Antifibrogenic Activities of CYP11A1-derived Vitamin D3-hydroxyderivatives Are Dependent on RORγ

Abstract

Previous studies showed that noncalcemic 20(OH)D₃, a product of CYP11A1 action on vitamin D₃, has antifibrotic activity in human dermal fibroblasts and in a bleomycin mouse model of scleroderma. In this study, we tested the role of retinoic acid-related orphan receptor γ (RORγ), which is expressed in skin, in the action of CYP11A1-derived secosteroids using murine fibroblasts isolated from the skin of wild-type (RORγ ^+/+), knockout (RORγ ^-/-), and heterozygote (RORγ ^+/-) mice. CYP11A1-derived 20(OH)D₃, 20,23(OH)₂D₃, 1,20(OH)₂D₃, and 1,20,23(OH)₃D₃ inhibited proliferation of RORγ ^+/+ fibroblasts in a dose-dependent manner with a similar potency to 1,25(OH)₂D₃. Surprisingly, this effect was reversed in RORγ ^+/- and RORγ ^-/- fibroblasts, with the most pronounced stimulatory effect seen in RORγ ^-/- fibroblasts. All analogs tested inhibited TGF-β1-induced collagen synthesis in RORγ ^+/+ fibroblasts and the expression of other fibrosis-related genes. This effect was curtailed or reversed in RORγ ^-/- fibroblasts. These results show that the antiproliferative and antifibrotic activities of the vitamin D hydroxy derivatives are dependent on a functional RORγ. The dramatic changes in the transcriptomes of fibroblasts of RORγ ^-/- versus wild-type mice following treatment with 20(OH)D₃ or 1,20(OH)₂D₃ provide a molecular basis to explain, at least in part, the observed phenotypic differences.

The retinoic acid-related orphan receptors (RORs), RORα, RORβ, and RORγ, constitute a subfamily of nuclear receptors (1). They regulate the transcription of target genes by binding to ROR-response elements (2). ROR receptors play a critical role in the control of various physiological and developmental processes and are implicated in the development of many diseases, including several malignancies, inflammatory and metabolic diseases, and psoriasis (2, 3). The RORγ gene (RORC) gives rise to 2 different receptor isoforms: RORγ1 and RORγt (RORγ2). RORγ1 is expressed in several metabolic tissues such as liver, kidney, skeletal muscles, small intestine, pancreas, and brown fat cells (4), and plays a role in insulin sensitivity and obesity (5, 6). RORγt expression is restricted to several immune cells (1). Mice lacking the RORγ gene lack lymph node organogenesis as well as an impaired allergic response (1) and increased susceptibility to thymic lymphoma (7). Treatment with RORγ inverse agonists diminishes the severity of autoimmune diseases such as diabetes and multiple sclerosis, and inflammatory lung and skin disease, in experimental mouse models (8-13). Thus, RORγ inverse agonists might potentially be useful in the treatment of collagen-induced arthritis, rheumatoid arthritis, asthma, and other autoimmune diseases, as well as obesity (14, 15) and psoriasis (11). Novel vitamin D3 derivatives used in this study are among the many endogenous ligands for RORs (16, 17).

New pathways of vitamin D activation through the hydroxylation of its side chain by CYP11A1 have recently been established (18-29). The products of these pathways, including 20(OH)D₃, 20,23(OH)₂D₃, 1,20(OH)₂D₃, and 1,20,23(OH)₃D₃, exhibit potent biological activity in different types of cells, including skin-derived cells (18-30). In addition, 20(OH)D₃ and 20,23(OH)₂D₃ are noncalcemic, whereas 1,20(OH)₂D₃ shows low calcemic activity, in contrast to calcitriol (1,25(OH)₂D₃) (24-26, 31). As with 1,25(OH)₂D₃, which is the major hormonally active form of vitamin D₃, the CYP11A1-derived hydroxymetabolites exhibit anticancer activities, antiproliferative, and pro-differentiation effects, and have immunomodulatory functions (32-34).

Vitamin D derivatives act mainly through the vitamin D receptor (VDR) (35). Our previous studies pertaining to CYP11A1-derived vitamin D₃ hydroxy-analogs show that in addition to acting on the VDR (17, 25, 33, 36-38), they also act as inverse agonists of RORα and RORγ receptors (16, 17), and as agonists on the aryl hydrocarbon receptor (AhR) (39). Both RORα and RORγ receptors are expressed in murine and human skin (16, 40-45).

As outlined previously, vitamin D and its derivatives exhibit antifibrotic properties (46, 47) including antifibrogenic activities by 20(OH)D₃ and 20,23(OH)₂D₃ in human dermal fibroblasts, and in mice when delivered IP (25). In this study, we explored the role of RORγ in response to CYP11A1-derived vitamin D derivatives in dermal fibroblasts isolated from RORγ ^-/- (knockout [KO]), RORγ ^-/+ (heterozygotes), and RORγ ^+/+ (wild-type) mice. Our aim was to identify a link between fibrosis, CYP11A1-derived vitamin D3 analogs and the RORγ receptor in relation to their therapeutic potential, and to expand our understanding of the actions of novel vitamin D3 derivatives. We also tested whether orally delivered 20(OH)D₃ exhibits antifibrogenic effects in mice.

