Elucidating the cellular mechanism for E2-induced dermal fibrosis

Background Both TGFβ and estradiol (E2), a form of estrogen, are pro-fibrotic in the skin. In the connective tissue disease, systemic sclerosis (SSc), both TGFβ and E2 are likely pathogenic. Yet the regulation of TGFβ in E2-induced dermal fibrosis remains ill-defined. Elucidating those regulatory mechanisms will improve the understanding of fibrotic disease pathogenesis and set the stage for developing potential therapeutics. Using E2-stimulated primary human dermal fibroblasts in vitro and human skin tissue ex vivo, we identified the important regulatory proteins for TGFβ and investigated the extracellular matrix (ECM) components that are directly stimulated by E2-induced TGFβ signaling. Methods We used primary human dermal fibroblasts in vitro and human skin tissue ex vivo stimulated with E2 or vehicle (ethanol) to measure TGFβ1 and TGFβ2 levels using quantitative PCR (qPCR). To identify the necessary cell signaling proteins in E2-induced TGFβ1 and TGFβ2 transcription, human dermal fibroblasts were pre-treated with an inhibitor of the extracellular signal-regulated kinase/mitogen-activated protein kinase (ERK/MAPK) pathway, U0126. Finally, human skin tissue ex vivo was pre-treated with SB-431542, a TGFβ receptor inhibitor, and ICI 182,780, an estrogen receptor α (ERα) inhibitor, to establish the effects of TGFβ and ERα signaling on E2-induced collagen 22A1 (Col22A1) transcription. Results We found that expression of TGFβ1, TGFβ2, and Col22A1, a TGFβ-responsive gene, is induced in response to E2 stimulation. Mechanistically, Col22A1 induction was blocked by SB-431542 and ICI 182,780 despite E2 stimulation. Additionally, inhibiting E2-induced ERK/MAPK activation and early growth response 1 (EGR1) transcription prevents the E2-induced increase in TGFβ1 and TGFβ2 transcription and translation. Conclusions We conclude that E2-induced dermal fibrosis occurs in part through induction of TGFβ1, 2, and Col22A1, which is regulated through EGR1 and the MAPK pathway. Thus, blocking estrogen signaling and/or production may be a novel therapeutic option in pro-fibrotic diseases.


Background
Transforming growth factor beta (TGFβ) is central to the production of extracellular matrix (ECM) [1], which promotes wound healing [2] but can also lead to organ fibrosis, especially in the connective tissue disease, systemic sclerosis (SSc) [3][4][5][6]. Several studies focus on TGFβ as an inducer of fibrosis [1], but few studies explore the regulation of TGFβ. Because TGFβ is a major inducer of fibrosis, its inhibition was the emphasis for a potential pharmacologic intervention in SSc, without success [7]. Since there are limited treatment options for SSc, it is important to understand the regulation of TGFβ isoforms to understand SSc disease pathogenesis and develop potential therapeutics.
Like TGFβ, the steroid hormone estrogen augments fibrosis. Because SSc has a female predominance [8], estrogen is hypothesized to be pathogenic in fibrosis and to contribute to sex differences noted in animal models of fibrosis. Female rats treated with bleomycin had more severe fibrosis and higher mortality than male rats [9]. In radiation-induced lung fibrosis, rats had less lung fibrosis if treated with an aromatase inhibitor that blocks estrogen production [10]. Likewise, mice with a heritable form of pulmonary hypertension that were treated with either an estrogen receptor inhibitor or an aromatase inhibitor showed significant disease improvement [11]. These results imply estrogen is influential in lung fibrosis and its blockade is a conceivable therapeutic target.
Estradiol (E2) is the most bioactive and abundant form of estrogen in non-pregnant women [12]. In the skin, E2 promotes ECM production. Women on hormone replacement therapy (HRT) have increased dermal thickness and collagen content than women not taking HRT [13][14][15][16][17][18]. Our group has shown that E2 levels are increased in postmenopausal women with SSc, and E2 induced dermal thickening and increased fibronectin (FN) in an ex vivo human skin organ culture model and in primary human dermal fibroblasts [19]. Elevated E2 levels in patients with SSc are associated with reduced survival [20], suggesting a role for E2 in SSc-fibrosis.
As with TGFβ, most studies regarding E2-induced fibrosis have centered on ECM production. However, few experiments have focused on the production of E2-induced pro-fibrotic mediators or the transcriptional regulation of these proteins. E2 induced the pro-fibrotic mediators connective tissue growth factor (CTGF) in rats [21] and TGFβ1 in a model of wound healing [22,23]. Yet, the cell signaling molecules used by E2 to induce TGFβ1 remain undefined. In this study, we seek to characterize E2-induced dermal fibrosis by describing the underlying cellular mechanisms that promote fibrosis.

