FSH Is Responsible for Androgen Deprivation Therapy-Associated Atherosclerosis in Mice by Exaggerating Endothelial Inflammation and Monocyte Adhesion

Abstract

Background: Androgen deprivation therapy (ADT) is the mainstay treatment for advanced prostate cancer. But ADTs with orchiectomy and gonadotropin-releasing hormone (GnRH) agonist are associated with increased risk of cardiovascular diseases, which appears less significant with GnRH antagonist. The difference of follicle-stimulating hormone (FSH) in ADT modalities is hypothesized to be responsible for ADT-associated cardiovascular diseases.

Methods: We administered orchiectomy, GnRH agonist, or GnRH antagonist in male ApoE^-/- mice fed with Western diet and manipulated FSH levels by testosterone and FSH supplementation or FSH antibody to investigate the role of FSH elevation on atherosclerosis. By combining lipidomics, in vitro study, and intraluminal FSHR (FSH receptor) inhibition, we delineated the effects of FSH on endothelium and monocytes and the underlying mechanisms.

Results: Orchiectomy and GnRH agonist, but not GnRH antagonist, induced long- or short-term FSH elevation and significantly accelerated atherogenesis. In orchiectomized and testosterone-supplemented mice, FSH exposure increased atherosclerosis. In GnRH agonist-treated mice, blocking of short FSH surge by anti-FSHβ antibody greatly alleviated endothelial inflammation and delayed atherogenesis. In GnRH antagonist-treated mice, FSH supplementation aggravated atherogenesis. Mechanistically, FSH, synergizing with TNF-α (tumor necrosis factor alpha), exacerbated endothelial inflammation by elevating VCAM-1 (vascular cell adhesion protein 1) expression through the cAMP/PKA (protein kinase A)/CREB (cAMP response element-binding protein)/c-Jun and PI3K (phosphatidylinositol 3 kinase)/AKT (protein kinase B)/GSK-3β (glycogen synthase kinase 3 beta)/GATA-6 (GATA-binding protein 6) pathways. In monocytes, FSH upregulated CD29 (cluster of differentiation 29) expression via the PI3K/AKT/GSK-3β/SP1 (specificity protein 1) pathway and promoted monocyte-endothelial adhesion both in vitro and in vivo. Importantly, FSHR knockdown by shRNA in endothelium of carotid arteries markedly reduced GnRH agonist-induced endothelial inflammation and atherosclerosis in mice.

Conclusions: FSH is responsible for ADT-associated atherosclerosis by exaggerating endothelial inflammation and promoting monocyte-endothelial adhesion.

Keywords: atherosclerosis; endothelial cells; follicle-stimulating hormone; monocytes; prostatic neoplasms.

Figure 1.

Long-term follicle-stimulating hormone (FSH) elevation is responsible for orchiectomy-accelerated atherogenesis. A, The experimental design in which ApoE^−/− mice were subjected to surgical castration (CAS) at 7 weeks and then subcutaneously injected with FSH, testosterone (T), or both every day from 8 weeks and euthanized at 20 weeks. During this time, mice were fed a Western diet. B and C, Serum levels of T and FSH in mice of the 4 groups as assessed by ELISA (n=8 mice). D and E, Oil red O staining of aortic roots (D) and aortic trees (E) of mice in the sham, CAS, CAS+T, and CAS+T+FSH groups. Representative images of aortic roots are shown in D (top) and E (left); scale bar, 500 μm. Two aortic trees are shown for each group (E); scale bar, 1 cm. The volumes of plaques in the aortic root and lesion sizes in the aortic tree were quantified and are shown (n=10 mice). F and G, Serum levels of soluble VCAM-1 (vascular cell adhesion protein 1) and MCP-1 (monocyte chemoattractant protein 1) in mice of the 4 groups as assessed by ELISA (n=8 mice). Data are presented as mean±SEM. Data were analyzed by 1-way ANOVA test followed by the Tukey multiple comparisons test (B–G).

Figure 2.