Material and Methods

Antifibrogenic effects of 20(OH)D₃ in a mouse model

Six-week-old female C57/BL6 mice, purchased from Jackson Laboratory, were treated as previously described (24) with minor differences as detailed in the legend to Table 1. Briefly, mice were injected subcutaneously, 6 days per week, with 50 µL sterile saline or an equal amount of saline containing 50 µg bleomycin (Fresenius Kabi USA, Lake Zurich, IL). Mice were further gavaged with 100 µL propylene glycol (1:5 dilution in saline) or with propylene glycol containing 20S-hydroxyvitamin D₃ (20(OH)D₃) at 3 different doses: 5, 15, or 30 µg/kg. Mice were sacrificed after 22 days. Skin was used for histological analyses. Briefly, skin samples were formalin-fixed paraffin embedded, sectioned, and stained with hematoxylin and eosin (24, 25). The Institutional Animal Care and Use Committee at the University of Tennessee Health Science Center approved experiments in mice under protocol number 15-111 (Dr. A. Postlethwaite, principal investigator).

Table 1.

20S(OH)D₃ Given via Oral Gavage Suppresses Bleomycin-induced Dermal Fibrosis

Condition		Dermal Thickness	P Value
Subcutaneous	Gavage	µM mean ± SEM
Saline	Vehicle	238 ± 9
Bleomycin	Vehicle	476 ± 33	<0.001
Bleomycin	20S(OH)D3   (5 µg/kg)	278 ± 15	0.004
Bleomycin	20S(OH)D3   (15 µg/kg)	314 ± 18	0.005
Bleomycin	20S(OH)D3   (30 µg/kg)	234 ± 31	0.004

Six-week-old C57BL/6 female mice were injected SC in a 1.5 cm × 1.5 cm area in the back in the lower thoracic region 6 days/wk with 50 µL sterile intravenous grade normal saline (“saline”) or 50 µL saline containing 50 µg bleomycin (Fresenius Kabi USA). Mice were gavaged 6 days/wk with either 100 µL propylene glycol diluted 1:5 with saline or propylene glycol diluted 1:5 with saline containing 20(OH)D₃ at a dose of 5 µg/kg, 15 µg/kg, or 30 µg/kg. On day 22, mice were euthanized and skin from the injection site was processed for histology and stained with Masson Trichrome. The degree of dermal thickening was measured by light microscopy by 2 examiners.

Experiments with fibroblasts from newborn mice

Skin tissue from newborn mice, RORγ wild-type (RORγ ^+/+), heterozygotes (RORγ ^-/+), or knock out (RORγ ^-/-) was collected was collected at the National Institute of Environmental Health Sciences. The protocols for the breeding of mice, genotyping, and organ collections are described elsewhere (48, 49). All animal studies followed guidelines outlined by the National Institutes of Health Guide for the Care and Use of Laboratory Animals and protocols were approved by the Institutional Animal Care and Use Committee at the National Institute of Environmental Health Sciences (approval date: April 26, 2018; animal welfare assurance number is A4149-1, Dr. A. Jetten, principal investigator). Fibroblasts were isolated from the skin of 1- to 2-day-old mice. Briefly, skin was rinsed in PBS (Gibco, Life Technologies, Carlsbad, CA), cut into smaller pieces and incubated in 2.4 U/mL of dispase II (Roche, Indianapolis, IN) overnight at 4°C. The dermis was first separated from the epidermis using a razor blade and further incubated in 0.1% collagenase (Roche) overnight at 4°C. The following day, the dermis was cut into smaller pieces and incubated in 0.25% Trypsin (CellGrow, Corning, NY) for 1 hour at 37°C. A suspension of single cells was prepared by pipetting and draining through a cell strainer. After centrifuging to remove trypsin, the cell pellet was resuspended in complete DMEM media (CellGrow) containing 10% serum (Sigma-Aldrich, St. Louis, MO), mouse fibroblast growth factor (Sigma-Aldrich), and 1% antibiotic (Penicillin-Streptomycin, CellGrow), and plated onto tissue culture-treated dishes (TPP, Midwest Scientific, Valley Park, MO) for expansion. Cells in their early passage were used for experiments and were treated with compounds diluted in DMEM and 10% charcoal-stripped fetal bovine serum (FBS; Sigma-Aldrich). In addition, cells isolated from different mice were pooled to provide more representative samples.

Cell proliferation assay (MTS assay)

Murine fibroblasts were plated onto 96-well plates at a confluence of about 70%. Cells were treated with the following vitamin D3 derivatives: 1,25(OH)₂D3 (from Sigma) or 20(OH)D3 synthesized chemically (50), or 20,23(OH)₂D3, 1,20(OH)₂D3, or 1,20,23(OH)₃D3 synthesized enzymatically (22, 51, 52), diluted in ethanol in DMEM containing charcoal-treated serum. Final concentrations of the vitamin D hydroxyderivatives were 0.1, 10, or 100 nM, with the final ethanol concentration being 0.1%, which was also used as a vehicle control. After 44 hours of incubation, MTS (Promega, Madison, WI) solution (20 μL) was added to the media, which was further incubated for 3 hours. The absorbance (490 nm) was recorded using an ELISA plate reader. The number of viable cells was measured in 6 replicates.

Collagen assay

Murine fibroblasts were grown to 90% confluence. Cells were treated with 1,25(OH)₂D3, 20(OH)D3, 20,23(OH)₂D3, 1,20(OH)₂D3, or 1,20,23(OH)₃D3 in 10% charcoal-treated FBS plus supplements, for 6 or 24 hours, in 60-mm Petri dishes in a total volume of 10 mL. The hydroxyvitamin D3 derivatives were used at a concentration of 100 nM. Supernatant (1 mL) was collected and used for collagen quantification by the Sircol collagen assay protocol (Biocolor, S1000). Total collagen concentration was calculated based on the absorbance of the samples in comparison to the standard curve absorbance.

