Androgen-sensitive hypertension associated with soluble guanylate cyclase-α₁ deficiency is mediated by 20-HETE

Via linkage analysis, we identified Cyp4a12a, encoding the predominant murine synthase of the vasoconstrictor 20-hydroxy-5,8,11,14-eicosatetraenoic acid (20-HETE), as a candidate sex-specific blood pressure-modifying gene in the context of deficient nitric oxide-soluble guanylate cyclase (NO-sGC) signaling. 20-HETE-generating CYP4 enzymes represent a potential therapeutic target for the treatment of hypertension.

Abstract

Dysregulated nitric oxide (NO) signaling contributes to the pathogenesis of hypertension, a prevalent and often sex-specific risk factor for cardiovascular disease. We previously reported that mice deficient in the α₁-subunit of the NO receptor soluble guanylate cyclase (sGCα₁^−/− mice) display sex- and strain-specific hypertension: male but not female sGCα₁^−/− mice are hypertensive on an 129S6 (S6) but not a C57BL6/J (B6) background. We aimed to uncover the genetic and molecular basis of the observed sex- and strain-specific blood pressure phenotype. Via linkage analysis, we identified a suggestive quantitative trait locus associated with elevated blood pressure in male sGCα₁^−/−S6 mice. This locus encompasses Cyp4a12a, encoding the predominant murine synthase of the vasoconstrictor 20-hydroxy-5,8,11,14-eicosatetraenoic acid (20-HETE). Renal expression of Cyp4a12a in mice was associated with genetic background, sex, and testosterone levels. In addition, 20-HETE levels were higher in renal preglomerular microvessels of male sGCα₁^−/−S6 than of male sGCα₁^−/−B6 mice. Furthermore, treating male sGCα₁^−/−S6 mice with the 20-HETE antagonist 20-hydroxyeicosa-6(Z),15(Z)-dienoic acid (20-HEDE) lowered blood pressure. Finally, 20-HEDE rescued the genetic background- and testosterone-dependent impairment of acetylcholine-induced relaxation in renal interlobar arteries associated with sGCα₁ deficiency. Elevated Cyp4a12a expression and 20-HETE levels render mice susceptible to hypertension and vascular dysfunction in a setting of sGCα₁ deficiency. Our data identify Cyp4a12a as a candidate sex-specific blood pressure-modifying gene in the context of deficient NO-sGC signaling.

NEW & NOTEWORTHY

Via linkage analysis, we identified Cyp4a12a, encoding the predominant murine synthase of the vasoconstrictor 20-hydroxy-5,8,11,14-eicosatetraenoic acid (20-HETE), as a candidate sex-specific blood pressure-modifying gene in the context of deficient nitric oxide-soluble guanylate cyclase (NO-sGC) signaling. 20-HETE-generating CYP4 enzymes represent a potential therapeutic target for the treatment of hypertension.

an estimated quarter of the adult population worldwide has hypertension, an important risk factor for myocardial infarction, stroke, and heart failure. Blood pressure control is achieved in less than half of hypertensive individuals (31), in spite of remarkable progress in our understanding of the etiology of hypertension, the identification and characterization of molecular signaling systems central to blood pressure regulation, and the development of novel therapies (26). Additional research remains necessary to fully elucidate the mechanisms responsible for hypertension and to develop successful hypertensive treatment strategies (28).

In recent years, genome-wide association studies have identified numerous loci associated with blood pressure, highlighting a substantial heritable component to the development of hypertension (27). With the exception of a few rare monogenic disorders, pathological changes in the regulation of systemic blood pressure are mostly the result of combined abnormalities involving several genes. The multiplicity of additive, perhaps even synergistic, genetic factors associated with blood pressure contribute to the complex etiology of essential hypertension (15). Genetic variants often have minor effect sizes and the impact of common variants on blood pressure can vary widely from patient to patient, suggesting that multiple signaling mechanisms are modifying blood pressure in a coordinated fashion, which remains to be elucidated.

Animal and population studies have implicated impaired signaling through the vasodilator nitric oxide (NO) and its downstream target, the cyclic 3′,5′-guanosine monophosphate (cGMP)-producing enzyme soluble guanylate cyclase (sGC), in the pathogenesis of hypertension. sGC is a heterodimeric enzyme consisting of an α₁- or α₂- and a β₁-subunit. Several loci associated with blood pressure contain genes encoding enzymes that, directly or indirectly, regulate cGMP levels (6, 14, 23, 24). The major variant of the single nucleotide polymorphism (SNP) rs13139571 in the locus containing the GUCY1A3 and GUCY1B3 genes, encoding sGCα₁ and sGCβ₁, respectively, is associated with higher systolic and diastolic blood pressure, highlighting the role of sGC in blood pressure regulation (7). We previously reported sex- and strain-specific systemic hypertension in mice deficient in sGCα₁ (sGCα₁^−/−) that still express a functional sGCα₂β₁ enzyme (4) and in mice deficient in sGCβ₁ (35), lacking all NO-inducible sGC activity. Sex-associated blood pressure differences are well documented in other animal models of hypertension (30) and, more clinically relevant, men are at greater risk for hypertension than premenopausal women (18).

