆LAMA1Enh and control PP2 cells (Figure 4—figure supplement 1F and Figure 4—source datas 1 and 2). To examine effects of complete LAMA1 loss-of-function, we additionally generated a hESC line harboring a deletion of the LAMA1 coding sequences (hereafter referred to as ∆LAMA1; Figure 4—figure supplement 2A,B), and produced PP2 stage cells. Similar to ∆LAMA1Enh cultures, ∆LAMA1 and control PP2 stage cultures contained similar numbers of PDX1- and NKX6.1-positive cells (Figure 4—figure supplement 2C,D). Likewise, mRNA expression of PDX1, NKX6.1, PROX1, PTF1A, and SOX9 was similar in ∆LAMA1 and control PP2 cells (p=4.3 × 10−2, 0.19, 0.16, 0.17, and 8.7 × 10−2, respectively; Figure 4—figure supplement 2E). These findings indicate that in vitro pancreatic lineage induction is unperturbed in both ∆LAMA1Enh cells exhibiting reduced LAMA1 expression, as well as ∆LAMA1 cells where LAMA1 coding sequences are disrupted.

Pancreatic progenitor-specific stretch enhancers at the CRB2 and PGM1 loci harbor T2D-associated variants

Multiple variants with evidence for T2D association in PSSE mapped outside of known risk loci, such as those mapping to CRB2 and PGM1 (Figure 3A). As with the LAMA1 PSSE, PSSE harboring variants at CRB2 and PGM1 were intronic to their respective genes, contained ATAC-seq peaks, and bound pancreatic lineage-determining transcription factors FOXA1, FOXA2, GATA4, GATA6, HNF6, SOX9, and PDX1 (Figure 5A,B and Figure 5—figure supplement 1A,B). Compared to the LAMA1 PSSE, CRB2 and PGM1 PSSE were less specific to pancreatic progenitors and exhibited significant H3K27ac signal in several other tissues and cell types, most notably brain, liver, and the digestive tract (Figure 5—figure supplement 1C,D).

Figure 5 with 3 supplements see all
A T2D risk-associated CRB2 pancreatic progenitor-specific stretch enhancer regulates CRB2 expression specifically in pancreatic progenitors.

(A) (Top) Locus plots showing T2D association p-values for variants in a 35 kb window (hg19 chr9:126,112,000–126,147,000) at the CRB2 locus and CRB2 PSSE (red box). Fine mapped variants within the 99% credible set for the novel CRB2 locus are colored black. All other variants are colored light gray. (Bottom) Chromatin states and ATAC-seq signal in ES, DE, GT, PP1, and PP2. TssA, active promoter; TssFlnk, flanking transcription start site; TssBiv, bivalent promoter; Repr, repressed; EnhA, active enhancer; EnhP, poised enhancer. (B) FOXA1, FOXA2, GATA4, GATA6, HNF6, SOX9, and PDX1 ChIP-seq profiles at the CRB2 PSSE in PP2. The variant rs2491353 (black) overlaps with transcription factor binding sites. (C) CRB2 mRNA expression at each developmental stage determined by RNA-seq, measured in fragments per kilobase per million fragments mapped (FPKM). Data shown as mean ± S.E.M. (n = 3 replicates from independent differentiations). Light blue and purple indicate classification of the CRB2 PSSE as typical enhancer (TE) and stretch enhancer (SE), respectively. Plotted points represent average of technical replicates for each differentiation. (D) CRB2 mRNA expression at each developmental stage determined by qPCR in control and ∆CRB2Enh cells. Data are shown as mean ± S.E.M. (n = 3 replicates from independent differentiations for control cells. ∆CRB2Enh cells represent combined data from two clonal lines with three replicates for each line from independent differentiations. n = 3 technical replicates for each sample; p=7.03 × 10−4,<1 × 10−6,<1 × 10−6, 1.46 × 10−2, and <1 × 10−6 for comparisons in ES, DE, GT, PP1, and PP2, respectively; student’s t-test, two sided; ***p<0.001 **p<0.01). Light blue and purple indicate classification of the CRB2 PSSE as TE and SE, respectively. Each plotted point represents the average of technical replicates for each differentiation. (E) mRNA expression determined by RNA-seq at PP2 of genes expressed in either control or ∆CRB2Enh cells (FPKM ≥ 1 at PP2) and located within the same topologically associated domain as CRB2. Data are shown as mean FPKM ± S.E.M. (n = 2 replicates from independent differentiations for control cells. ∆CRB2Enh cells represent combined data from two clonal lines with two replicates for each line from independent differentiations. p adj. = 0.158, 1.00, and 3.51 × 10−3, for MIR600HG, STRBP, and CRB2, respectively; DESeq2; **p<0.01, n.s., not significant). See also Figure 5—figure supplements 13.

Figure 5—source data 1

Genes downregulated in ∆CRB2Enh PP2 stage cells compared to control cells (p adj. <0.05).

https://cdn.elifesciences.org/articles/59067/elife-59067-fig5-data1-v2.xlsx

CRB2 is a component of the Crumbs protein complex involved in the regulation of cell polarity and neuronal, heart, retinal, and kidney development (Alves et al., 2013; Bulgakova and Knust, 2009; Dudok et al., 2016; Jiménez-Amilburu and Stainier, 2019; Slavotinek et al., 2015). However, its role in pancreatic development is unknown. To determine whether the CRB2 PSSE regulates CRB2 expression in pancreatic progenitors, we generated two independent hESC clones with homozygous deletions of the CRB2 PSSE (hereafter referred to as ∆CRB2Enh; Figure 5—figure supplement 2A) and performed pancreatic differentiation of ∆CRB2Enh and control hESC lines. In control cells, CRB2 was first expressed at the GT stage and increased markedly at the PP1 stage (Figure 5C). This pattern of CRB2 expression is consistent with H3K27ac deposition at the CRB2 PSSE in GT stage cells and classification as a SE at the PP1 and PP2 stages (Figure 5A and Figure 5—figure supplement 1C). In ∆CRB2Enh cells, we observed upregulation of CRB2 expression at earlier developmental stages, in particular at the DE and GT stages (p<1 × 10−6 at both stages; Figure 5D), suggesting that the CRB2 PSSE may be associated with repressive transcriptional complexes prior to pancreas induction. At the PP2 stage, CRB2 expression was significantly reduced in ∆CRB2Enh cells (p adj. = 3.51 × 10−3; Figure 5D), whereas the expression of other genes in the same TAD was not affected (p adj. ≥0.05; Figure 5E). Thus, the CRB2 PSSE specifically regulates CRB2 and is required for CRB2 expression in pancreatic progenitors.

Phenotypic characterization of PP2 stage ∆CRB2Enh cultures revealed similar percentages of PDX1- and NKX6.1-positive cells as in control cultures (Figure 5—figure supplement 2B,C). The expression of pancreatic transcription factors and global gene expression profiles were also similar in ∆CRB2Enh and control PP2 cells (Figure 5—figure supplement 2D,E and Figure 5—source data 1). Likewise, CRB2 deletion hESCs (∆CRB2) differentiated to the PP2 stage (Figure 5—figure supplement 3A,B) produced similar numbers of PDX1- and NKX6.1-positive cells and expressed pancreatic transcription factors at levels similar to control cells (Figure 5—figure supplement 3C–E). Thus, neither deletion of the CRB2 PSSE nor the CRB2 gene overtly impairs pancreatic lineage induction in the in vitro hESC differentiation system.

lama1 and crb2 zebrafish morphants display annular pancreas and decreased beta cell mass

Based on their classification as extracellular matrix and cell polarity proteins, respectively, Laminin (encoded by LAMA1) and CRB2 are predicted to regulate processes related to tissue morphogenesis, such as cell migration, tissue growth, and cell allocation within the developing organ. Furthermore, PSSE in general were enriched for proximity to genes involved in tissue morphogenesis (Figure 2E), suggesting that T2D risk variants acting within PSSE could have roles in pancreas morphogenesis. Since cell migratory processes and niche-specific signaling events are not fully modeled during hESC differentiation, we reasoned that the in vitro pancreatic differentiation system might not be suitable for studying Laminin and CRB2 function in pancreatic development.

To circumvent these limitations, we employed zebrafish as an in vivo vertebrate model to study the effects of reduced lama1 and crb2 levels on pancreatic development. The basic organization and cell types in the pancreas as well as the genes regulating endocrine and exocrine pancreas development are highly conserved between zebrafish and mammals (Dong et al., 2008; Field et al., 2003; Kimmel et al., 2015). To analyze pancreatic expression of Laminin and Crb proteins, we used Tg(ptf1a:eGFP)jh1 embryos to visualize pancreatic progenitor cells and the acinar pancreas by eGFP expression. At 48 hr post-fertilization (hpf), both Laminin and Crb proteins were detected in the eGFP and Nkx6.1 co-positive pancreatic progenitor cell domain (Figure 6—figure supplement 1A,B).

To determine the respective functions of lama1 and crb2 in pancreatic development, we performed knockdown experiments using anti-sense morpholinos directed against lama1 and the two zebrafish crb2 genes, crb2a and crb2b (Omori and Malicki, 2006; Pollard et al., 2006). Knockdown efficiency of each morpholino was validated using whole-mount immunohistochemistry. We observed significant reduction of Laminin staining throughout the pancreatic progenitor cell domain in embryos treated with morpholinos targeting lama1 (Figure 6—figure supplement 2A–D). In embryos treated with morpholinos targeting crb2a or crb2a and crb2b, we observed loss of staining in the pancreatic progenitor cell domain using antibodies specific to Crb2a or antibodies detecting all Crb proteins, respectively (Figure 6—figure supplement 3A–H) Residual panCrb protein signal was observed in the dorsal pancreas, which may be the result of expression of Crb proteins other than Crb2a and Crb2b in this region.

Consistent with prior studies (Pollard et al., 2006), lama1 morphants exhibited reduced body size and other gross anatomical defects at 78 hpf, whereas crb2a/b morphants appeared grossly normal. Both lama1 and crb2a/b morphants displayed an annular pancreas (15 out of 34 lama1 and 27 out of 69 crb2a/b morphants) characterized by pancreatic tissue partially or completely encircling the duodenum (Figure 6A–D), a phenotype indicative of impaired migration of pancreatic progenitors during pancreas formation. These findings suggest that both lama1 and crb2a/b control cell migratory processes during early pancreatic development and that reduced levels of lama1 or crb2a/b impair pancreas morphogenesis.

Figure 6 with 4 supplements see all
lama1 and crb2 regulate pancreas morphogenesis and beta cell differentiation.

(A,B) Representative 3D renderings of Tg(ptf1a:eGFP)jh1 control zebrafish embryos (A,A’) and lama1 morphants (B,B’) stained with DAPI (nuclei, blue) and antibody against insulin (red); n ≥ 15 embryos per condition. To account for reduced acinar pancreas size in lama1 morphants, control embryos were imaged at 50 hr post fertilization (hpf) and lama1 morphants at 78 hpf. 15 out of 34 lama1 morphants displayed an annular pancreas with two acinar pancreas domains (green) connected behind the presumptive intestine (B’, white arrow). Scale bar, 40 µM. (C,D) Representative 3D renderings of 78 hpf Tg(ptf1a:eGFP)jh1 control zebrafish embryos (C,C’) and crb2a/b morphants (D,D’) stained with DAPI (nuclei, blue) and antibodies against insulin (red); n ≥ 15 embryos per condition. Twenty-seven out of 69 crb2a/b morphants displayed an annular pancreas with the acinar pancreas (green) completely surrounding the presumptive intestine. Scale bar, 40 µM. (E) Representative 3D renderings of Tg(ptf1a:eGFP)jh1 control zebrafish embryos and crb2a/b, lama1, or crb2a/b + lama1 morphants stained with DAPI (nuclei, blue) and antibody against insulin (red). All embryos were imaged at 78 hpf except for controls to lama1 and crb2a/b + lama1 morphants, which were imaged at 50 hpf to account for reduced acinar pancreas size of lama1 morphants. Scale bar, 20 µM. (F) Quantification of beta (insulin+) cell nuclei per embryo from experiment in (E). p adj. = 4.0 × 10−3, 8.0 × 10−3, and 2.0 × 10−4 for comparison of hfp 78 control (n = 7 embryos) to hfp 78 crb2a/b (n = 8), hpf 50 control (n = 12) to hpf 78 lama1 (n = 10), or crb2a/b + lama1 (n = 12) morphants, respectively; ANOVA-Dunnett’s multiple comparison test; ***p<0.001 **p<0.01. 5 out of 8 crb2a/b, 3 out of 10 lama1, and 9 out of 12 crb2a/b + lama1 morphants displayed an annular pancreas. MO, morpholino; Control, standard control morpholino. See also Figure 6—figure supplements 14.

To gain insight into the effects of lama1 and crb2a/b knockdown on pancreatic endocrine cell development, we examined beta cell numbers (insulin+ cells) at 78 hpf. We also evaluated potential synergistic effects of combined lama1 and crb2a/b knockdown. To account for the reduction in body and pancreas size in lama1 morphants, we compared cell numbers in 78 hpf lama1 morphants with 50 hpf control embryos, which have a similarly sized acinar compartment as 78 hpf lama1 morphants. Beta cell numbers were significantly reduced in both lama1 and crb2a/b morphants (p=8.0 × 10−3 and 4.0 × 10−3 for comparisons of lama1 and crb2a/b morphants, respectively; Figure 6E,F), as well as in morphants with a combined knockdown of lama1 and crb2a/b (p=2.0 × 10−4; Figure 6F), showing that reduced lama1 and crb2a/b levels, both individually and in combination, impair beta cell development. Furthermore, we found that nearly all lama1, crb2a/b, and combined lama1 and crb2a/b morphants without an annular pancreas had reduced beta cell numbers, indicating independent roles of lama1 and crb2 in pancreas morphogenesis and beta cell differentiation. Finally, to investigate the contributions of individual crb2 genes to the observed phenotype, we performed knockdown experiments using morpholinos against crb2a and crb2b alone. Only crb2b morphants showed a significant reduction in beta cell numbers (p=4.4 × 10−2; Figure 6—figure supplement 4), suggesting that crb2b is the predominant crb2 gene required for beta cell development. Combined, these findings demonstrate that lama1 and crb2 are regulators of pancreas morphogenesis and beta cell development in vivo.