Ex vivo human skin organ culture
We received skin samples from healthy donors of various ages who underwent skin-resection procedures in the Division of Plastic Surgery at the Medical University of South Carolina under an approved IRB protocol. The demographic information of the donors is summarized in Table 1. The ex vivo human skin organ culture model was used as previously described [19,24,25]. For experimentation, we used a six-well tissue culture dish (Costar, Corning, NY) with six 3mm punches/well placed dermal side down in serumfree, phenol red-free DMEM (Corning, Corning, NY). In all experiments, skin punches and supernatants were harvested at the specified times and stored at − 80°C until further evaluation.

Cell culture
Primary fibroblasts were isolated from human dermal tissue using the previously described outgrowth method [19]. Primary dermal fibroblasts (passages 3-8) were plated in six-well culture dishes (Costar, Corning, NY) at a concentration of 2.0 × 10 5 cells/well. Prior to stimulation, the cells were serum-starved in serum-free, phenol redfree DMEM (Hyclone, South Logan, UT). Cells were then treated with 10 μM of a mitogen-activated protein kinase/ extracellular signal-regulated kinase (MAPK/ERK) inhibitor, U0126 [26], 100 nM of the estrogen receptor alpha (ERα) inhibitor fulvestrant (ICI 182,780) or 10 μM of the TGFβ receptor inhibitor, SB-431542 [27], 1 h before stimulation with E2 or ethanol (ETOH) as vehicle. U0126, SB-431542, and E2 were obtained from MilliporeSigma (St. Louis, MO), ICI 182,780 from Tocris (Minneapolis, MN), and the ETOH from Hyclone (South Logan, UT). For subcellular fractions, cells were grown in 10-cm culture dishes and maintained in complete media (DMEM supplemented with 1x antibiotic and antimycotic) from Hyclone. Once 90-95% confluent, the cells were serumstarved for at least 8 h in phenol red-free media then stimulated with ETOH or E2 for the indicated timepoints. The subcellular fractions were isolated using a subcellular protein fractionation kit for cultured cells (Thermofisher Scientific, Rockford, IL). Total MAP kinases were examined in the same immunoblots as their phosphorylated forms. Signal intensities of the phosphorylated MAP kinase bands were quantitated by densitometry, and the results, which were normalized against the intensities of the corresponding total MAP kinase bands in each sample, were expressed as the magnitude of increase compared with controls. Densitometry was calculated using Image J.

Statistical analysis
All data were tested for normal distribution using the Shapiro-Wilk normality test. If the data were normally distributed, a one-way ANOVA with Sidak's multiple comparisons post hoc test or paired, two-sample t test were used to determine statistical significance. For nonnormally distributed data, non-parametric one-way ANOVA using Dunn's multiple comparisons post hoc test or Wilcoxon matched-pairs signed rank tests were used to determine statistical significance, which was defined as a p value ≤ 0.05 using GraphPad Prism v.9.

Results
Demographics Table 1 includes the demographic information of healthy subjects who donated skin tissue used in the ex vivo human skin organ culture model experiments. The majority of the donors (63%) were women, with the donated tissue primarily isolated from the abdomen (74%).

E2 increased transcription of pro-fibrotic mediators in vitro
Pro-fibrotic mediators and ECM components were measured in primary human dermal fibroblasts post E2 stimulation. TGFβ1 and TGFβ2 transcript levels were 1.4-and 1.7-fold higher in E2-stimulated cells, respectively, than vehicletreated cells at 24 h, which reached statistical significance (Fig. 1a, b). TGFβ1 and TGFβ2 transcript levels showed an increasing trend 48 h post E2 stimulation (Supplementary Figure 1, Additional file 1). However, TGFβ3 did not increase significantly at the same time point (Additional file 2).