RNA sequencing preparation and analyses

Murine fibroblasts were grown as described previously. After supernatant was collected for the assay of collagen, cells were harvested for RNA isolation (RNAeasy Micro kit, Qiagen) and cDNA synthesis (High Capacity cDNA Reverse Transcription Kit with RNase Inhibitor, Applied Biosystems) following the manufacturers’ protocols. At least 200 ng of RNA for each sample (24-hour treatment) was used to create libraries for RNA sequencing by BGI Americas Corporation Service (Cambridge, MA). Sample quality and concentration were determined using Agilent 2100.

The raw sequence FASTQ files were aligned to Gencode’s GRCm38 p4 Release M11 genome using STAR, version 2.5.4b (parameters used --outReadsUnmapped Fastx; --outSAMtype BAM SortedByCoordinate; --outSAMattributes All; --outFilterIntronMotifs RemoveNoncanonicalUnannotated). Transcript abundances were then calculated from the alignments using Cufflinks, version 2.2.1 (parameters used -G; -L; --library-type fr-firststrand). Samples were then merged using default parameters with Cuffmerge. Finally, differential expression was calculated using Cuffdiff’s default parameters. The raw data are deposited at the NCBI GEO (GSE145818).

Pathway analysis using Ingenuity

Data were analyzed with Ingenuity Pathway Analysis (Ingenuity Systems, www.ingenuity.com). For generating networks, a data set containing gene identifiers and corresponding expression values was uploaded into the application. Each identifier was mapped to its corresponding object in Ingenuity’s Knowledge Base. A fold change cutoff of ±2 was set to identify molecules whose expression was significantly differentially regulated. These molecules, called Network Eligible molecules, were overlaid onto a global molecular network developed from information contained in Ingenuity’s Knowledge Base. Networks of Network Eligible Molecules were then algorithmically generated based on their connectivity. The Functional Analysis identified the biological functions and/or diseases that were most significant to the entire data set. Molecules from the dataset that met the fold change cutoff of ±2 and were associated with biological functions and/or diseases in Ingenuity’s Knowledge Base were considered for the analysis. Right-tailed Fisher’s exact test was used to calculate the P value determining the probability that each biological function and/or disease assigned to that data set is due to chance alone. The RNA sequencing (RNA-seq) raw data are deposited at the NCBI GEO (GSE145818).

Gene set enrichment analysis (GSEA (53)) was performed as follows. The fragments per kilobase of exon model per million reads mapped values were log2 transformed and the differentially expressing genes with an additional criterion of fold-change ±1.5 were selected for the enrichment analysis. GSEA was performed through METASCAPE (http://metascape.org) (54).

Quantitative RT-PCR

RT-PCR was carried out using Sybr green Master Mix (Thermo Scientific, Waltham, MA), in triplicates; as previously described (55). Approximately 750 ng of cDNA was used for PCR reactions. GAPDH and β-actin were used as internal controls. Primer sequences are listed in Table 2. The source of primers is as follows: Bmp7 (56), Mmp1b, Tgfb1, Col1A2, Col3A1 (57), Hgf (58), Cyp1b1 (59).

Table 2.

Primers and Sequences Used for RT-PCR

Gene name	Sequences
Bmp7	GATT- TCAGCCTGGACAACGAG GGGCAAC- CCTAAGATGGACAG
Mmp1b	AAGGCGATATTGTGCTCTCC CCTCATTGTTGTCGGTCCAC
Tgfb1	GCGGACTACTATGCTAAAGAGG TCAAAAGACAGCCACTCAGG
Col1A2	CCGTTGGCAAAGATGGTAG GTCAGCCCTGTAGAAGTATCC
Col3A1	AAAAACCCTGCTCGGAACT CAGCTGCACATCAACGACA
Hgf	CATTGGTAAAGGAGGCAGCTATAAA    GGATTTCGACAGTAGTTTTCCTGTAGG
Cyp1b1	AATGAGGAGTTCGGGCGCACA GGCGTGTGGAATGGTGACAGG
Gapdh	CCATTTTGTCTACGGGACGA AGTATGACTCCACTCACGGCAAAT
B-actin	CTAAGGCCAACCGTGAAAAG ACCAGAGGCATACAGGGACA

Data were analyzed using GraphPad Prism 8 (San Diego, CA) statistical software 1-way ANOVA or t-test, where appropriate; *P < 0.05; **P < 0.01; ***P < 0.001.

Results with Discussion

We previously reported that 20(OH)D₃ suppressed fibrogenesis in mice after IP treatment and that 20(OH)D3, 20,23(OH)₂D₃, and 17,20,23(OH)₃D₃ inhibited TGFβ-1-induced fibrinogenic effects in cultured dermal fibroblasts (24). In the current study, we analyzed the effect of 20(OH)D₃ on bleomycin-induced dermal fibrosis in mice after oral delivery by gavage to examine its usefulness for (pre)clinical studies on scleroderma. We focused on 20(OH)D₃, the first and major product of CYP11A1-mediated vitamin D₃ metabolism (20-22). Oral administration of 20(OH)D₃ resulted in significantly less bleomycin-induced dermal fibrosis and its thickening than in control mice, as shown by the representative sections in Fig. 1 and the quantification in Table 1. The effect was seen at doses of 5 µg/kg, 15 µg/kg, and 30 µg/kg with the maximum effect seen at 30 µg/kg (Table 1). These data demonstrated that oral treatment with 20(OH)D₃ attenuated the bleomycin-induced dermal fibrosis in mice. Hence, this route of administration can be considered for future clinical trials with noncalcemic vitamin D3 derivatives for the treatment of scleroderma, morphea, and other diseases involving fibrosis.

Figure 1.

