Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Jul 13;8:16058.
doi: 10.1038/ncomms16058.

Platelet Function Is Modified by Common Sequence Variation in Megakaryocyte Super Enhancers

Romina Petersen  1   2 John J Lambourne  1   2 Biola M Javierre  3 Luigi Grassi  1   2   4 Roman Kreuzhuber  1   2   5 Dace Ruklisa  1   2   6 Isabel M Rosa  1   2 Ana R Tomé  1   2 Heather Elding  7   8 Johanna P van Geffen  9 Tao Jiang  10 Samantha Farrow  1   2 Jonathan Cairns  3 Abeer M Al-Subaie  1   2   11 Sofie Ashford  1   2   4 Antony Attwood  1   2   4 Joana Batista  1   2 Heleen Bouman  7 Frances Burden  1   2 Fizzah A Choudry  1   2 Laura Clarke  5 Paul Flicek  5 Stephen F Garner  2 Matthias Haimel  4   12 Carly Kempster  1   2 Vasileios Ladopoulos  1 An-Sofie Lenaerts  13   14 Paulina M Materek  13   14 Harriet McKinney  1   2 Stuart Meacham  1   2   4 Daniel Mead  7 Magdolna Nagy  9 Christopher J Penkett  1   2   4 Augusto Rendon  1   2   15 Denis Seyres  1   2   4 Benjamin Sun  10 Salih Tuna  1   2   4 Marie-Elise van der Weide  1   2 Steven W Wingett  3 Joost H Martens  16 Oliver Stegle  5 Sylvia Richardson  6 Ludovic Vallier  14   17 David J Roberts  18   19   20 Kathleen Freson  21 Lorenz Wernisch  6 Hendrik G Stunnenberg  16 John Danesh  7   8   10   22 Peter Fraser  3   23 Nicole Soranzo  1   7   8   22 Adam S Butterworth  8   10   22 Johan W Heemskerk  9 Ernest Turro  1   2   4   6 Mikhail Spivakov  3 Willem H Ouwehand  1   2   7   8   22 William J Astle  1   2   6   10   22 Kate Downes  1   2 Myrto Kostadima  1   2   5 Mattia Frontini  1   2   22
Free PMC article

Platelet Function Is Modified by Common Sequence Variation in Megakaryocyte Super Enhancers

Romina Petersen et al. Nat Commun. .
Free PMC article


Linking non-coding genetic variants associated with the risk of diseases or disease-relevant traits to target genes is a crucial step to realize GWAS potential in the introduction of precision medicine. Here we set out to determine the mechanisms underpinning variant association with platelet quantitative traits using cell type-matched epigenomic data and promoter long-range interactions. We identify potential regulatory functions for 423 of 565 (75%) non-coding variants associated with platelet traits and we demonstrate, through ex vivo and proof of principle genome editing validation, that variants in super enhancers play an important role in controlling archetypical platelet functions.

Conflict of interest statement

P. Flicek is a member of the scientific advisory board of Fabric Genomics, Inc. All other authors declare no competing financial interests.