Discussion

In this study, we identify T2D-associated variants localized within chromatin active in pancreatic progenitors but not islets or other T2D-relevant tissues, suggesting a novel mechanism whereby a subset of T2D risk variants specifically alters pancreatic developmental processes. We link T2D-associated enhancers active in pancreatic progenitors to the regulation of LAMA1 and CRB2 and demonstrate a functional requirement in zebrafish for lama1 and crb2 in pancreas morphogenesis and endocrine cell formation. Furthermore, we provide a curated list of T2D risk-associated enhancers and candidate effector genes for further exploration of how the regulation of developmental processes in the pancreas can predispose to T2D.

Our analysis identified 11 loci where T2D-associated variants mapped in SE specifically active in pancreatic progenitors. Among these loci was LAMA1, which has stronger effects on T2D risk in lean compared to obese individuals (Perry et al., 2012). We also found evidence that variants in PSSE collectively have stronger enrichment for T2D in lean individuals, although the small number of PSSE and limited sample size of the BMI-stratified T2D genetic data prohibits a more robust comparison. There was also a notable lack of enrichment among PSSE variants for association with traits related to insulin secretion and beta cell function. If T2D-associated variants in PSSE indeed confer diabetes susceptibility by affecting beta cell development, the question arises as to why variants associated with traits related to beta cell function are not enriched within PSSE. As genetic association studies of endophenotypes are based on data from non-diabetic subjects, a possible explanation is that variants affecting beta cell developmental processes have no overt phenotypic effect under physiological conditions and contribute to T2D pathogenesis only during the disease process.

Since the genomic position of enhancers and transcription factor binding sites is not well conserved between species (Villar et al., 2015), a human cell model is necessary to identify target genes of enhancers associated with disease risk. By employing enhancer deletion in hESCs, we demonstrate that T2D-associated PSSE at the LAMA1 and CRB2 loci regulate LAMA1 and CRB2, respectively, and establish LAMA1 and CRB2 as the predominant target gene of their corresponding PSSE within TAD boundaries. By analyzing LAMA1 and CRB2 expression throughout the pancreatic differentiation time course, we show that the identified PSSE control LAMA1 and CRB2 expression in a temporal manner consistent with the activation pattern of their associated PSSE. While the specific T2D-relevant target genes of the majority of T2D-associated PSSE remain to be identified, it is notable that several are localized within TADs containing genes encoding transcriptional regulators. These include PROX1 and GATA4, which are known to regulate pancreatic development (Shi et al., 2017; Tiyaboonchai et al., 2017; Westmoreland et al., 2012), as well as HMGA2 and BCL6 with unknown functions in the pancreas. Our catalogue of T2D-associated PSSE provides a resource to fully characterize the gene regulatory program associated with developmentally mediated T2D risk in the pancreas. Our finding that predicted target genes of PSSE are similarly expressed in hESC-derived pancreatic progenitors and primary human embryonic pancreas (Figure 3B and Figure 3—figure supplement 1A) further underscores the utility of the hESC-based system for these studies.

In the embryo, endocrine cells differentiate by delaminating from a polarized epithelium of progenitors governed by local cell-cell and cell-matrix signaling events (Mamidi et al., 2018). These processes are not well-recapitulated in the hESC-based pancreatic differentiation system, highlighting a limitation of this system for studying the function of Laminin and CRB2, which are mediators of mechanical signals within an epithelium. Therefore, we analyzed their function in zebrafish as an in vivo model. We show that lama1 or crb2 knockdown leads to an annular pancreas and reduced beta cell numbers. The beta cell differentiation defect was also evident in embryos not displaying an annular pancreas, suggesting independent mechanisms.

Consistent with our findings in lama1 morphants, culture of pancreatic progenitors on Laminin-based substrates promotes endocrine cell differentiation (Mamidi et al., 2018). During in vivo pancreatic development, endothelial cells are an important albeit not the only source of Laminin in the pancreas (Heymans et al., 2019; Mamidi et al., 2018; Nikolova et al., 2006). While we do not know the respective contributions of endothelial cell- and pancreatic progenitor cell-derived Laminin to the phenotype of lama1 morphants, the T2D-associated LAMA1 PSSE is not active in endothelial cells (Figure 3—figure supplement 1C). Furthermore, we found no other T2D-associated variants at the LAMA1 locus mapping in endothelial cell enhancers or accessible chromatin sites in islets, suggesting that T2D risk is linked to LAMA1 regulation in pancreatic progenitors.

Similar to Laminin, CRB2 has been shown to regulate mechanosignaling (Varelas et al., 2010). Our observation that pancreatic progenitor cells express Crb proteins is consistent with the phenotype of crb2 morphants reflecting a progenitor-autonomous role of Crb2. Furthermore, the similarity in pancreatic phenotype between lama1 or crb2 morphants raises the possibility that signals from Laminin and Crb2 could converge on the same intracellular pathways in pancreatic progenitors.

Our findings suggest that variation in gene regulation during pancreatic development can predispose to T2D later in life. Several lines of evidence support the concept of a developmental impact on T2D risk. First, human genetic studies have shown a strong correlation between birth weight and adult cardiometabolic traits and disease (Horikoshi et al., 2016). Second, epidemiological studies provide evidence that offspring of mothers who were pregnant during a famine have a higher prevalence of T2D (Lumey et al., 2015). This phenomenon has been experimentally reproduced in rodents, where maternal malnutrition has been shown to cause reduced beta cell mass at birth and to render beta cells more prone to failure under stress (Nielsen et al., 2014). Together, our results provide a strong rationale for further exploration of how genetic variants affecting developmental gene regulation in the pancreas contribute to T2D risk.

Materials and methods

Key resources table
Reagent type (species)
or resource
DesignationSource or referenceIdentifiersAdditional information
AntibodyAPC Mouse monoclonal IgG1, κ Isotype ControlBD PharmingenCat# 555751, RRID:AB_398613Flow cytometry (1:100)
AntibodyChicken polyclonal anti-GFPAves LabsCat# GFP-1020, RRID:AB_10000240Immunohistochemistry (1:200)
AntibodyCy3-conjugated donkey polyclonal anti-mouseJackson ImmunoResearch LabsCat# 715-165-150, RRID:AB_2340813Immunofluorescence (1:1000)
AntibodyDyLight 488-conjugated donkey polyconal anti-goatJackson ImmunoResearch LabsCat# 705-545-003, RRID:AB_2340428Immunofluorescence (1:500)
AntibodyGoat polyclonal anti-CTCFSanta Cruz BiotechnologyCat# SC-15914X, RRID:AB_2086899ChIP-seq (4 ug)
AntibodyGoat polyclonal anti-FOXA1AbcamCat# ab5089, RRID:AB_304744ChIP-seq (4 ug)
AntibodyGoat polyclonal anti-FOXA2Santa Cruz BiotechnologyCat# sc-6554, RRID:AB_2262810ChIP-seq (4 ug)
AntibodyGoat polyclonal anti-GATA4Santa Cruz BiotechnologyCat# sc-1237, RRID:AB_2108747ChIP-seq (4 ug)
AntibodyGoat polyclonal anti-PDX1AbcamCat# ab47383, RRID:AB_2162359Immunofluorescence (1:500)
AntibodyGuinea pig polyclonal anti-InsulinBiomedaCat# v2024Immunohistochemistry (1:200)
AntibodyMouse monoclonal anti-Crb2aZIRCCat# Zs-4Immunohistochemistry (1:100)
AntibodyMouse polyclonal anti-GATA6Santa Cruz BiotechnologyCat# sc-9055, RRID:AB_2108768ChIP-seq (4 ug)
AntibodyMouse monoclonal anti-NKX6.1Developmental Studies Hybridoma BankCat# F64A6B4, RRID:AB_532380Immunofluorescence (1:300)
AntibodyMouse monoclonal anti-NKX6.1-Alexa Fluor 647BD BiosciencesCat# 563338, RRID:AB_2738144Flow cytometry (1:5)
AntibodyMouse monoclonal anti-NKX6.1Developmental Studies Hybridoma BankCat# F55A10, RRID:AB_532378Immunohistochemistry (1:10)
AntibodyMouse monoclonal anti-PDX1-PEBD BiosciencesCat# 562161, RRID:AB_10893589Flow cytometry (1:10)
AntibodyPE Mouse monoclonal IgG1, κ Isotype ControlBD PharmingenCat# 555749, RRID:AB_396091Flow cytometry (1:100)
AntibodyRabbit polyclonal anti-CRB2SigmaCat # SAB1301340Immunofluorescence (1:500)
AntibodyRabbit polyclonal anti-H3K27acActive MotifCat# 39133, RRID:AB_2561016ChIP-seq (4 ug)
AntibodyRabbit polyclonal anti-H3K4me1AbcamCat# ab8895, RRID:AB_306847ChIP-seq (4 ug)
AntibodyRabbit polyclonal anti-HNF6Santa Cruz BiotechnologyCat# sc-13050, RRID:AB_2251852ChIP-seq (4 ug)
AntibodyRabbit polyclonal anti-lamininSigmaCat# L9393, RRID:AB_477163Immunohistochemistry (1:100)
Immunofluorescence (1:30)
AntibodyRabbit monoclonal anti-panCrbJensen Laboratory, University of Massachusetts, AmherstN/AImmunohistochemistry (1:100)
AntibodyRabbit polyclonal anti-PDX1Beta Cell Biology ConsortiumAB1068ChIP-seq (4 ug)
AntibodyRabbit polyclonal anti-SOX9ChemiconCat# 5535, RRID:AB_2239761ChIP-seq (4 ug)
Cell line (Homo-sapiens)CyT49ViaCyte, IncNIHhESC-10–0041,
RRID:CVCL_B850
Male
Cell line (Homo-sapiens)H1WiCell Research
Institute
NIHhESC-10–0043,
RRID:CVCL_9771
Male
Chemical compound, drug2-MercaptoethanolThermo Fisher ScientificCat# 21985023
Chemical compound, drugAccutaseThermo Fisher ScientificCat# 00-4555-56
Chemical compound, drugB-27 supplementThermo Fisher ScientificCat# 17504044
Chemical compound, drugBovine Albumin Fraction VLife TechnologiesCat# 15260037
Chemical compound, drugD-(+)-Glucose Solution, 45%Sigma-AldrichCat# G8769
Chemical compound, drugDAPIInvitrogenCat# D1306Immunohistochemistry (1:200)
Chemical compound, drugDMEM High GlucoseVWRCat# 16750–082
Chemical compound, drugDMEM/F12 [-] L-glutamineVWRCat# 15–090-CV
Chemical compound, drugDMEM/F12 with L-Glutamine, HEPESCorningCat# 45000–350
Chemical compound, drugDMFEMD MilliporeCat# DX1730
Chemical compound, drugDPBSThermo Fisher ScientificCat# 21–031-CV
Chemical compound, drugDTTSigmaCat# D9779
Chemical compound, drugFetal Bovine SerumThermo Fisher ScientificCat# MT35011CV
Chemical compound, drugGlutamaxThermo Fisher ScientificCat# 35050–079
Chemical compound, drugGlutaMAXThermo Fisher ScientificCat# 35050061
Chemical compound, drugHoechst 33342Thermo Fisher ScientificCat# H3570
Chemical compound, drugHyClone Dulbecco’s Modified Eagles MediumThermo Fisher
Scientific
Cat# SH30081.FS
Chemical compound, drugIGEPAL-CA630SigmaCat# I8896
Chemical compound, drugIllumina tagmentation enzymeIlluminaCat# FC-121–1030
Chemical compound, drugInsulin-Transferrin-Selenium (ITS)Thermo Fisher ScientificCat# 41400045
Chemical compound, drugInsulin-Transferrin-Selenium-Ethanolamine (ITS-X)Thermo Fisher
Scientific
Cat# 51500–056
Chemical compound, drugKAAD-CyclopamineToronto Research ChemicalsCat# K171000
Chemical compound, drugK-acetateSigmaCat# P5708
Chemical compound, drugKnockOut SR XenoFreeThermo Fisher ScientificCat# A1099202
Chemical compound, drugLDN-193189StemgentCat# 04–0074
Chemical compound, drugMatrigelCorningCat# 356231
Chemical compound, drugMCDB 131Thermo Fisher
Scientific
Cat# 10372–019
Chemical compound, drugMg-acetateSigmaCat# M2545
Chemical compound, drugmTeSR1 Complete Kit - GMPSTEMCELL
Technologies
Cat# 85850
Chemical compound, drugNEBNext High-Fidelity 2X PCR Master MixNEBCat# M0541
Chemical compound, drugNon-Essential Amino AcidsThermo Fisher ScientificCat# 11140050
Chemical compound, drugO.C.T. CompoundSakura Finetek USACat# 25608–930
Chemical compound, drugPenicillin-StreptomycinThermo Fisher ScientificCat# 15140122
Chemical compound, drugPolyethylenimine (PEI)PolysciencesCat# 23966–1
Chemical compound, drugProtease inhibitorRocheCat# 05056489001
Chemical compound, drugRetinoic acidSigma-AldrichCat# R2625
Chemical compound, drugRNA ScreenTape Sample BufferAgilent TechnologiesCat# 5067–5577
Chemical compound, drugROCK Inhibitor Y-27632STEMCELL
Technologies
Cat# 72305
Chemical compound, drugRPMI 1640 [-] L-glutamineVWRCat# 15–040-CV
Chemical compound, drugSANT-1Sigma-AldrichCat# S4572
Chemical compound, drugSodium BicarbonateSigma-AldrichCat# NC0564699
Chemical compound, drugTamoxifenSigmaCat# T5648
Chemical compound, drugTGF-β RI Kinase Inhibitor IVCalbiochemCat# 616454
Chemical compound, drugTPBCalbiochemCat# 565740
Chemical compound, drugTranylcypromineCayman ChemicalCat# 10010494
Chemical compound, drugTris-acetateThermo Fisher ScientificCat# BP-152
Chemical compound, drugTTNPBEnzo Life SciencesCat# BML-GR105
Chemical compound, drugVectashield Antifade Mounting MediumVector LaboratoriesCat# H-1000
Chemical compound, drugXtremeGene 9RocheCat# 6365787001
Commercial assayHigh Sensitivity D1000 ScreenTapeAgilent TechnologiesCat# 5067–5584
Commercial assay, kitRNA ScreenTapeAgilent TechnologiesCat# 5067–5576
Commercial assay, kitRNA ScreenTape LadderAgilent TechnologiesCat# 5067–5578
Commercial assay, kitBD Cytofix/Cytoperm Plus Fixation/Permeabilization Solution KitBD BiosciencesCat# 554715
Commercial assay, kitChIP-IT High Sensitivity KitActive MotifCat# 53040
Commercial assay, kitiQ SYBR Green SupermixBio-RadCat# 1708884
Commercial assay, kitiScript cDNA Synthesis KitBio-RadCat# 1708891
Commercial assay, kitKAPA Library Preparation Kit (Illumina)Kapa BiosystemsCat# KK8234
Commercial assay, kitKAPA Stranded mRNA-Seq KitsKapa BiosystemsCat# KK8401
Commercial assay, kitMinElute PCR purification kitQIAGENCat# 28004
Commercial assay, kitQubit ssDNA assay kitThermo Fisher ScientificCat# Q10212
Commercial assay, kitRNeasy Micro KitQIAGENCat# 74004
Genetic reagent (D. rerio)Tg(ptf1a:eGFP)jh1PMID:16258076N/A
OtherSPRIselect beadBeckman CoulterCat# B23317
Recombinant proteinActivin AR and D SystemsCat# 338-AC/CF
Recombinant proteinHuman AB SerumValley BiomedicalCat# HP1022
Recombinant proteinRecombinant EGFR and D SystemsCat# 236-EG
Recombinant proteinRecombinant Heregulinβ−1PeprotechCat# 100–03
Recombinant proteinRecombinant KGF/FGF7R and D SystemsCat# 251 KG
Recombinant proteinRecombinant Mouse Wnt3AR and D SystemsCat# 1324-WN/CF
Recombinant proteinRecombinant NogginR and D SystemsCat# 3344 NG
Sequence-based reagentPx333 Plasmidhttp://www.addgene.org/64073/RRID:Addgene_64073
Sequence-based reagentLAMA1 ForwardThis paperqPCR primersGTG ATG GCA ACA GCG CAA A
Sequence-based reagentLAMA1 ReverseThis paperqPCR primersGAC CCA GTG ATA TTC TCT CCC A
Sequence-based reagentCRB2 ForwardThis paperqPCR primersACC ACT GTG CTT GTC CTG AG
Sequence-based reagentCRB2 ReverseThis paperqPCR primersTCC AGG GTC GCT AGA TGG AG
Sequence-based reagentTBP ForwardThis paperqPCR primersTGT GCA CAG GAG CCA AGA GT
Sequence-based reagentTBP ReverseThis paperqPCR primersATT TTC TTG CTG CCA GTC TGG
Sequence-based reagentLAMA1Enh Upstream GuideThis paperCRISPR sgRNAGTC AAA TTG CTA TAA CAC GG
Sequence-based reagentLAMA1Enh Downstream GuideThis paperCRISPR sgRNACCA CTT TAA GTA TCT CAG CA
Sequence-based reagentCRB2Enh Upstream GuideThis paperCRISPR sgRNAATA CAA AGC ACG TGA GA
Sequence-based reagentCRB2Enh Downstream GuideThis paperCRISPR sgRNAGAA TGC GGA TGA CGC CTG AG
Sequence-based reagentlama1-ATGPMID:16321372MorpholinoTCA TCC TCA TCT CCA TCA TCG CTC A
Obtained from GeneTools, LLC
Sequence-based reagentcrb2a-SPPMID:16713951MorpholinoACG TTG CCA GTA CCT GTG TAT CCT G
Obtained from GeneTools, LLC
Sequence-based reagentcrb2b-SPPMID:16713951MorpholinoTAA AGA TGT CCT ACC CAG CTT GAA C
Obtained from GeneTools, LLC
Sequence-based reagentstandard control MON/AMorpholinoCCT CTT ACC TCA GTT ACA ATT TAT A
Obtained from GeneTools, LLC
Software, algorithmAdobe Illustrator v 5.1http://www.adobe.com/products/illustrator.htmlRRID:SCR_014198
Software, algorithmAdobe Photoshop v 5.1http://www.adobe.com/products/photoshop.htmlRRID:SCR_014199
Software, algorithmBEDtools v 2.26.0https://github.com/arq5x/bedtools2RRID:SCR_006646
Software, algorithmBioconductorhttps://www.bioconductor.org/RRID: SCR_006442
Software, algorithmBurrows-Wheeler Aligner v 0.7.13http://bio-bwa.sourceforge.net/RRID:SCR_010910
Software, algorithmCENTIPEDE v 1.2http://centipede.uchicago.edu/N/A
Software, algorithmCufflinks v 2.2.1http://cole-trapnell-lab.github.io/cufflinks/RRID:SCR_014597
Software, algorithmdeepTools2 v 3.1.3https://deeptools.readthedocs.io/en/develop/content/installation.htmlN/A
Software, algorithmDESeq2 v 3.10https://bioconductor.org/packages/release/bioc/html/DESeq2.htmlRRID:SCR_015687
Software, algorithmFlowJo v10 softwarehttps://www.flowjo.com/solutions/flowjoRRID: SCR_008520
Software, algorithmGraphPad Prism v 8.1.2https://www.graphpad.com/scientific-software/prism/RRID: SCR_002798
Software, algorithmHOMER v 4.10.4http://homer.ucsd.edu/homer/RRID: SCR_010881
Software, algorithmJuicebox Tools v 1.4https://github.com/aidenlab/Juicebox/wiki/Juicebox-Assembly-ToolsN/A
Software, algorithmMACS2 v 2.1.4http://liulab.dfci.harvard.edu/MACS/RRID:SCR_013291
Software, algorithmMEME suite v 5.1.1http://meme-suite.org/RRID:SCR_001783
Software, algorithmMetascapehttp://metscape.ncibi.orgRRID:SCR_014687
Software, algorithmPicard Tools v 1.131http://broadinstitute.github.io/picard/RRID:SCR_006525
Software, algorithmR Project for Statistical Computing v 3.6.1http://www.r-project.org/RRID:SCR_001905
Software, algorithmSAMtools v 1.5http://samtools.sourceforge.netRRID:SCR_002105
Software, algorithmSTAR v 2.4https://github.com/alexdobin/STARN/A
Software, algorithmUCSC Genome Browserhttp://genome.ucsc.edu/RRID:SCR_005780
Software, algorithmvcf2diploid v 0.2.6ahttps://github.com/abyzovlab/vcf2diploidN/A
Software, algorithmZEISS ZEN Digital Imaging for Light Microscopyhttp://www.zeiss.com/microscopy/en_us/products/microscope-software/zen.html#introductionRRID:SCR_013672