MAPK/ERK inhibition reduced E2-induced TGFβ1 and TGFβ2 transcription and translation
We showed that E2 induced TGFβ1 and TGFβ2 transcription and ERK1/2 activation in vitro. Next, we tested the effects of blocking upstream ERK activation on TGFβ1 and TGFβ2 transcription in primary human dermal fibroblasts, using the inhibitor U0126 [26]. We first confirmed that U0126 inhibited E2-induced ERK activation (Fig. 2c). Pretreatment with the inhibitor 1 h before E2 stimulation prevented induction of TGFβ1 and TGFβ2 transcription after 24 h and TGFβ1 translation after 48 h, compared to E2 and vehicle ( Fig. 2d-f).

E2 increased transcription of ECM components and profibrotic mediators ex vivo
Even though both in vitro and in vivo evidence suggests that E2 induces FN protein levels ex vivo [19], its effects on the steady-state mRNA levels of other ECM products are unknown. We measured steady-state mRNA levels of known genes implicated in fibrosis in an ex vivo human skin organ culture model 24, 48, and 72 h after E2 stimulation. At 24 h post E2 stimulation, TGFβ2 transcript levels increased in E2-stimulated vs. vehicletreated skin (Fig. 4a), and the effect subsided 48 h post E2 stimulation (Additional file 6). Col22A1 is an early, TGFβ-responsive gene that contributes to the ECM [36]. E2 also induced Col22A1 transcription at 48 h post stimulation (Fig. 4b), not at 72 h (Additional file 6). Interestingly, similar increases were observed for steadystate mRNA levels of TGFβ1, albeit at 48 h post E2 stimulation (Fig. 4c)

Inhibiting E2 signaling through estrogen receptor alpha inhibited Col22A1 transcription ex vivo
We demonstrated that E2-induced FN production and dermal thickness ex vivo are inhibited by fulvestrant (ICI 182,780), which inhibits estrogen signaling through one of its major receptors, estrogen receptor alpha (ERα) [19]. We tested if E2 inhibition through ERα blockade would prevent the E2-induced increase in Col22A1 transcription ex vivo. After 1 h pre-treatment of human tissue with the ERα inhibitor, ICI 182,780, then subsequent ETOH and E2 treatment, E2-induced Col22A1 transcription was significantly decreased (Fig. 6).