20-Hydroxy-5,8,11,14-eicosatetraenoic acid (20-HETE), a primary arachidonic acid metabolite and a key regulator of the microcirculation, is produced by enzymes from the cytochrome P450-4 (CYP4) family (32, 44). Elevated production of 20-HETE was previously observed in cardiovascular diseases, including hypertension (38, 41). A functional haplotype of CYP4F2, together with CYP4A11 considered to be responsible for most of the renal 20-HETE formation in humans (33), was characterized by increased transcription of CYP4F2 and was associated with elevated urinary 20-HETE levels and enhanced susceptibility to hypertension (16). This suggested that regulation of transcription is particularly important in controlling 20-HETE signaling (16). Furthermore, in animal models of hypertension, inhibiting 20-HETE signaling normalized blood pressure (37, 43). Increased 20-HETE signaling has also been implicated in the development of androgen-mediated vascular dysfunction and hypertension (36). In both rats (21) and mice (11, 42), increased expression of Cyp4A8 and the male-specific androgen-regulated Cyp4a12a, respectively, was reported to contribute to the hypertension associated with increased androgen levels.

Here, we report the identification of a suggestive quantitative trait locus (QTL) associated with elevated blood pressure in male sGCα₁^−/−S6 mice. We provide evidence that sex- and strain-specific hypertension in sGCα₁^−/−S6 mice is associated with elevated Cyp4a12a gene expression and higher 20-HETE levels. In addition, hypertension and vascular dysfunction in sGCα₁^−/−S6 mice are attenuated by 20-HETE inhibition. Together, our results identify Cyp4a12a as a candidate blood pressure modifying gene in the context of deficient NO-sGC signaling.

METHODS

Animals.

This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institute of Health. Housing of and all procedures involving mice [on the 129S6 (S6), C57BL6/J (B6), or mixed S6-B6 background] were approved by the Institutional Animal Care and Use Committees of Massachusetts General Hospital (Subcommittee on Research Animal Care).

Generation and blood pressure measurement of F2 offspring from an sGCα₁^−/−S6 X sGCα₁^−/−B6 intercross.

Generation of sGCα₁^−/−S6 mice, sGCα₁^−/−B6 mice (back-crossed 8 generations with B6 mice from Jackson Laboratories), and F2 offspring from an sGCα₁^−/−S6 X sGCα₁^−/−B6 intercross were described previously (4). Wild-type (WT)B6 and WTS6 mice were purchased from the Jackson Laboratory and Taconic Farms, respectively. Mean arterial pressure (MAP) was measured invasively, as previously described (2) in 284 male F2 offspring at 4 mo of age and in mice treated acutely with the selective 20-HETE antagonist 20-hydroxyeicosa-6(Z),15(Z)-dienoic acid (20-HEDE, administered at a dose of 10 mg/kg iv) (40). Briefly, mice were treated with buprenorphine (buprenex, 0.1 mg/kg sc) 30 min before surgery; anesthetized by intraperitoneal injection with ketamine (120 mg/kg), fentanyl (90 mg/kg), and pancuronium (2 mg/kg); intubated; and mechanically ventilated (Fi_{O2 = 1, 10 μl/g, 120 breaths/min). A fluid-filled catheter was inserted into the left carotid artery for infusion of saline solution (1.2 ml/h) and for measurement of MAP. We previously reported that hypertension associated with sGC deficiency was detectable both in anesthetized mice (assessed with an indwelling catheter) and awake mice (assessed via tailcuff plethysmography and telemetry) (4). At the end of the experiment, mice were euthanized by injecting pentobarbital (100 mg/kg iv).}

Genotyping.

Isolation of genomic DNA and the protocol used to perform the genome scans were described previously (2). Briefly, 284 sGCα₁^−/−F2 mice were genotyped using 120 B6-S6 informative SNPs [including 8 on chromosome 4 (2)] covering the murine autosomes (average intermarker spacing = 22.2 Mb, largest intermarker gap = 54.6 Mb). All SNPs were assayed using Sequenom MassARRAY iPLEX GOLD chemistry, and SpectroCHIPs were analyzed in automated mode by a MassArray MALDI-TOF Compact system 2 with a solid phase laser mass spectrometer (Bruker Daltonics). Variants were called by real-time SpectroCaller algorithm and analyzed by SpectroTyper v.4.0 software and clusters were manually reviewed for validation of genotype calls. Subsequently, 284 sGCα₁^−/−F2 mice were genotyped using 24 SNPs spanning a region on chromosome 4 extending from 46 to 136 Mb, resulting in the identification of a suggestive QTL. Reported genetic map positions for the markers were retrieved from the SNP database (build 37) of the National Center for Biotechnology Information (NCBI).

Linkage and sequence analysis.

As described previously (2), parametric linkage analysis was performed using MAPMAKER/QTL to identify QTL associated with blood pressure. For an F2 intercross, a logarithm of the odds score (LOD) of ∼2.8 is considered “suggestive” (expected to be seen by chance once per genome scan). A region within 1.5 LOD score units of the maximum LOD was used to represent the 95% interval for the QTL. The genome-wide scans were plotted with the use of the J/QTL mapping program (version 0.8). Genomic, cDNA, and amino acid sequences for the B6 Cyp4a12a loci were obtained from publicly available databases (ENSEMBL and NCBI). SNPs, indels, and structural variants between B6 and the corresponding Cyp4a12a loci from the S5 genome (the S6 genome is not available) were obtained from the Mouse Genome Project at the Wellcome Trust Sanger Institute (Release 1410 based on GRCm38). Short-read S5 sequence data was obtained from ENA (Study: ERP000036). Alignments were obtained with Blast (NCBI).