Maintenance and differentiation of CyT49 hESCs

Request a detailed protocol

Genomic and gene expression analyses (ChIP-seq, ATAC-seq, RNA-seq) for generation of chromatin maps and target gene identification were performed in CyT49 hESCs (male). Propagation of CyT49 hESCs was carried out by passing cells every 3 to 4 days using Accutase (eBioscience) for enzymatic cell dissociation, and with 10% (v/v) human AB serum (Valley Biomedical) included in the hESC media the day of passage. hESCs were seeded into tissue culture flasks at a density of 50,000 cells/cm2. hESC research was approved by the University of California, San Diego, Institutional Review Board and Embryonic Stem Cell Research oversight committee.

Pancreatic differentiation was performed as previously described (Schulz et al., 2012; Wang et al., 2015; Xie et al., 2013). Briefly, a suspension-based culture format was used to differentiate cells in aggregate form. Undifferentiated aggregates of hESCs were formed by re-suspending dissociated cells in hESC maintenance medium at a concentration of 1 × 106 cells/mL and plating 5.5 mL per well of the cell suspension in 6-well ultra-low attachment plates (Costar). The cells were cultured overnight on an orbital rotator (Innova2000, New Brunswick Scientific) at 95 rpm. After 24 hr the undifferentiated aggregates were washed once with RPMI medium and supplied with 5.5 mL of day 0 differentiation medium. Thereafter, cells were supplied with the fresh medium for the appropriate day of differentiation (see below). Cells were continually rotated at 95 rpm, or 105 rpm on days 4 through 8, and no media change was performed on day 10. Both RPMI (Mediatech) and DMEM High Glucose (HyClone) medium were supplemented with 1X GlutaMAX and 1% penicillin/streptomycin. Human activin A, mouse Wnt3a, human KGF, human noggin, and human EGF were purchased from R and D systems. Other added components included FBS (HyClone), B-27 supplement (Life Technologies), Insulin-Transferrin-Selenium (ITS; Life Technologies), TGFβ R1 kinase inhibitor IV (EMD Bioscience), KAAD-Cyclopamine (KC; Toronto Research Chemicals), and the retinoic receptor agonist TTNPB (RA; Sigma Aldrich). Day-specific differentiation media formulations were as follows:

  • Days 0 and 1: RPMI + 0.2% (v/v) FBS, 100 ng/mL Activin, 50 ng/mL mouse Wnt3a, 1:5000 ITS. Days 1 and 2: RPMI + 0.2% (v/v) FBS, 100 ng/mL Activin, 1:5000 ITS

  • Days 2 and 3: RPMI + 0.2% (v/v) FBS, 2.5 mM TGFβ R1 kinase inhibitor IV, 25 ng/mL KGF, 1:1000 ITS

  • Days 3–5: RPMI + 0.2% (v/v) FBS, 25 ng/mL KGF, 1:1000 ITS

  • Days 5–8: DMEM + 0.5X B-27 Supplement, 3 nM TTNPB, 0.25 mM KAAD-Cyclopamine, 50 ng/mL Noggin

  • Days 8–10: DMEM/B-27, 50 ng/mL KGF, 50 ng/mL EGF

Cells at D0 correspond to the embryonic stem cell (ES) stage, cells at D2 correspond to the definitive endoderm (DE) stage, cells at D5 correspond to the gut tube (GT) stage, cells at D7 correspond to the early pancreatic progenitor (PP1) stage, and cells at D10 correspond to the late pancreatic progenitor (PP2) stage.

Maintenance and differentiation of H1 hESCs

Request a detailed protocol

∆LAMA1Enh and ∆CRB2Enh clonal lines were derived by targeting H1 hESCs (male). Cells were maintained and differentiated as described with some modifications (Jin et al., 2019; Rezania et al., 2014). In brief, hESCs were cultured in mTeSR1 media (Stem Cell Technologies) and propagated by passaging cells every 3–4 days using Accutase (eBioscience) for enzymatic cell dissociation. hESC research was approved by the University of California, San Diego, Institutional Review Board and Embryonic Stem Cell Research Oversight Committee.

For differentiation, cells were dissociated using Accutase for 10 min, then reaggregated by plating the cells at a concentration of ~5.5 e6 cells/well in a low attachment six-well plate on an orbital shaker (100 rpm) in a 37°C incubator. The following day, undifferentiated cells were washed in base media (see below) and then differentiated using a multi-step protocol with stage-specific media and daily media changes.

All stage-specific base media were comprised of MCDB 131 medium (Thermo Fisher Scientific) supplemented with NaHCO3, GlutaMAX, D-Glucose, and BSA using the following concentrations:

  • Stage 1/2 medium: MCDB 131 medium, 1.5 g/L NaHCO3, 1X GlutaMAX, 10 mM D-Glucose, 0.5% BSA

  • Stage 3/4 medium: MCDB 131 medium, 2.5 g/L NaHCO3, 1X GlutaMAX, 10 mM D-glucose, 2% BSA

Media compositions for each stage were as follows:

  • Stage 1 (day 0–2): base medium, 100 ng/ml Activin A, 25 ng/ml Wnt3a (day 0). Day 1–2: base medium, 100 ng/ml Activin A

  • Stage 2 (day 3–5): base medium, 0.25 mM L-Ascorbic Acid (Vitamin C), 50 ng/mL FGF7

  • Stage 3 (day 6–7): base medium, 0.25 mM L-Ascorbic Acid, 50 ng/mL FGF7, 0.25 µM SANT-1, 1 µM Retinoic Acid, 100 nM LDN193189, 1:200 ITS-X, 200 nM TPB

  • Stage 4 (day 8–10): base medium, 0.25 mM L-Ascorbic Acid, 2 ng/mL FGF7, 0.25 µM SANT-1, 0.1 µM Retinoic Acid, 200 nM LDN193189, 1:200 ITS-X, 100 nM TPB

Cells at D0 correspond to the embryonic stem cell (ES) stage, cells at D3 correspond to the definitive endoderm (DE) stage, cells at D6 correspond to the gut tube (GT) stage, cells at D8 correspond to the early pancreatic progenitor (PP1) stage, and cells at D11 correspond to the late pancreatic progenitor (PP2) stage.

Generation of ∆LAMA1Enh, ∆CRB2Enh, ∆LAMA1, and ∆CRB2 hESC lines

Request a detailed protocol

To generate clonal homozygous LAMA1Enh and CRB2Enh deletion hESC lines, sgRNAs targeting each relevant enhancer were designed and cloned into Px333-GFP, a modified version of Px333 (Addgene, #64073). To generate clonal homozygous LAMA1 and CRB2 deletion hESC lines, sgRNAs targeting the second exon of each gene were designed and cloned into Px458 (Addgene, #48138). Plasmids expressing the sgRNAs were transfected into H1 hESCs with XtremeGene 9 (Roche). Twenty-four hr later, 8000 GFP+ cells were sorted into a well of six-well plate. Individual colonies that emerged within 5–7 days after transfection were subsequently transferred manually into 48-well plates for expansion, genomic DNA extraction, PCR genotyping, and Sanger sequencing. sgRNA oligos are listed below.

  • LAMA1Enh Upstream Guide: GTCAAATTGCTATAACACGG

  • LAMA1Enh Downstream Guide: CCACTTTAAGTATCTCAGCA

  • CRB2Enh Upstream Guide: ATACAAAGCACGTGAGA

  • CRB2Enh Downstream Guide: GAATGCGGATGACGCCTGAG

  • LAMA1 Exon 2 Guide: ATCAGCACCAATGCCACCTG

  • CRB2 Exon 2 Guide: TCGATGTCCAGCTCGCAGCG

Human tissue

Request a detailed protocol

Human embryonic pancreas tissue was obtained from the Birth Defects Research Laboratory of the University of Washington. Studies for use of embryonic human tissue were approved by the Institutional Review Board of the University of California, San Diego. A pancreas from a 54- and 58-day gestation embryo each were pooled for RNA-seq analysis.