Atherosclerotic lesions increased with reduced stabilities in gonadotropin-releasing hormone (GnRH) agonist–treated mice. A, The experimental design in which 8-week-old ApoE^−/− mice were fed a Western diet and subcutaneously injected with saline, leuprolide (Leu), or degarelix (Deg) every month (8, 12, and 16 wk). One group of Leu-injected mice was also fed bicalutamide (Bica). All mice were euthanized at 20 weeks. B and C, Serum levels of testosterone and follicle-stimulating hormone (FSH) in mice of the 4 groups as assessed by ELISA (n=8 mice). D, Oil red O staining of aortic roots of mice in the saline, Leu, Deg, and Leu+Bica groups. Representative images are shown with a 500-μm scale bar. The volumes of plaques were quantified (n=8–10 mice). E, Oil red O staining of aortic trees of mice in the 4 groups. Three aortic trees were shown for each group. Scale bar, 1 cm. The lesion sizes were quantified (n=7 mice). F, SmαA (smooth muscle α-actin) immunostaining (red) on the plaques of aortic roots of mice. Representative images are shown in the left (scale bar, 300 µm) and higher magnification images for boxed areas are on the right (scale bar, 50 µm). The SMαA⁺ areas were measured to indicate smooth muscle cell content (n=7 mice). G, F4/80 immunostaining (red) on the plaques of aortic roots of mice. Representative images are shown in the left (scale bar, 300 µm) and higher magnification images for boxed areas are on the right (scale bar, 50 µm). The F4/80⁺ areas were measured to indicate macrophage content (n=7 mice). H, Heat map of significantly altered serum lipids of mice in the saline, Leu, and Deg groups. The color of each section is proportional to the significance of the change in metabolites (red, upregulated; blue, downregulated; n=5 mice). I, Principal component (PC) analysis score plots based on phospholipid profiles of the saline, Leu, and Deg groups (n=5 mice). Data are presented as mean±SEM. Data were analyzed by 1-way ANOVA test followed by the Tukey multiple comparisons test (D–G).

Figure 3.

Short-term follicle-stimulating hormone (FSH) elevation is critical to gonadotropin-releasing hormone (GnRH) agonist–associated atherogenesis. A, The experimental design in which 8-week-old ApoE^−/− mice were subcutaneously injected with leuprolide (Leu) every month (8, 12, and 16 wk) and supplemented with the anti-FSHβ antibody or control IgG intraperitoneally during the first 2 weeks (8–10 wk). Mice were fed a Western diet and euthanized at 12 or 20 weeks. B and C, Oil red O staining of aortic trees (B) and aortic roots (C) of mice in the saline, Leu+Ctrl-IgG, and Leu+anti-FSHβ groups. Representative images of aortic roots are shown in the left (B). Scale bar, 500 μm. Two aortic trees were shown for each group (C). Scale bar, 1 cm. The volumes of plaques in the aortic root and the lesion sizes in the aortic tree were quantified and are shown (n=7–8 mice). D, The experimental design in which 8-week-old ApoE^−/− mice were subcutaneously injected with degarelix (Deg) every month (8, 12, and 16 wk) and supplemented with FSH or saline intraperitoneally during the first 2 weeks (8–10 wk). Mice were fed a Western diet and euthanized at 20 weeks. E, Serum levels of FSH in mice of the Deg+FSH group as assessed by ELISA before and after FSH supplementation (n=4 mice). F and G, Oil red O staining of aortic trees (F) and aortic roots (G) of mice in the Deg+saline and Deg+FSH groups. Representative images are shown in the left. Scale bars, 1 cm (F) and 500 µm (G). The lesion sizes in the aortic tree and the plaques in the aortic root were quantified and are shown in the right (n=6 mice). Data are presented as mean±SEM. Data were analyzed by 1-way ANOVA test followed by the Tukey multiple comparisons test (B and C) and by the unpaired, 2-tailed Student t test (E–G). Ctrl indicates control.

Figure 4.