Measurement of gene expression.

Total RNA was extracted from whole kidneys harvested from mice using TRIzol reagent (Life Technologies), and cDNA was synthesized using Maloney's murine leukemia virus reverse transcriptase (Promega). Cyp4a10, Cyp4a12a, Cyp4a14, and Rn18s transcript levels were measured by real-time PCR using the SYBR or TaqMan PCR master mix (Life Technologies) in a Mastercycler ep realplex 2 (Eppendorf). Primers were designed for Cyp4a10 (5′-TTCCCTGATGGACGCTCTTTA-3′ and 5′-GCAAACCTGGAAGGGTCAAAC-3′), Cyp4a12a (5′-AGCTCCCTGGATTGGGCGTGG-3′ and 5′-TGCAGGCACTGTTAGCCTTGCA-3′), Cyp4a14 (5′-TTTAGCCCTACAAGGTACTTGGA-3′ and 5′-GCAGCCACTGCCTTCGTAA-3′), and Rn18s (5′-CGGCTACCACATCCAAGGAA-3′ and 5′-GCTGGAATTACCGCGGCT-3′) for normalization. Changes in the relative gene expression normalized to Rn18s levels were determined using the relative cycle threshold method.

Orchiectomy.

Male 8- to 11-wk-old mice were treated with buprenorphine (buprenex, 0.1 mg/kg sc) 30 min before surgery; next, mice were anaesthetized with ketamine (100 mg/kg) and xylazine (10 mg/kg). A small abdominal incision was made in the skin and the peritoneum. The testes were exteriorized together with the connected fat pad. After ligation, the arteries were cut, and the testes and epididymis were removed. When the procedure was completed on both sides, the peritoneum and the skin were closed with silk suture. Animals were treated with buprenorphine (buprenex, 0.1 mg/kg sc q12 h) for 1 day. Eight weeks later, mice were euthanized, and tissue was harvested.

Testosterone treatment.

5α-Dihydro-testosterone pellets (15 mg, 21-day release, Innovative Research of America) or vehicle pellets were implanted subcutaneously. Three weeks later, mice were euthanized, and tissue was harvested.

Measurements of 20-HETE.

Renal preglomerular microvessels were dissected under stereomicroscopy (magnification: approximately ×60–210) in ice-cold Krebs bicarbonate buffer using no. 5 dissecting forceps and iris scissors. Vascular segments were dissected at a length of 0.5 mm and incubated in oxygenated Krebs bicarbonate buffer, pH 7.4, with 1 × 10⁻³ mol/l NADPH for 1 h at 37°C with gentle shaking. Deuterated-20-HETE was added as an internal standard and 20-HETE was extracted and quantified by liquid chromatography-tandem mass spectrometry (LC-MS/MS) as previously described (42). WT samples were assessed using a Shimadzu Triple Quadrupole Mass Spectrometer LCMS-8050 equipped with a Nexera UHPLC using the multiple reaction-monitoring mode.

Measurement of vascular reactivity on isolated interlobar arteries.

Renal interlobar arteries (∼100-μm diameter) were mounted on wires in the chambers of a multivessel myograph (JP Trading) filled with Krebs buffer (37°C) gassed with 95% O₂-5% CO₂. After mounting, the vessels were set to an internal circumference equivalent to 90% of the determined internal circumference necessary to simulate its anticipated physiological intraluminal pressure in vivo when placed under a transmural pressure of 100 mmHg, followed by 30–60 min of equilibration. Isometric tension was monitored continuously before and after the experimental interventions. Experiments were conducted to determine concentration-response curves for acetylcholine-induced relaxation after contraction with 1 × 10⁻⁶ mol/l phenylephrine. A cumulative relaxation-response curve to acetylcholine (1 × 10⁻⁸-1 × 10⁻⁴ mol/l) was constructed and the maximal relaxation recorded. In the experiments depicted (see Fig. 4B), 1 × 10⁻⁶ mol/l 20-HETE was added after pretreatment of vessels with 1 × 10⁻⁶ mol/l 20-HEDE before the challenge with acetylcholine.

Fig. 4.

20-HETE mediates genotype- and strain-specific vascular dysfunction in sGCα₁^−/− mice. Acetylcholine-induced relaxation was studied in phenylephrine-precontracted renal interlobar arteries from male WT (squares) and sGCα₁-deficient mice (sGCα₁^−/−) on the S6 (sGCα₁^−/−S6, circles) or B6 background (sGCα₁^−/−B6, triangles), in the absence or presence of 20-HEDE (grey symbols) and/or 20-HETE (open symbols). A: 20-HEDE restored the ability of acetylcholine to induce vasorelaxation in sGCα₁^−/−S6 mice; n = 6/group. *P < 0.05 vs. WTS6 and vs. sGCα₁^−/−S6 20-HEDE; #P < 0.05 vs. WTS6. B: 20-HETE impairs endothelium-dependent vasorelaxation; n = 6/group. *P < 0.05 sGCα₁^−/−S6 20-HEDE + 20-HETE vs. sGCα₁^−/−S6 20-HEDE and WTS6 20-HEDE + 20-HETE vs. WTS6 20-HEDE. C: 20-HEDE restores the ability of acetylcholine to induce vasorelaxation in sGCα₁^−/−S6 to the level observed in sGCα₁^−/−B6 mice; n = 9/group. *P < 0.05 vs. sGCα₁^−/−B6 and vs. sGCα₁^−/−S6 20-HEDE. All statistics were performed using repeated measures ANOVA, followed by one-way ANOVA for each concentration.

