Blood pressure-associated polymorphism controls ARHGAP42 expression via serum response factor DNA binding

Abstract

We recently demonstrated that selective expression of the Rho GTPase-activating protein ARHGAP42 in smooth muscle cells (SMCs) controls blood pressure by inhibiting RhoA-dependent contractility, providing a mechanism for the blood pressure-associated locus within the ARHGAP42 gene. The goals of the current study were to identify polymorphisms that affect ARHGAP42 expression and to better assess ARHGAP42's role in the development of hypertension. Using DNase I hypersensitivity methods and ENCODE data, we have identified a regulatory element encompassing the ARHGAP42 SNP rs604723 that exhibits strong SMC-selective, allele-specific activity. Importantly, CRISPR/Cas9-mediated deletion of this element in cultured human SMCs markedly reduced endogenous ARHGAP42 expression. DNA binding and transcription assays demonstrated that the minor T allele variation at rs604723 increased the activity of this fragment by promoting serum response transcription factor binding to a cryptic cis-element. ARHGAP42 expression was increased by cell stretch and sphingosine 1-phosphate in a RhoA-dependent manner, and deletion of ARHGAP42 enhanced the progression of hypertension in mice treated with DOCA-salt. Our analysis of a well-characterized cohort of untreated borderline hypertensive patients suggested that ARHGAP42 genotype has important implications in regard to hypertension risk. Taken together, our data add insight into the genetic mechanisms that control blood pressure and provide a potential target for individualized antihypertensive therapies.

Figure 1. ARHGAP42 expression in SMCs is regulated by allele-specific mechanisms and controls BP.

(A) Total RNA isolated from HuAoSMCs heterozygous at the rs604723 SNP (C/T) was subjected to first-strand cDNA synthesis using reverse transcriptase. Reaction products were then subjected to a TaqMan-based PCR assay using allele-specific primers to the ARHGAP42 rs604723 variation. Data represent mean ± SEM of n = 4 experiments; *P < 0.01 vs. the major C allele (Student’s t test). (B) ARHGAP42 mRNA levels were measured by the Genotype-Tissue Expression (GTEx) Consortium. The minor T ARHGAP42 allele at the rs604723 polymorphism was significantly associated with increased ARHGAP42 expression in aortic and coronary artery samples. (C) Schematic of the Arhgap42 gene-trap and SM-MHC^CreERT2 mice used for SMC-specific ARHGAP42 rescue experiments. (D) WT and Arhgap42^gt/gt SM-MHC^CreERT2 mice were injected i.p. with vehicle (corn oil) or tamoxifen (100 mg/kg) for 5 consecutive days as indicated. Two weeks after the last injection, BP was measured by tail cuff method, and Arhgap42 mRNA levels in the aorta were measured by semiquantitative RT-PCR analysis using primers to exons 1 and 4. Data are expressed as mean ± SEM; n = 6 for WT and Arhgap42^gt/gt SM-MHC^CreERT2 mice with vehicle treatment, n = 5 for Arhgap42^gt/gt SM-MHC^CreERT2 mice with tamoxifen treatment. *P < 0.05 vs. WT; **P < 0.05 vs. corn oil–treated (ANOVA). Note that tamoxifen treatment restored Arhgap42 expression and reduced BP to WT levels (representative of 3 separate experiments).

Figure 2. An enhancer within the ARHGAP42 first intron displays strong SMC-specific and allele-specific activity and is required for endogenous ARHGAP42 expression.

(A) Map of the chromatin determinations used to characterize potential regulatory elements near the ARHGAP42 BP-associated locus. The SNPs that define the BP-associated allele (r² > 0.8) are shown at the top. (B) The indicated DNase-hypersensitive (DHS) regions were cloned into the pGL3 luciferase vector and transfected into primary human bronchial SMCs and mouse ECs. Luciferase activity in cell lysates was measured 2 days later and is expressed as fold over the promoterless pGL3 vector. Data represent mean ± SEM of n = 6 experiments; *P < 0.001 vs. in ECs (Student’s t test). (C) Site-directed mutagenesis was used to test the effects of the major/minor alleles on DHS1 and DHS2 enhancer activity. Data represent mean ± SEM of n = 6 experiments; *P < 0.01 vs. the major allele (Student’s t test). (D) Schematic of the 102-bp deletion (in red) generated by our CRISPR/Cas9–mediated gene editing protocol. (E) ARHGAP42 message was measured by semiquantitative RT-PCR in human bronchial SMC cultures transfected with expression plasmids encoding Cas9 and the guide RNAs shown in D (n = 5). The reduction in ARHGAP42 expression was normalized to the efficiency of DHS2 deletion, which ranged from 45% to 95%. Data represent mean ± SEM of n = 5 separate experiments; *P < 0.05 vs. cells transfected with empty guide RNA expression plasmid (Student’s t test).

Figure 3. The minor T allele at rs604723 promotes SRF binding.