Follicle-stimulating hormone (FSH) synergized with TNF-α (tumor necrosis factor alpha) to enhance the expression of adhesion molecules in the endothelium. Human umbilical vein endothelial cells (HUVECs) were treated with dimethyl sulfoxide (DMSO), TNF-α (10 ng/mL), FSH (50 ng/mL), or FSH and TNF-α for 24 hours (A–I); or TNF-α (10 ng/mL) for 24 or 48 hours, the combination of FSH and TNF-α for 6 hours following only TNF-α for 24 or 48 hours (J and K). A through E, The effects of DMSO, TNF-α, FSH, or FSH and TNF-α on the expression of adhesion molecules were determined by qRT-PCR (quantitative reverse transcription-polymerase chain reaction; A–D, n=3 biological replicates) and Western blotting (E, n=4 biological replicates). GAPDH was used as a loading control. F, Representative images of VCAM-1 (vascular cell adhesion protein 1; green) in HUVECs (CD31 [cluster of differentiation 31] positive, red) with the indicated treatment were determined using immunofluorescence (IF). Scale bar, 50 µm. G, Levels of MCP-1 (monocyte chemoattractant protein 1) in culture media of HUVECs with the indicated treatment as assessed by ELISA; n=5 biological replicates. H and I, The adhesion of GFP (green fluorescent protein)-labeled human monocytic THP-1 cells to HUVECs with the indicated treatment was observed under a fluorescence microscope (scale bar, 300 µm). The adhered cell numbers were counted from images of 3 biological replicates, and the data are shown as the fold change (ratio to DMSO). J, Effect of transient stimulation with FSH (6 h) on VCAM-1 and E-selectin expression by Western blotting. n=4 biological replicates. K, Representative images of VCAM-1 (green) in HUVECs (CD31 positive, red) with the indicated treatment (24 h) were determined using IF. Scale bar, 50 µm. Data are presented as mean±SEM. Data were analyzed by 1-way ANOVA test followed by the Tukey multiple comparisons test (A–D, G, and H) and by 2-way ANOVA test followed by the Tukey multiple comparisons test (E and J). Sample coll. indicates sample collection.

Figure 5.

Follicle-stimulating hormone (FSH) promoted adhesion of monocytes to endothelium. Human monocytic THP-1 cells and human primary CD11b⁺CD14⁺ cells were infected with GFP (green fluorescent protein) adenovirus for 48 hours, treated with FSH (50 ng/mL) for 24 hours, and then cocultured with human umbilical vein endothelial cells (HUVECs; A–E, and K). THP-1 cells were exposed to FSH (50 ng/mL) for 24 hours (F–J). A and B, The effect of FSH on the adhesion of THP-1 cells to resting HUVEC monolayers (scale bar, 300 µm); n=3 biological replicates. C and D, The effects of FSH, anti-FSHR (FSH receptor) antibody on THP-1 cell adhesion. GFP-labeled THP-1 cells with anti-FSHR antibody or control (Ctrl) IgG pretreatment were treated with FSH and then cocultured with HUVECs. Scale bar, 300 µm; n=6 mice. E, The effect of FSH on the adhesion of CD11b⁺CD14⁺ monocytes (n=5 biological replicates; scale bar, 300 µm). F, The effect of FSH on the adhesion of THP-1 cells in the normal blood flow of wild-type C57BL/6 mice. Representative statistics about the number of firm adhesion and rolling adhesion THP-1 cells are shown on the right. Black arrows indicate rolling adhesion THP-1 cells, and blue arrows indicate firm adhesion THP-1 cells (n=6 mice; scale bar, 50 µm). G and H, The effects of FSH on the expression of adherent molecules were determined by qRT-PCR (quantitative reverse transcription polymerase chain reaction; G, n=5 biological replicates) and Western blotting (H, n=3 biological replicates). I and J, The effects of FSH on membrane CD29 were determined by immunofluorescence (I) and flow cytometry (J); n=3 biological replicates. K, The effect of anti-CD29 antibody on the adhesion of THP-1 cells. THP-1 cells with the indicated FSH treatment were exposed to anti-CD29 antibody or Ctrl-IgG antibody for 1 hour and then cocultured with HUVECs (scale bar, 300 µm); n=3 biological replicates. Data are presented as mean±SEM. Data were analyzed by unpaired, 2-tailed Student t test (B and E–J) and by 1-way ANOVA test followed by the Tukey multiple comparisons test (K). CD indicates cluster of differentiation; DMSO, dimethyl sulfoxide; FITC, fluorescein isothiocyanate; and PBMC, peripheral blood mononuclear cell.