H&E stained sections of murine skin. Bleomycin (BM) was administered, 50 µg/mouse SC, for 22 days. 20(OH)D₃ or its solvent, propylene glycol (PG) was given as an oral gavage as described in the legend to Table 1. (A) Vehicle control (saline SC and propylene glycol gavage); (B) bleomycin SC and propylene glycol gavage; (C) bleomycin SC and 5 µg/kg 20(OH)D₃ gavage; (D) bleomycin SC and 15 µg/kg 20(OH)D₃ gavage; and (E) bleomycin SC and 30 µg/kg 20(OH)D₃ gavage.

Previous studies have shown that 1,25(OH)₂D₃ inhibits fibroblast proliferation and decreases the synthesis of collagen (60); effects also seen with CYP11A1-derived vitamin D3 hydroxyderivatives (24, 61, 62). To ascertain the role of RORγ in these responses, we examined the effect of these derivatives on cells isolated from the skin of wild type (RORγ ^+/+), RORγ ^-/-, and RORγ ^-/+ mice. The proliferation of fibroblasts treated for 48 hours with 1,25(OH)₂D₃ or with the novel vitamin D3 derivatives 20(OH)D3, 20,23(OH)₂D₃, 1,20(OH)₂D₃, and 1,20,23(OH)₃D₃, was significantly reduced in wild-type fibroblasts (P < 0.01 or P < 0.001) compared with ethanol-treated control cells. This inhibition was observed with both 0.1 and 100 nM secosteroid, concentrations selected based on a previous study (33). In contrast, this inhibitory effect on proliferation was reversed in both homozygous KO (RORγ ^-/-) and heterozygous (RORγ ^-/+) fibroblasts (Fig. 2) and instead a significant increase in proliferation was observed in response to every secosteroid tested. The effects were generally more pronounced in homozygous fibroblasts than heterozygous fibroblasts indicating a dosing effect (Fig. 2) and were observed at 24 and 72 hours posttreatment (data not shown).

Figure 2.

The effect of secosteroids on the proliferation of fibroblast cells from RORγ ^+/+, RORγ ^-/+, and RORγ ^-/- mice. Fibroblasts were isolated from murine skin and grown until the third passage. After they reached 70% confluence, cells were treated with secosteroids at the indicated concentrations for 48 hours. MTS was added and absorbance read after 3 h of incubation. Data were analyzed using the t test and represent means ± SD (n ≥ 6), **P < 0.01, ***P < 0.001, ****P < 0.0001 by t test. ROR, retinoic acid-related orphan receptor.

It is known that VDR is required for proliferation and differentiation of epidermal cells and the action of 1,25(OH)₂D₃ (63), as well as epidermal stem cells (64) and plays a role in inflammation and injury. Dermal fibroblast function was altered in VDR KO mice (65). The effect of 1,25(OH)₂D₃ has been tested in mouse models of tissue fibrosis, showing that this vitamin D3 derivative can prevent fibrosis independently of VDR-mediated transcriptional activity (66). Because vitamin D₃ derivatives can act on fibrosis independently of VDR, we proposed that that action might be RORγ receptor dependent. There is a shortage of information in the literature to support the role of RORγ receptor in the action of vitamin D in murine skin. The presented results support an important role of RORγ in the inhibition of dermal fibroblast proliferation by secosteroids.

Next, we examined the role of RORγ in the inhibition of collagen synthesis by vitamin D3 derivatives in dermal fibroblasts. Fibroblasts, isolated from the skin of RORγ wild type or KO mice, were treated with 100 nM 1,25(OH)₂D₃, 20(OH)D₃, 20,23(OH)₂D₃, 1,20(OH)₂D₃, or 1,20,23(OH)₃D₃ and the amount of collagen in the supernatant measured after 6 hours (data not shown) or 24 hours of treatment (Fig. 3). There was a significant decrease in the collagen concentration in the supernatants from wild-type fibroblasts after treatment with all of the secosteroids compared with the ethanol-treated control cells. As observed for proliferation, the effect was reversed for the RORγ ^-/- fibroblasts, which showed a significant increase in the collagen concentration in response to all the secosteroids tested (Fig. 3).

Figure 3.

The effect of secosteroids on collagen production by fibroblast cells from RORγ ^+/+ and RORγ ^-/- mice. Fibroblasts were isolated from RORγ ^+/+ (wild-type [WT]) and RORγ ^-/- (knockout [KO]) mouse skin and were grown in DMEM supplemented with 10% FBS, testosterone, and mouse fibroblast growth factor to 90% confluence. Cells were then treated with secosteroids (100 nM) in 10% charcoal-treated FBS plus supplements for 24 hours in 60-mm Petri dishes in total volume of 10 mL. The supernatant was collected and used for collagen quantification using the Sircol collagen assay Biocolor, S1000. Data were analyzed using the t test and represent means ± SD, *P < 0.05, **P < 0.01. FBS, fetal bovine serum; ROR, retinoic acid-related orphan receptor.

To relate these effects to global transcriptomic activity, we compared the gene expression profiles between RORγ ^+/+ and RORγ ^-/- fibroblasts by RNA-seq analysis. Pathway analysis showed that several proliferation-related functions, DNA damage responses, and hepatic fibrosis were among the top control pathways (Table 3, Supplemental Table 1, located in a digital research materials repository (67)). In addition, the main regulators for which expression was inhibited in RORγ ^-/- fibroblasts compared with the RORγ ^+/+ included Tgfb1, Vegf, TNnfa, dextran sulfate, and Erbb2 (Table 3, Supplemental Table 1) (67). A dramatic change in the expression of upstream genes, with predominant inhibition in RORγ ^-/- fibroblasts compared to RORγ ^+/+ fibroblasts, was observed. This included genes for many cytokines, chemokines, growth factors, enzymes, and transcriptional regulators with only a few being upregulated (Table 3, Supplemental Table 2, located in a digital research materials repository (67)).

