A Positive Feedback Loop Exists between Estradiol and IL-6 and Contributes to Dermal Fibrosis

Abstract

Systemic sclerosis (SSc) is characterized by dermal fibrosis with a female predominance, suggesting a hormonal influence. Patients with SSc have elevated interleukin (IL)-6 levels, and post-menopausal women and older men also have high estradiol (E2) levels. In the skin, IL-6 increases the enzymatic activity of aromatase, thereby amplifying the conversion of testosterone to E2. Therefore, we hypothesized that an interplay between E2 and IL-6 contributes to dermal fibrosis. We used primary dermal fibroblasts from healthy donors and patients with diffuse cutaneous (dc)SSc, and healthy donor skin tissues stimulated with recombinant IL-6 and its soluble receptor (sIL-6R) or E2. Primary human dermal fibroblasts and tissues from healthy donors stimulated with IL-6+sIL-6R produced E2, while E2-stimulated dermal tissues and fibroblasts produced IL-6. Primary dermal fibroblasts from healthy donors treated with IL-6+sIL-6R and the aromatase inhibitor anastrozole (ANA) and dcSSc fibroblasts treated with ANA produced less fibronectin (FN), type III collagen A1 (Col IIIA1), and type V collagen A1 (Col VA1). Finally, dcSSc dermal fibroblasts treated with the estrogen receptor inhibitor fulvestrant also generated less FN, Col IIIA1, and Col VA1. Our data show that IL-6 exerts its pro-fibrotic influence in human skin in part through E2 and establish a positive feedback loop between E2 and IL-6.

Keywords: IL-6; aromatase; dermal fibrosis; estradiol; systemic sclerosis.

Figure 1

Relative quantity (RQ) IL-6 steady-state transcript and protein levels in E2-treated dermal tissue and fibroblasts. (a) RQ IL-6 steady-state mRNA 24 or 48 h post-vehicle (ETOH) or E2 treatment [10 nM] in human skin tissue ex vivo. n = 7–8 individual samples, two-way ANOVA. (b) RQ of secreted IL-6 protein 24, 48, or 72 h post-ETOH vs. E2 treatment in skin tissue ex vivo. n = 9 individual samples. (c) RQ steady-state mRNA levels of IL-6 post-vehicle vs. E2 treatment in human dermal fibroblasts from healthy donors. (d). RQ of secreted IL-6 protein 48 h post-ETOH vs. E2 treatment in human dermal fibroblasts in vitro. n = 5 individual cell lines, two-way ANOVA with post−hoc analysis (a–c) after log transformation (c) or paired t-test (d). Error bars = SEM. * p < 0.05, ** p < 0.01.

Figure 2

Steady-state CYP19A1 transcript levels in dermal tissue ex vivo and fibroblasts in vitro. (a) Steady-state CYP19A1 mRNA levels 24 and 48 h post-vehicle vs. E2 treatment in healthy donor human skin tissue ex vivo. n = 8 individual samples, two-way ANOVA. (b) Steady-state CYP19A1 mRNA levels 8, 16, or 24 h post-vehicle vs. E2 treatment [10 nM] in human primary dermal fibroblasts. n = 4–5 individual cell lines, two-way ANOVA. (c) Steady-state mRNA levels CYP19A1 24 h post-vehicle (phosphate-buffered saline (PBS)) vs. IL-6+sIL-6R [20 ng/mL] treatment in primary human dermal fibroblasts. n = 5 individual cell lines, paired t-test. Error bars = SEM. * p < 0.05.

Figure 3

IL-6-induced aromatase activity in dermal tissue ex vivo and dermal fibroblasts in vitro. (a) Relative quantity (RQ) of secreted E2 production 72 h post-IL-6+sIL-6R [20 ng/mL] treatment in human skin tissue ex vivo. n = 9 individual samples, one-way ANOVA after log transformation with post−hoc analysis. (b) RQ of secreted E2 production 96 h post-IL-6+sIL-6R [20 ng/mL] treatment in human primary dermal fibroblasts in vitro. n = 5 individual cell lines with 3 measured in duplicate, one-way ANOVA after log transformation with post−hoc analysis. In all conditions, 10 nM of testosterone was added to both skin tissue and fibroblast cultures. V1 = PBS, V2 = DMSO. Error bars = SEM. * p < 0.05, *** p < 0.005, **** p < 0.0001.

