Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Pancreatic islet enhancer clusters enriched in type 2 diabetes risk-associated variants

Subjects

Abstract

Type 2 diabetes affects over 300 million people, causing severe complications and premature death, yet the underlying molecular mechanisms are largely unknown. Pancreatic islet dysfunction is central in type 2 diabetes pathogenesis, and understanding islet genome regulation could therefore provide valuable mechanistic insights. We have now mapped and examined the function of human islet cis-regulatory networks. We identify genomic sequences that are targeted by islet transcription factors to drive islet-specific gene activity and show that most such sequences reside in clusters of enhancers that form physical three-dimensional chromatin domains. We find that sequence variants associated with type 2 diabetes and fasting glycemia are enriched in these clustered islet enhancers and identify trait-associated variants that disrupt DNA binding and islet enhancer activity. Our studies illustrate how islet transcription factors interact functionally with the epigenome and provide systematic evidence that the dysregulation of islet enhancers is relevant to the mechanisms underlying type 2 diabetes.

This is a preview of subscription content, access via your institution

Access options

Rent or buy this article

Prices vary by article type

from$1.95

to$39.95

Prices may be subject to local taxes which are calculated during checkout

Figure 1: Integrative regulatory maps of human pancreatic islet cells.
Figure 2: Transcription factor networks establish distinct types of interactions with the epigenome.
Figure 3: Enhancer clusters form functional three-dimensional chromatin domains.
Figure 4: Known and new transcription factor motifs are enriched in clustered islet enhancers.
Figure 5: Islet enhancers are enriched in loci associated with T2D and fasting glycemia.
Figure 6: A T2D risk variant at ZFAND3 disrupts islet enhancer activity.

Similar content being viewed by others

Accession codes

Primary accessions

ArrayExpress

References

  1. Rada-Iglesias, A. et al. A unique chromatin signature uncovers early developmental enhancers in humans. Nature 470, 279–283 (2011).

    Article  CAS  PubMed  Google Scholar 

  2. Heintzman, N.D. et al. Histone modifications at human enhancers reflect global cell-type-specific gene expression. Nature 459, 108–112 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Creyghton, M.P. et al. Histone H3K27ac separates active from poised enhancers and predicts developmental state. Proc. Natl. Acad. Sci. USA 107, 21931–21936 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Bonn, S. et al. Tissue-specific analysis of chromatin state identifies temporal signatures of enhancer activity during embryonic development. Nat. Genet. 44, 148–156 (2012).

    Article  CAS  PubMed  Google Scholar 

  5. Gaulton, K.J. et al. A map of open chromatin in human pancreatic islets. Nat. Genet. 42, 255–259 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Shen, Y. et al. A map of the cis-regulatory sequences in the mouse genome. Nature 488, 116–120 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. ENCODE Project Consortium. An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74 (2012).

  8. Kahn, S.E. Clinical review 135: the importance of β-cell failure in the development and progression of type 2 diabetes. J. Clin. Endocrinol. Metab. 86, 4047–4058 (2001).

    CAS  PubMed  Google Scholar 

  9. Lyssenko, V. et al. Clinical risk factors, DNA variants, and the development of type 2 diabetes. N. Engl. J. Med. 359, 2220–2232 (2008).

    Article  CAS  PubMed  Google Scholar 

  10. Ashcroft, F.M. & Rorsman, P. Diabetes mellitus and the β cell: the last ten years. Cell 148, 1160–1171 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Bhandare, R. et al. Genome-wide analysis of histone modifications in human pancreatic islets. Genome Res. 20, 428–433 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Stitzel, M.L. et al. Global epigenomic analysis of primary human pancreatic islets provides insights into type 2 diabetes susceptibility loci. Cell Metab. 12, 443–455 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Khoo, C. et al. Research resource: the Pdx1 cistrome of pancreatic islets. Mol. Endocrinol. 26, 521–533 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Bernstein, B.E. et al. The NIH Roadmap Epigenomics Mapping Consortium. Nat. Biotechnol. 28, 1045–1048 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Zhu, J. et al. Genome-wide chromatin state transitions associated with developmental and environmental cues. Cell 152, 642–654 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Tennant, B.R. et al. Identification and analysis of murine pancreatic islet enhancers. Diabetologia 56, 542–552 (2013).

    Article  CAS  PubMed  Google Scholar 

  17. Artner, I. et al. MafB is required for islet β cell maturation. Proc. Natl. Acad. Sci. USA 104, 3853–3858 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Ahlgren, U., Jonsson, J., Jonsson, L., Simu, K. & Edlund, H. β-cell-specific inactivation of the mouse Ipf1/Pdx1 gene results in loss of the β-cell phenotype and maturity onset diabetes. Genes Dev. 12, 1763–1768 (1998).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Sander, M. et al. Homeobox gene Nkx6.1 lies downstream of Nkx2.2 in the major pathway of β-cell formation in the pancreas. Development 127, 5533–5540 (2000).