Figure 6.

The cAMP/PKA (protein kinase A)/CREB (cAMP response element-binding protein) and the PI3K (phosphatidylinositol 3 kinase)/AKT (protein kinase B)/GSK-3β (glycogen synthase kinase 3 beta) pathways were involved in the synergistic action of follicle-stimulating hormone (FSH) and TNF-α (tumor necrosis factor alpha) on endothelial inflammation, tightly regarding the transcription factors c-Jun and GATA-6 (GATA-binding protein 6). Human umbilical vein endothelial cells (HUVECs) were pretreated with the FSHR (FSH receptor) antibody, PKA inhibitor H89 (10 μM), PI3K inhibitor LY294002 (10 μM), or FSHR antibody for 1 hour and then treated with TNF-α (10 ng/mL), FSH (50 ng/mL), or FSH+TNF-α for 6 hours. A, The effects of FSHR blockade on the expression of VCAM-1 (vascular cell adhesion protein 1), E-selectin, phosphorylated CREB (p-CREB), and total CREB in HUVECs with the indicated treatment. GAPDH served as a loading control (Ctrl); n=5 biological replicates. B and C, The effects of PKA inhibitor H89 (B) and PI3K inhibitor LY204002 (C) on the expression of VCAM-1 and E-selectin in HUVECs. p-CREB, total CREB, phosphorylated AKT (p-AKT), and total AKT were also detected to indicate the effectiveness of inhibitors; n=5 biological replicates. D and E, The effects of PKAα knockdown (D) and p85α knockdown (E) on the Gαs (G-protein alpha s)/cAMP/PKA/CREB and the PI3K/AKT/GSK-3β pathways in HUVECs; n=5 biological replicates. F, The effect of simultaneous c-Jun and GATA-6 knockdown on VCAM-1 and FSHR expression in HUVECs; n=5 biological replicates. G and H, Enrichment of c-Jun and GATA-6 on VCAM-1 promoter as determined by the chromatin immunoprecipitation assay in HUVECs with indicated treatment for 6 hours. IgG was served as negative Ctrl; n=3 biological replicates. I and J, Representative images of c-Jun (I, green) and GATA-6 (J, green) in the endothelium (CD31 positive, red) of the ascending aorta of mice (n=8) were determined using immunofluorescence (IF). Scale bar, 50 µm. Higher magnification images for boxed areas are shown in the right. Scale bar, 25 µm. White arrows indicate c-Jun or GATA-6+ cells. Data are presented as mean±SEM. Data were analyzed by 2-way ANOVA test followed by the Tukey multiple comparisons test (A–F). CAS indicates castration; CD, cluster of differentiation; DMSO, dimethyl sulfoxide; p-c-Jun, phosphorylated c-Jun; p-GSK-3β, phosphorylated GSK-3β; and T, testosterone.

Figure 7.