Zebrafish husbandry

Request a detailed protocol

Adult zebrafish and embryos were cared for and maintained under standard conditions. All research activity involving zebrafish was reviewed and approved by SBP Medical Discovery Institute Institutional Animal Care and Use Committee. The following transgenic lines were used: Tg(ptf1a:eGFP)jh1 (Godinho et al., 2005).

Morpholino injections in zebrafish

Request a detailed protocol

The following previously validated morpholinos were injected into the yolk at the one-cell stage in a final volume of either 0.5 or 1 nl: 0.75 ng lama1-ATG (5’- TCATCCT CATCTCCATCATCGCTCA −3’); 3 ng crb2a-SP, (5’-ACGTTGCCAGTACCTGTGTATCCTG-3’) (Omori and Malicki, 2006; Watanabe et al., 2010); 3 ng crb2b-SP, (5’-TAAAGATGTCCTACCCAGCTTGAAC-3’) (Omori and Malicki, 2006); 6.75 ng standard control MO (5’- CCTCTTACCTCAGTTACAATTTATA −3’). All morpholinos were obtained from GeneTools, LLC.

Chromatin immunoprecipitation sequencing (ChIP-seq)

Request a detailed protocol

ChIP-seq was performed using the ChIP-IT High-Sensitivity kit (Active Motif) according to the manufacturer’s instructions. Briefly, for each cell stage and condition analyzed, 5–10 × 106 cells were harvested and fixed for 15 min in an 11.1% formaldehyde solution. Cells were lysed and homogenized using a Dounce homogenizer and the lysate was sonicated in a Bioruptor Plus (Diagenode), on high for 3 × 5 min (30 s on, 30 s off). Between 10 and 30 µg of the resulting sheared chromatin was used for each immunoprecipitation. Equal quantities of sheared chromatin from each sample were used for immunoprecipitations carried out at the same time. A total of 4 µg of antibody were used for each ChIP-seq assay. Chromatin was incubated with primary antibodies overnight at 4°C on a rotator followed by incubation with Protein G agarose beads for 3 hr at 4°C on a rotator. Antibodies used were rabbit anti-H3K27ac (Active Motif 39133), rabbit anti-H3K4me1 (Abcam ab8895), rabbit anti-H3K4me3 (Millipore 04–745), rabbit anti-H3K27me3 (Millipore 07–499), goat anti-CTCF (Santa Cruz Biotechnology SC-15914X), goat anti-GATA4 (Santa Cruz SC-1237), rabbit anti-GATA6 (Santa Cruz SC-9055), goat anti-FOXA1 (Abcam Ab5089), goat-anti-FOXA2 (Santa Cruz SC-6554), rabbit anti-PDX1 (BCBC AB1068), rabbit anti-HNF6 (Santa Cruz SC-13050), and rabbit anti-SOX9 (Chemicon AB5535). Reversal of crosslinks and DNA purification were performed according to the ChIP-IT High-Sensitivity instructions, with the modification of incubation at 65°C for 2–3 hr, rather than at 80°C for 2 hr. Sequencing libraries were constructed using KAPA DNA Library Preparation Kits for Illumina (Kapa Biosystems) and library sequencing was performed on either a HiSeq 4000 System (Illumina) or NovaSeq 6000 System (Illumina) with single-end reads of either 50 or 75 base pairs (bp). Sequencing was performed by the Institute for Genomic Medicine (IGM) core research facility at the University of California at San Diego (UCSD). Two replicates from independent hESC differentiations were generated for each ChIP-seq experiment.

ChIP-seq data analysis

Request a detailed protocol

ChIP-seq reads were mapped to the human genome consensus build (hg19/GRCh37) and visualized using the UCSC Genome Browser (Kent et al., 2002). Burrows-Wheeler Aligner (BWA) (Li and Durbin, 2009) version 0.7.13 was used to map data to the genome. Unmapped and low-quality (q < 15) reads were discarded. SAMtools (Li et al., 2009) was used to remove duplicate sequences and HOMER (Heinz et al., 2010) was used to call peaks using default parameters and to generate tag density plots. Stage- and condition-matched input DNA controls were used as background when calling peaks. The BEDtools suite of programs (Quinlan and Hall, 2010) was used to perform genomic algebra operations. For all ChIP-seq experiments, replicates from two independent hESC differentiations were generated. Tag directories were created for each replicate using HOMER. Directories from each replicate were then combined, and peaks were called from the combined replicates. For histone modifications and CTCF peaks, pearson correlations between each pair of replicates were calculated over the called peaks using the command multiBamSummary from the deepTools2 package (Ramírez et al., 2016). For pancreatic lineage-determining transcription factors (GATA4, GATA6, FOXA1, FOXA2, HNF6, PDX1, SOX9), correlations were calculated for peaks overlapping PSSE. Calculated Pearson correlations are as follow:

H3K4me1H3K27acCTCFH3K4me3H3K27me3
ES0.900.910.870.811.00
DE0.970.840.860.990.99
GT0.970.870.890.970.99
PP10.970.850.890.960.99
PP20.980.870.870.971.00
GATA4GATA6FOXA1FOXA2HNF6PDX1SOX9
PP20.860.820.870.800.950.640.86

RNA isolation and sequencing (RNA-seq) and qRT-PCR

Request a detailed protocol

RNA was isolated from cell samples using the RNeasy Micro Kit (Qiagen) according to the manufacturer instructions. For each cell stage and condition analyzed between 0.1 and 1 × 106 cells were collected for RNA extraction. For qRT-PCR, cDNA synthesis was first performed using the iScript cDNA Synthesis Kit (Bio-Rad) and 500 ng of isolated RNA per reaction. qRT-PCR reactions were performed in triplicate with 10 ng of template cDNA per reaction using a CFX96 Real-Time PCR Detection System and the iQ SYBR Green Supermix (Bio-Rad). PCR of the TATA binding protein (TBP) coding sequence was used as an internal control and relative expression was quantified via double delta CT analysis. For RNA-seq, stranded, single-end sequencing libraries were constructed from isolated RNA using the TruSeq Stranded mRNA Library Prep Kit (Illumina) and library sequencing was performed on either a HiSeq 4000 System (Illumina) or NovaSeq 6000 System (Illumina) with single-end reads of either 50 or 75 base pairs (bp). Sequencing was performed by the Institute for Genomic Medicine (IGM) core research facility at the University of California at San Diego. A complete list of RT-qPCR primer sequences can be found below.

LAMA1 forwardGTG ATG GCA ACA GCG CAA A
LAMA1 reverseGAC CCA GTG ATA TTC TCT CCC A
CRB2 forwardACC ACT GTG CTT GTC CTG AG
CRB2 reverseTCC AGG GTC GCT AGA TGG AG
PDX1 forwardAAG TCT ACC AAA GCT CAC GCG
PDX1 reverseGTA GGC GCC GCC TGC
NKX6.1 forwardCTG GCC TGT ACC CCT CAT CA
NKX6.1 reverseCTT CCC GTC TTT GTC CAA CA
PROX1 forwardAAC ATG CAC TAC AAT AAA GCA AAT GAC
PROX1 reverseCAG GAA TCT CTC TGG AAC CTC AAA
PTF1A forwardGAA GGT CAT CAT CTG CCA TC
PTF1A reverseGGC CAT AAT CAG GGT CGC T
SOX9 forwardAGT ACC CGC ACT TGC ACA AC
SOX9 reverseACT TGT AAT CCG GGT GGT CCT T
TBP forwardTGT GCA CAG GAG CCA AGA GT
TBP reverseATT TTC TTG CTG CCA GTC TGG

RNA-seq data analysis

Request a detailed protocol

Reads were mapped to the human genome consensus build (hg19/GRCh37) using the Spliced Transcripts Alignment to a Reference (STAR) aligner v2.4 (Dobin et al., 2013). Normalized gene expression (fragments per kilobase per million mapped reads; FPKM) for each sequence file was determined using Cufflinks v2.2.1 (Trapnell et al., 2010) with the parameters: --library-type fr-firststrand --max-bundle-frags 10000000. For all RNA-Seq experiments, replicates from two independent hESC differentiations were generated. Pearson correlations between bam files corresponding to each pair of replicates were calculated, and are as follow:

LAMA1Enh clone 1 PP21.00
LAMA1Enh clone 2 PP20.99
CRB2Enh clone 1 PP20.98
CRB2Enh clone 2 PP20.90
LAMA1Enh control PP20.92
CRB2Enh control PP20.99

Assay for transposase accessible chromatin sequencing (ATAC-seq)

Request a detailed protocol

ATAC-seq (Buenrostro et al., 2013) was performed on approximately 50,000 nuclei. The samples were permeabilized in cold permeabilization buffer 0.2% IGEPAL-CA630 (I8896, Sigma), 1 mM DTT (D9779, Sigma), Protease inhibitor (05056489001, Roche), 5% BSA (A7906, Sigma) in PBS (10010–23, Thermo Fisher Scientific) for 10 min on the rotator in the cold room and centrifuged for 5 min at 500 × g at 4°C. The pellet was resuspended in cold tagmentation buffer (33 mM Tris-acetate (pH = 7.8) (BP-152, Thermo Fisher Scientific), 66 mM K-acetate (P5708, Sigma), 11 mM Mg-acetate (M2545, Sigma), 16% DMF (DX1730, EMD Millipore) in Molecular biology water (46000 CM, Corning)) and incubated with tagmentation enzyme (FC-121–1030; Illumina) at 37°C for 30 min with shaking at 500 rpm. The tagmented DNA was purified using MinElute PCR purification kit (28004, QIAGEN). Libraries were amplified using NEBNext High-Fidelity 2X PCR Master Mix (M0541, NEB) with primer extension at 72°C for 5 min, denaturation at 98°C for 30 s, followed by 8 cycles of denaturation at 98°C for 10 s, annealing at 63°C for 30 s and extension at 72°C for 60 s. After the purification of amplified libraries using MinElute PCR purification kit (28004, QIAGEN), double size selection was performed using SPRIselect bead (B23317, Beckman Coulter) with 0.55X beads and 1.5X to sample volume. Finally, libraries were sequenced on HiSeq4000 (Paired-end 50 cycles, Illumina).

ATAC-seq data analysis

Request a detailed protocol

ATAC-seq reads were mapped to the human genome (hg19/GRCh37) using Burrows-Wheeler Aligner (BWA) version 0.7.13 (Li and Durbin, 2009), and visualized using the UCSC Genome Browser (Kent et al., 2002). SAMtools (Li et al., 2009) was used to remove unmapped, low-quality (q < 15), and duplicate reads. MACS2 (Zhang et al., 2008) was used to call peaks, with parameters ‘shift set to 100 bps, smoothing window of 200 bps’ and with ‘nolambda’ and ‘nomodel’ flags on. MACS2 was also used to call ATAC-Seq summits, using the same parameters combined with the ‘call-summits’ flag.

For all ATAC-Seq experiments, replicates from two independent hESC differentiations were generated. Bam files for each pair of replicates were merged for downstream analysis using SAMtools, and Pearson correlations between bam files for each individual replicate were calculated over a set of peaks called from the merged bam file. Correlations were performed using the command multiBamSummary from the deepTools2 package (Ramírez et al., 2016) with the ‘—removeOutliers’ flag and are as follow:

ES0.95
DE0.83
GT1.00
PP11.00
PP21.00

For downstream analysis, ATAC-seq peaks were merged from two independent differentiations for ES, DE, GT, PP1, and PP2 stage cells and from four donors for primary islets. Primary islet ATAC-seq data was obtained from previously published datasets (Greenwald et al., 2019).

Hi-C data analysis

Request a detailed protocol

Hi-C data were processed as previously described with some modifications (Dixon et al., 2015). Read pairs were aligned to the hg19 reference genome separately using BWA-MEM with default parameters (Li and Durbin, 2009). Specifically, chimeric reads were processed to keep only the 5’ position and reads with low mapping quality (<10) were filtered out. Read pairs were then paired using custom scripts. Picard tools were then used to remove PCR duplicates. Bam files with alignments were further processed into text format as required by Juicebox tools (Durand et al., 2016). Juicebox tools were then applied to generate hic files containing normalized contact matrices. All downstream analysis was based on 10 Kb resolution KR normalized matrices.

Chromatin loops were identified by comparing each pixel with its local background, as described previously (Rao et al., 2014) with some modifications. Specifically, only the donut region around the pixel was compared to model the expected count. Briefly, the KR-normalized contact matrices at 10 Kb resolution were used as input for loop calling. For each pixel, distance-corrected contact frequencies were calculated for each surrounding bin and the average of all surrounding bins. The expected counts were then transformed to raw counts by multiplying the counts with the raw-to-KR normalization factor. The probability of observing raw expected counts was calculated using Poisson distribution. All pixels with p-value<0.01 and distance less than 10 Kb were selected as candidate pixels. Candidate pixels were then filtered to remove pixels without any neighboring candidate pixels since they were likely false positives. Finally, pixels within 20 Kb of each other were collapsed and only the most significant pixel was selected. The collapsed pixels with p-value<1e-5 were used as the final list of chromatin loops.

A full set of scripts used for processing Hi-C data (Qiu, 2021) is available at https://github.com/MSanderlab/Pancreatic-progenitor-epigenome-maps-prioritize-type-2-diabetes-risk-genes-with-roles-in-development/tree/master (copy archived at swh:1:rev:ba79c687523c2696ea0ef30d8476e28a0d860f18).

Definition of chromatin states

Request a detailed protocol

We collected or generated H3K4me1, H3K27ac, H3K4me1, H3K4me3, H3K27me3, and CTCF ChIP-seq data at each developmental stage and in mature islets. Data corresponding to mature islets was downloaded from previously published studies (Bhandare et al., 2010; Parker et al., 2013; Pasquali et al., 2014). Sequence reads were mapped to the human genome hg19 using bwa (version 0.7.12) (Li and Durbin, 2009), and low quality and duplicate reads were filtered using samtools (version 1.3) (Li et al., 2009). Using these reads, we then called chromatin states jointly across all data using chromHMM (version 1.12) (Ernst and Kellis, 2012) and used a 10-state model and 200 bp bin size, as models with larger state numbers did not empirically resolve any additional informative states. We then assigned state names based on patterns defined by the NIH Epigenome Roadmap (Kundaje et al., 2015), which included active promoter/TssA (high H3K4me3, high H3K27ac), flanking TSS/TssFlnk1 (high H3K4me3), flanking TSS/TssFlnk 2 (high H3K4me3, high H3K27ac, high H3K4me1), bivalent Tss/TssBiv (high H3K27me3, high H3K4me3), poised enhancer/EnhP (high H3K4me1), insulator/CTCF (high CTCF), active enhancer/EnhA (high H3K27ac, high H3K4me1), repressor (high H3K27me3), and two quiescent (low signal for all assays) states. The state map with assigned names is shown in Figure 1—figure supplement 1A.

