(a) Regional association plot of TCF21 locus showing results from CARDIoGRAM meta-analysis in Caucasians, identifying rs12190287, and also depicting rs12524865. (b) Integrated workflow to identify CHD risk-associated mechanisms, with emphasis of approaches used in the study (blue boxes). CEU: Caucasians of European descent from UT, USA; EA: East Asians; eQTL: Expression quantitative trait loci; ASE: Allele-specific expression, TFBS: Transcription factor binding site. (c) Linkage disequilibrium (LD) plot of the TCF21 locus at 6q23.2 from 1568 Metabochip genotyped Europeans (ADVANCE replication cohort) with imputation from HapMap Phase II and III CEU, showing eSNPs for TCF21, and rs2327429. White to black squares represent increasing r² values. (d) The risk haplotype block containing all of the eSNPS (in red) at the TCF21 locus has a combined frequency of 0.366. r² values in LD with rs12190287 shown as a percentage in parentheses. Risk and protective alleles determined from CARDIoGRAM meta-analysis phenotypic data. (e) ENCODE ChIP-seq active chromatin histone modification data from 7 cell lines including promoter and enhancer marks, H3K4me1, H3K4me3, and H3K27ac. ENCODE tracks for DNase hypersensitivity, RNA-seq, and TF binding ChIP-seq data. Green box surrounds peaks overlapping rs12190287.

(a) Transcriptional activity of rs12190287-C and G variants were determined in heterozygous primary human coronary artery smooth muscle (HCASM), rat aortic smooth muscle cells (RASM) and HEK, HepG2, and A7r5 cell lines. pLuc-MCS vector containing the putative enhancer region for each rs12190287 variant (C-Luc and G-Luc) was transfected for 24 hours and the ratio of firefly and Renilla luciferase activities were normalized to empty reporter (pLuc). *P<0.01 versus G-Luc for each condition. (b) Dual-luciferase assay of wildtype rs12190287 enhancer or mutants, DelT or T/A (as shown), transfected in A7r5 as described above. *P<0.001 versus G-Luc for each condition. (c) Electrophoretic mobility shift assays (EMSA) showing protein binding to [γ³²P]ATP-labeled rs12190287 C/G probes incubated with nuclear extract (NE) from various cell types, along with 100× excess C, G, or negative control (C, G, or Ctrl comp) unlabeled probe as competitor. Arrows and bar-headed lines represent specific and non-specific shifted complexes, respectively. (d) Dual-luciferase assay of rs12190287 C/G enhancer co-transfected with empty vector (Empty) or constitutively active protein kinase A (PKA) or mitogen-activated protein kinase (MEKK) in A7r5 using consensus CRE and AP-1 reporters as positive controls. *P<0.001 versus C-Luc+Empty or CRE-Luc+Veh or AP1-Luc+Veh where indicated. (e) EMSA of rs12190287 C/G with respective competition showing protein binding using control A7r5 or PMA treated A7r5 NE. Values are mean ± SD from triplicates. Similar results were observed from three independent experiments.

(a) Dual-luciferase assay of rs12190287-C and G variants determined in A7r5 cells transfected with rs12190287 enhancer (C-Luc and G-Luc), consensus AP-1 reporter (AP1-Luc) or empty reporter (pLuc), along with human c-JUN, JUND, or ATF3 expression vectors and measured after 24 hours. *P<0.0001 versus C-Luc, G-Luc or −1-Luc+Empty where indicated. (b) Dual-luciferase assay of rs12190287 enhancer in A7r5 cells transfected with empty vector (Empty) or dominant negative expression constructs for human CREB (CREBΔ), c-JUN (JunΔ), or ATF3 (ATF3Δ). *P<0.01 versus C-Luc+Empty. (c) Dual-luciferase assay of rs12192087 enhancer in heterozygous HCASMC transfected with small interfering RNA duplexes (siRNA) against c-JUN, JUND, ATF3 or a negative control (Neg si). *P<0.01 versus C-Luc or G-Luc+Neg si. (d) EMSA of rs12190287 C/G enhancer showing specific shifted complexes (arrow) and JunD and c-Jun super-shifted complexes (bold arrows). Bar-head line denotes non-specific shifted complexes. Values are normalized mean ± SD from triplicates. Similar results were observed from three independent experiments.

(a) TaqMan based qRT-PCR results showing relative mRNA expression levels of total TCF21 in heterozygous HCASMC treated with TGF-β1, PDGF-BB or PMA for indicated times. Expression levels were normalized to 18S and expressed as fold change from 0 h. *P<0.0001 versus 0 h. (b) TaqMan qRT-PCR expression results of c-JUN, JUND, and ATF3 in heterozygous HCASMC treated TGF-β1 or PDGF-BB for the indicated times. (c) Allele-specific TaqMan qRT-PCR based expression of TCF21 at rs12190287 shown as normalized allelic ratio C/G in heterozygous HCASMC treated with either TGF-β1 or PDGF-BB for the indicated times. *P<0.01 versus 0 h. Values are mean ± SD of triplicates. Similar results were observed from three independent experiments.