Table 3.

Gene Set Enrichment Analysis for Signaling, Proliferation, Differentiation, or Healing Processes Based on RNA-Seq Data Obtained After 24 Hours of in vitro Incubation of Murine Fibroblasts, RORγ^+/+ (WT), or RORγ^-/- (KO) with Ethanol (vehicle), 1,20(OH)2D₃, or 20(OH)D₃

Pathway Name	Sample Sets
	KO_EtOH vs WT-EtOH		KO_1,20(OH)2D3 vs KO_EtOH		KO_20(OH)D3 vs KO_EtOH		WT_1,202D3 vs WT_EtOH		WT_20(OH)D3 vs WT_EtOH
	logp	Expression
Positive regulation of collagen biosynthetic process	-3.4	UP
Vascular endothelial growth factor receptor signaling pathway	-1.8	UP
Wound healing	-1.8	UP
Epithelial cell development	-1.7	UP
Regulation of epidermal cell differentiation	-1.6	UP
Cellular response to vascular endothelial growth factor stimulus	-1.5	UP
Regulation of wound healing	-1.3	UP
Epidermal cell differentiation	-1.3	UP							-1.5	DOWN
Cell cycle	-31	DOWN
Cell division	-30	DOWN					-1.4	UP
DNA replication	-27	DOWN
Cell cycle, mitotic	-26	DOWN
Regulation of cell-cycle process	-22	DOWN
Regulation of cell-cycle phase transition	-16	DOWN
DNA repair	-15	DOWN
Cell cycle	-15	DOWN
Epithelial cell proliferation	-14	DOWN					-2	UP
Inflammatory response	-12	DOWN							-2.1	UP
Positive regulation of cell cycle	-10	DOWN
DNA replication	-9.5	DOWN
Regulation of reactive oxygen species metabolic process	-6.7	DOWN
Regulation of vascular endothelial growth factor production	-6.3	DOWN
IL-17 signaling pathway	-5.9	DOWN			-2	DOWN
Response to growth factor	-5.8	DOWN
Nitric oxide metabolic process	-5.2	DOWN
Regulation of DNA repair	-4.6	DOWN
TP53 regulates transcription of genes involved in G1 cell-cycle arrest	-4.6	DOWN
TNF signaling pathway	-4.4	DOWN
Regulation of TNF superfamily cytokine production	-4.4	DOWN
Response to oxidative stress	-4.2	DOWN
Endothelial cell proliferation	-4.2	DOWN
Cellular response to reactive oxygen species	-4.2	DOWN	-1.6	UP
Response to UV	-3.6	DOWN
Positive regulation of inflammatory response	-3.6	DOWN
Response to ionizing radiation	-3.6	DOWN
Keratinocyte proliferation	-3.2	DOWN
Epithelial cell differentiation	-3	DOWN					-1.5	DOWN	-1.5	DOWN
Cellular response to ionizing radiation	-2.9	DOWN
NF-kappa B signaling pathway	-2.9	DOWN
Collagen degradation	-2.7	DOWN
Positive regulation of fibroblast proliferation	-2.6	DOWN
Wound healing	-2.6	DOWN
Stem cell differentiation	-2.6	DOWN					-1.5	DOWN	-2	DOWN
Response to X-ray	-2.5	DOWN
Cellular response to hydrogen peroxide	-2.4	DOWN
Positive regulation of nitric oxide metabolic process	-2.4	DOWN
Collagen formation	-2.3	DOWN
Regulation of insulin-like growth factor (IGF) transport and uptake by IGF binding proteins	-2.3	DOWN
TGF-beta signaling pathway	-2	DOWN
Regulation of response to wounding	-2	DOWN
Response to epidermal growth factor	-2	DOWN
BMP signaling pathway	-1.9	DOWN
Nucleotide excision repair	-1.9	DOWN
Base excision repair	-1.8	DOWN							-2.5	UP
Response to gamma radiation	-1.8	DOWN
Fibroblast proliferation	-1.8	DOWN					-2	UP
Negative regulation of NF-kappaB transcription factor activity	-1.8	DOWN
Positive regulation of BMP signaling pathway	-1.6	DOWN
Vascular wound healing	-1.6	DOWN
Response to oxygen radical	-1.6	DOWN

Abbreviations: BMP, bone morphogenetic protein; KO, knockout; RNA-seq, RNA sequencing; ROR, ROR, retinoic acid-related orphan receptor; WT, wild-type.

Among ingenuity’s predicted associations with biological functions and diseases, cancer, gastrointestinal diseases, organismal injury and abnormalities, reproductive system diseases, and hematological diseases were significantly affected in RORγ ^-/- fibroblasts compared with the RORγ ^+/+ (Supplemental Table 1 (67)). Molecular and cellular functions, such as cell death and survival, cellular movement, cell cycle, cellular growth and proliferation, cellular development, as well as organismal survival, development and tissue morphology also appear to be affected in RORγ ^-/- fibroblasts compared to the RORγ ^+/+ (Supplemental Tables 1, 2, 3) (67). The main cellular functions that were affected are pertinent to liver, kidney, and heart and included hepatotoxicity, renal toxicity, and cardiotoxicity (Supplemental Tables 1, 2, 3, 4, located in a digital research materials repository (67)). The analysis of the canonical (Supplemental Table 3) (67) and toxicity pathway (Supplemental Table 4, located in a digital research materials repository (67)) also reveals dramatic changes in a number of genes and signaling pathways with predominant inhibition in RORγ ^-/- fibroblasts as compared with the RORγ ^+/+. Among the pathways/genes affected are nuclear receptor signaling including VDR, retinoid X receptor (RXR), AhR, peroxisome proliferator-activated receptor, inflammatory signaling including nuclear factor-ƙB and interleukins, growth factors including vascular endothelial growth factor (VEGF), TGF-B1, platelet-derived growth factor, and others such as NRF2 and bone morphogenetic protein (BMP). Thus, deletion of RORγ leads to dramatic changes not only in the gene expression pattern, but also in signaling pathways and their regulators, which would affect the fibroblast phenotype and cell responsiveness to secosteroids (see the following section).