We next defined stretch enhancer elements at each developmental stage and in mature islets. For each active enhancer (EnhA) element, we determined the number of consecutive 200 bp bins covered by the enhancer. We then modeled the resulting bin counts for enhancers in each cell type using a Poisson distribution. Enhancers with a p-value less than. 001 were labeled as stretch enhancers and otherwise labeled as traditional enhancers.

Permutation-based significance

Request a detailed protocol

A random sampling approach (10,000 iterations) was used to obtain null distributions for enrichment analyses, in order to obtain p-values. Null distributions for enrichments were obtained by randomly shuffling enhancer regions using BEDTools (Quinlan and Hall, 2010) and overlapping with ATAC-seq peaks. p-values<0.05 were considered significant.

Assignment of enhancer target genes

Request a detailed protocol

Transcriptomes were filtered for genes expressed (FPKM ≥1) at each relevant stage, and BEDTools (Quinlan and Hall, 2010) was used to assign each enhancer to the nearest annotated TSS.

Gene ontology

Request a detailed protocol

All gene ontology analyses were performed using Metascape (Zhou et al., 2019) with default parameters.

Motif enrichment analysis

Request a detailed protocol

The findMotifsGenome.pl. command in HOMER (Heinz et al., 2010) was used to identify enriched transcription factor binding motifs. de novo motifs were assigned to transcription factors based on suggestions generated by HOMER.

T2D-relevant trait enrichment analysis

Request a detailed protocol

GWAS summary statistics for T2D (Mahajan et al., 2018; Perry et al., 2012), metabolic traits (HOMA-B, HOMA-IR [Dupuis et al., 2010], fasting glucose, fasting insulin [Manning et al., 2012], fasting proinsulin [Strawbridge et al., 2011], 2 hr glucose adjusted for BMI [Saxena et al., 2010], HbA1c, insulin secretion rate, disposition index, acute insulin response, peak insulin response [Wood et al., 2017]), and developmental traits (head circumference [Taal et al., 2012], birth length [van der Valk et al., 2015], birth weight [Horikoshi et al., 2016]) conducted with individuals of European ancestry were obtained from various sources including the MAGIC consortium, EGG consortium, and authors of the studies. Custom LD score annotation files were created for PSSE, PP2 stretch enhancers, and islet stretch enhancers using LD score regression version 1.0.1 (Bulik-Sullivan et al., 2015). Enrichments for GWAS trait-associated variants within PSSE, PP2 stretch enhancers, and islet stretch enhancers were estimated with stratified LD score regression (Finucane et al., 2015). We next determined enrichment in the proportion of variants in accessible chromatin sites within islet SE and PSSE with nominal association to beta cell-related glycemic traits. For each trait, we calculated a 2 × 2 table of variants mapping in and outside of islet SE or PSSE and with or without nominal association and then determined significance using a chi-square test.

Adipocyte differentiation analysis

Request a detailed protocol

Chromatin states for human adipose stromal cell (hASC) differentiation stages (1-4) were obtained from a published study (Varshney et al., 2017). PSSE were intersected with hASC chromatin states using BEDTools intersect (version 2.26.0) (Quinlan and Hall, 2010) with default parameters.

Identification of T2D risk loci intersecting PSSE

Request a detailed protocol

T2D GWAS summary statistics were obtained from the DIAMANTE consortium (Mahajan et al., 2018). Intersection of variants and PSSE was performed using BEDTools intersect (version 2.26.0) (Quinlan and Hall, 2010) with default parameters. The adjusted significance threshold was set at p<4.66 × 10−6 (Bonferroni correction for 10,738 variants mapping in PSSE). Putative novel loci were defined as those with (1) at least one variant in a PSSE reaching the adjusted significance threshold and (2) mapping at least 500 kb away from a known T2D locus.

ATAC-seq footprinting analysis

Request a detailed protocol

ATAC-seq footprinting was performed as previously described (Aylward et al., 2018). In brief, diploid genomes for CyT49 were created using vcf2diploid (version 0.2.6a) (Rozowsky et al., 2011) and genotypes called from whole genome sequencing and scanned for a compiled database of TF sequence motifs from JASPAR (Mathelier et al., 2016) and ENCODE (ENCODE Project Consortium, 2012) with FIMO (version 4.12.0) (Grant et al., 2011) using default parameters for p-value threshold and a 40.9% GC content based on the hg19 human reference genome. Footprints within ATAC-seq peaks were discovered with CENTIPEDE (version 1.2) (Pique-Regi et al., 2011) using cut-site matrices containing Tn5 integration counts within a ± 100 bp window around each motif occurrence. Footprints were defined as those with a posterior probability ≥0.99.

Generation of similarity matrices for total transcriptomes

Request a detailed protocol

For each replicate, FPKM values corresponding to total transcriptome were filtered for genes expressed (FPKM ≥1) in ≥1 replicate. For expressed genes, log(FPKM+1) values were used to calculate Pearson correlations.

Immunofluorescence analysis

Request a detailed protocol

Cell aggregates derived from hESCs were allowed to settle in microcentrifuge tubes and washed twice with PBS before fixation with 4% paraformaldehyde (PFA) for 30 min at room temperature. Fixed samples were washed twice with PBS and incubated overnight at 4°C in 30% (w/v) sucrose in PBS. Samples were then loaded into disposable embedding molds (VWR), covered in Tissue-Tek O.C.T. Sakura Finetek compound (VWR) and flash frozen on dry ice to prepare frozen blocks. The blocks were sectioned at 10 µm and sections were placed on Superfrost Plus (Thermo Fisher) microscope slides and washed with PBS for 10 min. Slide-mounted cell sections were permeabilized and blocked with blocking buffer, consisting of 0.15% (v/v) Triton X-100 (Sigma) and 1% (v/v) normal donkey serum (Jackson Immuno Research Laboratories) in PBS, for 1 hr at room temperature. Slides were then incubated overnight at 4°C with primary antibody solutions. The following day slides were washed five times with PBS and incubated for 1 hr at room temperature with secondary antibody solutions. Cells were washed five times with PBS before coverslips were applied.

All antibodies were diluted in blocking buffer at the ratios indicated below. Primary antibodies used were goat anti-PDX1 (1:500 dilution, Abcam ab47383), mouse anti-NKX6.1 (1:300 dilution, Developmental Studies Hybridoma Bank F64A6B4), rabbit anti-Laminin (1:30, Sigma L-9393), and rabbit anti-CRB2 (1:500, Sigma SAB1301340). Secondary antibodies against goat and mouse were Alexa488- and Cy3-conjugated donkey antibodies, respectively (Jackson Immuno Research Laboratories 705-545-003 and 715-165-150, respectively), and were used at dilutions of 1:500 (anti-goat Alexa488) or 1:1000 (anti-mouse Cy3). Cell nuclei were stained with Hoechst 33342 (1:3000, Invitrogen). Representative images were obtained with a Zeiss Axio-Observer-Z1 microscope equipped with a Zeiss ApoTome and AxioCam digital camera. Figures were prepared in Adobe Creative Suite 5.

Flow cytometry analysis

Request a detailed protocol

Cell aggregates derived from hESCs were allowed to settle in microcentrifuge tubes and washed with PBS. Cell aggregates were incubated with Accutase at room temperature until a single-cell suspension was obtained. Cells were washed with 1 mL ice-cold flow buffer comprised of 0.2% BSA in PBS and centrifuged at 200 g for 5 min. BD Cytofix/Cytoperm Plus Fixation/Permeabilization Solution Kit was used to fix and stain cells for flow cytometry according to the manufacturer's instructions. Briefly, cell pellets were re-suspended in ice-cold BD Fixation/Permeabilization solution (300 µL per microcentrifuge tube). Cells were incubated for 20 min at 4°C. Cells were washed twice with 1 mL ice-cold 1 × BD Perm/Wash Buffer and centrifuged at 10°C and 200 × g for 5 min. Cells were re-suspended in 50 µL ice-cold 1 × BD Perm/Wash Buffer containing diluted antibodies, for each staining performed. Cells were incubated at 4°C in the dark for 1–3 hr. Cells were washed with 1.25 mL ice-cold 1X BD Wash Buffer and centrifuged at 200 × g for 5 min. Cell pellets were re-suspended in 300 µL ice-cold flow buffer and analyzed in a FACSCanto II (BD Biosciences). Antibodies used were PE-conjugated anti-PDX1 (1:10 dilution, BD Biosciences); and AlexaFluor 647-conjugated anti-NKX6.1 (1:5 dilution, BD Biosciences). Data were processed using FlowJo software v10.

Whole mount immunohistochemistry

Request a detailed protocol

Zebrafish larvae were fixed and stained according to published protocols (Lancman et al., 2013) using the following antibodies: chicken anti-GFP (1:200; Aves Labs; GFP-1020), guinea pig anti-insulin (1:200; Biomeda; v2024), mouse anti-Crb2a (1:100; ZIRC; zs-4), rabbit anti-panCrb (1:100; provided by Dr. Abbie M. Jensen at University of Massachusetts, Amherst; Hsu and Jensen, 2010), rabbit anti-Laminin (1:100; Sigma;L9393), mouse anti-Nkx6.1 (1:10; DSHB; F55A10), and DAPI (1:200; 500 mg/ml, Invitrogen; D1306).

Imaging and quantification of beta cell numbers in zebrafish

Request a detailed protocol

To quantify beta cell numbers, 50 and 78 hpf zebrafish larvae were stained for confocal imaging using DAPI and guinea pig anti-insulin antibody (1:200; Biomeda; v2024). Whole mount fluorescent confocal Z-stacks (0.9 μm steps) images were collected for the entire islet with optical slices captured at a focal depth of 1.8 μm. Samples were imaged using a Zeiss 710 confocal microscope running Zen 2010 (Black) software. Final images were generated using Adobe Photoshop CS6 and/or ImageJ64 (vs.1.48b).

Data sources

Request a detailed protocol

The following datasets used in this study were downloaded from the GEO and ArrayExpress repositories:

RNA-seq: Pancreatic differentiation of CyT49 hESC line (E-MTAB-1086); primary islet data (GSE115327).

ChIP-seq: H3K27ac data in primary islets (E-MTAB-1919 and GSE51311); H3K27ac data in pancreatic differentiation of CyT49 hESC line (GSE54471); H3K4me1 data in pancreatic differentiation of CyT49 hESC line (GSE54471); H3K4me1 data in primary islets (E-MTAB-1919 and E-MTAB 189); H3K27me3 and H3K4me3 in pancreatic differentiation of CyT49 hESC line (E-MTAB-1086); H3K4me3 and H3K27me3 in primary islets (E-MTAB-189); CTCF in primary islets (E-MTAB-1919); PDX1 in CyT49 PP2 (GSE54471); samples from ROADMAP consortium: http://ncbi.nlm.nih.gov/geo/roadmap/epigenomics.

ATAC-seq: primary islet data (PRJN527099); CyT49 PP2 (GSE115327).

Hi-C datasets were generated in collaboration with the Ren laboratory at University of California, San Diego as a component of the 4D Nucleome Project (Dekker et al., 2017) under accession number 4DNES0LVRKBM.

Quantification and statistical analyses

Request a detailed protocol

Statistical analyses were performed using GraphPad Prism (v8.1.2), and R (v3.6.1). Statistical parameters, such as the value of n, mean, standard deviation (SD), standard error of the mean (SEM), significance level (*p<0.05, **p<0.01, and ***p<0.001), and the statistical tests used, are reported in the figures and figure legends. The ‘n’ refers to the number of independent pancreatic differentiation experiments analyzed (biological replicates).

Statistically significant gene expression changes were determined with DESeq2 (Love et al., 2014).

Data availability

All mRNA-seq, ChIP-seq, and ATAC-seq datasets generated for this study have been deposited at GEO under the accession number GSE149148. Source data files have been provided for Figures 2, 4, and 5.

The following data sets were generated
    1. Geusz RJ
    2. Wang A
    3. Chiou J
    4. Lancman JJ
    5. Wetton N
    6. Kefalopoulou S
    7. Wang J
    8. Qiu Y
    9. Yan J
    10. Aylward A
    11. Ren B
    12. Si Dong PD
    13. Gaulton KJ
    14. Sander M
    (2021) NCBI Gene Expression Omnibus
    ID GSE149148. Pancreatic progenitor epigenome maps prioritize type 2 diabetes risk genes with roles in development.
The following previously published data sets were used
    1. Xie R
    2. Everett LJ
    3. Lim HW
    4. Patel NA
    5. Schug J
    6. Kroon E
    7. Kelly OG
    8. Wang A
    9. D'Amour KA
    10. Robins AJ
    11. Won KJ
    12. Kaestner KH
    13. Sander M
    (2013) ArrayExpress
    ID E-MTAB-1086. ChIP-seq and RNA-seq of coding RNA of the progression of human embryonic stem cells to beta cells to characterize the epigenetic programs that underlie pancreas differentiation.
    1. Jin W
    2. Mulas F
    3. Gaertner B
    4. Sui Y
    5. Wang J
    6. Matta I
    7. Zeng C
    8. Vinckier N
    9. Wang A
    10. Nguyen-Ngoc K
    11. Chiou J
    12. Kaestner KH
    13. Frazer KA
    14. Carrano AC
    15. Shih H
    (2019) NCBI Gene Expression Omnibus
    ID GSE115327. Identification of microRNA-dependent gene regulatory networks driving human pancreatic endocrine cell differentiation.
    1. Parker SC
    2. Stitzel ML
    3. Taylor DL
    4. Orozco JM
    (2013) NCBI Gene Expression Omnibus
    ID GSE51311. Chromatin stretch enhancer states drive cell-specific gene regulation and harbor human disease risk variants (ChIP-seq).
    1. Wang A
    2. Yue F
    3. Li Y
    4. Xie R
    (2015) NCBI Gene Expression Omnibus
    ID GSE54471. Developmental Competence Encoded at the Level of Enhancers.
    1. Bhandare R
    2. Schug J
    3. Lay JL
    4. Fox A
    5. Smirnova O
    6. Liu C
    7. Naji A
    8. Kaestner KH
    (2010) ArrayExpress
    ID E-MTAB-189. ChIP-Seq of human normal pancreatic islets with anti-histone antibodies to analyse histone modifications.
    1. 4DN Network
    2. Ren Laboratory
    (2020) 4D Nucleome
    ID Sample4DNES0LVRKBM. Replicates of Hi-C on CyT49 cells differentiated to pancreatic endoderm.