    Article  CAS  PubMed  Google Scholar 

  20. Sund, N.J. et al. Tissue-specific deletion of Foxa2 in pancreatic β cells results in hyperinsulinemic hypoglycemia. Genes Dev. 15, 1706–1715 (2001).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Sussel, L. et al. Mice lacking the homeodomain transcription factor Nkx2.2 have diabetes due to arrested differentiation of pancreatic β cells. Development 125, 2213–2221 (1998).

    Article  CAS  PubMed  Google Scholar 

  22. Servitja, J.M. & Ferrer, J. Transcriptional networks controlling pancreatic development and β cell function. Diabetologia 47, 597–613 (2004).

    Article  CAS  PubMed  Google Scholar 

  23. Wilson, M.E., Scheel, D. & German, M.S. Gene expression cascades in pancreatic development. Mech. Dev. 120, 65–80 (2003).

    Article  CAS  PubMed  Google Scholar 

  24. Oliver-Krasinski, J.M. & Stoffers, D.A. On the origin of the β cell. Genes Dev. 22, 1998–2021 (2008).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Gerstein, M.B. et al. Integrative analysis of the Caenorhabditis elegans genome by the modENCODE project. Science 330, 1775–1787 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Kyrmizi, I. et al. Plasticity and expanding complexity of the hepatic transcription factor network during liver development. Genes Dev. 20, 2293–2305 (2006).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Jin, C. et al. H3.3/H2A.Z double variant–containing nucleosomes mark 'nucleosome-free regions' of active promoters and other regulatory regions. Nat. Genet. 41, 941–945 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Ong, C.T. & Corces, V.G. Enhancer function: new insights into the regulation of tissue-specific gene expression. Nat. Rev. Genet. 12, 283–293 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Morán, I. et al. Human β cell transcriptome analysis uncovers lncRNAs that are tissue-specific, dynamically regulated, and abnormally expressed in type 2 diabetes. Cell Metab. 16, 435–448 (2012).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  30. Stefflova, K. et al. Cooperativity and rapid evolution of cobound transcription factors in closely related mammals. Cell 154, 530–540 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Zinzen, R.P., Girardot, C., Gagneur, J., Braun, M. & Furlong, E.E. Combinatorial binding predicts spatio-temporal cis-regulatory activity. Nature 462, 65–70 (2009).

    Article  CAS  PubMed  Google Scholar 

  32. Zhou, Q., Brown, J., Kanarek, A., Rajagopal, J. & Melton, D.A. In vivo reprogramming of adult pancreatic exocrine cells to β-cells. Nature 455, 627–632 (2008).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Noordermeer, D. et al. The dynamic architecture of Hox gene clusters. Science 334, 222–225 (2011).

    Article  CAS  PubMed  Google Scholar 

  34. Ernst, J. et al. Mapping and analysis of chromatin state dynamics in nine human cell types. Nature 473, 43–49 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Smith, S.B. et al. Rfx6 directs islet formation and insulin production in mice and humans. Nature 463, 775–780 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Zaret, K.S. et al. Pioneer factors, genetic competence, and inductive signaling: programming liver and pancreas progenitors from the endoderm. Cold Spring Harb. Symp. Quant. Biol. 73, 119–126 (2008).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Scott, R.A. et al. Large-scale association analyses identify new loci influencing glycemic traits and provide insight into the underlying biological pathways. Nat. Genet. 44, 991–1005 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Morris, A.P. et al. Large-scale association analysis provides insights into the genetic architecture and pathophysiology of type 2 diabetes. Nat. Genet. 44, 981–990 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Hindorff, L.A. et al. A Catalog of Published Genome-Wide Association Studies http://www.genome.gov/26525384.

  40. El-Assaad, W., et al. Saturated fatty acids synergize with elevated glucose to cause pancreatic β-cell death. Endocrinology 144, 4154–4163 (2003).

    Article  CAS  PubMed  Google Scholar 

  41. Shimabukuro, M., Zhou, Y.T., Levi, M. & Unger, R.H. Fatty acid–induced β cell apoptosis: a link between obesity and diabetes. Proc. Natl. Acad. Sci. USA 95, 2498–2502 (1998).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Hu, L. et al. Regulation of lipolytic activity by long-chain acyl-coenzyme A in islets and adipocytes. Am. J. Physiol. Endocrinol. Metab. 289, E1085–E1092 (2005).