Statistical analysis.

Statistical analyses were performed with either GraphPad Prism 6.0 (GraphPad Software) or Stata 11 (Statacorp). All continuous measurements were expressed as mean ± SE. Multiway ANOVA with post hoc testing using Sidak's correction for multiple comparisons was used when comparing gene expression. One-way ANOVA was used when comparing blood pressure among three groups. When comparing between only two groups (MAP in mice treated with vehicle or 20-HETE inhibitor; 20-HETE levels; Cyp4a12a expression in testosterone-treated mice), the independent Student's t-test was used. Vascular reactivity in interlobar arteries was initially analyzed with repeated-measures ANOVA; however, given the finding of significant interaction P values, we subsequently utilized one-way ANOVA with post hoc testing using Sidak's correction for multiple comparisons for each concentration of acetylcholine used. The effects on blood pressure of the genomic modifiers were assessed using Strata 11 to determine whether their exerted effects were independent of one another and whether their effects were additive or synergistic.

RESULTS

Identification of a blood pressure quantitative trait locus on chromosome 4 and of Cyp4a12a as a candidate modifier gene for blood pressure in sGCα₁^−/− mice.

We previously reported that genetic background affects the blood pressure phenotype in sGCα₁^−/− mice (2). Linkage analysis identified a putative QTL (Chr4D1-QTL) linked to MAP in the context of sGCα₁ deficiency that is located close to Mm37-4-112419040 (LOD = 3.6). The 1.5 LOD confidence interval spanned an ∼42-MB region on chromosome 4 between Mm37-4-92132381 and Mm37-4-134165380. Chr4D1-QTL is syntenic with previously identified blood pressure-related QTLs in the human (Homo sapiens 1p33) and rat (Rattus norvegicus 5q36) genome and encompasses the Cyp4abx cluster (22, 29, 34).

The murine Cyp4abx locus encodes multiple Cyp4 isoforms, including Cyp4a10, Cyp4a12a, and Cyp4a14, the three murine Cyp4a isoforms that convert arachidonic acid to 20-HETE. Renal Cyp4a10 mRNA expression levels were similar in WT and sGCα₁^−/− mice, similar in B6 and S6, but higher in female than male mice (Fig. 1A). Renal Cyp4a14 mRNA expression levels were also similar in WT and sGCα₁^−/− mice and in female B6 and S6 mice, but higher in female than male mice, and higher in male B6 than male S6 mice (Fig. 1B). Importantly, renal expression of Cyp4a12a, the predominant 20-HETE synthase in murine kidneys (20), correlated with the sex- and strain-specific blood pressure phenotype observed in sGCα₁^−/− mice (29): Cyp4a12a expression was higher in hypertensive male S6 mice than in normotensive female S6 mice or normotensive B6 mice of either sex (Fig. 1C).

Fig. 1.

Renal expression levels [relative to male B6 wild type (WT)] of Cyp4a10, Cyp4a14, and Cyp4a12a. mRNAs encoding Cyp4a10 (A), Cyp4a14 (B), and Cyp4a12a (C) were measured via quantitative (q)RT-PCR in kidneys of male and female, WT, and mice deficient in the α₁-subunit of the nitric oxide (NO) receptor soluble guanylate cyclase (sGCα₁^−/− mice) on the S6 or B6 genetic background; n = 6–20/group. Cyp4a10 and Cyp4a14 mRNA expression levels are similar in WT and sGCα₁^−/− mice (P = 0.22 for both), similar in B6 and S6 female mice (P = 0.26 and 0.99, respectively), but higher in female than male mice. Cyp4a14 mRNA expression levels are higher in male B6 than S6 mice. Cyp4a12a expression is higher in male S6 mice than in B6 or female S6 mice, and tends to be higher in male B6 mice than female B6 mice. *P < 0.0001, §P < 0.05, and †P = 0.07 vs. male mice of same strain and genotype; #P < 0.01 vs. male B6 mice of the same genotype; all by three-way ANOVA.

These differences in Cyp4a12a expression levels in male S6 vs. male B6 mice were associated with higher 20-HETE levels in renal preglomerular microvessels (PGMV) isolated from male sGCα₁^−/−S6 mice than from male sGCα₁^−/−B6 mice (Fig. 2). Similar results were obtained in PGMV isolated from WT mice (450 ± 317 and 192 ± 63 pg 20-HETE/mg protein in S6 and B6 WT mice, respectively, n = 10 each, P = 0.021).

Fig. 2.

20-Hydroxy-5,8,11,14-eicosatetraenoic acid (20-HETE) levels are higher in renal preglomerular microvessels isolated from male sGCα₁^−/−S6 mice than from male sGCα₁^−/−B6 mice. *P < 0.05 by Student's t-test vs. B6; n = 3–5/group.

Activity of the 20-HETE signaling system modulates blood pressure in sGCα₁^−/− mice.