References

    1. Dupuis J
    2. Langenberg C
    3. Prokopenko I
    4. Saxena R
    5. Soranzo N
    6. Jackson AU
    7. Wheeler E
    8. Glazer NL
    9. Bouatia-Naji N
    10. Gloyn AL
    11. Lindgren CM
    12. Mägi R
    13. Morris AP
    14. Randall J
    15. Johnson T
    16. Elliott P
    17. Rybin D
    18. Thorleifsson G
    19. Steinthorsdottir V
    20. Henneman P
    21. Grallert H
    22. Dehghan A
    23. Hottenga JJ
    24. Franklin CS
    25. Navarro P
    26. Song K
    27. Goel A
    28. Perry JR
    29. Egan JM
    30. Lajunen T
    31. Grarup N
    32. Sparsø T
    33. Doney A
    34. Voight BF
    35. Stringham HM
    36. Li M
    37. Kanoni S
    38. Shrader P
    39. Cavalcanti-Proença C
    40. Kumari M
    41. Qi L
    42. Timpson NJ
    43. Gieger C
    44. Zabena C
    45. Rocheleau G
    46. Ingelsson E
    47. An P
    48. O'Connell J
    49. Luan J
    50. Elliott A
    51. McCarroll SA
    52. Payne F
    53. Roccasecca RM
    54. Pattou F
    55. Sethupathy P
    56. Ardlie K
    57. Ariyurek Y
    58. Balkau B
    59. Barter P
    60. Beilby JP
    61. Ben-Shlomo Y
    62. Benediktsson R
    63. Bennett AJ
    64. Bergmann S
    65. Bochud M
    66. Boerwinkle E
    67. Bonnefond A
    68. Bonnycastle LL
    69. Borch-Johnsen K
    70. Böttcher Y
    71. Brunner E
    72. Bumpstead SJ
    73. Charpentier G
    74. Chen YD
    75. Chines P
    76. Clarke R
    77. Coin LJ
    78. Cooper MN
    79. Cornelis M
    80. Crawford G
    81. Crisponi L
    82. Day IN
    83. de Geus EJ
    84. Delplanque J
    85. Dina C
    86. Erdos MR
    87. Fedson AC
    88. Fischer-Rosinsky A
    89. Forouhi NG
    90. Fox CS
    91. Frants R
    92. Franzosi MG
    93. Galan P
    94. Goodarzi MO
    95. Graessler J
    96. Groves CJ
    97. Grundy S
    98. Gwilliam R
    99. Gyllensten U
    100. Hadjadj S
    101. Hallmans G
    102. Hammond N
    103. Han X
    104. Hartikainen AL
    105. Hassanali N
    106. Hayward C
    107. Heath SC
    108. Hercberg S
    109. Herder C
    110. Hicks AA
    111. Hillman DR
    112. Hingorani AD
    113. Hofman A
    114. Hui J
    115. Hung J
    116. Isomaa B
    117. Johnson PR
    118. Jørgensen T
    119. Jula A
    120. Kaakinen M
    121. Kaprio J
    122. Kesaniemi YA
    123. Kivimaki M
    124. Knight B
    125. Koskinen S
    126. Kovacs P
    127. Kyvik KO
    128. Lathrop GM
    129. Lawlor DA
    130. Le Bacquer O
    131. Lecoeur C
    132. Li Y
    133. Lyssenko V
    134. Mahley R
    135. Mangino M
    136. Manning AK
    137. Martínez-Larrad MT
    138. McAteer JB
    139. McCulloch LJ
    140. McPherson R
    141. Meisinger C
    142. Melzer D
    143. Meyre D
    144. Mitchell BD
    145. Morken MA
    146. Mukherjee S
    147. Naitza S
    148. Narisu N
    149. Neville MJ
    150. Oostra BA
    151. Orrù M
    152. Pakyz R
    153. Palmer CN
    154. Paolisso G
    155. Pattaro C
    156. Pearson D
    157. Peden JF
    158. Pedersen NL
    159. Perola M
    160. Pfeiffer AF
    161. Pichler I
    162. Polasek O
    163. Posthuma D
    164. Potter SC
    165. Pouta A
    166. Province MA
    167. Psaty BM
    168. Rathmann W
    169. Rayner NW
    170. Rice K
    171. Ripatti S
    172. Rivadeneira F
    173. Roden M
    174. Rolandsson O
    175. Sandbaek A
    176. Sandhu M
    177. Sanna S
    178. Sayer AA
    179. Scheet P
    180. Scott LJ
    181. Seedorf U
    182. Sharp SJ
    183. Shields B
    184. Sigurethsson G
    185. Sijbrands EJ
    186. Silveira A
    187. Simpson L
    188. Singleton A
    189. Smith NL
    190. Sovio U
    191. Swift A
    192. Syddall H
    193. Syvänen AC
    194. Tanaka T
    195. Thorand B
    196. Tichet J
    197. Tönjes A
    198. Tuomi T
    199. Uitterlinden AG
    200. van Dijk KW
    201. van Hoek M
    202. Varma D
    203. Visvikis-Siest S
    204. Vitart V
    205. Vogelzangs N
    206. Waeber G
    207. Wagner PJ
    208. Walley A
    209. Walters GB
    210. Ward KL
    211. Watkins H
    212. Weedon MN
    213. Wild SH
    214. Willemsen G
    215. Witteman JC
    216. Yarnell JW
    217. Zeggini E
    218. Zelenika D
    219. Zethelius B
    220. Zhai G
    221. Zhao JH
    222. Zillikens MC
    223. Borecki IB
    224. Loos RJ
    225. Meneton P
    226. Magnusson PK
    227. Nathan DM
    228. Williams GH
    229. Hattersley AT
    230. Silander K
    231. Salomaa V
    232. Smith GD
    233. Bornstein SR
    234. Schwarz P
    235. Spranger J
    236. Karpe F
    237. Shuldiner AR
    238. Cooper C
    239. Dedoussis GV
    240. Serrano-Ríos M
    241. Morris AD
    242. Lind L
    243. Palmer LJ
    244. Hu FB
    245. Franks PW
    246. Ebrahim S
    247. Marmot M
    248. Kao WH
    249. Pankow JS
    250. Sampson MJ
    251. Kuusisto J
    252. Laakso M
    253. Hansen T
    254. Pedersen O
    255. Pramstaller PP
    256. Wichmann HE
    257. Illig T
    258. Rudan I
    259. Wright AF
    260. Stumvoll M
    261. Campbell H
    262. Wilson JF
    263. Bergman RN
    264. Buchanan TA
    265. Collins FS
    266. Mohlke KL
    267. Tuomilehto J
    268. Valle TT
    269. Altshuler D
    270. Rotter JI
    271. Siscovick DS
    272. Penninx BW
    273. Boomsma DI
    274. Deloukas P
    275. Spector TD
    276. Frayling TM
    277. Ferrucci L
    278. Kong A
    279. Thorsteinsdottir U
    280. Stefansson K
    281. van Duijn CM
    282. Aulchenko YS
    283. Cao A
    284. Scuteri A
    285. Schlessinger D
    286. Uda M
    287. Ruokonen A
    288. Jarvelin MR
    289. Waterworth DM
    290. Vollenweider P
    291. Peltonen L
    292. Mooser V
    293. Abecasis GR
    294. Wareham NJ
    295. Sladek R
    296. Froguel P
    297. Watanabe RM
    298. Meigs JB
    299. Groop L
    300. Boehnke M
    301. McCarthy MI
    302. Florez JC
    303. Barroso I
    304. DIAGRAM Consortium
    305. GIANT Consortium
    306. Global BPgen Consortium
    307. Anders Hamsten on behalf of Procardis Consortium
    308. MAGIC investigators
    (2010) New genetic loci implicated in fasting glucose homeostasis and their impact on type 2 diabetes risk
    Nature Genetics 42:105–116.
    https://doi.org/10.1038/ng.520
    1. Flannick J
    2. Mercader JM
    3. Fuchsberger C
    4. Udler MS
    5. Mahajan A
    6. Wessel J
    7. Teslovich TM
    8. Caulkins L
    9. Koesterer R
    10. Barajas-Olmos F
    11. Blackwell TW
    12. Boerwinkle E
    13. Brody JA
    14. Centeno-Cruz F
    15. Chen L
    16. Chen S
    17. Contreras-Cubas C
    18. Córdova E
    19. Correa A
    20. Cortes M
    21. DeFronzo RA
    22. Dolan L
    23. Drews KL
    24. Elliott A
    25. Floyd JS
    26. Gabriel S
    27. Garay-Sevilla ME
    28. García-Ortiz H
    29. Gross M
    30. Han S
    31. Heard-Costa NL
    32. Jackson AU
    33. Jørgensen ME
    34. Kang HM
    35. Kelsey M
    36. Kim BJ
    37. Koistinen HA
    38. Kuusisto J
    39. Leader JB
    40. Linneberg A
    41. Liu CT
    42. Liu J
    43. Lyssenko V
    44. Manning AK
    45. Marcketta A
    46. Malacara-Hernandez JM
    47. Martínez-Hernández A
    48. Matsuo K
    49. Mayer-Davis E
    50. Mendoza-Caamal E
    51. Mohlke KL
    52. Morrison AC
    53. Ndungu A
    54. Ng MCY
    55. O'Dushlaine C
    56. Payne AJ
    57. Pihoker C
    58. Post WS
    59. Preuss M
    60. Psaty BM
    61. Vasan RS
    62. Rayner NW
    63. Reiner AP
    64. Revilla-Monsalve C
    65. Robertson NR
    66. Santoro N
    67. Schurmann C
    68. So WY
    69. Soberón X
    70. Stringham HM
    71. Strom TM
    72. Tam CHT
    73. Thameem F
    74. Tomlinson B
    75. Torres JM
    76. Tracy RP
    77. van Dam RM
    78. Vujkovic M
    79. Wang S
    80. Welch RP
    81. Witte DR
    82. Wong TY
    83. Atzmon G
    84. Barzilai N
    85. Blangero J
    86. Bonnycastle LL
    87. Bowden DW
    88. Chambers JC
    89. Chan E
    90. Cheng CY
    91. Cho YS
    92. Collins FS
    93. de Vries PS
    94. Duggirala R
    95. Glaser B
    96. Gonzalez C
    97. Gonzalez ME
    98. Groop L
    99. Kooner JS
    100. Kwak SH
    101. Laakso M
    102. Lehman DM
    103. Nilsson P
    104. Spector TD
    105. Tai ES
    106. Tuomi T
    107. Tuomilehto J
    108. Wilson JG
    109. Aguilar-Salinas CA
    110. Bottinger E
    111. Burke B
    112. Carey DJ
    113. Chan JCN
    114. Dupuis J
    115. Frossard P
    116. Heckbert SR
    117. Hwang MY
    118. Kim YJ
    119. Kirchner HL
    120. Lee JY
    121. Lee J
    122. Loos RJF
    123. Ma RCW
    124. Morris AD
    125. O'Donnell CJ
    126. Palmer CNA
    127. Pankow J
    128. Park KS
    129. Rasheed A
    130. Saleheen D
    131. Sim X
    132. Small KS
    133. Teo YY
    134. Haiman C
    135. Hanis CL
    136. Henderson BE
    137. Orozco L
    138. Tusié-Luna T
    139. Dewey FE
    140. Baras A
    141. Gieger C
    142. Meitinger T
    143. Strauch K
    144. Lange L
    145. Grarup N
    146. Hansen T
    147. Pedersen O
    148. Zeitler P
    149. Dabelea D
    150. Abecasis G
    151. Bell GI
    152. Cox NJ
    153. Seielstad M
    154. Sladek R
    155. Meigs JB
    156. Rich SS
    157. Rotter JI
    158. Altshuler D
    159. Burtt NP
    160. Scott LJ
    161. Morris AP
    162. Florez JC
    163. McCarthy MI
    164. Boehnke M
    165. Broad Genomics Platform
    166. DiscovEHR Collaboration
    167. CHARGE
    168. LuCamp
    169. ProDiGY
    170. GoT2D
    171. ESP
    172. SIGMA-T2D
    173. T2D-GENES
    174. AMP-T2D-GENES
    (2019) Exome sequencing of 20,791 cases of type 2 diabetes and 24,440 controls
    Nature 570:71–76.
    https://doi.org/10.1038/s41586-019-1231-2
    1. Fuchsberger C
    2. Flannick J
    3. Teslovich TM
    4. Mahajan A
    5. Agarwala V
    6. Gaulton KJ
    7. Ma C
    8. Fontanillas P
    9. Moutsianas L
    10. McCarthy DJ
    11. Rivas MA
    12. Perry JRB
    13. Sim X
    14. Blackwell TW
    15. Robertson NR
    16. Rayner NW
    17. Cingolani P
    18. Locke AE
    19. Tajes JF
    20. Highland HM
    21. Dupuis J
    22. Chines PS
    23. Lindgren CM
    24. Hartl C
    25. Jackson AU
    26. Chen H
    27. Huyghe JR
    28. van de Bunt M
    29. Pearson RD
    30. Kumar A
    31. Müller-Nurasyid M
    32. Grarup N
    33. Stringham HM
    34. Gamazon ER
    35. Lee J
    36. Chen Y
    37. Scott RA
    38. Below JE
    39. Chen P
    40. Huang J
    41. Go MJ
    42. Stitzel ML
    43. Pasko D
    44. Parker SCJ
    45. Varga TV
    46. Green T
    47. Beer NL
    48. Day-Williams AG
    49. Ferreira T
    50. Fingerlin T
    51. Horikoshi M
    52. Hu C
    53. Huh I
    54. Ikram MK
    55. Kim BJ
    56. Kim Y
    57. Kim YJ
    58. Kwon MS
    59. Lee J
    60. Lee S
    61. Lin KH
    62. Maxwell TJ
    63. Nagai Y
    64. Wang X
    65. Welch RP
    66. Yoon J
    67. Zhang W
    68. Barzilai N
    69. Voight BF
    70. Han BG
    71. Jenkinson CP
    72. Kuulasmaa T
    73. Kuusisto J
    74. Manning A
    75. Ng MCY
    76. Palmer ND
    77. Balkau B
    78. Stančáková A
    79. Abboud HE
    80. Boeing H
    81. Giedraitis V
    82. Prabhakaran D
    83. Gottesman O
    84. Scott J