Follicle-stimulating hormone (FSH) upregulated CD29 (cluster of differentiation 29) expression through the PI3K (phosphatidylinositol 3 kinase)/AKT (protein kinase B)/GSK-3β (glycogen synthase kinase 3 beta)/SP1 (specificity protein 1) signal transduction pathway. A through E and J, Human monocytic THP-1 cells were pretreated with anti-FSHR (FSH receptor) antibody for 1 hour or with the PI3K inhibitor LY294002 (10 μM) for 3 hours and then treated with FSH (50 ng/mL) for 24 hours. F and J, THP-1 cells were transfected with SP1 siRNAs for 24 hours and then treated with or without FSH. A, Effects of FSHR blockade on CD29 expression in THP-1 cells with the indicated treatment. β-actin served as a loading control (Ctrl); n=3 biological replicates. B, Effects of PI3K inhibition on the expression of CD29 in THP-1 cells. Phosphorylated AKT (p-AKT) and total AKT were also detected to indicate the effectiveness of inhibitors; n=3 biological replicates. C, The effects of PI3K inhibition on THP-1 cell adhesion to human umbilical vein endothelial cells (HUVECs). Scale bar, 300 µm. Data were from images of 3 biological replicates. D, Effects of PI3K inhibition on membrane CD29 distribution in THP-1 cells. Data were from images of 3 biological replicates. E, The effects of PI3K inhibition on the expression of phosphorylated GSK-3β (p-GSK-3β), total GSK-3β, phosphorylated SP1 (p-SP1), and total SP1 in THP-1 cells; n=3 biological replicates. F, Effects of SP1 knockdown on the expression of CD29 in THP-1 cells. β-actin served as a loading Ctrl; n=3 biological replicates. G, Effects of FSH on SP1 transcriptional activity. THP-1 cells were transfected with a specific dual-luciferase system that contains multiple SP1-binding sites. After 24 hours of transfection, THP-1 cells were exposed to FSH for 24 hours. The relative luciferase represents SP1 transcriptional activity; n=3 biological replicates. H, The effects of FSH on the association of GSK-3β and SP1 in THP-1 cells. GSK-3β antibody was used to coimmunoprecipitate SP1 with or without FSH treatment in THP-1 cells. I, Enrichment of SP1 on the CD29 promoter as determined by the chromatin immunoprecipitation assay in THP-1 cells with indicated treatment for 6 hours. IgG was served as negative Ctrl; n=3 biological replicates. J, Effects of FSH on SP1 cellular location. Immunofluorescence staining showed SP1 localization in THP-1 cells treated with FSH for 12 hours. Scale bar, 10 μm. Data were from images of 3 biological replicates. Data are presented as mean±SEM. Data were analyzed by 2-way ANOVA test followed by the Tukey multiple comparisons test (A, B, E, F, and I) and by 1-way ANOVA test followed by the Tukey multiple comparisons test (C and D). Data in G were analyzed by unpaired, 2-tailed Student t test. IB indicates immunoblotting; and IP, immunoprecipitation.

Figure 8.

Intraluminal inhibition of FSHR (follicle-stimulating hormone receptor) ameliorated endothelial inflammation and atherosclerosis. A, The experimental design in which ApoE^−/− mice were subjected to partial carotid ligation incubated with adenovirus vector with FSHR-targeted shRNA (Ad-sh-FSHR) or control virus at 7 weeks and then subcutaneously injected with saline or leuprolide (Leu) at 8 weeks. Mice were fed a Western diet and euthanized at 12 weeks. B and C, Representative images of hematoxylin & eosin (H&E)-stained ligated carotid arteries (B) and quantification of the intima/media ratio (C) of mice in the saline+Ad-sh-CL, saline+Ad-sh-FSHR, Leu+Ad-sh-CL, and Leu+Ad-sh-FSHR groups (n=8 mice). Scale bar, 100 μm. D and E, Representative images of oil red O–stained ligated carotid arteries (D) and quantification of plaques in the common carotid arteries of each group (E); n=8 mice. F, Representative images of VCAM-1 (vascular cell adhesion protein 1; red) expression in the endothelium (CD31 [cluster of differentiation 31] positive, red) from the common carotid arteries of 12-week-old mice (n=8) were determined using immunofluorence (IF). Scale bar, 100 µm. G, Representative images of GATA-6 (GATA-binding protein 6; green) expression colocalized with DAPI (4′,6-diamidino-2-phenylindole) from the common carotid arteries using IF. Scale bar, 50 µm. H, Proposed mechanisms of follicle-stimulating hormone (FSH) in atherosclerosis progression. Data are presented as mean±SEM. Data were analyzed by 2-way ANOVA test followed by the Tukey multiple comparisons test (C and E). Ad-sh-CL indicates adenovirus vector with control-targeted shRNA; AKT, protein kinase B; CREB, cAMP response element-binding protein; Gαs, G-protein alpha s; GSK-3β, glycogen synthase kinase 3 beta; PKA, protein kinase A; SP1, specificity protein 1; TNF-α, tumor necrosis factor alpha; and TNFR1, tumor necrosis factor receptor 1.