To evaluate the impact of the Cyp4abx genotype on blood pressure, we assessed MAP in F2 offspring from the sGCα₁^−/−S6 X sGCα₁^−/−B6 intercross (sGCα₁^−/−F2 mice), according to their Cyp4abx genotype. The presence of the S6 genotype for Mm37-4-112419040 was associated with increased blood pressure in a recessive manner: MAP was ∼10 mmHg higher in mice homozygous for the S6 locus than in mice carrying at least one B6 allele (Fig. 3A). Heart rate did not differ between groups. In addition, MAP stratified by genotypes of both the cyp4abx and renin loci revealed the blood pressure to be lowest in mice homozygous for the renin and Cyp alleles associated with lower blood pressure and highest in mice homozygous for the renin and Cyp alleles associated with higher blood pressure. The two genes exerted their effects on blood pressure independently and in an additive manner (Fig. 3B). Finally, treatment of male sGCα₁^−/−S6 mice, known to develop hypertension (4), with 20-HEDE, a selective 20-HETE antagonist, decreased blood pressure (Fig. 3C) without affecting heart rate. Together, these data suggest that 20-HETE signaling modifies blood pressure in a setting of impaired NO-cGMP signaling.

Fig. 3.

A: mice homozygous for the S6 Cyp4abx locus have higher blood pressure than heterozygous mice or mice homozygous for the B6 Cyp4abx locus. Mean arterial blood pressure (MAP) was measured invasively in anesthetized male sGCα₁^−/−F2 mice homozygous for the S6 allele of mm37-4-112419040 (n = 53), homozygous for the B6 allele of mm37-4-112419040 (n = 49), or heterozygous for mm37-4-112419040 (n = 73). *P < 0.001 vs. all other groups by one-way ANOVA. B: mice were stratified according to renin and cyp4abx genotype. Mice were grouped according to the number of risk alleles they expressed, ranging from 0 (homozygous for the low blood pressure variant in both the renin and Cyp4abx locus), 1 (heterozygous in either the renin or the Cyp4abx locus), 2 (either homozygous for the high blood pressure variant in one locus and homozygous for the low blood pressure variant in the other locus, or heterozygous at both loci), 3 (homozygous for the high blood pressure variant in one locus and heterozygous for the high blood pressure variant in the other locus), to 4 (homozygous for the high blood pressure variant in both the renin and Cyp4abx loci) risk alleles. MAP was measured invasively in anesthetized male sGCα₁^−/−F2 mice with 0 (n = 19), 1 (n = 44), 2 (n = 59), 3 (n = 38), and 4 (n = 17) risk alleles. The interaction P value was not significant between the 2 genes (P = 0.83). Therefore, the 2 genes exert their effects on blood pressure independently. For each risk allele, the blood pressure increases by 5.9 mmHg (95% confidence interval 3.4–8.4, P < 0.001). C: antagonizing 20-HETE with 20-hydroxyeicosa-6(Z),15(Z)-dienoic acid (20-HEDE) normalizes blood pressure in sGCα₁^−/−S6 mice. MAP was measured invasively in anesthetized male WT and in sGCα₁^−/− mice on the S6 background treated with vehicle or with 20-HEDE (n = 5/group). *P < 0.05 vs. WT; #P = 0.01 vs. vehicle in sGCα₁^−/−S6 by two-way ANOVA.

Activity of the 20-HETE signaling system modulates endothelium-dependent vascular reactivity in sGCα₁^−/− mice.

To test whether differences in 20-HETE signaling contribute to the vascular dysfunction observed in sGCα₁^−/−S6 mice (23), we compared the effect of inhibiting 20-HETE signaling on endothelial-dependent vasorelaxation in renal interlobar arteries isolated from male WTS6 and sGCα₁^−/−S6 mice. We previously reported that the vasodilatory effect of acetylcholine was attenuated in sGCα₁^−/−S6 mice compared with WTS6 mice (25). Incubating renal interlobar arteries isolated from sGCα₁^−/−S6 mice with 20-HEDE restored the ability of acetylcholine to induce vasorelaxation to levels similar to those observed in arteries isolated from WTS6 mice (Fig. 4A). Adding 20-HEDE to the organ bath did not affect acetylcholine-induced vasorelaxation in arteries isolated from WTS6 mice. In contrast, adding 20-HETE to the organ bath reversed the ability of 20-HEDE to improve acetylcholine-induced vasorelaxation in sGCα₁^−/−S6 arteries and reduced endothelium-dependent vasorelaxation in WT arteries (Fig. 4B).

NO-dependent vasorelaxation is attenuated to a greater extent in sGCα₁^−/−S6 than in sGCα₁^−/−B6 mice (4). Treatment with 20-HEDE rescued the ability of acetylcholine to induce vasorelaxation in renal interlobar arteries isolated from sGCα₁^−/−S6 to the level observed in sGCα₁^−/−B6 (Fig. 4C). These results highlight a direct role for the 20-HETE signaling in endothelium-dependent vasorelaxation. Differences in 20-HETE signaling may therefore be, at least in part, responsible for the greater impairment of vascular relaxation and hypertension in sGCα₁^−/− mice than WT mice and in sGCα₁^−/−S6 mice than in sGCα₁^−/−B6 mice.

Activity of the 20-HETE signaling system contributes to sex-specific hypertension associated with impaired NO-cGMP signaling.