    Article  CAS  PubMed  Google Scholar 

  43. Helgason, A. et al. Refining the impact of TCF7L2 gene variants on type 2 diabetes and adaptive evolution. Nat. Genet. 39, 218–225 (2007).

    Article  CAS  PubMed  Google Scholar 

  44. Cho, Y.S. et al. Meta-analysis of genome-wide association studies identifies eight new loci for type 2 diabetes in east Asians. Nat. Genet. 44, 67–72 (2012).

    Article  CAS  Google Scholar 

  45. Zaret, K.S. & Grompe, M. Generation and regeneration of cells of the liver and pancreas. Science 322, 1490–1494 (2008).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Whyte, W.A. et al. Master transcription factors and mediator establish super-enhancers at key cell identity genes. Cell 153, 307–319 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Lovén, J. et al. Selective inhibition of tumor oncogenes by disruption of super-enhancers. Cell 153, 320–334 (2013).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  48. Maurano, M.T. et al. Systematic localization of common disease-associated variation in regulatory DNA. Science 337, 1190–1195 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  49. Trynka, G. et al. Chromatin marks identify critical cell types for fine mapping complex trait variants. Nat. Genet. 45, 124–130 (2013).

    Article  CAS  PubMed  Google Scholar 

  50. Heinz, S. et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell 38, 576–589 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. Seo, S., Lim, J.W., Yellajoshyula, D., Chang, L.W. & Kroll, K.L. Neurogenin and NeuroD direct transcriptional targets and their regulatory enhancers. EMBO J. 26, 5093–5108 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  52. Bucher, P. et al. Assessment of a novel two-component enzyme preparation for human islet isolation and transplantation. Transplantation 79, 91–97 (2005).

    Article  PubMed  Google Scholar 

  53. McCulloch, L.J. et al. GLUT2 (SLC2A2) is not the principal glucose transporter in human pancreatic β cells: implications for understanding genetic association signals at this locus. Mol. Genet. Metab. 104, 648–653 (2011).

    Article  CAS  PubMed  Google Scholar 

  54. Servitja, J.M. et al. Hnf1α (MODY3) controls tissue-specific transcriptional programs and exerts opposed effects on cell growth in pancreatic islets and liver. Mol. Cell. Biol. 29, 2945–2959 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  55. van Arensbergen, J. et al. Derepression of Polycomb targets during pancreatic organogenesis allows insulin-producing β-cells to adopt a neural gene activity program. Genome Res. 20, 722–732 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  56. Langmead, B., Trapnell, C., Pop, M. & Salzberg, S.L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 10, R25 (2009).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  57. Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137 (2008).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  58. de Hoon, M.J., Imoto, S., Nolan, J. & Miyano, S. Open source clustering software. Bioinformatics 20, 1453–1454 (2004).

    Article  CAS  PubMed  Google Scholar 

  59. McLean, C.Y. et al. GREAT improves functional interpretation of cis-regulatory regions. Nat. Biotechnol. 28, 495–501 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  60. Blechinger, S.R. et al. The heat-inducible zebrafish hsp70 gene is expressed during normal lens development under non-stress conditions. Mech. Dev. 112, 213–215 (2002).

    Article  CAS  PubMed  Google Scholar 

  61. Meng, A., Tang, H., Ong, B.A., Farrell, M.J. & Lin, S. Promoter analysis in living zebrafish embryos identifies a cis-acting motif required for neuronal expression of GATA-2. Proc. Natl. Acad. Sci. USA 94, 6267–6272 (1997).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  62. Ravassard, P. et al. A genetically engineered human pancreatic β cell line exhibiting glucose-inducible insulin secretion. J. Clin. Invest. 121, 3589–3597 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  63. Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA 102, 15545–15550 (2005).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  64. Tena, J.J. et al. An evolutionarily conserved three-dimensional structure in the vertebrate Irx clusters facilitates enhancer sharing and coregulation. Nat. Commun. 2, 310 (2011).

    Article  CAS  PubMed  Google Scholar 

  65. Splinter, E., de Wit, E., van de Werken, H.J., Klous, P. & de Laat, W. Determining long-range chromatin interactions for selected genomic sites using 4C-seq technology: from fixation to computation. Methods 58, 221–230 (2012).

    Article  CAS  PubMed  Google Scholar 

  66. Rozen, S. & Skaletsky, H. Primer3 on the WWW for general users and for biologist programmers. Methods Mol. Biol. 132, 365–386 (2000).

