Modulation of neointimal lesion formation by endogenous androgens is independent of vascular androgen receptor

Abstract

Aims: Low androgen levels have been linked with an increased risk of cardiovascular disease in men. Previous studies have suggested that androgens directly inhibit atherosclerotic lesion formation although the underlying mechanisms for this remain unclear. This study addressed the hypothesis that endogenous androgens inhibit arterial remodelling by a direct action on the androgen receptor (AR) in the vascular wall.

Methods and results: We studied a series of novel mouse lines with cell-specific deletion of the AR in either the endothelium or in smooth muscle cells or both cell types. Findings were compared with a model of global androgen deficiency in wild-type mice (castrated). We characterized the cardiovascular phenotype, vascular pharmacology and histology, and assessed neointimal lesion formation following vascular injury to the femoral artery. Cell-specific AR deletion did not alter body weight, circulating testosterone levels or seminal vesicle weight, but caused limited alterations in arterial contractility and blood pressure. Neointimal lesion formation was unaltered by selective deletion of AR from the vascular endothelium, smooth muscle, or both cell types. Castration in wild-type mice increased neointimal lesion volume (Sham vs. Castration: 2.4 × 10(7) ± 4.5 × 10(6) vs. 3.9 × 10(7) ± 4.9 × 10(6) µm(3), P = 0.04, n = 9-10).

Conclusion: Vascular cell-specific AR deletion had no effect on neointimal lesion formation, while low systemic androgen levels adversely affect neointimal lesion size. These findings suggest that the cardio-protective effects of androgens are mediated either by AR outside the vasculature or by AR-independent mechanisms.

Keywords: Androgen receptor; Arterial injury; Neointima; Testosterone.

Figure 1

Identification of AR in murine vascular cells. AR is expressed in EC (red arrows) and SMC (green arrows) in healthy mouse aorta (A). The dashed yellow line indicates the external elastic lamina. AR expression was up-regulated by testosterone (1 × 10⁻⁷ M) in cultured mouse aortic EC (B) and SMC (C). EC were identified using antibodies against vWF and CD31; SMCs were identified using an antibody against SMA. Nuclei were counter-stained with DAPI.

Figure 2

Characterization of mice with vascular cell-specific AR deletion. Cell-specific AR deletion was confirmed (A; n = 4 for each genotype) using PCR on genomic DNA from freshly isolated aortic EC and SMC. Mouse genotypes were confirmed (B) with PCR using genomic DNA from ear clip samples. Deletion of vascular AR did not alter circulating testosterone levels (C; n = 14–23) or seminal vesicle weight (D; n = 8–15) but deletion of AR from EC (VE-ARKO) produced a small increase in systolic blood pressure (E). *P < 0.05 (n = 7–12) by one-way ANOVA plus Bonferroni post-hoc test. (WT = wild-type litter mates carrying floxed-AR; SM-ARKO = AR ablated in SMC, VE-ARKO = AR ablated in EC, SM/VE-ARKO = AR ablated in both EC and SMC.)

Figure 3

Agonist-dependent vascular dysfunction in mice with selective deletion of vascular AR. In isolated femoral (A, B, E) and mesenteric (C, D, F) arteries cumulative concentration-response curves were produced using phenylephrine (PhE; A and C) or endothelin-1 (ET-1; B and D). Acetylcholine (ACh; E and F) induced-relaxation was obtained after contraction with a sub-maximal concentration of PhE (3 × 10⁻⁶ M). ^$$P < 0.01 vs. WT; *P < 0.05, **P < 0.01 vs. corresponding WT concentration; ^#P < 0.05, ^##P < 0.01 vs. corresponding VE-ARKO concentration; two-way ANOVA with the Bonferroni post-hoc test. (WT = wild-type litter mates carrying floxed-AR; SM-ARKO = AR ablated in SMC, VE-ARKO = AR ablated in EC, SM/VE-ARKO = AR ablated in both EC and SMC. n = 5–9.)

Figure 4

Testosterone induces relaxation in vascular ARKO arteries. Supra-physiological concentration of testosterone induced vascular relaxation both in femoral (A) and mesenteric (B) arteries from all genotypes, which was independent of the type of pre-constriction. Vascular ARKO produced no dramatic changes in testosterone-mediated relaxation despite some small differences in relaxation at specific concentrations. *P < 0.05, **P < 0.01 vs. corresponding WT concentration; ^#P < 0.05 vs. corresponding VE-ARKO concentration; two-way ANOVA plus Bonferroni post-hoc test. (WT = wild-type litter mates carrying floxed-AR; SM-ARKO = AR ablated in SMC, VE-ARKO = AR ablated in EC, SM/VE-ARKO = AR ablated in both EC and SMC. n = 7–9.)

Figure 5

Castration reduces vascular AR expression. Plasma testosterone concentrations (A), seminal vesicle weight (B), and AR expression in lesion-bearing femoral arteries (C) were reduced following castration. **P < 0.01. Data were analysed by Student's t-test. (A: n = 9–11; B: n = 8–15; C: n = 9).

Figure 6

Castration increases neointimal lesion formation following wire injury but not following arterial ligation. In arteries subjected to either (A) wire-induced injury or (B) ligation, neointimal lesion distribution (i) and neointimal volume (ii) were determined by OPT. Maximal cross-sectional narrowing (iii) was measured in serial sections stained with Masson's trichrome. Panels A(i) and B(i) show mean neointimal lesion volumes for each group; error bars have been omitted for clarity. Panels A(ii & iii) and B(ii & iii) show individual data points from each animal in the group with lines and error bars indicating mean ± SEM. *P < 0.05, by Student's t-test (A: n = 8–10; B: n = 6–8 ).

Figure 7

Effect of vascular-specific AR ablation on neointimal lesion formation. Lesion formation following (A) wire-induced injury (n = 7–14) or (B) ligation-induced injury (n = 6–14) was determined by optical projection tomography (OPT) and histology. Panels A(i) and B(i) show mean neointimal lesion volumes for each genotype; error bars have been omitted for clarity. Panels A(ii and iii) an B(ii and iii) show individual data points from each animal with lines and error bars indicating mean ± SEM. Vascular AR deletion had no effect on lesion formation in response to wire-induced injury. Selective deletion of AR from ECs produced a small increase in neointimal lesion volume following ligation injury (*P < 0.05 by one-way ANOVA plus Bonferroni post-hoc test) but did not alter maximal cross-sectional narrowing. (WT = wild-type litter mates carrying floxed-AR; SM-ARKO = AR ablated in SMC, VE-ARKO = AR ablated in EC, SM/VE-ARKO = AR ablated in both EC and SMC.)