(A) Schematic of sequence conservation at the center of the DHS2 region and of CArG homology at the rs604723 SNP. (B) Gel shift assays were performed by combining recombinant SRF (rSRF) with radiolabeled 100-bp oligonucleotide probes containing the major or minor alleles at rs633185 and rs604723. Representative image shown from n = 2. (C) Biotin-labeled 20-bp oligonucleotides containing the major C or minor T allele at rs604723 or a consensus CArG element were conjugated to streptavidin beads and incubated with HuAoSMC nuclear extracts. Washed immunoprecipitates were analyzed for the presence of SRF by Western blotting. Data are representative of 2 separate experiments. (D) ChIP assays were used to measure SRF binding to the DHS2 region in cultured HuAoSMCs and HuCoSMCs that are heterozygous (CT) and homozygous major (CC) at the rs604723 SNP, respectively. Data represent mean ± SEM of n = 4 experiments; *P < 0.05 vs. IgG in HuAoSMCs (Student’s t test). (E) SRF-ChIP immunoprecipitates from HuAoSMCs were subjected to a TaqMan-based assay that discriminates between the major and minor alleles at the rs604723 SNP. Data represent mean ± SEM of n = 4 experiments; *P < 0.01 vs. major allele (Student’s t test). (F) Increasing amounts of DNase I (0–1 μg) were added to permeabilized nuclei isolated from HuAoSMCs. Following genomic DNA isolation, allele-specific primers were used to amplify a 300-bp region containing the rs604723 SNP. Data represent mean ± SEM of n = 3 experiments; *P < 0.05 (Student’s t test).

Figure 4. The allele-specific activity of the DHS2 enhancer is SRF-dependent.

(A) Major and minor DHS2 luciferase constructs were transfected into HuBrSMCs along with myocardin or empty expression vector. Data represent mean ± SEM of n = 5 experiments; *P < 0.05 vs. major allele plus myocardin (Student’s t test). (B) DHS2-luciferase activity was measured in HuBrSMCs treated with control (NTC) or SRF siRNA. Data represent mean ± SEM of n = 6 experiments; *P < 0.01 vs. the minor allele (Student’s t test). (C) Allele-specific GRAF3 mRNA levels were measured by semiquantitative RT-PCR in control and SRF knockdown HuAoSMCs. Data represent mean ± SEM of n = 3 experiments; *P < 0.05 vs. the major C allele in control cells; **P < 0.05 vs. minor T allele in control cells (Student’s t test). (D) Confirmation of SRF knockdown in SMCs treated with control or SRF siRNAs. Data are representative of 3 separate experiments.

Figure 5. ARHGAP42 expression is activated by RhoA signaling and cell stretch.

(A) Primary rat aortic SMCs were treated with 10 μM sphingosine 1-phosphate (S1P) with or without the ROCK inhibitor Y-27632. Arhgap42 expression was measured after 72 hours by semiquantitative PCR. Data represent mean ± SEM of 4 experiments; *P < 0.001 vs. control; **P < 0.001 vs. S1P-treated (ANOVA). n = 4. (B) Using the FX-4000T Flexcell system, primary rat aortic SMCs were subjected to 0 (Ctrl) or 20% equibiaxial elongation at 1 Hz (cyclic strain). Arhgap42 message was measured at 18 hours by quantitative PCR. Data represent mean ± SEM; n = 3 from 2 independent experiments; *P < 0.001 vs. no cyclic strain; **P < 0.05 vs. minus Y-27632 (ANOVA). (C) Rat portal veins placed in ex vivo culture were subjected to 0 or 600 mg of static stretching force. At 72 hours Arhgap42 message was measured by semiquantitative RT-PCR. Graph shows ImageJ-based quantification of 3 independent experiments. *P < 0.05 (Student’s t test). (D) Portal veins isolated from Arhgap42^+/gt mice were cultured ex vivo and subjected to 0 or 300 mg of static stretching force for 5 days. After LacZ staining, tissues were processed for standard microscopy including H&E staining. Scale bar: 200 μm. Data are representative of 3 independent experiments.

Figure 6. Arhgap42 expression limits the development of hypertension.

(A) WT mice were treated with l-NAME (450 mg/l in drinking water). After 14 days, BP was measured by tail cuff method, and Arhgap42 message in isolated mesenteric arteries was measured by quantitative PCR. Data represent mean ± SEM; n = 4 per group; *P < 0.05 vs. untreated (Student’s t test). (B) WT mice were implanted with a 50-mg slow-release DOCA pellet and then fed 0.9% NaCl in drinking water. After 3 weeks, BP was measured by tail cuff method, and Arhgap42 message in aorta was measured by quantitative PCR. *P < 0.05 vs. untreated; n = 6 per group (Student’s t test). (C) Following radiotelemeter implantation and equilibration, 3 WT and 5 Arhgap42^gt/gt SM-MHC^CreERT2 mice were implanted with a 50-mg slow-release DOCA pellet and then fed 0.9% NaCl in drinking water for 3 weeks. Ten days after the start of the DOCA-salt regimen, both groups were treated with tamoxifen (100 mg/kg) by oral gavage for 3 consecutive days. Graph represents average mean arterial BP averaged over each 24-hour period. P < 0.05 (slopes test).