Figure 4

IL-6-induced type III Col A1 and type V Col A1 protein levels in dermal fibroblasts. Representative immunoblots and densitometry of Col IIIA1 (a) and Col VA1 (b) protein levels in primary human dermal fibroblasts from healthy donors treated for 96 h with IL-6+sIL-6R [20 ng/mL]. Fibroblasts were treated with DMSO or ANA [300 nM] 1 h prior to PBS or IL-6+sIL-6R. For Col VA1 (b), both the 220 and 140 kD bands were used for calculation. Protein levels were normalized to GAPDH and presented as fold change. V1 = PBS, V2 = DMSO. n = 4 individual cell lines used in 9 independent experiments. One-way ANOVA with post hoc analysis. Error bars = SEM. * p < 0.05, ** p < 0.01.

Figure 5

CYP19A1 relative quantity and aromatase enzymatic activity in dcSSc dermal fibroblasts in vitro. (a) Steady-state mRNA levels of CYP19A1 at baseline in primary dermal fibroblasts from healthy donors vs. dcSSc patients. CYP19A1 gene expression was normalized to beta 2 micoglobulin (B2M). n = 6 individual healthy donor cell lines (one healthy control donor measured in duplicate), n = 5 SSc donor fibroblast cell lines, p-value calculated using unpaired t-test after log transformation. (b) Relative quantity (RQ) of secreted E2 produced 96 h post-IL-6+sIL-6R [20 ng/mL] treatment in SSc primary dermal fibroblasts with vehicle or ANA inhibition [300 nM]. n = 6 individual cell lines, 4 measured in duplicate, one-way ANOVA with post hoc analysis. (c) Comparison of E2 production in primary dermal fibroblasts from healthy donors vs. dcSSc patients post-IL-6+sIL-6R [20 ng/mL] treatment and ANA inhibition [300 nM]. Data are relative to fibroblasts treated with their respective vehicles. n = 5 individual cell lines, measured in duplicate. Unpaired t-test. Error bars = SEM. * p < 0.05.

Figure 6

Steady-state ECM transcript and protein levels in ANA-treated dcSSc dermal fibroblasts. FN (top a), Col IIIA1 (top b), and Col VA1 (top c) fold change transcript levels were quantified in primary dermal fibroblasts from dcSSc patients 48 h (a,c) or 96 h (b) post-treatment with ANA [300 nM]. mRNA levels were normalized to GAPDH. Representative immunoblots and densitometry of FN (bottom a), Col IIIA1 (bottom b), and Col VA1 (bottom c) protein levels in primary dermal fibroblasts from dcSSc patients 120 h (a,c) or 96 h (b) post-treatment with ANA [300 nM]. For Col VA1 (c), both the 220 and 140 kD bands were used in the calculation. Protein levels were normalized to GAPDH and presented as fold change. n = 5 individual cell lines used in 6–14 independent experiments. Error bars = SEM. Paired t-test, protein data log transformed, * p < 0.05.

Figure 7

Reduction in FN and type IIIA1 and VA1 collagen protein levels produced by dcSSc dermal fibroblasts after fulvestrant treatment. Representative immunoblots and densitometry of FN and type IIIA1 and VA1 collagen protein levels in primary dermal fibroblasts from dcSSc patients 96 (a,c) or 120 (b) hours post-treatment with fulvestrant (ICI) [100 nM]. Protein levels were normalized to GAPDH. For Col VA1 (c), both the 220 and 140 kD bands were used in the calculation. n = 7 individual cell lines in 7 (a,c) and 11 (b) independent experiments. Error bars = SEM. Paired t-test, * p < 0.05.