Figure 1
Figure 1. Unique three-dimensional regulatory landscapes define megakaryopoiesis and erythropoiesis.
(a) Top panel, MK ATAC-seq peak (126,428) dynamics from HSCs through CMPs and MEPs, as well as EBs open chromatin as determined by DNase-seq (light green and grey, open and closed chromatin, respectively). H3K27ac in CD34+ haematopoietic stem and progenitor cells (HSPCs, data from ROADMAP), enhancer regions (Enh) and CTCF binding sites in MKs have been added for comparison (dark green, present). Categories: (I) Open chromatin regions present in all five cell types. In MKs 24,318/47,502 (51.2%) of ATAC-seq peaks were CTCF-binding sites and 25,548/47,502 (53.8%) of these were enhancers. (II) Open chromatin regions present from HSCs to MKs, but absent from EBs. (III) Open chromatin regions present either only in MKs or (IV) only in MKs and EBs. Bottom panel, representative examples of open chromatin peaks for the four categories. (b) Categorization of elements based on differences in H3K27ac signal intensities: black, nonsignificantly different (n=∼57,000); blue and red, significantly higher in MKs (n=6,810) and EBs (n=5,237), respectively. (c) Heatmap of 1,546 genes differentially expressed (DE) in RNA-seq analysis of MKs (left) and EBs (right). (d) Circular plot representing the interactions between DE genes (MK-DE, light blue; EB-DE, red), differentially acetylated (DA) elements (MK-DA, green; EB-DA, brown) and differentially interacting (DI) elements (MK-DI, dark blue; EB-DI, orange) on the outer arcs. Inner arc colours follow the same colour scheme and indicate overlap of attributes for these categories. Connections reflect a concordance of fold changes: DE genes in MKs tend to interact with regions specifically acetylated in MKs compared with EBs and vice versa.
Figure 2
Figure 2. Identification of SEs their effects on gene expression and their opening dynamics.
(a) Schematic of the stitching process to identify enhancer clusters and ranking based on H3K27ac signal intensities. (b) Overlap of SE sets in MKs and EBs. (c) Gene expression, in MKs, for genes connected to TEs only (blue), SE constituents only (pink), or a combination of TEs and SE constituents (yellow) (box plot: line indicates median, upper and lower box margins indicate first and third quartile). Top row of schematic shows a gene regulated by five TEs, second row shows a gene regulated by five SE constituents and the bottom rows show genes regulated by different combinations of five TEs and SE constituents. P-values for Wilcoxon test between different categories are in Supplementary Table 3. (d) Opening dynamics of MK SEs constituents during HSC differentiation. Open chromatin regions overlapping with MK SE constituents in HSCs, CMPs, MEPs and EBs. H3K27ac in CD34+ haematopoietic stem and progenitor cells (HSPCs) and CTCF-binding sites in MKs added for comparison (colour legend as in Fig. 1a).
Figure 3
Figure 3. GWAS non-coding sentinel variants associated with platelet traits are enriched in SEs of MKs.
(a) Categorization of sentinel variants associated with CBC-P (count, mean volume, volume width distribution and platelet crit (mean volume × count)) by location; exonic or splice site (light blue), intronic or intergenic (yellow) and promoter (green). Number of intronic or intergenic SNPs localized to SE constituents and TEs, detailed description of annotation in Supplementary Fig. 8a. (b) Venn diagram showing the overlap of the sets of genes to which the CBC-P-associated variants were assigned by variant effect predictor (VEP, green) and by the analysis reported in this study (orange). (c) Density distribution of the genomic distance between a CBC-P sentinel SNP and the transcriptional start site (TSS) of the gene it has been assigned to by VEP (green) and the approach used in this study (orange). For genes with several TSSs, the mean position of all TSSs was used. (d) P-values characterizing the significance of difference between the prevalence of CBC-P versus CBC-red cell trait-associated non-coding sentinel variants within SE and other enhancers. All P-values are based on a permutation test involving 999,999 simulations of locations of significantly associated sentinel variants. Each dot corresponds to a comparison of two categories of enhancers—the cell types of both enhancers are indicated on y axis and the enhancer type is denoted on x axis. The surface area of each dot is proportional to the number of significant association signals either for CBC-P or CBC-red cell traits residing within either of the two enhancers being compared (pleiotropic variants are not counted). Number of variants tested for each category available in Supplementary Table 10.