Discussion
Both E2 and TGFβ1 are central to ECM production in the skin; however, little is known about the cellular mechanisms underlying E2-induced dermal fibrosis or the transcriptional and translational regulation of TGFβ.
We are the first to report that, in the skin, E2 induces TGFβ1 and TGFβ2 through ERα signaling, MAPK activation, and EGR1 induction. TGFβ1 and TGFβ2, in turn, induce Col22A1. Based on these observations, we propose a cellular mechanism that mediates E2-induced dermal fibrosis. Col22A1 has been identified in patients with SSc and is known to contribute to fibrosis. Whole exome sequencing of patients with diffuse cutaneous SSc showed an enrichment of genes in the ECM pathway, including Col22A1 [37]. Additionally, whole exome sequencing of African American patients with SSc identified Col22A1 as a rare variant that may increase African American susceptibility to SSc [38]. To characterize the regulation of Col22A1, our group reported that primary human dermal fibroblasts from patients with SSc and healthy controls release Col22A1 in direct response to TGFβ1 stimulation. We also showed that Col22A1 mediates the transition of fibroblasts to myofibroblasts [36]. These data indicate that Col22A1 is detected in patients with SSc and likely contributes to fibrosis. Our current study extends the understanding of Col22A1 induction by demonstrating that it can occur in direct response to a pro-fibrotic mediator that induces TGFβ, namely E2. Thus, E2 contributes to dermal fibrosis through inducing Col22A1.
We show that both TGFβ1 and TGFβ2 are induced by E2, as they are in a wound healing model [22,23], and both likely contribute to dermal fibrosis. Even though the TGFβ1 and TGFβ2 isoforms are encoded by different genes, both use the same signaling receptors and cascades [39][40][41] and likely contribute to E2-induced Col22A1 transcription.
This report also extends the number of E2-induced ECM proteins and mediators known to impact dermal fibrosis to include Col IA2, Col IIIA1, Col22A1, TGFβ1, and TGFβ2 in our ex vivo human skin organ culture model. Yet, Col IA2 and Col IIIA1 steady-state transcript levels were not increased in primary dermal fibroblast single-cell culture in vitro. The discrepancy between ex vivo and in vitro results may be explained by the inclusion of other cell types in ex vivo skin tissue, such as keratinocytes, which are responsive to E2 and can contribute to ECM formation [42][43][44].
While primary human dermal fibroblasts produce less Col IA1, Col IA2, FN, and Col22A1 when pre-treated with the TGFβ receptor inhibitor SB-431542 before stimulation with TGFβ1 [36,45], E2-induced Col IA2, Col IIIA1, and FN transcription was not prevented by pre-treatment with the inhibitor. This is likely because E2 can induce other pro-fibrotic mediators to influence ECM production. Our unpublished findings suggest that E2 increases the transcript and protein levels of the profibrotic cytokine IL-6, which can then increase collagen and FN levels through trans-signaling [46]. IL-6 also activates STAT3, which promotes Col IA2 synthesis posttranscriptionally [47]. Thus, E2-induced fibrosis likely occurs through both the non-canonical TGFβ-ERK and IL-6-STAT3 signaling pathways.
We investigated whether the MAPK pathway, which is central to E2-induced Col22A1 signaling, regulates TGFβ1 and TGFβ2. We report that E2-induced ERK1/2 phosphorylation in primary human dermal fibroblasts is vital to TGFβ1 and TGFβ2 expression, since blockade of the MAPK/ERK pathway was inhibitory. In apoptotic cells, TGFβ transcriptional and translational regulation also occurs through the MAPK pathway, in addition to the RhoA and PI-3K/AKT pathways [48]. Specifically, TGFβ1 synthesis is induced by activation of CD36 on macrophages in response to apoptotic cells [49]. Further studies are needed to elucidate the role of CD36 in E2- Transcription factors are crucial to comprehend E2induced TGFβ1 and TGFβ2 transcription. EGR1 is a candidate transcription factor in this pathway because it has been reported downstream of TGFβ1 signaling in human dermal fibroblasts [32], and the EGR1-TGFβ1 relationship is important in a murine model of pulmonary fibrosis [50]. We report that EGR1 is also upstream of E2-induced TGFβ1 transcription in primary human dermal fibroblasts, which is not surprising as EGR1 binds to and activates the TGFβ1 promoter [31] and exogenous expression of EGR1 in the cancer cell line HT1080 led to increased secretion of TGFβ1 [51]. Additionally, EGR1 increased E2-induced TGFβ2 transcription in primary dermal fibroblasts, which parallels the findings in experiments focused on tendon repair [52,53].
ERα is one of the classical estrogen receptors involved in E2 signaling [54]. Fulvestrant is FDA-approved for use in breast cancer and is effective at ameliorating bleomycin-induced lung fibrosis [55]. We extend those findings in this report to suggest that ERα signaling blockade can also prevent E2-induced Col22A1 transcription ex vivo. These results, in conjunction with our previous findings showing that circulating E2 levels are elevated in patients with SSc [20] suggest that fulvestrant is a promising therapeutic option in patients with SSc.

Conclusions
Dermal fibrosis is a key feature of some pro-fibrotic diseases, such as SSc, making it critical that we understand its underlying mechanism. Here, we suggest a cell signaling mechanism for E2-induced dermal fibrosis and TGFβ regulation. We found that E2-induced TGFβ1 and TGFβ2 directly increased Col22A1 and contributed to ECM accumulation. Currently, several FDA-approved medications exist that inhibit E2 signaling and production. However, these medications have been limited to use in hormonal cancers. Our data suggest that therapies that inhibit E2 signaling may reduce dermal fibrosis and present an interesting treatment alternative in pro-fibrotic diseases.