Data obtained from RNA-seq Ingenuity pathway analysis are confirmed with GSEA. Knocking out RORγ receptor affected differentiation and proliferation of a variety of cells, including lymphocytes, T and B cells, mononuclear cells, erythrocytes, stem cells, osteoblasts, mesenchymal, chondrocytes, cardiomyocytes, hepatocytes, muscle, fat cells, epidermal and endodermal cells, etc. (data not shown). Proliferation of fibroblasts, as well as endothelial cell development is impaired in cells lacking RORγ receptor (Table 3). Collagen synthesis and degradation and wound healing processes are downregulated by the lack of RORγ in murine fibroblasts (Table 3). In addition, processes that are known to be affected by vitamin D3 derivatives are altered by the lack of RORγ in murine fibroblasts, including cell cycle and cell division, DNA replication and repair, nucleotide excision repair, reaction to radical oxygen species, response to radiation (X-ray, UV, ionizing radiation, and gamma radiation) (Table 3). Conversely, inflammatory responses, including interleukin-17 pathway, p53, TNFα, and nuclear factor-ƙB, along with the BMPB and VEGF signaling are downregulated in cells lacking RORγ receptor in comparison to wt (Table 3).

In RORγ ^+/+ fibroblasts, expression of many genes was altered by 20(OH)D₃ compared with ethanol-treated control cells, of which 337 were up-regulated and 238 down-regulated (Fig. 4). Biologically relevant target genes were determined based on a fold change (FC) threshold of 1.5. Ingenuity pathway analyses shows that 20(OH)D₃ affects many immune, bone/cartilage, atherosclerosis, antioxidative and growth factor related, and other signaling pathways with significant (P < 0.05) VDR/RXR activation (Supplemental Tables 5, 6, located in a digital research materials repository (67)). This is consistent with previously published data in human keratinocytes (28, 30, 34, 39, 68).

Figure 4.

Venn diagrams. Genes with a fold change of +1.5 or -1.5 were compared between fibroblasts treated with 1,20(OH)₂D₃ and 20(OH)D₃ to identify the genes that were common/unique.

In wild-type fibroblasts treated with 1,20(OH)₂D₃ many target genes were affected by the secosteroid compared with ethanol-treated control cells, with the expression of 376 genes up-regulated (FC ≥1.5) including Fbxw7, Nup210, Trim5, Cr2, Rpl, Rps, Tdrd, and Stub, and 215 and genes down-regulated (FC ≤-1.5) including Pde4dip, Tnrc6b, Ptpn6, Rpl21, Chn2, Ehf, and Erbb4 (Fig. 4). Identified pathways regulated by 1,20(OH)₂D3 were similar to those regulated by 20(OH)D3, and included interleukins, VDR/RXR, nuclear factor kB, VEGF, and NRF2 (Supplemental Tables 7, 8, located in a digital research materials repository (67)). Identification of these pathways is consistent with our functional data indicating their immunomodulatory functions, involvement of VDR/RXR, and NRF2 signaling (23, 30, 34, 36, 37, 39, 68).

Further analysis revealed that the expression of 215 genes was downregulated by only 1,20(OH)₂D₃ and 238 genes were downregulated only by 20(OH)D3, with a further 235 genes downregulated by both vitamin D derivatives for the RORγ ^+/+ fibroblasts, as presented in the Venn diagram (Fig. 4A). The major functions and the relevance of these genes are described in Table 4. In addition, the expression of 376 genes was up-regulated by only 1,20(OH)₂D₃ and 337 genes by only 20(OH)D3, with 147 genes up regulated by both vitamin D derivatives (Fig. 4B). For the RORγ ^-/- fibroblasts, similar pathways were affected by treatment with 20(OH)D3 compared with ethanol treated RORγ ^-/-cells (Supplemental Tables 9, 10, located in a digital research materials repository (67)) or with 1,20(OH)₂D₃ (Supplemental Tables 11, 12, located in a digital research materials repository (67)) with some differences. Importantly, VDR/RXR signaling was not detected among pathways activated by 1,20(OH)₂D3 in RORγ ^-/- fibroblasts. In conclusion, knocking out the RORγ significantly affected the responsiveness of cells to 20(OH)D₃ and 1,20(OH)₂D₃.

Table 4.

Gene Expression in Murine Fibroblasts: Comparisons Between RORγ^-/- and Wild-Type

Expression (Fold Change)	Gene Name
-1.5	MMP1b	Interstitial collagenase b
-1.23782	MMP2	Matrix metallopeptidase 2 or type IV collagenase
-3	MMP8	Matrix metallopeptidase 8 or neutrophil collagenase
-2.86067	MMP13	Matrix metallopeptidase 13 or collagenase 3
-2.89474	BMP6	Bone morphogenic protein
1.738077	TGFb1	Transforming growth factor beta
-6.14074	TGFbr3	Transforming growth factor beta receptor
-1.43419	Col1A1	Collagen type I alpha 1 chain or collagen of skin, tendon, and bone, Alpha-1 chain
-6.07225	Col1A2	Collagen type I alpha 2 chain or collagen of skin, tendon, and bone, alpha-2 chain
-4.09706	Col3A1	Collagen type III alpha 1 chain or Ehlers-Danlos syndrome type IV
24.464	Col7A1	Collagen type VII alpha 1 chain
-61.0003	Col28A1	Collagen type XXVIII alpha 1 chain
-21.1603	HGF	Hepatocyte growth factor (HGF) or scatter factor (SF)
-9.12445	Cyp1B1	Cytochrome P450 family 1 subfamily B member 1

RNA-sequencing analysis was performed using RNA from murine fibroblasts. Gene expression is shown as fold change.