(a) Total enrichment of c-Jun, JunD, and ATF3 at rs12190287 determined by chromatin immunoprecipitation (ChIP) in heterozygous HCASMC treated with PDGF-BB (20 ng/ml) or vehicle (Control) for 6 hrs. (b) Allele-specific enrichment of c-Jun, JunD and ATF3 at rs12190287 determined by HaploChIP in heterozygous HCASMC treated with PDGF-BB, shown as normalized allelic-ratio C/G. (c–d) AP-1 positive and negative control regions for c-Jun, JunD, and ATF3 enrichment at FOSB (c) and MYOG (d) promoters, respectively. (e) Total enrichment of histone modifications H3K4me1, H3K4me3, H3K27ac, and H3K27me1 at rs12190287 determined by ChIP in heterozygous HCASMC treated with PDGF-BB. (f) Allele-specific enrichment of H3K4me1, H3K4me3, H3K27ac, and H3K27me1 at rs12190287 determined by HaploChIP in heterozygous HCASMC treated with PDGF-BB, shown as normalized allelic ratio C/G. *P<0.001 versus Control for each condition. Values are mean ± SD from triplicates. Similar results were observed from 3–4 independent experiments. For HaploChIP experiments, separate heterozygous individual HCASMCs were used for each replicate experiment.

(a) Dual-luciferase assay of rs12190287 C/G enhancer transfected with human WT1-B (−KTS), WT1-D (+KTS) expression constructs in A7r5 cells, measured after 24 hours. *P<0.01 versus C-Luc or G-Luc+Empty. (b) Dual-luciferase assay of rs12190287 C/G enhancer co-transfected with human c-JUN and WT1-B or WT1-D expression constructs in A7r5 cells. AP1-Luc reporter was used as a positive control. *P<0.05 versus C-Luc or G-Luc+cJun. (c) Dual-luciferase assay of rs12190287 C/G enhancer transfected in heterozygous HCASMC with siRNA against WT1 (−/+KTS) compared to negative control (Neg si). *P<0.05 versus C-Luc or G-luc+Neg si. (d) TaqMan based qRT-PCR results showing relative human WT1 mRNA expression levels in heterozygous HCASMC treated with TGF-β1 or PDGF-BB for the indicated times. (e) Total enrichment of WT1 at rs12190287 enhancer, FOSB or MYOG promoter regions determined by chromatin immunoprecipitation (ChIP) in heterozygous HCASMC treated with PDGF-BB for 6 hrs. Values represent fold change relative to enrichment with IgG control. *P<0.01 versus Control WT1. (f) Allele-specific enrichment of WT1 at rs12190287 determined by HaploChIP in heterozygous HCASMC treated with PDGF-BB, shown as normalized allelic-ratio C/G. *P<0.0005 versus Control WT1. Values are mean ± SD from triplicates. Similar results were observed from three independent experiments.

(a) Linkage disequilibrium (LD) plot from ∼2400 Metabochip genotyped East Asian samples from the TAICHI study (HALST cohort) with imputation from HapMap Phase II and III Han Chinese (CHB), showing eSNPs for TCF21 and distinct haplotype blocks containing rs12524865 and rs12190287. White to black squares represent increasing r² values, also shown in blocks. (b) Dual-luciferase assay of rs12524865 C/A enhancer transfected in A7r5 and treated with adenylyl cyclase activator, forskolin (Fsk) or PKC activator, phorbol-12-myristate-13-acetate (PMA) for 4 hours. Relative luciferase activities measured after 24 hours. *P<0.01 versus C-Luc+Veh. (c) Total enrichment of c-Jun, JunB, JunD, and ATF3 at rs12524865 determined by chromatin immunoprecipitation (ChIP) in HCASMC treated with PDGF-BB or vehicle (Control) for 6 hours. *P<0.005 versus Control for each condition. (d) Total enrichment of histone modifications H3K4me1, H3K4me3, H3K27ac, and H3K27me1 at rs12524865 determined by ChIP in HCASMC treated with PDGF-BB. *P<0.001 versus Control for each condition. Values are mean ± SD from triplicates. Similar results were observed from three independent experiments.

Individuals carrying risk alleles for rs12190287 or rs12524865 at 6q23.2 are expected to have increased TCF21 expression upon stimulation of PDGFR-β by PDGF-BB in coronary artery smooth muscle cells, due to increased enrichment of active histone modifications (represented by closed and open diamonds) leading to an open chromatin conformation, allowing binding of an active AP-1 TF complex containing various combinations of c-Jun, JunD, and ATF3. WT1 functions as a transrepressor of this active complex at rs12190287, whereby WT1 may fine-tune the spatial and temporal activation of TCF21 expression. Multiple kinases other than mitogen-activated kinase (MEKK) may be involved in the activation and recruitment of AP-1 complexes to these binding sites.