To test whether 20-HETE signaling contributes to the sex-specific hypertension in a setting of impaired NO-cGMP signaling (4), we assessed the impact of testosterone on Cyp4a12a expression levels in sGCα₁^−/−S6 mice. We previously demonstrated that the hypertension associated with sGCα₁ deficiency is testosterone dependent: orchiectomy or treatment with a testosterone receptor antagonist normalized blood pressure in male sGCα₁^−/−S6 mice while treating normotensive female sGCα₁^−/−S6 mice with testosterone increased blood pressure (4). Cyp4a12a expression was higher in kidney of female sGCα₁^−/−S6 mice treated with testosterone than female sGCα₁^−/−S6 mice treated with vehicle (Fig. 5A). In addition, Cyp4a12a expression was lower in orchiectomized WTS6 and sGCα₁^−/−S6 mice than in sham-operated controls (Fig. 5B). Together, these findings indicate that expression of Cyp4a12a is regulated by testosterone and that testosterone-induced Cyp4a12a expression may contribute to the sex-specific testosterone-induced hypertension observed in sGCα₁^−/− mice.

Fig. 5.

Renal expression levels of Cyp4a12a in testosterone-treated or orchiectomized mice. mRNA levels of Cyp4a12a were measured via qRT-PCR in kidneys of female sGCα₁^−/−S6 mice treated with vehicle or testosterone and in sham-operated or orchiectomized male WTS6 or sGCα₁^−/−S6 mice. A: Cyp4a12a expression levels (relative to sGCα₁^−/−S6 vehicle) are higher in testosterone-treated than in vehicle-treated mice. n = 7/group; *P < 0.0001 by Student's t-test vs. vehicle-treated mice. B: Cyp4a12a expression levels (relative to WTS6 Sham) are also higher in sham-operated than orchiectomized WTS6 and sGCα₁^−/−S6 mice; n = 7/group. *P < 0.0001 by two-way ANOVA vs. sham-operated mice.

As discussed above, the ability of acetylcholine to induce vasorelaxation was impaired in arteries isolated from male sGCα₁^−/−S6 mice. Renal interlobar arteries from orchiectomized sGCα₁^−/−S6 mice displayed higher sensitivity to acetylcholine-induced vasorelaxation than interlobar arteries from sham-operated sGCα₁^−/−S6 mice (Fig. 6A). The improvement in endothelium-derived vasorelaxation in orchiectomized sGCα₁^−/−S6 mice was comparable to that observed in 20-HEDE-treated renal interlobar arteries isolated from sham-operated sGCα₁^−/−S6 mice (Fig. 6B). 20-HETE treatment of arteries isolated from orchiectomized sGCα₁^−/−S6, but not WTS6 mice, impaired acetylcholine-induced vasorelaxation (Fig. 6C). These results suggest that the susceptibility of sGCα₁^−/−S6 mice to the vasoconstrictive effects of 20-HETE is higher than that of WT mice and provide a mechanism for the testosterone-dependent hypertension in mice with impaired NO-cGMP signaling.

Fig. 6.

20-HETE mediates sex-specific vascular dysfunction in sGCα₁^−/−S6 mice. Acetylcholine-induced relaxation was studied in phenylephrine-precontracted renal interlobar arteries from male wild-type (WTS6, squares) and sGCα₁-deficient mice (sGCα₁^−/−S6, circles), either sham-operated (black symbols) or orchiectomized (open crossed out symbols), and/or in the presence of 20-HETE (grey crossed out symbols). A: orchiectomy restored the ability of acetylcholine to induce vasorelaxation in sGCα₁^−/−S6 mice; n = 6/group. *P < 0.05 vs. sGCα₁^−/−S6 orchiectomy; #P < 0.05 vs. WTS6. B: orchiectomy restored the ability of acetylcholine to induce vasorelaxation in sGCα₁^−/−S6 mice to the same extent as 20-HEDE did; n = 6/group. *P < 0.05 sGCα₁^−/−S6 vs. sGCα₁^−/−S6 orchiectomy and sGCα₁^−/−S6 vs. sGCα₁^−/−S6 20-HEDE. C: incubating arteries isolated from orchiectomized sGCα₁^−/−S6, but not WTS6 mice, with 20-HETE impaired acetylcholine-induced vasorelaxation; n = 6/group. *P < 0.05 vs. sGCα₁^−/−S6 orchiectomy. All statistics were performed using repeated-measures ANOVA, followed by one-way ANOVA for each concentration.

DISCUSSION

This study highlights a role for the 20-HETE-vasoconstricting pathway in a sex-specific and strain-specific mouse model of hypertension: sGCα₁^−/− mice (1, 4). Our data suggest that higher Cyp4a12a expression and associated 20-HETE levels in S6 mice than in B6 mice increase susceptibility to hypertension in S6 mice under conditions of impaired NO signaling. In addition, lower expression of Cyp4a12a in female than in male S6 mice may explain the resistance of female S6 sGCα₁^−/− mice to the development of hypertension.