Comparing the effects of 20(OH)D₃ and 1,20(OH)₂D₃ on RORγ ^-/- fibroblasts, analysis revealed that the expression of 262 genes was downregulated by only 1,20(OH)₂D₃ and 379 genes by only 20(OH)D3, with the expression of 273 genes being down regulated by both secosteroids (Fig. 4C). There were 498 genes whose expression was up-regulated by only 1,20(OH)₂D₃ and 412 genes by only 20(OH)D₃, with the expression of 158 genes being up-regulated by both secosteroids (Fig. 4D). The genes whose expression was regulated by both 1,20(OH)₂D₃ and 20(OH)D₃ are presented in Table 3. These genes are primarily involved in regulation of cell proliferation, differentiation, and tumorigenesis, cell signaling, and metal transport. Thus, treatment with 20(OH)D₃ and 1,20(OH)₂D₃ activate distinct and overlapping patterns of gene expression in RORγ ^-/- fibroblasts.

We performed further RNA-seq analyses to determine whether RORγ is important for transcriptional regulation of key genes involved in fibrosis. We selectively checked the expression of different collagenases and genes from the family of TGF-β-related growth factors. Collagenases/matrix metallopeptidases (MMPs) play roles in degradation of extracellular matrix proteins, as well as differentiation, apoptosis, angiogenesis, and host defense (69). MMPs were expressed at only low levels in cells lacking RORγ (Table 4). MMP8 is involved in skin tumors, arthritis, and lung fibrosis (69), whereas MMP13 (collagenase 3), is the main collagenase in mouse tissue involved in regulation of interstitial collagen and growth plate cartilage defects (70). MMP8 and MMP13 expression changed, respectively 3- and 2.86-fold, in RORγ ^-/- fibroblasts compared with wild-type, whereas MMP2 and MMP1b expression was not significantly different (-1.5 and -1.2, respectively) (Table 4).

Collagen is the most abundant protein in connective tissue. In skin, the collagen is produced by fibroblasts, which play an important role in wound healing. Collagen type I alpha 1 chain (71) and collagen type I alpha 2 chain (Col1A2) have a ubiquitous distribution, including bones, dermis of the skin, ligaments, and tendons (71). Mutations in these genes are associated with various diseases in humans, including osteogenesis imperfecta, Ehlers-Danlos syndrome, and osteoporosis (71). Collagen type 3 alpha 1 chain (Col3A1) is found in skin, blood vessels, and intestine, and mutations in this type of collagen are involved in aneurysms in arteries, as well as in Ehlers-Danlos syndrome (71). Collagen type XXVIII alpha 1 chain (Col28A1) is found in skin dermis and the sciatic nerve (71). Collagen type VII alpha 1 chain (Col7A1) is commonly found in the dermis of the skin and the bladder, and is associated with epidermolysis bullosa acquisita (71). The expression of the collagen genes, Col1A2 and Col3A1, in RORγ ^-/- fibroblasts was reduced, respectively, 6- and 4-fold compared with wild-type fibroblasts, whereas Col1A1 showed only a -1.4-fold change (Table 4). Interestingly, Col28A1 expression was decreased 61-fold in RORγ ^-/- fibroblasts compared with wild-type cells. Of all the collagen genes analyzed, only Col7A1 showed an increase in expression (24-fold) in RORγ ^-/- fibroblasts compared with wild-type.

BMP6 belongs to the family of TGF-β-related growth factors, plays an important role in the proliferation and development of the epidermis, and mutations have been implicated in the development of psoriasis in humans (72). BMP6 is strongly expressed in mouse skin under pathological conditions, such as inflammation (72). BMP6 mRNA was expressed at a 2.9-fold lower level in RORγ ^-/- fibroblasts compared with wild-type cells. TGF-β receptor 3 mRNA expression decreased 6.1-fold, whereas TGFβ1 expression was not significantly changed (1.7-fold) (Table 4).

Hepatocyte growth factor (HGF) or scatter factor plays an anti-inflammatory and anti-apoptotic role (73). Transgenic HGF/scatter factor mice are vulnerable to UV light and serve as a model for pigmentation and skin pathology (73). HGF was expressed at 21-fold lower levels in RORγ ^-/- cells than wild-type cells (Table 4). Cyp1b1 expression was also down-regulated in RORγ ^-/- fibroblasts (9-fold) when compared with wild-type cells (Table 4). The expression of Cyb1b1 was stimulated by secosteroids tested in wild-type fibroblasts, but not in cells lacking RORγ (Fig. 5). Cytochrome P450 CYP1b1 (CYP1b1) belongs to the cytochrome P450 family of heme-containing monooxygenases and plays a key role the metabolic activation of polycyclic aromatic hydrocarbons (74). Cyp1b1 is found in skin keratinocytes, and it can produce sterols that affect tumorigenesis (74).