    85. Carey J
    86. Kwan P
    87. Grant G
    88. Smith JD
    89. Neale BM
    90. Purcell S
    91. Butterworth AS
    92. Howson JMM
    93. Lee HM
    94. Lu Y
    95. Kwak SH
    96. Zhao W
    97. Danesh J
    98. Lam VKL
    99. Park KS
    100. Saleheen D
    101. So WY
    102. Tam CHT
    103. Afzal U
    104. Aguilar D
    105. Arya R
    106. Aung T
    107. Chan E
    108. Navarro C
    109. Cheng CY
    110. Palli D
    111. Correa A
    112. Curran JE
    113. Rybin D
    114. Farook VS
    115. Fowler SP
    116. Freedman BI
    117. Griswold M
    118. Hale DE
    119. Hicks PJ
    120. Khor CC
    121. Kumar S
    122. Lehne B
    123. Thuillier D
    124. Lim WY
    125. Liu J
    126. van der Schouw YT
    127. Loh M
    128. Musani SK
    129. Puppala S
    130. Scott WR
    131. Yengo L
    132. Tan ST
    133. Taylor HA
    134. Thameem F
    135. Wilson G
    136. Wong TY
    137. Njølstad PR
    138. Levy JC
    139. Mangino M
    140. Bonnycastle LL
    141. Schwarzmayr T
    142. Fadista J
    143. Surdulescu GL
    144. Herder C
    145. Groves CJ
    146. Wieland T
    147. Bork-Jensen J
    148. Brandslund I
    149. Christensen C
    150. Koistinen HA
    151. Doney ASF
    152. Kinnunen L
    153. Esko T
    154. Farmer AJ
    155. Hakaste L
    156. Hodgkiss D
    157. Kravic J
    158. Lyssenko V
    159. Hollensted M
    160. Jørgensen ME
    161. Jørgensen T
    162. Ladenvall C
    163. Justesen JM
    164. Käräjämäki A
    165. Kriebel J
    166. Rathmann W
    167. Lannfelt L
    168. Lauritzen T
    169. Narisu N
    170. Linneberg A
    171. Melander O
    172. Milani L
    173. Neville M
    174. Orho-Melander M
    175. Qi L
    176. Qi Q
    177. Roden M
    178. Rolandsson O
    179. Swift A
    180. Rosengren AH
    181. Stirrups K
    182. Wood AR
    183. Mihailov E
    184. Blancher C
    185. Carneiro MO
    186. Maguire J
    187. Poplin R
    188. Shakir K
    189. Fennell T
    190. DePristo M
    191. de Angelis MH
    192. Deloukas P
    193. Gjesing AP
    194. Jun G
    195. Nilsson P
    196. Murphy J
    197. Onofrio R
    198. Thorand B
    199. Hansen T
    200. Meisinger C
    201. Hu FB
    202. Isomaa B
    203. Karpe F
    204. Liang L
    205. Peters A
    206. Huth C
    207. O'Rahilly SP
    208. Palmer CNA
    209. Pedersen O
    210. Rauramaa R
    211. Tuomilehto J
    212. Salomaa V
    213. Watanabe RM
    214. Syvänen AC
    215. Bergman RN
    216. Bharadwaj D
    217. Bottinger EP
    218. Cho YS
    219. Chandak GR
    220. Chan JCN
    221. Chia KS
    222. Daly MJ
    223. Ebrahim SB
    224. Langenberg C
    225. Elliott P
    226. Jablonski KA
    227. Lehman DM
    228. Jia W
    229. Ma RCW
    230. Pollin TI
    231. Sandhu M
    232. Tandon N
    233. Froguel P
    234. Barroso I
    235. Teo YY
    236. Zeggini E
    237. Loos RJF
    238. Small KS
    239. Ried JS
    240. DeFronzo RA
    241. Grallert H
    242. Glaser B
    243. Metspalu A
    244. Wareham NJ
    245. Walker M
    246. Banks E
    247. Gieger C
    248. Ingelsson E
    249. Im HK
    250. Illig T
    251. Franks PW
    252. Buck G
    253. Trakalo J
    254. Buck D
    255. Prokopenko I
    256. Mägi R
    257. Lind L
    258. Farjoun Y
    259. Owen KR
    260. Gloyn AL
    261. Strauch K
    262. Tuomi T
    263. Kooner JS
    264. Lee JY
    265. Park T
    266. Donnelly P
    267. Morris AD
    268. Hattersley AT
    269. Bowden DW
    270. Collins FS
    271. Atzmon G
    272. Chambers JC
    273. Spector TD
    274. Laakso M
    275. Strom TM
    276. Bell GI
    277. Blangero J
    278. Duggirala R
    279. Tai ES
    280. McVean G
    281. Hanis CL
    282. Wilson JG
    283. Seielstad M
    284. Frayling TM
    285. Meigs JB
    286. Cox NJ
    287. Sladek R
    288. Lander ES
    289. Gabriel S
    290. Burtt NP
    291. Mohlke KL
    292. Meitinger T
    293. Groop L
    294. Abecasis G
    295. Florez JC
    296. Scott LJ
    297. Morris AP
    298. Kang HM
    299. Boehnke M
    300. Altshuler D
    301. McCarthy MI
    (2016) The genetic architecture of type 2 diabetes
    Nature 536:41–47.
    https://doi.org/10.1038/nature18642
    1. Gaulton KJ
    2. Ferreira T
    3. Lee Y
    4. Raimondo A
    5. Mägi R
    6. Reschen ME
    7. Mahajan A
    8. Locke A
    9. Rayner NW
    10. Robertson N
    11. Scott RA
    12. Prokopenko I
    13. Scott LJ
    14. Green T
    15. Sparso T
    16. Thuillier D
    17. Yengo L
    18. Grallert H
    19. Wahl S
    20. Frånberg M
    21. Strawbridge RJ
    22. Kestler H
    23. Chheda H
    24. Eisele L
    25. Gustafsson S
    26. Steinthorsdottir V
    27. Thorleifsson G
    28. Qi L
    29. Karssen LC
    30. van Leeuwen EM
    31. Willems SM
    32. Li M
    33. Chen H
    34. Fuchsberger C
    35. Kwan P
    36. Ma C
    37. Linderman M
    38. Lu Y
    39. Thomsen SK
    40. Rundle JK
    41. Beer NL
    42. van de Bunt M
    43. Chalisey A
    44. Kang HM
    45. Voight BF
    46. Abecasis GR
    47. Almgren P
    48. Baldassarre D
    49. Balkau B
    50. Benediktsson R
    51. Blüher M
    52. Boeing H
    53. Bonnycastle LL
    54. Bottinger EP
    55. Burtt NP
    56. Carey J
    57. Charpentier G
    58. Chines PS
    59. Cornelis MC
    60. Couper DJ
    61. Crenshaw AT
    62. van Dam RM
    63. Doney AS
    64. Dorkhan M
    65. Edkins S
    66. Eriksson JG
    67. Esko T
    68. Eury E
    69. Fadista J
    70. Flannick J
    71. Fontanillas P
    72. Fox C
    73. Franks PW
    74. Gertow K
    75. Gieger C
    76. Gigante B
    77. Gottesman O
    78. Grant GB
    79. Grarup N
    80. Groves CJ
    81. Hassinen M
    82. Have CT
    83. Herder C
    84. Holmen OL
    85. Hreidarsson AB
    86. Humphries SE
    87. Hunter DJ
    88. Jackson AU
    89. Jonsson A
    90. Jørgensen ME
    91. Jørgensen T
    92. Kao WH
    93. Kerrison ND
    94. Kinnunen L
    95. Klopp N
    96. Kong A
    97. Kovacs P
    98. Kraft P
    99. Kravic J
    100. Langford C
    101. Leander K
    102. Liang L
    103. Lichtner P
    104. Lindgren CM
    105. Lindholm E
    106. Linneberg A
    107. Liu CT
    108. Lobbens S
    109. Luan J
    110. Lyssenko V
    111. Männistö S
    112. McLeod O
    113. Meyer J
    114. Mihailov E
    115. Mirza G
    116. Mühleisen TW
    117. Müller-Nurasyid M
    118. Navarro C
    119. Nöthen MM
    120. Oskolkov NN
    121. Owen KR
    122. Palli D
    123. Pechlivanis S
    124. Peltonen L
    125. Perry JR
    126. Platou CG
    127. Roden M
    128. Ruderfer D
    129. Rybin D
    130. van der Schouw YT
    131. Sennblad B
    132. Sigurðsson G
    133. Stančáková A
    134. Steinbach G
    135. Storm P
    136. Strauch K
    137. Stringham HM
    138. Sun Q
    139. Thorand B
    140. Tikkanen E
    141. Tonjes A
    142. Trakalo J
    143. Tremoli E
    144. Tuomi T
    145. Wennauer R
    146. Wiltshire S
    147. Wood AR
    148. Zeggini E
    149. Dunham I
    150. Birney E
    151. Pasquali L
    152. Ferrer J
    153. Loos RJ
    154. Dupuis J
    155. Florez JC
    156. Boerwinkle E
    157. Pankow JS
    158. van Duijn C
    159. Sijbrands E
    160. Meigs JB
    161. Hu FB
    162. Thorsteinsdottir U
    163. Stefansson K
    164. Lakka TA
    165. Rauramaa R
    166. Stumvoll M
    167. Pedersen NL
    168. Lind L
    169. Keinanen-Kiukaanniemi SM
    170. Korpi-Hyövälti E
    171. Saaristo TE
    172. Saltevo J
    173. Kuusisto J
    174. Laakso M
    175. Metspalu A
    176. Erbel R
    177. Jöcke KH
    178. Moebus S
    179. Ripatti S
    180. Salomaa V
    181. Ingelsson E
    182. Boehm BO
    183. Bergman RN
    184. Collins FS
    185. Mohlke KL
    186. Koistinen H
    187. Tuomilehto J
    188. Hveem K
    189. Njølstad I
    190. Deloukas P
    191. Donnelly PJ
    192. Frayling TM
    193. Hattersley AT
    194. de Faire U
    195. Hamsten A
    196. Illig T
    197. Peters A
    198. Cauchi S
    199. Sladek R
    200. Froguel P
    201. Hansen T
    202. Pedersen O
    203. Morris AD
    204. Palmer CN
    205. Kathiresan S
    206. Melander O
    207. Nilsson PM
    208. Groop LC
    209. Barroso I
    210. Langenberg C
    211. Wareham NJ
    212. O'Callaghan CA
    213. Gloyn AL
    214. Altshuler D
    215. Boehnke M
    216. Teslovich TM
    217. McCarthy MI
    218. Morris AP
    219. DIAbetes Genetics Replication And Meta-analysis (DIAGRAM) Consortium
    (2015) Genetic fine mapping and genomic annotation defines causal mechanisms at type 2 diabetes susceptibility loci
    Nature Genetics 47:1415–1425.
    https://doi.org/10.1038/ng.3437
    1. Horikoshi M
    2. Beaumont RN
    3. Day FR
    4. Warrington NM
    5. Kooijman MN
    6. Fernandez-Tajes J
    7. Feenstra B
    8. van Zuydam NR
    9. Gaulton KJ
    10. Grarup N
    11. Bradfield JP
    12. Strachan DP
    13. Li-Gao R
    14. Ahluwalia TS
    15. Kreiner E
    16. Rueedi R
    17. Lyytikäinen LP
    18. Cousminer DL
    19. Wu Y
    20. Thiering E
    21. Wang CA
    22. Have CT
    23. Hottenga JJ
    24. Vilor-Tejedor N
    25. Joshi PK
    26. Boh ETH
    27. Ntalla I
    28. Pitkänen N
    29. Mahajan A
    30. van Leeuwen EM
    31. Joro R
    32. Lagou V
    33. Nodzenski M
    34. Diver LA
    35. Zondervan KT
    36. Bustamante M
    37. Marques-Vidal P
    38. Mercader JM
    39. Bennett AJ
    40. Rahmioglu N
    41. Nyholt DR
    42. Ma RCW
    43. Tam CHT
    44. Tam WH
    45. Ganesh SK
    46. van Rooij FJ
    47. Jones SE
    48. Loh PR
    49. Ruth KS
    50. Tuke MA
    51. Tyrrell J
    52. Wood AR
    53. Yaghootkar H
    54. Scholtens DM
    55. Paternoster L
    56. Prokopenko I
    57. Kovacs P
    58. Atalay M
    59. Willems SM
    60. Panoutsopoulou K
    61. Wang X
    62. Carstensen L
    63. Geller F
    64. Schraut KE
    65. Murcia M
    66. van Beijsterveldt CE
    67. Willemsen G
    68. Appel EVR
    69. Fonvig CE
    70. Trier C
    71. Tiesler CM
    72. Standl M
    73. Kutalik Z
    74. Bonas-Guarch S
    75. Hougaard DM
    76. Sánchez F
    77. Torrents D
    78. Waage J
    79. Hollegaard MV
    80. de Haan HG
    81. Rosendaal FR
    82. Medina-Gomez C
    83. Ring SM
    84. Hemani G
    85. McMahon G
    86. Robertson NR
    87. Groves CJ
    88. Langenberg C
    89. Luan J
    90. Scott RA
    91. Zhao JH
    92. Mentch FD
    93. MacKenzie SM
    94. Reynolds RM
    95. Lowe WL
    96. Tönjes A
    97. Stumvoll M
    98. Lindi V
    99. Lakka TA
    100. van Duijn CM
    101. Kiess W
    102. Körner A
    103. Sørensen TI
    104. Niinikoski H
    105. Pahkala K
    106. Raitakari OT
    107. Zeggini E
    108. Dedoussis GV
    109. Teo YY
    110. Saw SM
    111. Melbye M
    112. Campbell H
    113. Wilson JF
    114. Vrijheid M
    115. de Geus EJ
    116. Boomsma DI
    117. Kadarmideen HN
    118. Holm JC
    119. Hansen T
    120. Sebert S
    121. Hattersley AT
    122. Beilin LJ
    123. Newnham JP
    124. Pennell CE
    125. Heinrich J
    126. Adair LS
    127. Borja JB
    128. Mohlke KL
    129. Eriksson JG
    130. Widén EE
    131. Kähönen M
    132. Viikari JS