    CAS  PubMed  Google Scholar 

  67. Boj, S.F., Parrizas, M., Maestro, M.A. & Ferrer, J. A transcription factor regulatory circuit in differentiated pancreatic cells. Proc. Natl. Acad. Sci. USA 98, 14481–14486 (2001).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  68. Gupta, S., Stamatoyannopoulos, J.A., Bailey, T.L. & Noble, W.S. Quantifying similarity between motifs. Genome Biol. 8, R24 (2007).

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  69. Rosenbloom, K.R. et al. ENCODE data in the UCSC Genome Browser: year 5 update. Nucleic Acids Res. 41, D56–D63 (2013).

    Article  CAS  PubMed  Google Scholar 

  70. ENCODE Project Consortium. A user's guide to the encyclopedia of DNA elements (ENCODE). PLoS Biol. 9, e1001046 (2011).

  71. Hansen, K.D., Irizarry, R.A. & Wu, Z. Removing technical variability in RNA-seq data using conditional quantile normalization. Biostatistics 13, 204–216 (2012).

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank J. Rios (IDIBAPS) for expert statistical advice, M. Parsons (Johns Hopkins University) for Tg(ins:mCherry)jh2 transgenics and R. Stein (Vanderbilt University), J. Habener (Harvard University) and G. Gradwohl (Institute of Genetics and Molecular and Cellular Biology) for MAFA, IDX1, and NGN3 DNA constructs. We thank the DIAGRAM and MAGIC consortia, the Singapore Prospective Study Program, the Singapore Consortium of Cohort Studies, the Singapore Indian Eye Study, the Singapore Malay Eye Study and Y.Y. Teo, E.S. Tai, T.Y. Wong, W.Y. Lim and X. Wang (National University of Singapore; funded by the National Medical Research Council of Singapore, Singapore Translational Researcher Award schemes, the Biomedical Research Council of Singapore and the National Research Foundation (NRF) Fellowship scheme). This work was carried out in part at the Centre Esther Koplowitz. This work was funded by grants from a European Foundation for the Study of Diabetes Lilly fellowship (L. Pasquali), the Ministerio de Economía y Competitividad (SAF2011-27086 to J.F., BFU2010-14839 and CSD2007-00008 to J.L.G.S.), the Innovative Medicines Initiative (DIRECT to M.I.M. and J.F.), the Andalusian Government (CVI-3488 to J.L.G.S.), the Biology of Liver and Pancreatic Development and Disease Marie Curie Initial Training Network (F.M. and J.F.), the Wellcome Trust (090532, 098381 and 090367 to M.I.M., 095101 to A.L.G., 101033 to J.F.), Juvenile Diabetes Research Foundation (31-2012-783 to T.B., F.P. and L. Piemonti) and Framework Programme 7 (HEALTH-F4-2007-201413 to M.I.M.).

Author information

Authors and Affiliations

Authors

Contributions

L. Pasquali, K.J.G., J.L.G.S., I.M.-E., F.M., M.I.M. and J.F. analyzed integrated data and wrote and edited the manuscript, which all authors have approved. İ.A., S.A.R.-S.,J.P.-C., J.G.-H., T.N., I.M.-E., C.G.-M., I.C., N.C., M.A.M. and J.J.T. performed and analyzed experiments. L. Pasquali, K.J.G., L.M., J.P.-C. and I.M. performed computational analysis. A.L.G., M.v.d.B., F.P., L. Piemonti, T.B. and P.R. provided materials and reagents. L. Pasquali and J.F. conceived and coordinated the project.

Corresponding author

Correspondence to Jorge Ferrer.

Ethics declarations

Competing interests

P.R. is a shareholder and consultant for Endocells. None of the other authors report a conflict of interest relevant to the contents of this report.

Integrated supplementary information

Supplementary Figure 1 Transcription factor ChIP-seq signals are specific and consistent in biological replicates.

(a) Schematic of transcription factor expression in three major pancreatic islet-cell types. (b) Assessment of antibody specificity by western blot using human islet nuclear extracts revealed expected migration patterns for all 5 transcription factors. As a control, western blots using the same antibodies but nuclear extracts obtained from Hela cells did not show specific bands (not shown). Molecular-mass markers (in kDa) are shown on the left. (c) Antibody assessment by immunofluorescence confocal microscopy of formaldehyde-fixed human pancreatic sections, showing expected islet-cell type nuclear localization of the tested transcription factors. FOXA2 was also found to localize to the pancreatic acinar tissue, as expected from its known expression profile. (d) Normalized ChIP-Seq signal correlation between two biological human islet replicates computed genome-wide over 1 Kb bins. r values represent Pearson correlation coefficients. (e) Transcription factor binding peaks identified at a stringent threshold (P < 10-10) in both biological replicates were retained for subsequent analysis.

Source data

Supplementary Figure 2 A transcription factor network in human islet cells.