Previously, using a nonbiased genetic approach to reveal genetic factors that modulate blood pressure in the context of sGCα₁-deficiency, we identified a blood pressure quantitative trait locus (Hsgcq), a candidate blood pressure modifier gene (renin), and a signaling mechanism that contributes to the strain-specific hypertension in sGCα₁^−/− mice (2). The effect size of Hsgcq accounted for approximately one-third of the observed hypertension in sGCα₁^−/−S6 mice (4), implying that other modifier genes contribute to the hypertension. In the same genome-wide linkage scan for blood pressure, another locus (on chromosome 4, designated Chr4D1-QTL) displayed suggestive linkage to blood pressure (LOD of 3.6). Our results suggest that both the renin and Cyp4abx loci have independent and additive effects on blood pressure, with an effect on blood pressure for each risk allele of ∼6 mmHg. The association on chromosome 4 became less significant upon additional typing with a panel of fine-mapping markers across Chr4D1-QTL (LOD = 3.2). This observation could be due to decreased recombinatorial degrees of freedom because of increased marker density or genotyping errors. When fine-mapping markers only were analyzed, the LOD decreased further to 2.8. This drop in significance may be explained by the fact that the original peak marker (Mm37-4-112419040, located at 112.4 Mb) was not retyped in the fine-mapping set (with nearest flanking markers located at 103.8 and 115.6 Mb). A further attempt to refine the linkage peak using the fine-mapping markers only, adding a third independent cohort of 96 F2 mice to the original 284 mice, did not improve the LOD score.

Chr4D1-QTL is syntenic with previously identified blood pressure QTLs in humans (Homo sapiens 1p33) and rats (Rattus norvegicus 5q36) (29, 34). In addition, genetic variants associated with essential hypertension have been identified in loci containing genes that encode CYP4 isoforms thought to be responsible for the majority of 20-HETE production in humans, including CYP4A11 and CYP4F2 (9, 10, 16, 39). Similar to the genetic make-up of the human CYP4ABX cluster, the mouse Cyp4abx cluster contains multiple Cyp4 isoforms, including the male-specific androgen-regulated Cyp4A12a, the predominant 20-HETE synthase in murine kidneys. Other Cyp4 isoforms in the locus (including Cyp4a14 and Cyp4a10) have limited capability to convert arachidonic acid into 20-HETE (20).

Expression of Cyp4a12a was previously shown to be strain dependent, with higher gene expression and corresponding protein and activity levels in the kidneys of WTS6 mice than of WTB6 mice (20). Studies also demonstrated that expression of Cyp4a12a and associated 20-HETE production were sex specific and androgen regulated and that 20-HETE-induced vascular dysfunction contributed to testosterone-dependent hypertension (40, 42). We hypothesized that the sex- and strain-dependent hypertension, previously reported (2) in sGCα₁^−/− mice, was the result of impaired NO-cGMP signaling combined with an increased 20-HETE signaling activity. Here, we show that the higher Cyp4a12a mRNA expression levels in the kidneys of male mice on the S6 background than male mice on the B6 background were associated with higher levels of 20-HETE in renal PGMV of male sGCα₁^−/−S6 mice than male sGCα₁^−/−B6 mice and male WTS6 mice than male WTB6 mice. Inhibiting 20-HETE signaling with 10 mg/kg 20-HEDE lowered blood pressure in hypertensive male sGCα₁^−/−S6 but not WTS6 mice. The observation that Cyp4a12a mRNA expression levels are similar between male sGCα₁^−/− and WT mice on the S6 background suggests that the hypertensive effect of 20-HETE is more pronounced in a setting of impaired NO-cGMP signaling. Based on the fact that WTS6 mice and sGCα₁^−/−S6 mice share the same Cyp4abx genotype, that 20-HETE levels are higher in WTS6 mice than in WTB6 mice, and that blood pressure is higher in WTS6 mice than in WTB6 mice (3, 17), it is conceivable that a higher dose of 20-HEDE may lower blood pressure in WTS6 mice.

20-HEDE also restored endothelium-dependent vasorelaxation in interlobar arteries isolated from sGCα₁^−/−S6 mice to the level observed in sGCα₁^−/−B6 and WTS6 mice. Concomitantly, addition of 20-HETE impaired the ability of acetylcholine to relax WTS6 interlobar arteries to the level observed in sGCα₁^−/−S6 interlobar arteries. Finally, endothelium-dependent vasorelaxation was restored in sGCα₁^−/−S6 mice lacking testosterone to the levels observed in WTS6 mice. This vasodilatory rescue was abolished upon adding 20-HETE to interlobar arteries isolated from orchiectomized sGCα₁^−/−S6 mice but not from WTS6 mice. Together, these results suggest that testosterone-induced Cyp4a12a expression and associated increases in 20-HETE levels contribute to the vascular dysfunction and hypertension observed in male sGCα₁^−/−S6 mice. Furthermore, these findings suggest a role for testosterone in increasing susceptibility of S6 mice to the hypertensive and vasoconstrictive effects of 20-HETE in a setting of impaired sGC signaling via a yet-to-be-determined mechanism. For example, it is conceivable that testosterone sensitizes the vasculature to 20-HETE-induced vasoconstriction and increased myogenic tone. Previous studies suggested that 20-HETE can both mediate androgen-induced hypertension and cause hypertension independent of androgens (42) and that 20-HETE contributes to enhanced myogenic constriction in NO synthase 3-deficient mice (12). Our data suggest that the absence of functional NO signaling sensitizes the vasculature to the vasoconstricting effects of 20-HETE, even in the absence of testosterone. Whether 20-HETE, which can be released from vascular smooth muscle (12), alters vascular function via the endothelium and/or smooth muscle remains to be determined.