    133. Lehtimäki T
    134. Vollenweider P
    135. Bønnelykke K
    136. Bisgaard H
    137. Mook-Kanamori DO
    138. Hofman A
    139. Rivadeneira F
    140. Uitterlinden AG
    141. Pisinger C
    142. Pedersen O
    143. Power C
    144. Hyppönen E
    145. Wareham NJ
    146. Hakonarson H
    147. Davies E
    148. Walker BR
    149. Jaddoe VW
    150. Jarvelin MR
    151. Grant SF
    152. Vaag AA
    153. Lawlor DA
    154. Frayling TM
    155. Davey Smith G
    156. Morris AP
    157. Ong KK
    158. Felix JF
    159. Timpson NJ
    160. Perry JR
    161. Evans DM
    162. McCarthy MI
    163. Freathy RM
    164. CHARGE Consortium Hematology Working Group
    165. Early Growth Genetics (EGG) Consortium
    (2016) Genome-wide associations for birth weight and correlations with adult disease
    Nature 538:248–252.
    https://doi.org/10.1038/nature19806
    1. Manning AK
    2. Hivert MF
    3. Scott RA
    4. Grimsby JL
    5. Bouatia-Naji N
    6. Chen H
    7. Rybin D
    8. Liu CT
    9. Bielak LF
    10. Prokopenko I
    11. Amin N
    12. Barnes D
    13. Cadby G
    14. Hottenga JJ
    15. Ingelsson E
    16. Jackson AU
    17. Johnson T
    18. Kanoni S
    19. Ladenvall C
    20. Lagou V
    21. Lahti J
    22. Lecoeur C
    23. Liu Y
    24. Martinez-Larrad MT
    25. Montasser ME
    26. Navarro P
    27. Perry JR
    28. Rasmussen-Torvik LJ
    29. Salo P
    30. Sattar N
    31. Shungin D
    32. Strawbridge RJ
    33. Tanaka T
    34. van Duijn CM
    35. An P
    36. de Andrade M
    37. Andrews JS
    38. Aspelund T
    39. Atalay M
    40. Aulchenko Y
    41. Balkau B
    42. Bandinelli S
    43. Beckmann JS
    44. Beilby JP
    45. Bellis C
    46. Bergman RN
    47. Blangero J
    48. Boban M
    49. Boehnke M
    50. Boerwinkle E
    51. Bonnycastle LL
    52. Boomsma DI
    53. Borecki IB
    54. Böttcher Y
    55. Bouchard C
    56. Brunner E
    57. Budimir D
    58. Campbell H
    59. Carlson O
    60. Chines PS
    61. Clarke R
    62. Collins FS
    63. Corbatón-Anchuelo A
    64. Couper D
    65. de Faire U
    66. Dedoussis GV
    67. Deloukas P
    68. Dimitriou M
    69. Egan JM
    70. Eiriksdottir G
    71. Erdos MR
    72. Eriksson JG
    73. Eury E
    74. Ferrucci L
    75. Ford I
    76. Forouhi NG
    77. Fox CS
    78. Franzosi MG
    79. Franks PW
    80. Frayling TM
    81. Froguel P
    82. Galan P
    83. de Geus E
    84. Gigante B
    85. Glazer NL
    86. Goel A
    87. Groop L
    88. Gudnason V
    89. Hallmans G
    90. Hamsten A
    91. Hansson O
    92. Harris TB
    93. Hayward C
    94. Heath S
    95. Hercberg S
    96. Hicks AA
    97. Hingorani A
    98. Hofman A
    99. Hui J
    100. Hung J
    101. Jarvelin MR
    102. Jhun MA
    103. Johnson PC
    104. Jukema JW
    105. Jula A
    106. Kao WH
    107. Kaprio J
    108. Kardia SL
    109. Keinanen-Kiukaanniemi S
    110. Kivimaki M
    111. Kolcic I
    112. Kovacs P
    113. Kumari M
    114. Kuusisto J
    115. Kyvik KO
    116. Laakso M
    117. Lakka T
    118. Lannfelt L
    119. Lathrop GM
    120. Launer LJ
    121. Leander K
    122. Li G
    123. Lind L
    124. Lindstrom J
    125. Lobbens S
    126. Loos RJ
    127. Luan J
    128. Lyssenko V
    129. Mägi R
    130. Magnusson PK
    131. Marmot M
    132. Meneton P
    133. Mohlke KL
    134. Mooser V
    135. Morken MA
    136. Miljkovic I
    137. Narisu N
    138. O'Connell J
    139. Ong KK
    140. Oostra BA
    141. Palmer LJ
    142. Palotie A
    143. Pankow JS
    144. Peden JF
    145. Pedersen NL
    146. Pehlic M
    147. Peltonen L
    148. Penninx B
    149. Pericic M
    150. Perola M
    151. Perusse L
    152. Peyser PA
    153. Polasek O
    154. Pramstaller PP
    155. Province MA
    156. Räikkönen K
    157. Rauramaa R
    158. Rehnberg E
    159. Rice K
    160. Rotter JI
    161. Rudan I
    162. Ruokonen A
    163. Saaristo T
    164. Sabater-Lleal M
    165. Salomaa V
    166. Savage DB
    167. Saxena R
    168. Schwarz P
    169. Seedorf U
    170. Sennblad B
    171. Serrano-Rios M
    172. Shuldiner AR
    173. Sijbrands EJ
    174. Siscovick DS
    175. Smit JH
    176. Small KS
    177. Smith NL
    178. Smith AV
    179. Stančáková A
    180. Stirrups K
    181. Stumvoll M
    182. Sun YV
    183. Swift AJ
    184. Tönjes A
    185. Tuomilehto J
    186. Trompet S
    187. Uitterlinden AG
    188. Uusitupa M
    189. Vikström M
    190. Vitart V
    191. Vohl MC
    192. Voight BF
    193. Vollenweider P
    194. Waeber G
    195. Waterworth DM
    196. Watkins H
    197. Wheeler E
    198. Widen E
    199. Wild SH
    200. Willems SM
    201. Willemsen G
    202. Wilson JF
    203. Witteman JC
    204. Wright AF
    205. Yaghootkar H
    206. Zelenika D
    207. Zemunik T
    208. Zgaga L
    209. Wareham NJ
    210. McCarthy MI
    211. Barroso I
    212. Watanabe RM
    213. Florez JC
    214. Dupuis J
    215. Meigs JB
    216. Langenberg C
    217. DIAbetes Genetics Replication And Meta-analysis (DIAGRAM) Consortium
    218. Multiple Tissue Human Expression Resource (MUTHER) Consortium
    (2012) A genome-wide approach accounting for body mass index identifies genetic variants influencing fasting glycemic traits and insulin resistance
    Nature Genetics 44:659–669.
    https://doi.org/10.1038/ng.2274
    1. Saxena R
    2. Hivert MF
    3. Langenberg C
    4. Tanaka T
    5. Pankow JS
    6. Vollenweider P
    7. Lyssenko V
    8. Bouatia-Naji N
    9. Dupuis J
    10. Jackson AU
    11. Kao WH
    12. Li M
    13. Glazer NL
    14. Manning AK
    15. Luan J
    16. Stringham HM
    17. Prokopenko I
    18. Johnson T
    19. Grarup N
    20. Boesgaard TW
    21. Lecoeur C
    22. Shrader P
    23. O'Connell J
    24. Ingelsson E
    25. Couper DJ
    26. Rice K
    27. Song K
    28. Andreasen CH
    29. Dina C
    30. Köttgen A
    31. Le Bacquer O
    32. Pattou F
    33. Taneera J
    34. Steinthorsdottir V
    35. Rybin D
    36. Ardlie K
    37. Sampson M
    38. Qi L
    39. van Hoek M
    40. Weedon MN
    41. Aulchenko YS
    42. Voight BF
    43. Grallert H
    44. Balkau B
    45. Bergman RN
    46. Bielinski SJ
    47. Bonnefond A
    48. Bonnycastle LL
    49. Borch-Johnsen K
    50. Böttcher Y
    51. Brunner E
    52. Buchanan TA
    53. Bumpstead SJ
    54. Cavalcanti-Proença C
    55. Charpentier G
    56. Chen YD
    57. Chines PS
    58. Collins FS
    59. Cornelis M
    60. J Crawford G
    61. Delplanque J
    62. Doney A
    63. Egan JM
    64. Erdos MR
    65. Firmann M
    66. Forouhi NG
    67. Fox CS
    68. Goodarzi MO
    69. Graessler J
    70. Hingorani A
    71. Isomaa B
    72. Jørgensen T
    73. Kivimaki M
    74. Kovacs P
    75. Krohn K
    76. Kumari M
    77. Lauritzen T
    78. Lévy-Marchal C
    79. Mayor V
    80. McAteer JB
    81. Meyre D
    82. Mitchell BD
    83. Mohlke KL
    84. Morken MA
    85. Narisu N
    86. Palmer CN
    87. Pakyz R
    88. Pascoe L
    89. Payne F
    90. Pearson D
    91. Rathmann W
    92. Sandbaek A
    93. Sayer AA
    94. Scott LJ
    95. Sharp SJ
    96. Sijbrands E
    97. Singleton A
    98. Siscovick DS
    99. Smith NL
    100. Sparsø T
    101. Swift AJ
    102. Syddall H
    103. Thorleifsson G
    104. Tönjes A
    105. Tuomi T
    106. Tuomilehto J
    107. Valle TT
    108. Waeber G
    109. Walley A
    110. Waterworth DM
    111. Zeggini E
    112. Zhao JH
    113. Illig T
    114. Wichmann HE
    115. Wilson JF
    116. van Duijn C
    117. Hu FB
    118. Morris AD
    119. Frayling TM
    120. Hattersley AT
    121. Thorsteinsdottir U
    122. Stefansson K
    123. Nilsson P
    124. Syvänen AC
    125. Shuldiner AR
    126. Walker M
    127. Bornstein SR
    128. Schwarz P
    129. Williams GH
    130. Nathan DM
    131. Kuusisto J
    132. Laakso M
    133. Cooper C
    134. Marmot M
    135. Ferrucci L
    136. Mooser V
    137. Stumvoll M
    138. Loos RJ
    139. Altshuler D
    140. Psaty BM
    141. Rotter JI
    142. Boerwinkle E
    143. Hansen T
    144. Pedersen O
    145. Florez JC
    146. McCarthy MI
    147. Boehnke M
    148. Barroso I
    149. Sladek R
    150. Froguel P
    151. Meigs JB
    152. Groop L
    153. Wareham NJ
    154. Watanabe RM
    155. GIANT consortium
    156. MAGIC investigators
    (2010) Genetic variation in GIPR influences the glucose and insulin responses to an oral glucose challenge
    Nature Genetics 42:142–148.
    https://doi.org/10.1038/ng.521
    1. van der Valk RJ
    2. Kreiner-Møller E
    3. Kooijman MN
    4. Guxens M
    5. Stergiakouli E
    6. Sääf A
    7. Bradfield JP
    8. Geller F
    9. Hayes MG
    10. Cousminer DL
    11. Körner A
    12. Thiering E
    13. Curtin JA
    14. Myhre R
    15. Huikari V
    16. Joro R
    17. Kerkhof M
    18. Warrington NM
    19. Pitkänen N
    20. Ntalla I
    21. Horikoshi M
    22. Veijola R
    23. Freathy RM
    24. Teo YY
    25. Barton SJ
    26. Evans DM
    27. Kemp JP
    28. St Pourcain B
    29. Ring SM
    30. Davey Smith G
    31. Bergström A
    32. Kull I
    33. Hakonarson H
    34. Mentch FD
    35. Bisgaard H
    36. Chawes B
    37. Stokholm J
    38. Waage J
    39. Eriksen P
    40. Sevelsted A
    41. Melbye M
    42. van Duijn CM
    43. Medina-Gomez C
    44. Hofman A
    45. de Jongste JC
    46. Taal HR
    47. Uitterlinden AG
    48. Armstrong LL
    49. Eriksson J
    50. Palotie A
    51. Bustamante M
    52. Estivill X
    53. Gonzalez JR
    54. Llop S
    55. Kiess W
    56. Mahajan A
    57. Flexeder C
    58. Tiesler CM
    59. Murray CS
    60. Simpson A
    61. Magnus P
    62. Sengpiel V
    63. Hartikainen AL
    64. Keinanen-Kiukaanniemi S
    65. Lewin A
    66. Da Silva Couto Alves A
    67. Blakemore AI
    68. Buxton JL
    69. Kaakinen M
    70. Rodriguez A
    71. Sebert S
    72. Vaarasmaki M
    73. Lakka T
    74. Lindi V
    75. Gehring U
    76. Postma DS
    77. Ang W
    78. Newnham JP
    79. Lyytikäinen LP
    80. Pahkala K
    81. Raitakari OT
    82. Panoutsopoulou K
    83. Zeggini E
    84. Boomsma DI
    85. Groen-Blokhuis M
    86. Ilonen J
    87. Franke L
    88. Hirschhorn JN
    89. Pers TH
    90. Liang L
    91. Huang J
    92. Hocher B
    93. Knip M
    94. Saw SM
    95. Holloway JW
    96. Melén E
    97. Grant SF
    98. Feenstra B
    99. Lowe WL
    100. Widén E
    101. Sergeyev E
    102. Grallert H
    103. Custovic A
    104. Jacobsson B
    105. Jarvelin MR
    106. Atalay M
    107. Koppelman GH
    108. Pennell CE
    109. Niinikoski H
    110. Dedoussis GV
    111. Mccarthy MI
    112. Frayling TM
    113. Sunyer J
    114. Timpson NJ
    115. Rivadeneira F
    116. Bønnelykke K
    117. Jaddoe VW
    118. Early Genetics and Lifecourse Epidemiology (EAGLE) Consortium
    119. Genetic Investigation of ANthropometric Traits (GIANT) Consortium
    120. Early Growth Genetics (EGG) Consortium
    (2015) A novel common variant in DCST2 is associated with length in early life and height in adulthood
    Human Molecular Genetics 24:1155–1168.
    https://doi.org/10.1093/hmg/ddu510