(a) Number of high-confidence transcription factor peaks and logo of the enriched sequence motifs compared with the in vitro reported motif of each transcription factor, along with p-values for enrichment using HOMER. As expected, most enriched transcription factor sequence motifs underlying bound sites included those previously reported in vitro for the same transcription factors. High-confidence transcription factor peaks were defined as those called at P < 1×10-10 in replicate islet samples. (b) Transcription factor binding site distribution relative to gene annotations. Transcription factor binding sites were enriched near the 5′ end of annotated genes, although in absolute terms most binding sites were more distant. The relative distribution of the different compartments in the entire genome is shown as black bars. (c-f) Examples of transcription factor binding patterns at loci harboring the PDX1, FOXA2, NKX6-1, and MAFB genes, showing frequent co-occupancy and clusters of binding sites of the five transcription factors at multiple sites near their own genes and near each other's genes. Arrows indicate transcription factor binding to known enhancers (Area I-IV) in PDX1. (g) Over 90% of PDX1 high-confidence binding sites show co-occupancy by at least one other factor in at least one replicate sample, illustrating that islet transcription factors very frequently bind to shared locations. Co-occupancy is defined by overlap of peaks by at least 1 bp. Comparable findings were encountered when we assessed co-occupancy relative to the binding sites of the four other transcription factors (not shown). (h) Transcription factor co-occupancy was also computed genome-wide by correlating binding signals between all transcription factor pairs, or between replicates for the same transcription factor. This analysis is restricted to sites bound by NKX6.1 and/or PDX1, because both are predominantly present in β-cells and therefore ensure that the correlation analysis is restricted to binding events occurring in the same cell type. Comparisons with ChIP-Seq data for MEIS1 in an umbilical cord blood cell line is shown as a reference. Numbers represent the Pearson's r correlation coefficient. NA: unavailable data. (i) Enrichment of histone modification marks (H3K4me3, H3K27ac and H3K4me1) at the INS (insulin) locus highlight the high purity and appropriate differentiated state of the human pancreatic islets used throughout this study.

Supplementary Figure 3 Sample consistency and genomic location of accessible chromatin subclasses.

(a) Read density for FAIRE, H2A.Z, different active histone modification marks and CTCF in 6 Kb windows centered on the 5 different classes of accessible chromatin defined by K-median clustering as shown in Fig. 2a. C1 sites showed strong H3K4me3 enrichment, C2 sites showed monomodal H3K4me1 enrichment without H3K4me3 or H3K27ac enrichment, C3 sites showed bimodal H3K4me1 enrichment without H3K4me3 and strong H3K27ac enrichment, C4 sites showed strong CTCF occupancy, and C5 sites lacked enriched active histone modifications or CTCF binding. This analysis was performed with data from sample HI32, for which all marks were available. (b) FAIRE, H2A.Z, histone modification signals, and CTCF signals were comparable in replicate human islet samples. Read densities in 6 Kb windows were centered on the 5 different classes of open chromatin defined by K-median clustering of sample HI32 as shown in Fig. 2a, and the read density in the same genomic sites is shown for replicate samples. Pearson's correlation coefficients for read densities between replicate samples in these sites are shown beneath. (c) Distribution relative to gene TSS of accessible chromatin classes shows that as expected C1 (promoter-like) accessible chromatin sites are preferentially located at the 5'end of annotated genes, unlike other accessible chromatin sites.

Supplementary Figure 4 Genomic features and chromatin state at PDX1-binding sites.

A large fraction of PDX1 binding sites are associated with active enhancer chromatin, although PDX1 binds to all major classes of accessible chromatin, and in all cases shows a similarly high evolutionary sequence conservation or DNA-binding recognition motif enrichment. From left to right: Number and percentage of PDX1-bound sites that overlap with each accessible chromatin class, percentage of PDX1-bound sites that fall into each distance interval from the transcriptional start site (TSS), average PhastConst sequence conservation scores at -3 to +3 Kb relative to the binding site center, and fold enrichment of the PDX1 consensus recognition motif in PDX1-bound sites vs. non-bound sites of the same accessible chromatin class.

Supplementary Figure 5 Transcription factor binding at non-C3 accessible chromatin sites does not occur predominantly in islet-specific transcribed genes.