Renal mRNA expression levels of Cyp4a10 and Cyp4a14 mRNA were higher in female than male mice, as was previously reported (20), regardless of genetic background or genotype. Genetic deletion of Cyp4a14, resulting in male-specific and testosterone-dependent systemic hypertension, was accompanied with increased renal Cyp4a12a expression levels (11). The inverse, seemingly sex-driven interaction between Cyp4a12a and Cyp4a14 implies that androgens are an upstream regulator of this relationship. Alternatively, it cannot be excluded that sex-specific expression of Cyp4a14 contributes to the sex-specific expression of Cyp4a12a and therefore that Cyp4a14 is a modifier gene in the identified QTL.

Both of the pathways identified in our linkage analysis [the renin (2) and 20-HETE signaling pathways] are likely centrally involved in the hypertension associated with impaired NO-cGMP signaling, and can cross regulate each other (5). In addition, rat 5q36, the rat locus syntenic with mouse Chr4D1-QTL, also contains a QTL for serum renin concentrations, possibly highlighting yet another layer of interaction between these pathways. The relevance of these interactions among the NO-cGMP, the androgens-regulated 20-HETE, and the renin signaling systems in human disease remains to be determined. For example, does the combination of known blood pressure variants in the GUCY (7), REN (8, 19), and CYP4A11 (9, 10) genes (encoding sGC, renin, and cytochrome P450, family 4, subfamily A, polypeptide 11, respectively), which are associated with relatively small increases in blood pressure, have an additive or even synergistic effect on blood pressure in humans? Even if statistical and logistical roadblocks impair our ability to study gene-gene interactions, well-designed population studies (stratifying by sex and GUCY, REN, and CYP4A11 alleles that are associated with higher blood pressure) may provide answers.

Taken together, our results identify a blood pressure QTL containing the Cyp4a12a gene, which is expressed in a sex- and strain-specific manner. The observation, that inhibiting 20-HETE signaling improves NO-mediated vasorelaxation and lowers blood pressure in hypertensive sGCα₁^−/− mice, identifies Cyp4a12a as candidate modifier gene in Chr4D1-QTL. These results corroborate our hypothesis that testosterone-induced Cyp4a12a expression exacerbates vascular dysfunction associated with impaired NO-cGMP signaling and, thereby, contributes to the hypertension observed in male sGCα₁^−/−S6 mice.

This study reflects an integrative approach to understanding the etiology of hypertension in which multiple recognized and well-studied blood pressure modulating pathways [including NO-cGMP (7, 13), renin (19), and 20-HETE signaling (9, 10)] coregulate blood pressure (Fig. 7). Such an approach is necessary to further our knowledge of the etiology of essential hypertension and to improve our understanding of the molecular mechanisms that contribute to the development of hypertension, likely including altered activity of multiple blood pressure-modifying pathways. We suggest that the presented data may eventually inform the identification of at-risk populations of humans that suffer from essential hypertension and would benefit from treatment targeting multiple pathways.

Fig. 7.

An interactive role for the NO-cGMP, renin-angiotensin, and 20-HETE signaling systems in androgen-dependent hypertension. Our data suggest that a testosterone-induced increase in 20-HETE levels exacerbates systemic vascular dysfunction associated with impaired NO-cGMP signaling, and thereby contributes to the hypertension observed in male sGCα₁^−/− mice on the S6 background (sGCα₁^−/−S6). We previously implicated increased renin-angiotensin-aldosterone system (RAAS) signaling in the gender- and strain-specific hypertension associated with sGC deficiency (2). Genetic variation in both the 20-HETE and RAAS-signaling system modifies the impact of impaired NO-cGMP signaling on blood pressure, having both independent and additive effects. How NO-cGMP signaling, the RAAS, and 20-HETE signaling cross regulate each other to modulate blood pressure remains to be determined.

GRANTS

This work was supported by the American Heart Association (10SDG2610313 to E. S. Buys, Philips Resuscitation Fellowship Award to P. Y. Sips), an Eleanor and Miles Shore 50th Anniversary Fellowship program for Scholars in Medicine from Harvard Medical School (to E. S. Buys), National Institutes of Health (P01-HL-034300 to M. L. Schwartzman, K08-HL-111210 to R. Malhotra, R01-HL-74352 to K. D. Bloch, HL-111392 to J. R. Falck, HL-34300-26A1S1 to V. Garcia, and 5R01-EY-022746 to E. S. Buys), a grant from the Foundation Leducq [to K. D. Bloch], grants from the FWO-Vlaanderen and the UGent-GOA Programs (FWOG093810N and BOF10/GOA/024, respectively, to P. Bloch), and a fellowship from the Belgian American Education Foundation (to S. Vandenwijngaert).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

A.C.D., S.V., V.G., R.E.T.T., D.I.N., K.A., M.J.R., L.T.T., F.F.Z., W.S.L., S.M., A.K., C.S., R.T., A.G.H., P.Y.S., M.J.D., and P.B. performed experiments; A.C.D., S.V., V.G., R.E.T.T., D.I.N., K.A., M.J.R., L.T.T., F.F.Z., and R.M. analyzed data; A.C.D., S.V., C.S., and R.M. interpreted results of experiments; A.C.D. and S.V. prepared figures; A.C.D. and S.V. drafted manuscript; A.K., C.S., A.G.H., J.R.F., M.J.D., K.D.B., D.B.B., M.L.S., and E.S.B. edited and revised manuscript; J.R.F., K.D.B., D.B.B., M.L.S., and E.S.B. conception and design of research; E.S.B. approved final version of manuscript.