Figure 4
Figure 4. Association between SE-localized sentinel variant rs3557 and thrombus phenotypes.
(a) Chr1 1q23.3 locus view comprising FCER1G and three other genes. From top to bottom: H3K27ac signal track and SE location in MKs (blue) and EBs (red); *position of sentinel variant rs3557 and genes in green. Scale bar in bottom right corner represents 2 kb. Maximum read signal scale 60 for each track. (b) Schematic representation of the glycoprotein (GP)VI/Fc receptor γ-chain signalling receptor complex for collagen on platelets. (ch) Associations of genotypes of rs1613662 and rs3557 with the residuals of platelet function phenotypes, after adjustment for covariates. Dots show distribution of the phenotypic residuals; central lines show genotype-specific mean estimates and whiskers represent 95% confidence intervals. (c,e) Associations with platelet membrane level of GPVI after linear adjustement for the interaction of logged mean platelet volume and sex (rs1613662: GG=36, GA=221, AA=587, likelihood ratio additive P=1.6 × 10−27; rs3557, TT=696, TG=139, GG=9, likelihood ratio additive P=4.6 × 10−5). (d,f) Associations of fibrinogen binding to integrin αIIbβ3 after platelet activation with CRP-XL, adjusted for sex (rs1613662: GG=49, GA=381, AA=992, likelihood ratio additive P=1.6 × 10−7; rs3557, TT=1,175, TG=229, GG=18, likelihood ratio additive P=4.6 × 10−72). (g,h) Associations for rs1613662 and rs3557 with thrombus formation upon flowing whole blood over collagen III in microchambers, measured by quantile-normalized sex-adjusted platelet surface area coverage (PltSac; GG=1, GA=29, AA=63, likelihood ratio additive P=1.8 × 10−2) and quantile-normalized sex-adjusted activation of integrin αIIbβ3 (ITG; TT=67, TG=24, GG=2, likelihood ratio additive P=3.4 × 10−3), respectively.
Figure 5
Figure 5. Effect of the SE-localized platelet trait associated sentinel variant rs2363877 on VWF and CD9 protein abundance.
(a) Chr12p13.31 locus view comprising VWF, CD9 and two other genes. From top to bottom: H3K27ac signal track and SE locations in MKs (blue) and EBs (red). Region of SE deleted by genome-editing (black); positions of sentinel variant rs2363877(*) and genes (green). Scale bar in bottom right corner represents 10 kb. Maximum read signal scale 60 for each track. (b,c) Associations of variant rs2363877 with (b) concentration of VWF in platelets (Y axis ng μl−1 normalized against total protein content; for subjects of genotypes: GG, n=20; GA, n=47; AA, n=26; likelihood ratio, P=10.0 × 10−5) and (c) CD9 abundance on platelet surface (y axis mean fluorescence intensity (MFI) adjusted for mean platelet volume; for subjects of genotypes: GG, n=122; GA, n=165; AA, n=78; likelihood ratio, P=1.3 × 10−6). Lines indicate mean, whiskers indicate 95% confidence interval. (d) Transcript levels of VWF and CD9 in MKs obtained by forward programming of wild type and genome-edited pluripotent stem cells (n=3 biological replicates each in triplicate; error bars generated from s.e. calculated from delta Ct value across technical and biological replicates, Student’s t-test *P=2.2 × 10−2 and ***P=5.0 × 10−4).

Similar articles

See all similar articles

Cited by 13 articles

See all "Cited by" articles


    1. Astle W. J. et al. . The allelic landscape of human blood cell trait variation and links to common complex disease. Cell 167, 1415–1429 e19 (2016). - PMC - PubMed
    1. Chen L. et al. . Transcriptional diversity during lineage commitment of human blood progenitors. Science 345, 1251033 (2014). - PMC - PubMed
    1. Kautz L. & Nemeth E. Molecular liaisons between erythropoiesis and iron metabolism. Blood 124, 479–482 (2014). - PMC - PubMed
    1. Kaushansky K. Lineage-specific hematopoietic growth factors. N. Engl. J. Med. 354, 2034–2045 (2006). - PubMed
    1. Maurano M. T. et al. . Systematic localization of common disease-associated variation in regulatory DNA. Science 337, 1190–1195 (2012). - PMC - PubMed

Publication types