(a) Examples of islet-specific transcription factors binding to the 5' flanking regions of two ubiquitously expressed genes. Expression in non-islet tissues is depicted as an average signal for all 14 tissues for simplicity, yet shows comparably high levels in all individual tissues. (b) Genes that are bound by one or more islet-specific transcription factor at their promoter (C1 chromatin), but lack active enhancer (C3) chromatin at <25 Kb from TSS, tend to be transcriptionally active at a similar level in human islets and non-islet tissues. The boxes show Log2 ratios of quantile-normalized expression levels in islets vs. indicated non-islet tissues, and depict the interquartile range (IQR). Whiskers extend to either the maximum value or to 1.5 times the IQR, and notches indicate 95% confidence intervals of the median. (c) Density of C4 (CTCF-bound) and C5 (no active histone modifications) accessible chromatin sites bound by two or more transcription factors in the vicinity of the TSSs of 1,000 most islet-specific genes, ubiquitously active genes, or islet inactive genes. The results are depicted at the same scale as Fig 2c, and show that, unlike C1 and C3, transcription factor binding to C4 and C5 is not enriched in any of the three gene subsets. (d) Genes that are bound by three or more islet-specific transcription factor at promoter (C1) accessible chromatin, but lack active enhancers (C3) chromatin <25 Kb from TSS, tend to be transcriptionally active at similar level in human islets and non islet tissues. The results are presented as described above for panel b. The statistical analysis of this data was performed with pooled non-islet data, and is shown in Fig 2g.

Supplementary Figure 6 Pancreatic islet enhancer clusters.

(a) Distribution of distances between adjacent C3 sites, compared to adjacent randomized C3 sites. We calculated the distances between all adjacent C3 sites, and between adjacent randomized C3 sites taken from 1,000 iterations that used only the mappable genome of each individual chromosome separately, excluding sex chromosomes. The graph depicts the distribution of intersite distances for chromosome 1 as an example, with a vertical black line marking the 25th percentile of intersite distances for the randomized C3 sites. This threshold was used to create clusters, which were formed by three or more C3 sites that were maximally separated by that distance. (b) Size distribution of the 3,677 clusters of islet C3 sites, formed by 3-54 C3 sites each. The size range of the C3 clusters is 2 to 322 Kb (median 23 Kb). The clusters formed by randomized C3 sites are shown in black for comparison. (c) Gene expression in β-cells, islets, and 14 non-pancreatic tissues for genes that do not contain a C3 site, those that have a clustered C3 site, or an orphan C3 within < 25 Kb from the TSS. The boxes show RNA expression levels in islets or the collection of non-islet tissues, expressed as the interquartile range (IQR). Whiskers extend to either the maximum value or to 1.5 times the IQR, and notches indicate 95% confidence intervals of the median. All p values resulted from a Wilcoxon Rank Sum test.

Source data

Supplementary Figure 7 Pancreatic islet transcription factor–bound enhancer clusters.

(a) Enriched transcription in human islets is associated with clusters of enhancers that exhibit high transcription factor occupancy. We devised a transcription factor occupancy score that measures the extent to which the enhancers that form part of clusters are bound by multiple transcription factors (see methods), and used it to divide clusters in quartiles. Clusters were linked to a gene if they were located within < 25 Kb from the gene's TSS. The box plots show the expression ratio in islets vs. non-islet tissues for all genes in each category, expressed as the IQR. Whiskers extend to either the maximum value or to 1.5 times the IQR, and notches indicate 95% confidence intervals of the median. (b) Genes that were linked to clusters of enhancers that exhibit high transcription factor occupancy (clusters showing the upper 50% average transcription factor occupancy scores) show islet-enriched expression. The boxes show Log2 ratios of quantile-normalized expression levels in islets vs. indicated non-islet tissues, and depict the IQR. Whiskers extend to either the maximum value or to 1.5 times the IQR, and notches indicate 95% confidence intervals of the median. (c) Genes associated with clusters of enhancers that are highly bound by islet-specific transcription factors are enriched in typical β-cell functions. The analysis was performed by linking clusters to nearby genes with Genomic Regions Enrichment of Annotations Tool (GREAT) using default parameters, and the results illustrate the most significantly enriched functional annotation categories. (d) Most genes that have established roles in β-cell function and identity are associated with islet transcription factor bound enhancer clusters. A complete list with references is shown in Supplementary Table 2.

Source data

Supplementary Figure 8 In vivo functional validation of islet enhancers.

Human islet C3 (active enhancer) sites drove cell-specific enhancer activity in zebrafish. (a) Merged images of YFP and brightfield channels of zebrafish embryos injected with five human C3 (active enhancer) sites, or a control construct containing the hsp70 promoter and a region lacking enhancer function. YFP expression is observed in the pancreatic islet (pi), neurons (ne), lens (lens), and floor plate (fp). We injected each construct in >200 eggs in at least 3 independent experiments. The quantitative analysis is provided in Supplementary Table 9. (b) Islet-specific patterns driven by sites C3-3, C3-4 and C3-5 were confirmed by establishing stable transgenic lines (white arrows point to pancreatic islet expression). C3-3 derived lines show expression in the islet, lens (lens), ventral neural tube (nt) including floor plate (fp) and hindbrain (hi). C3-4 derived lines show expression in the islet and a subset of midbrain neurons (ne). C3-5 derived lines show activity in the pancreatic islet, pronephric ducts (pd) and broad pancentral nervous system activity (cns) including hindbrain (hi) and olfactory bulbs (ob). Expression in the lens is ectopic activity from hsp70 core promoter. All embryos are 72 hpf, oriented dorsally or laterally anterior to the right. The table shown in (c) provides information on tested C3 sites.

