Influence of Androgen Receptor in Vascular Cells on Reperfusion following Hindlimb Ischaemia

Abstract

Aims: Studies in global androgen receptor knockout (G-ARKO) and orchidectomised mice suggest that androgen accelerates reperfusion of the ischaemic hindlimb by stimulating angiogenesis. This investigation used novel, vascular cell-specific ARKO mice to address the hypothesis that the impaired hindlimb reperfusion in G-ARKO mice was due to loss of AR from cells in the vascular wall.

Methods and results: Mice with selective deletion of AR (ARKO) from vascular smooth muscle cells (SM-ARKO), endothelial cells (VE-ARKO), or both (SM/VE-ARKO) were compared with wild type (WT) controls. Hindlimb ischaemia was induced in these mice by ligation and removal of the femoral artery. Post-operative reperfusion was reduced in SM-ARKO and SM/VE-ARKO mice. Immunohistochemistry indicated that this was accompanied by a reduced density of smooth muscle actin-positive vessels but no change in the density of isolectin B4-positive vessels in the gastrocnemius muscle. Deletion of AR from the endothelium (VE-ARKO) did not alter post-operative reperfusion or vessel density. In an ex vivo (aortic ring culture) model of angiogenesis, AR was not detected in vascular outgrowths and angiogenesis was not altered by vascular ARKO or by exposure to dihydrotestosterone (DHT 10(-10)-10(-7)M; 6 days).

Conclusion: These results suggest that loss of AR from vascular smooth muscle, but not from the endothelium, contributes to impaired reperfusion in the ischaemic hindlimb of G-ARKO. Impaired reperfusion was associated with reduced collateral formation rather than reduced angiogenesis.

Fig 1. Impact of vascular AR deletion on pre- and post-operative hindlimb perfusion.

(A) Laser Doppler was used to monitor perfusion of the foot pad immediately before and after femoral artery ligation, which was quantified by software and normalised to tail. (Areas of interest are indicated by yellow boxes. The grey area is excluded by software automatically.) (B) Perfusion in the un-operated hindlimb was unaffected by selective deletion of vascular AR. Femoral artery removal produced a dramatic reduction in hindlimb perfusion that was similar in wild type and ARKO mice. ** p<0.01 compared with corresponding pre-operative measurements (n = 14–16).

Fig 2. Deletion of AR from smooth muscle, but not from endothelium, reduced post-operative reperfusion of the ischaemic hindlimb.

(A) Laser Doppler imaging demonstrated partial restoration of hindlimb perfusion over 21 days following femoral artery removal. The blood flow in the foot pad was quantified and normalised to flow in the tail. (Areas of interest are indicated by yellow boxes. The grey area is excluded by software automatically.) (B) Deletion of AR from smooth muscle impaired this recovery, whereas deletion of AR from endothelial cells did not. $ p<0.01 difference among genotypes; * p<0.01 versus WT; # p<0.01 versus VE-ARKO. Data were analysed by two way ANOVA plus Tukey post hoc test (n = 14–16).

Fig 3. AR deletion from vascular smooth muscle, but not from endothelial cells, altered vascular density in ischaemic gastrocnemius.

(A) Detection of endothelial cells (IB4-positive) and smooth muscle cells (SMA-positive) in vascular structures was achieved using immunohistochemistry on cross-sections of gastrocnemius muscle. High resolution tile-scanning (@ x200) using slide scanner Axio Scan.Z1 (Zeiss) allowed quantification of vascular phenotype and density of the whole muscle section. (B) Quantitation of vessel density indicated that deletion of AR from neither vascular smooth muscle nor endothelium altered the rise in density of IB4-positive vessels in ischaemic hindlimb. **, p<0.01 by two way ANOVA. (C) Deletion of AR from vascular smooth muscle, but not endothelium, prevented the ischaemia-induced increase in SMA-positive vessels. * p<0.05; ** p<0.01; by two way ANOVA plus Tukey post hoc test. ## P<0.01, by Student’s t-test (n = 7–10).

Fig 4. Localisation of vascular ARKO in murine ischaemic gastrocnemius.

(A) AR was detected in the vascular smooth muscle (AR = green, SMA = red, DAPI = blue) in the ischaemic gastrocnemius muscle at day 21 from WT and VE-ARKO, but not SM-ARKO and SM/VEARKO. No positive staining of AR could be identified on the luminal (endothelial) side of the vessel in any genotypes. (Yellow arrows indicate AR-positive nuclei) (B) Double staining of AR and CD31 in WT confirmed that AR was not expressed in vascular endothelium in the ischaemic hindlimb. (White arrows indicate nuclei of endothelial cells.) Some AR positive cells were also found in the perivascular connective tissue.

Fig 5. Androgen signalling did not regulate ex vivo angiogenesis in the aortic ring assay.

Descending thoracic aorta were from all four genotypes were freshly dissected, cut into aortic rings and embedded in collagen gel. Endothelial cell sprouting and tube-like structure formation was stimulated by culture medium with 1% FBS. Tube-like structures were counted manually at day 6. (A) Representative phase-contrast image of tube-like structure (arrows) growing from a WT aortic ring. (B) Representative images of AR expression in tube-like structure from WT aortic rings with/without DHT treatment. Yellow lines highlight a tube-like structure. (C) Neither DHT nor vascular ARKO significantly altered angiogenesis in the aortic ring assay (By two way ANOVA). VEGF (5ng/ml) treatment in WT rings served as positive control for maximum angiogenesis (*, p<0.01 by Student's t-test). Exposures were performed in triplicates, and repeated 6 times, with the exception of the VEGF group, which was repeated 4 times.