Figure 5.

mRNA levels of genes involved in fibrosis after treatment of fibroblasts from RORγ ^+/+ and RORγ ^-/- mice. Fibroblasts from RORγ ^+/+ (WT) and RORγ ^-/- (KO) mice were treated with 100 nM of the secosteroids indicated, or ethanol vehicle (control). Data are presented as mean ± SD, n = 3. Analysis was done using the t-test: *P < 0.05, **P < 0.01, ***P < 0.001 or ****P < 0.0001 versus control (ethanol). KO, knockout; ROR, retinoic acid-related orphan receptor; WT, wild-type.

In summary, the RNA-seq analysis shows dramatic differences in the gene transcriptome between wild-type and RORγ ^-/- fibroblasts in response to several secosteroids. The differences at least in part correlate with some of the phenotypic differences described in this paper. However, it is possible that this effect is not solely related to the RORγ receptor because different factors may also contribute to the observed differences, which represents a challenge for future investigation.

Quantitative RT-PCR was used to determine whether the presence of RORγ in murine fibroblasts is important for the expression of genes involved in fibrotic processes upon treatment with the CYP11A1-derived vitamin D3 metabolites. The expression of genes encoding collagen (Col1A2 and Col3A1) was decreased in wild-type fibroblasts treated with 1,25(OH)₂D₃ or 20(OH)D₃ (and also by 1,20(OH)₂D₃ in the case of Col3A1) (Fig. 5). This effect was reversed in RORγ ^-/- cells, in which an increase in the expression of the collagen genes was seen after treatment with all the secosteroids tested, 11- to 13-fold with 20(OH)D₃ and 20,23(OH)₂D₃) (Fig. 5).

The expression of TGF-β and its receptor (Tgfbr) is upregulated in tumors of the skin (72). TGFβ1 plays an important role in skin cell proliferation, inflammation, and apoptosis, specifically in the epidermis (72) and Tgfβ11 KO mice exhibit delayed wound healing (75). The expression of theTgfb1 gene was not significantly altered by the treatment of wild-type fibroblasts with any of the secosteroids, whereas all secosteroids stimulated its expression in the RORγ ^-/- fibroblasts, 11- and 20-fold, respectively, for 20(OH)D₃ and 20,23(OH)₂D₃ (Fig. 5). Our data suggest that besides the VDR (76), RORγ may also be involved in the action of vitamin D₃ derivatives in the wound healing process.

MMP1b or fibroblast collagenase is an interstitial collagenase that is involved in tissue repair, remodeling of skin and wound healing, tumor invasion, and atherosclerosis (70). The expression of Mmp1b, as well as Hgf and Bmp7, were significantly higher in wild-type fibroblasts after treatment with secosteroids with only 1 exception (1,20,23(OH)₃D₃ on Hgf) (Fig. 5). In contrast, Hgf 1 expression was not significantly altered in RORγ ^-/- fibroblasts treated with the secosteroids, whereas the expression of Mmp1b and Bmp7 was significantly reduced. These results suggest a possible role for RORγ in the anti-inflammatory action of vitamin D₃ derivatives.

Concluding Discussion

This study shows that oral delivery of 20(OH)D₃, a product of CYP11A1 action on vitamin D₃, can attenuate bleomycin-induced dermal fibrosis in mice, similar to what we reported before for IP administration (24). CYP11A-derived secosteroids, like 1,25(OH)₂D₃, inhibited the proliferation of wild-type dermal fibroblasts, but increased the proliferation of RORγ ^-/- and RORγ ^+/- fibroblasts. Similarly, collagen production was decreased by the secosteroids in the wild-type fibroblasts, but increased in RORγ ^-/- cells. Thus, the inhibition of fibroblast proliferation and collagen synthesis by secosteroids is dependent on an adequate level of RORγ gene expression as one copy of RORγ (as in heterozygotes) is not sufficient to generate the same effects as in wild-type fibroblasts.

Comparison of gene expression profiles by RNA-seq identified numerous genes that were differentially expressed in RORγ ^-/- fibroblasts compared with wild-type cells. Treatment of wild-type fibroblasts with secosteroids revealed large differences in gene expression between 20(OH)D3- and 1,20(OH)₂D₃-treated cells, with a subset of 382 genes regulated by both secosteroids. This was also seen with RORγ ^-/- fibroblasts, with both 20(OH)D₃ and 1,20(OH)₂D₃ regulating the expression of 431 genes. These observations suggested different mechanisms of action and may be related to differences in the ability of these 2 secosteroids to activate receptors and in the case of wild-type, fibroblasts likely include RORγ, VDR, and AhR as reported for keratinocytes (68). The involvement of some of these receptors in RORγ ^-/- fibroblasts requires further study.

We also identified a correlation between fibroblasts lacking RORγ expression and the expression of genes related to fibrosis. We further demonstrated that the effects of CYP11A1-derived hydroxyvitamin D₃ derivatives were different between RORγ ^-/- and wild-type dermal fibroblasts, with reduced expression of fibrosis-related genes generally only occurring in fibroblasts with functional RORγ. Our findings indicate that RORγ facilitates the inhibition of the profibrotic activities by secosteroids in murine fibroblasts. In summary, our study identifies a role for RORγ in the regulation of skin fibrosis.

Glossary

Abbreviations

AhR

aryl hydrocarbon receptor

BMP

bone morphogenetic protein

FBS

fetal bovine serum

FC

fold change

GSEA

gene set enrichment analysis

HGF

hepatocyte growth factor

KO

knockout

MMP

matrix metallopeptidase

RNA-seq

RNA sequencing

ROR

retinoic acid-related orphan receptor

RXR

retinoid X receptor

VDR

vitamin D receptor

VEGF

vascular endothelial growth factor

Additional Information

Disclosure Summary: The authors have declared that no conflict of interest exists.

Data Availability

All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.