Supplementary Figure 9 Functional MAFB targets.

Human β cells (EndoC-βH1) were transduced with lentiviral vectors expressing two independent RNA hairpins that target MAFB RNA (MAFB shRNAs) or four independently transduced negative control non-targeting shRNA sequences (NT). (a) Quantitative PCR showed that MAFB knockdown with the two shRNAs (black and grey dots) led to 64% and 55% inhibition of MAFB mRNA, respectively, relative to the pooled data from NT shRNAs (white dots). The medians are indicated by an horizontal line, p-values were obtained with Student's t test. (b-c) Islet-enriched genes ROBO2 and G6PC2 are depicted as examples that support the observation that genes associated with MAFB-bound enhancer clusters were enriched among downregulated genes after knockdown of MAFB. Note that both loci contained prominent enhancer clusters, were bound by MAFB, and transcripts showed significant downregulation with two MAFB shRNAs, compared with NT shRNAs. (d,e) Examples of genes bound by MAFB at promoter open chromatin, which like other MAFB targets of this class showed no changes in expression upon MAFB knockdown. NS: P > 0.05. UNT: untreated cells. NT: non-targeting shRNA sequences.

Supplementary Figure 10 Examples of 4C-seq analysis.

4C-Seq analysis at different loci that contain enhancer clusters showing frequent interactions between distal C3 enhancers and promoters of islet-enriched genes. The red triangle indicates the promoter viewpoint for each 4C-Seq experiment. Genomic sites that interact with viewpoints are represented by a track labeled 4C-Seq sites. The islet regulome track depicts accessible chromatin sites following the colour code shown in the upper right quadrant.

Source data

Supplementary Figure 11 Motif combinations enriched in human islet enhancers.

(a) Enrichment of 3-motif combinations in clustered islet enhancers vs. non-islet enhancers derived from 9 non-islet cell types. We examined clustered islet C3 sites (or islet enhancers) vs. non-islet enhancers and computed the frequency of all possible combinations of 3 motifs from the 46 islet-enriched motifs, restricting the motif search window to +/- 250 bp from the center of the genomic site. The data is rank-ordered so that combinations that are most enriched in islet enhancers are shown on top. The red line represents the median enrichment, and orange lines depict the range of enrichment values for all nine islet/non-islet ratios. For comparison, we identified combinations of 3 motifs enriched in keratinocyte (NHEK) enhancers, and studied their enrichment in islet enhancers (blue line). (b) The most enriched 3-motif combinations in islet active enhancers showed markedly higher enrichment than the individual motifs that compose them. The box plots depict the motif enrichment distribution values in islet vs. active enhancer elements in the 9 non-pancreatic cell types. Motifs for RFX and/or pioneer FOXA factors are invariably present in the most enriched combinations. (c) Aggregation plot that represents the frequency density of the 100 most islet-enriched 3-motif combinations in enhancers from different cell types (islets and 4 four non-islet cell lines for which FAIRE and H2A.Z datasets were available to create plots centered on accessible chromatin with the same criteria). Randomized clustered islet enhancers are shown for comparison. (d) All instances of the 10 most enriched motif combinations in human islet enhancers where identified in the mouse genome without considering direct sequence orthology, and the resulting sites were associated to nearby genes with Genomic Regions Enrichment of Annotations Tool (GREAT) using default parameters. The graph depicts the median mRNA expression in mouse islets and non-islet tissues for genes that were linked to the motif combinations. The notches surrounding each median line interval represent +/-1.58 IQR/sqrt(n) where IQR is the difference between the 1th and the 3rd quantiles. When notches from two distributions do not overlap this is taken as strong evidence for a significant difference between two medians. On the x-axis gene expression is shown as RPKM values quantile-normalized across all tissues. MEF: mouse embryonic fibroblast. Wilcoxon Rank Sum P < 9.6×10-31 for the comparsion between pancreatic islets vs. non-islet tissues.

Supplementary Figure 12 Enrichment of pruned T2D and fasting glycemia association in variants overlapping islet regulome sites.

HapMap association data for T2D (DIAGRAMv3) and FG level (MAGIC) were pruned retaining only the most significant variant in each LD block (r2>.2). Variants overlapping each class of islet sites (C1-C5, clustered C3, orphan C3) were then evaluated for fold enrichment over matched background variants at several p value thresholds. An increasing percentage of variants overlap C3 and clustered C3 sites at more significant p value thresholds for both T2D (top left panel) and FG (bottom left panel). Similar patterns of enrichment were observed for C1 sites in T2D data. When removing known European T2D/FG loci, variants overlapping C3, clustered C3 and C1 sites remain enriched at the most significant p-value thresholds for T2D only (right top panel).

Source data

Supplementary Figure 13 ACSL1: a novel candidate locus harboring a common fasting glycemia and T2D risk variant in a clustered enhancer.

(a) Regional plot of DIAGRAM variants and islet regulome elements at ACSL1 (r2 values based on 1KG CEU with rs735949). SNP rs735949 is strongly associated with T2D (DIAGRAM and Metabochip combined P = 3.7×10-6) and independently with FG level (MAGIC FG P=1.6x10-5), yet does not surpass conventional GWA significance thresholds. (b) SNP rs735949 is the lead SNP in this locus and overlaps an intronic, isletselective C3 enhancer that is bound by multiple islet transcription factors. Non-islet chromatin state data is taken from Ernst et al.. This SNP is in strong linkage disequilibrium with another SNP, within the same enhancer, which overlaps a nuclear receptor-like motif that is highly enriched in islet enhancers (rs72695654; 1KG CEU r2= .94) (c). (d) Electrophoretic mobility shift assay shows that the nucleotide change at rs72695654 abolishes sequence-specific binding of a protein complex in MIN6 β cells, supporting an islet regulatory function of this variant. Competition gradients identified by the grey triangle correspond to 5, 50, and 100-fold excess of “cold” competitor probe.

Supplementary Figure 14 Transcription factor and chromatin maps of the locus containing rs7903146 at TCF7L2.

(a) T2D associated SNP rs7903146 at TCF7L2 locus, previously shown to affect islet chromatin accessible state and enhancer function maps to a C3 active enhancer site bound by NKX2.2, FOXA2 and MAFB in human pancreatic islets. (b) Active enhancer chromatin at rs7903146 is specific to human pancreatic islets, as it is not observed in Roadmap Epigenomics datasets including >200 human tissues or cellular types.

Supplementary Figure 15 The human islet regulome browser.

This open access browser enables viewing of human islet transcription factors binding, human islet open chromatin states, clusters of enhancers, islet motifs, and DIAGRAM T2D association p-values, or MAGIC FG p-values at desired levels of resolution throughout the genome. In addition to these standard tracks, it is also possible for users to upload their own variants or regions sets for temporary display. (a) Front panel of the browser. (b) Example of a locus, depicting the T2D-associated region at CDKN2B/CDKN2B-AS1 that has a highly associated SNP mapping to a transcription factor bound C3 site. Tables with information on the coordinates of the regulatory elements and transcription factor binding for the browsed regions are visualized on the bottom panel and are available for download. DIAGRAM T2D-association p-values are represented by red dots, whereas MAGIC FG-association p-values are represented by blue dots. For each variant the color intensity of the dot is proportional to –Log p-value of association. Vertical colored lines depict different chromatin states. Black lines point to transcription factors binding sites and their intensity is proportional to the number of bound transcription factors. Islet-specific genes are shown in dark grey. (c) At a zoom-in resolution of less than 1 Kb per window, islet enriched motifs described in Supplementary Table 3 are shown in C3 sites and their sequence and genomic location are visualized in a table format in the middle panel. The human islet regulome browser is available at www.isletregulome.org.

Supplementary Figure 16 Examples of T2D- and fasting glycemia–associated variants located in islet active enhancers.

Screenshots from the human islet regulome browser showing T2D and/or FG-associated loci in which the associated variants directly map to C3 active enhancer sites (grey boxes). GWAs DIAGRAM T2D-associated (in red) and MAGIC FG-associated variants (in blue) are shown along with transcription factors binding and chromatin states as described in Supplementary Figure 15.

Supplementary information

Supplementary Text and Figures

Supplementary Figures 1–16 and Supplementary Tables 1–10 (PDF 27597 kb)

Source data

Rights and permissions

Reprints and permissions

About this article

Cite this article

Pasquali, L., Gaulton, K., Rodríguez-Seguí, S. et al. Pancreatic islet enhancer clusters enriched in type 2 diabetes risk-associated variants. Nat Genet 46, 136–143 (2014). https://doi.org/10.1038/ng.2870

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/ng.2870

This article is cited by

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing