Exposure to Progestin 17-OHPC Induces Gastrointestinal Dysfunction through Claudin-1 Suppression in Female Mice with Increased Anxiety-Like Behaviors.

We examined the potential impact of 17-OHPC on the expression of tight junction genes, namely CLDN1 , OCLN , and ZO1 . CSC cells were initially treated with 17-OHPC for 3 days, followed by a 3-day period without 17-OHPC. Our findings reveal that ZO1 mRNA levels remained unchanged, while OCLN mRNA levels were significantly reduced during the initial 3-day period of 17-OHPC treatment, with expression returning to baseline during the subsequent 3-day incubation without 17-OHPC. In contrast, 3 days of 17-OHPC treatment led to sustained suppression of CLDN1 expression during the subsequent 3-day incubation period in its absence (see Fig. 1 a). We further explored the potential role and mechanism underlying 17-OHPC-induced CLDN1 suppression. Progressive 5′-deletion constructs were generated and transfected into CSC cells for luciferase reporter assays. Our results demonstrate that the reporter activity for deletion constructs of pCLDN1-100 and pCLDN1-0 significantly increased compared to the pCLDN1-2000 group, suggesting that the 17-OHPC-responsive element may reside within the range of −200 to 0 on the CLDN1 promoter (see Fig. 1 b). Subsequently, we investigated potential binding motifs within this region and created mutation reporter constructs. Among these motifs, mutations were introduced into two VDR motifs located at −102 and −71 (changing from red to green), respectively (see Fig. 1 c). The reporter activities for these mutation constructs revealed that only mutations in VDR, specifically M-102/VDR and M-71/VDR, significantly increased compared to the pCLDN1-2000 group (see Fig. 1 d). Furthermore, the VDR double mutation M-102/71-VDR completely reversed the 17-OHPC-mediated suppression of reporter activity (see Fig. 1 e), indicating that 17-OHPC suppresses CLDN1 through two VDR binding elements on the CLDN1 promoter. We also investigated the binding abilities of potential motifs, including RXRα, AP2, VDR, and Sp1, through ChIP techniques. Our findings indicate a significant reduction in VDR binding ability in the presence of 17-OHPC, while there was no difference observed with E2 and P4 treatments compared to the control (CTL) group (see Fig. 1 f). Detailed statistical information is shown below. Effects of 17-OHPC on CLDN1 expression and VDR binding. a Human CSC were treated with 10 μ m 17-OHPC for 3 days, followed by a switch to vehicle (CTL) for another 3 days. mRNA levels were determined. n = 4. * , p < 0.05 versus day 0 treatment; ¶ , p < 0.05 versus day 2 treatment. b CLDN1 reporter activities for deletion constructs. n = 5. * , p < 0.05 versus pCLDN1-2000 treatment; ¶ , p < 0.05 versus pCLDN1-100 treatment. c Schematic model illustrating potential binding motifs on the CLDN1 promoter; VDR binding motif (in red) and its mutation (in green). d CLDN1 reporter activities for single mutation constructs. n = 5. * , p < 0.05 versus pCLDN1-2000 treatment. e CLDN1 reporter activities for double mutation constructs. n = 5. * , p < 0.05 versus pCLDN1-2000/CTL treatment; ¶ , p < 0.05 versus pCLDN1-2000/OHPC treatment. f CSC were incubated with 10 μ m CTL, 17-OHPC, E2, or P4 for 3 days, then switched to CTL for another 3 days, and cells were harvested for ChIP assay. n = 4. * , p < 0.05 versus CTL+CTL treatment. Data presented as mean ± SD , analyzed by one-way ANOVA. In Figure 1 a, one-way ANOVA revealed a significant effect on OCLN (F[6,27] = 3.916, p = 0.032); a significant effect on CLDN1 (F[6,27] = 4.762, p = 0.027) and no significant effect on ZO1. Subsequent Turkey analysis revealed that 18-OHPC treatment decreased OCLN expression on day 2, 3, and 4 (58, 46, and 51% vs. day 0 group, respectively, p < 0.01) and decreased CLDN1 expression on day 2 (66% vs. day 0 group, p < 0.01) and day 3, 4, 5, and 6 (58, 68, 52, and 62% vs. day 2 group, p < 0.05). In Figure 1 b, one-way ANOVA revealed a significant effect on CLDN1 reporter activity (F[9,49] = 2.841, p = 0.041). Subsequent Turkey analysis revealed that 18-OHPC treatment increased activity at pCLDN1-100 mutant (142% vs. pCLDN1-2000 group, p < 0.01) and increased activity at pCLDN1-0 mutant (121% vs. pCLDN1-100 group, p = 0.037). In Figure 1 d, one-way ANOVA revealed a significant effect on CLDN1 reporter activity (F[7,39] = 4.725, p = 0.018). Subsequent Turkey analysis revealed that 18-OHPC treatment increased activity at M-102/VDR and M-71/VDR mutants (145 and 156% vs. pCLDN1-2000 group, respectively, p < 0.01). In Figure 1 e, one-way ANOVA revealed a significant effect on CLDN1 reporter activity (F[6,29] = 6.183, p < 0.01). Subsequent Turkey analysis revealed that 18-OHPC treatment decreased activity at pCLDN1-2000/OHPC, M-102/VDR/OHPC, M-71/VDR/OHPC, and pCLDN1-2000/CTL/shVDR mutants (31, 62, 54, and 41% vs. pCLDN1-2000/CTL group, respectively, p < 0.01) and increased activities at M-102/VDR/OHPC and M-71/VDR/OHPC mutants (200 and 174% vs. pCLDN1-2000/OHPC group, respectively, p < 0.05). In Figure 1 f, one-way ANOVA revealed a significant effect on VDR ChIP analysis (F[3,15] = 4.816, p = 0.027), while showed no significant effect on RXRα, AP2, and Sp1 ChIP analysis. Subsequent Turkey analysis revealed that 18-OHPC treatment (OHPC+CTL) decreased binding ability (41% vs. CTL+CTL group, p < 0.01). We investigated the impact of 17-OHPC-mediated epigenetic changes on the CLDN1 promoter. CSC cells were treated with 10 μ m of either VEH (CTL), 17-OHPC, E2, or P4 for 3 days, followed by a switch to CTL for an additional 3 days before being harvested for chromatin immunoprecipitation (ChIP) assay. The results revealed that 17-OHPC treatment (OHPC+CTL) led to increased DNA methylation (see Fig. 2 a), as well as elevated levels of H3K27me2 and H3K27me3 on the CLDN1 promoter (see Fig. 2 b), while E2 and P4 treatments showed no significant effect. Neither treatment had any discernible impact on H4 methylation (see Fig. 2 c) or histone acetylation (see Fig. 2 d). Detailed statistical information is shown below: Potential impact of progestin exposure on epigenetic changes on CLDN1 promoter. Human CSC were treated with 10 μ m of vehicle (CTL), 17-OHPC, E2, or P4 for 3 days, followed by a switch to CTL for another 3 days, and then cells were collected for ChIP analysis: ( a ) DNA methylation. b H3 methylation on the CLDN1 promoter. c H4 methylation on the CLDN1 promoter. d Histone acetylation on the CLDN1 promoter. Data presented as mean ± SD, analyzed by one-way ANOVA. n = 4. * , p < 0.05 versus CTL+CTL treatment. In Figure 2 a, one-way ANOVA revealed a significant effect on DNA methylation (F(3,15) = 5.183, p = 0.018). Subsequent Turkey analysis revealed that 18-OHPC treatment (OHPC+CTL) increased DNA methylation (168% vs. CTL+CTL group, p < 0.01). In Figure 2 b, one-way ANOVA revealed a significant effect on H3K27me2 ChIP analysis (F(3,15) = 3.291, p = 0.031) and H3K27me3 ChIP analysis (F[3,15] = 3.825, p = 0.022), while showed no significant effect on H3K9me2 and H3K9me3 ChIP analysis. Subsequent Turkey analysis revealed that 18-OHPC treatment (OHPC+CTL) increased binding ability on H3K27me2 and H3K27me3 (154 and 167% vs. CTL+CTL group, respectively, p < 0.01). In Figure 2 c, one-way ANOVA revealed no significant effect on ChIP analysis of H4K20me1, H4K20me3, and H4R3me1. In Figure 2 d, one-way ANOVA revealed no significant effect on ChIP analysis of H3K9,14,18,23,27ac and H3K5,8,12,16ac. CSC cells were treated with either VEH (CTL) or 17-OHPC for 3 days before being infected with lentivirus carrying empty (EMP), shVDR, VDR expression (↑VDR), or ERβ expression (↑ERβ). Subsequently, they were further incubated with CTL for an additional 3 days for biological assays. Analysis of mRNA expression confirmed the efficacy of VDR and ERβ manipulation by lentivirus. Complete restoration of 17-OHPC-mediated CLDN1 suppression was observed with VDR expression, while only partial restoration was achieved with ERβ expression (see Fig. 3 a). Corresponding protein expression mirrored the mRNA findings (see Fig. 3 b, c, online suppl. S1a). Assessment of cellular redox balance revealed that 17-OHPC treatment significantly increased 8-oxo-dG generation (see Fig. 3 d, e) and reduced the GSH/GSSG ratio compared to CTL+CTL/EMP treatment (see Fig. 3 f). VDR knockdown (shVDR) partly mimicked the effects of 17-OHPC, whereas VDR expression (↑VDR) partially reversed these effects, and ERβ expression (↑ERβ) completely reversed them. Evaluation of pro-inflammatory cytokine gene expression showed that mRNA levels of IL1β, IL6, IL17A, and MCP1 were augmented by 17-OHPC treatment and partially potentiated by VDR manipulation. However, ERβ expression completely reversed this effect (see online suppl. Fig. S2a). Similar secretion patterns were observed for these cytokines at the cellular level, including IL1β, IL6, IL17A, and MCP1 (see online suppl. Fig S2b, S2c, S2d, S2e). Detailed statistical information is shown below. 17-OHPC induces persistent CLDN1 suppression and oxidative stress, with partial rby VDR and ERβ expression. CSC were treated with either CTL or 17-OHPC for 3 days before being infected with empty (EMP), shVDR, VDR expression (↑VDR), or ERβ expression (↑ERβ) lentivirus, followed by treatment with CTL for another 3 days for assays. a mRNA expression ( n = 4). b protein expression ( n = 5). c Representative full blots for ( b ). d 8-oxo-dG generation ( n = 5). e Representative pictures for ( d ). f GSH/GSSG ratio ( n = 5). Data presented as mean ± SD, analyzed by one-way ANOVA. * , p < 0.05 versus CTL+CTL/EMP treatment; ¶ , p < 0.05 versus OHPC+CTL/EMP treatment. In Figure 3 a, one-way ANOVA revealed a significant effect on mRNA levels of CLDN1 (F[4,19] = 6.372, p < 0.01), VDR (F[4,19] = 4.829, p = 0.037), and ERβ (F[4,19] = 5.265, p = 0.029). For CLDN1, subsequent Turkey analysis revealed that treatment of OHPC+CTL/EMP, CTL+CTL/shVDR, and OHPC+CTL/↑ERβ decreased mRNA (42% [ p < 0.01], 37% [ p < 0.01], and 76% [ p < 0.05] vs. CTL+CTL/EMP group, respectively), but treatment of OHPC+CTL/↑ERβ increased mRNA (181% vs. OHPC+CTL/EMP group, p < 0.01). For VDR, subsequent Turkey analysis revealed that treatment of CTL+CTL/shVDR decreased mRNA (31% vs. CTL+CTL/EMP group, p < 0.01), but treatment of OHPC+CTL/↑VDR increased mRNA (231% vs. OHPC+CTL/EMP group, p < 0.01). For ERβ, subsequent Turkey analysis revealed that treatment of OHPC+CTL/EMP and OHPC+CTL//↑VDR decreased mRNA (54% and 61% vs. CTL+CTL/EMP group, respectively, p < 0.01), but treatment of OHPC+CTL/↑ERβ increased mRNA (211% vs. OHPC+CTL/EMP group, p < 0.01). In Figure 3 b, one-way ANOVA revealed a significant effect on protein levels of CLDN1 (F[4,24] = 6.426, p < 0.01), VDR (F[4,24] = 3.946, p = 0.043), and ERβ (F[4,24] = 4.827, p = 0.031). For CLDN1, subsequent Turkey analysis revealed that treatment of OHPC+CTL/EMP, CTL+CTL/shVDR, and OHPC+CTL/↑ERβ decreased protein (38% [ p < 0.01], 42% [ p < 0.01], and 72% [ p < 0.05] vs. CTL+CTL/EMP group, respectively), but treatment of OHPC+CTL/↑ERβ increased protein (189% vs. OHPC+CTL/EMP group, p < 0.01). For VDR, subsequent Turkey analysis revealed that treatment of CTL+CTL/shVDR decreased protein (39% vs. CTL+CTL/EMP group, p < 0.01), but treatment of OHPC+CTL/↑VDR increased protein (178% vs. OHPC+CTL/EMP group, p < 0.01). For ERβ, subsequent Turkey analysis revealed that treatment of OHPC+CTL/EMP and OHPC+CTL//↑VDR decreased protein (46% and 54% vs. CTL+CTL/EMP group, respectively, p < 0.01), but treatment of OHPC+CTL/↑ERβ increased protein (164% vs. OHPC+CTL/EMP group, p < 0.01). In Figure 3 d, one-way ANOVA revealed a significant effect on 8-oxo-dG formation (F(4,24) = 6.891, p < 0.01). Subsequent Turkey analysis revealed that treatment of OHPC+CTL/EMP, CTL+CTL/shVDR, and OHPC+CTL/↑VDR increased 8-oxo-dG (194% [ p < 0.01], 139% [ p < 0.05], and 146% [ p < 0.05] vs. CTL+CTL/EMP group, respectively), but treatment of CTL+CTL/shVDR and OHPC+CTL/↑VDR decreased 8-oxo-dG (72% and 75% vs. OHPC+CTL/EMP group, respectively, p < 0.05). In Figure 3 f, one-way ANOVA revealed a significant effect on GSH/GSSG ratio (F[4,24] = 9.146, p < 0.001). Subsequent Turkey analysis revealed that treatment of OHPC+CTL/EMP, CTL+CTL/shVDR, OHPC+CTL/↑VDR, and OHPC+CTL/↑ERβ decreased GSH/GSSG ratio (32% [ p < 0.01], 59% [ p < 0.05], 63% [ p < 0.05], and 40% [ p < 0.01] vs. CTL+CTL/EMP group, respectively), but treatment of CTL+CTL/shVDR and OHPC+CTL/↑VDR increased GSH/GSSG ratio (184% and 197% vs. OHPC+CTL/EMP group, respectively, p < 0.01). We investigated the roles of VDR and 17-OHPC in IEC and demonstrated that 17-OHPC treatment (WT/OHPC) significantly reduced CLDN1 mRNA levels compared to the WT/VEH group, with VDR deficiency replicating this effect. VDR mRNA levels were notably decreased in the VDR deficiency (shVDR) group compared to the WT group, indicating successful VDR knockdown (see Fig. 4 a). Subsequent protein expression analysis of VDR and CLDN1 revealed similar patterns to mRNA expression (see Fig. 4 b, c, online suppl. S1b). Immunostaining for CLDN1 protein corroborated these findings showing consistent expression levels with mRNA levels (see Fig. 4 d, e). Measurement of cellular redox balance across the groups demonstrated that 17-OHPC treatment (WT/OHPC) decreased the GSH/GSSG ratio and increased 8-oxo-dG generation compared to WT/VEH treatment. VDR deficiency either partially or completely replicated this effect (see Fig. 4 f, g). Detailed statistical information is shown below. In Figure 4 a, for CLDN1 mRNA level, two-way ANOVA revealed a significant interaction of shVDR and OHPC treatment (F[1,12] = 17.634, p < 0.01), simple main effects analysis showed a significant effect of OHPC treatment on CLDN1 mRNA, p < 0.001, and a significant effect of shVDR treatment on CLDN1 mRNA, p < 0.01; for VDR mRNA level, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,12] = 1.254, p = 0.857), and simple main effects analysis showed no significant effect of OHPC treatment, while a significant effect of shVDR treatment, p < 0.001. VDR knockdown in intestine mimics 17-OHPC treatment-induced CLDN1 suppression, with minor impact on redox balance in IEC. VDR deficiency in intestine (shVDR) or wild-type (WT) female mice were injected with either vehicle (VEH) or 17-OHPC, and IEC cells were subsequently purified for analysis. a mRNA expression ( n = 4). b Protein expression ( n = 5). c Representative full blots for ( b ). d CLDN1 quantitation by immunostaining ( n = 5). e Representative pictures for ( d ). f GSH/GSSG ratio ( n = 5). g 8-OHdG generation ( n = 5). Data presented as mean ± SD, analyzed by two-way ANOVA. * , p < 0.05 versus WT/VEH treatment; # , p < 0.05 versus shVDR/VEH treatment. In Figure 4 b, for CLDN1 protein level, two-way ANOVA revealed a significant interaction of shVDR and OHPC treatment (F[1,16] = 14.815, p < 0.01), and simple main effects analysis showed a significant effect of OHPC treatment on CLDN1 protein, p < 0.001, and a significant effect of shVDR treatment on CLDN1 protein, p < 0.01; for VDR protein level, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,16] = 0.965, p = 0.857), and simple main effects analysis showed no significant effect of OHPC treatment, while a significant effect of shVDR treatment, p < 0.01. In Figure 4 d, for CLDN1 staining quantitation, two-way ANOVA revealed a significant interaction of shVDR and OHPC treatment (F[1,16] = 15.936, p < 0.01), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and a significant effect of shVDR treatment, p < 0.01. In Figure 4 f, for GSH/GSSG ratio, two-way ANOVA revealed a significant interaction of shVDR and OHPC treatment (F[1,16] = 7.581, p = 0.024), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and a significant effect of shVDR treatment, p = 0.018. In Figure 4 g, for 8-OHdG formation, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,16] = 1.362, p = 0.925), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and no significant effect of shVDR treatment. We investigated the influence of 17-OHPC and VDR on redox balance and inflammation in serum, revealing that 17-OHPC treatment reduced the GSH/GSSG ratio (see Fig. 5 a) and increased 8-OHdG formation (see Fig. 5 b) compared to the WT/VEH group. Moreover, 17-OHPC treatment significantly elevated cytokine levels of IL1β (see Fig. 5 c), IL6 (see Fig. 5 d), MCP1 (see Fig. 5 e), and IL17A (see Fig. 5 f) compared to the WT/VEH group, with VDR deficiency completely recapitulating these effects. Detailed statistical information is shown below. VDR knockdown in intestine has minimal impact on 17-OHPC treatment-mediated redox balance and inflammation in serum. VDR deficiency in intestine (shVDR) or wild-type (WT) female mice were injected with either vehicle (VEH) or 17-OHPC, and serum was isolated for analysis. a GSH/GSSG ratio. b 8-OHdG generation. c – f Serum levels of cytokines, including IL1β ( c ), IL6 ( d ), MCP1 ( e ), and IL17A ( f ). Data presented as mean ± SD, analyzed by two-way ANOVA. n = 5. * , p < 0.05 versus WT/VEH; # , p < 0.05 versus shVDR/VEH treatment. In Figure 5 a, for GSH/GSSG ratio, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,16] = 0.984, p = 0.492), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and no significant effect of shVDR treatment. In Figure 5 b, for 8-OHdG formation, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,16] = 1.118, p = 0.947), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and no significant effect of shVDR treatment. In Figure 5 c, for IL1β in serum, two-way ANOVA revealed a significant interaction of shVDR and OHPC treatment (F[1,16] = 3.486, p = 0.046), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and a significant effect of shVDR treatment, p = 0.042. In Figure 5 d, for IL6 in serum, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,16] = 1.209, p = 0.518), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and no significant effect of shVDR treatment. In Figure 5 e, for MCP1 in serum, two-way ANOVA revealed a significant interaction of shVDR and OHPC treatment (F[1,16] = 3.816, p = 0.041), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and a significant effect of shVDR treatment, p = 0.037. In Figure 5 f, for IL17A in serum, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,16] = 1.352, p = 0.683), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and no significant effect of shVDR treatment. We investigated the potential impact of 17-OHPC and VDR on GI dysfunction and observed that 17-OHPC treatment increased intestinal permeability compared to the WT/VEH group, with similar effects observed with VDR knockdown (shVDR) treatment (see Fig. 6 a). Subsequently, we assessed the gut microbiota and found that all treatments had minimal effects on both species richness (see Fig. 6 b) and diversity (see Fig. 6 c). Analysis of the relative frequencies of different phyla revealed that Actinobacteria and Firmicutes were dominant in the WT/VEH group. However, 17-OHPC treatment decreased the abundance of Actinobacteria while increasing the abundance of Firmicutes, with VDR deficiency replicating this pattern (see Fig. 6 d). Furthermore, the relative abundance of g_Mucispirillum was significantly reduced by 17-OHPC exposure, with VDR knockdown showing a similar decrease (see Fig. 6 e). Finally, we examined the relative abundance of specific bacteria and found that 17-OHPC exposure attenuated the abundance of p_Deferribacteres while increasing the abundance of p_Tenericutes and p_Proteobacteria. VDR deficiency mirrored these effects, with no significant differences observed for any other bacteria among the groups. Detailed statistical information is shown below. VDR knockdown in intestine resembles 17-OHPC treatment-mediated GI dysfunction. VDR deficiency in intestine (shVDR) or wild-type (WT) female mice were injected with either vehicle (VEH) or 17-OHPC, and the treated mice were utilized for GI dysfunction analysis. a Intestinal permeability ( n = 5). b – f Gut microbiota analysis: species richness ( b ); diversity ( c ); frequencies of different phyla ( d ); abundance of mucispirillum ( e ); abundance of bacteria ( f ). n = 9. Data presented as mean ± SD, analyzed by two-way ANOVA. * , p < 0.05 versus WT/VEH treatment. In Figure 6 a, for intestinal permeability assay, two-way ANOVA revealed a significant interaction of shVDR and OHPC treatment (F[1,16] = 9.164, p < 0.01), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and a significant effect of shVDR treatment, p < 0.01. In Figure 6 b, for observed species, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,32] = 0.924, p = 0.896), and simple main effects analysis showed no significant effect of OHPC treatment and no significant effect of shVDR treatment. In Figure 6 c, for observed species, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,32] = 0.849, p = 0.912), and simple main effects analysis showed no significant effect of OHPC treatment and no significant effect of shVDR treatment. In Figure 6 e, for relative abundance of g- Mucispirillum , two-way ANOVA revealed a significant interaction of shVDR and OHPC treatment (F[1,32] = 18.792, p < 0.01), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and a significant effect of shVDR treatment, p < 0.01. In Figure 6 f, for relative abundance of p_Bacteroidetes, p_Firmicutes, p_Saccharibacteria, p_Actinobacteria, p_Cyanobacteria, and p_Verrucomicrobia, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment, simple main effects analysis showed no significant effect of OHPC treatment and shVDR treatment; for relative abundance of p_Proteobacteria, two-way ANOVA revealed a significant interaction of shVDR and OHPC treatment (F[1,32] = 9.716, p < 0.01), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.01, and a significant effect of shVDR treatment, p < 0.01; for relative abundance of p_Deferribacteres, two-way ANOVA revealed a significant interaction of shVDR and OHPC treatment (F[1,32] = 23.624, p < 0.01), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and a significant effect of shVDR treatment, p < 0.01; for relative abundance of p_Tenericutes, two-way ANOVA revealed a significant interaction of shVDR and OHPC treatment (F[1,32] = 31.925, p < 0.01), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.01, and a significant effect of shVDR treatment, p < 0.01. We investigated the potential roles of VDR and 17-OHPC in animal behaviors. In terms of gene expression in the hippocampus, 17-OHPC exposure significantly decreased mRNA levels of ERβ and SOD2 compared to the WT/VEH group. VDR deficiency mirrored this effect, with no significant difference observed in SYP mRNA levels among the groups (see Fig. 7 a). Protein expression patterns were consistent with mRNA levels (see Fig. 7 b, c, S1c). Further analysis of mRNA levels in the amygdala revealed that VDR deficiency replicated 17-OHPC treatment-induced suppression of SOD2 expression, while showing no effect on the expression of ERβ and SYP among the groups (see Fig. 7 d). Additionally, no significant effect on gene expression was observed in the hypothalamus among the groups (see Fig. 7 e). Regarding animal behaviors, the results of anxiety-like behavior tests indicated that 17-OHPC treatment significantly reduced buried marbles in the marble-burying test (see Fig. 7 f). In the elevated plus maze (EPM) test, mice spent less time in the Open Arm but more time in the Closed Arm after 17-OHPC treatment (see Fig. 7 g); intestinal-specific VDR deficiency showed no effect. Autism-like behaviors were evaluated using the three-chambered social test, revealing that neither 17-OHPC nor VDR had any significant effect on sociability (see Fig. 7 h) or social novelty (see Fig. 7 i) among the groups. Detailed statistical information is shown below. VDR knockdown in intestine exhibits limited impact on 17-OHPC exposure-induced anxiety-like behaviors and gene suppression in the brain. VDR deficiency in intestine (shVDR) or wild-type (WT) female mice were injected with either vehicle (VEH) or 17-OHPC, and the treated mice were used for assays. Tissues from the hippocampus, amygdala, and hypothalamus were isolated for gene expression assays. a mRNA expression in the hippocampus ( n = 4). b Protein quantitation in the hippocampus ( n = 5). c Representative full blots for ( b , d ) mRNA expression in the amygdala ( n = 4). e mRNA levels in the hypothalamus ( n = 4). Animal behavior assays ( n = 9). f Marble-burying test. g Elevated plus maze test. Three-chambered social test for sociability ( h ) and social novelty ( i ). Data presented as mean ± SD, analyzed by two-way ANOVA. * , p < 0.05 versus WT/VEH; # , p < 0.05 versus shVDR/VEH treatment. In Figure 7 a, for SOD2 mRNA level, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,12] = 1.281, p = 0.827), simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and no significant effect of shVDR treatment; for ERβ mRNA level, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,12] = 1.107, p = 0.749), simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and no significant effect of shVDR treatment; for SYP mRNA level, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,12] = 0.915, p = 0.764), and simple main effects analysis showed no significant effect of OHPC treatment and shVDR treatment. In Figure 7 b, for SOD2 protein level, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,16] = 1.562, p = 0.783), simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001 and no significant effect of shVDR treatment; for ERβ protein level, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,16] = 1.829, p = 0.883), simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and no significant effect of shVDR treatment; for SYP protein level, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,16] = 1.326, p = 0.947), simple main effects analysis showed no significant effect of OHPC treatment and shVDR treatment. In Figure 7 d, for SOD2 mRNA level, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,12] = 1.086, p = 0.786), simple main effects analysis showed a significant effect of OHPC treatment, p < 0.01, and no significant effect of shVDR treatment; for ERβ mRNA level, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,12] = 0.926, p = 0.817), and simple main effects analysis showed a significant effect of OHPC treatment and shVDR treatment; for SYP mRNA level, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,12] = 1.183, p = 0.917), and simple main effects analysis showed no significant effect of OHPC treatment and shVDR treatment. In Figure 7 e, for mRNA levels of SOD2, ERβ, and SYP, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment, and simple main effects analysis showed no significant effect of OHPC treatment and shVDR treatment. In Figure 7 f, for buried marbles, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment ( F (1, 32) = 1.329, p = 0.716), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and no significant effect of shVDR treatment. In Figure 7 g, for open arm, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,32] = 1.387, p = 0.957), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and no significant effect of shVDR treatment; for closed arm, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment (F[1,32] = 1.093, p = 0.825), and simple main effects analysis showed a significant effect of OHPC treatment, p < 0.001, and no significant effect of shVDR treatment. In Figure 7 h, for sociability at both stranger 1 side and empty side, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment, and simple main effects analysis showed no significant effect of OHPC treatment and shVDR treatment. In Figure 7 i, for social novelty at both stranger 1 side and stranger 2 side, two-way ANOVA revealed no significant interaction of shVDR and OHPC treatment, and simple main effects analysis showed no significant effect of OHPC treatment and shVDR treatment. We devised a model to illustrate that 17-OHPC treatment induces GI dysfunction mediated by VDR in female mice. Moreover, progestin treatment triggers anxiety-like behaviors by modulating the gene expression of SOD2 and ERβ in the brain through oxidative stress and epigenetic modifications (see Fig. 8 ). Progestin 17-OHPC treatment induces GI dysfunction and anxiety-like behaviors in female mice. ERβ, estrogen receptor β; CLDN1, claudin-1; 17-OHPC, 17-hydroxyprogesterone caproate; SOD2, superoxide dismutase 2; VDR, vitamin D receptor.

We illustrate the potential effect of progestin 17-OHPC treatment-mediated CLDN1 inhibition through epigenetic modifications and reduced association of VDR with the CLDN1 promoter in CRC cells. While intestine-specific VDR deficiency can partially replicate the altered redox balance, inflammatory response, and GI dysfunction induced by 17-OHPC treatment, it exhibits minimal impact on the anxiety-like behaviors triggered by 17-OHPC. Regarding the potential effect of progestin on brain function, prenatal exposure to progestin induces epigenetic modifications leading to persistent gene suppression in the brain, thereby eliciting autism-like behaviors in rodent offspring [ 21 , 28 ]. Our findings in female mice reveal that 17-OHPC exposure directly suppresses the expression of SOD2 [ 22 ] and ERβ [ 21 , 22 ] genes in the hippocampus [ 3 ] and amygdala but not in the hypothalamus. However, there is no discernible effect on SYP expression in these areas [ 29 ]. While 17-OHPC exposure induces anxiety-like behaviors in mice, it does not trigger autism-like behaviors, suggesting that progestin exposure may modulate women’s mood, including anxiety, fear, and depression [ 4 , 30 , 31 ]. Regarding the potential effect of progestin on GI function, progestin exposure induces GI dysfunction in rodent offspring, characterized by altered gut microbiota and intestinal permeability. Intestinal VDR deficiency partially mimics this effect, although the precise mechanism of progestin-induced GI dysfunction through VDR remains unclear [ 16 ]. Our study demonstrates that 17-OHPC treatment leads to significant epigenetic modifications on the CLDN1 promoter, including potentiated DNA methylation and histone 3 methylation, resulting in VDR dissociation from the CLDN1 promoter, subsequent CLDN1 suppression, and GI dysfunction in female mice exhibiting anxiety-like behaviors. This underscores the critical role of progestin and VDR in modulating GI function [ 7 ], consistent with recent reports implicating VDR in the regulation of GI tract inflammation and immune function [ 32 , 33 ]. Regarding the role of VDR in brain and GI function, previous studies have shown that vitamin D and VDR signaling pathways are involved in maintaining the integrity of the brain and gut microbiome through the gut-microbiome-brain axis [ 34 ]. Our findings indicate that progestin exposure-mediated epigenetic changes cause dissociation of VDR from the CLDN11 promoter, resulting in suppression of CLDN1 expression and subsequent GI dysfunction. Intestine-specific VDR deficiency suppresses CLDN1 expression, exacerbates intestinal permeability, and alters the intestinal microbiome, mirroring progestin-mediated GI dysfunction. However, VDR-mediated GI dysfunction showed little effect on progestin-induced anxiety-like behaviors. This suggests that VDR in the GI system may still influence brain function, such as neurodevelopmental disorders [ 10 ], even though it does not manifest in anxiety-like behaviors [ 35 , 36 ]. On the other hand, it remains unclear whether the anxiety-like behaviors are a direct result of progestin’s impact on the brain or if they are mediated by the gut. Limitations of this study include the relatively small number ( n = 9) of animals used for each experimental treatment and the selection of 17-OHPC as the representative of progestins without investigating other types of progestins, which may compromise the accuracy of the conclusions. Furthermore, the absence of intestine-specific VDR overexpression mice precludes the investigation of the potential rescuing effect of VDR on progestin exposure-mediated GI dysfunction in in vivo studies. In conclusion, exposure to progestin 17-OHPC disrupts redox balance and induces epigenetic modifications on the CLDN1 promoter, leading to the suppression of CLDN1 expression through VDR dissociation. Intestine-specific VDR deficiency partially mimics 17-OHPC treatment-induced alterations in intestinal permeability and gut microbiota but has minimal impact on anxiety-like behaviors. In general, progestin treatment triggers GI dysfunction via VDR-mediated suppression of CLDN1, in addition to eliciting anxiety-like behaviors in female mice.

Since the 1940s, progestin has been introduced and widely utilized for hormonal contraceptives and the prevention of preterm birth, providing significant benefits in birth control and alleviating symptoms related to sex hormones, such as endometriosis and menstrual irregularities [ 1 , 2 ]. However, numerous synthetic progestins exhibit various off-target effects on neural and gastrointestinal (GI) function [ 3 , 4 ]. Progestin influences the brain through the modulation of epigenetic changes and gene expression [ 5 , 6 ]. Furthermore, it may affect brain function and structure via the gut-brain axis. Synthetic progestins frequently disturb the gut microbiota, resulting in gut dysbiosis. This disruption contributes to impaired bidirectional communication between the gut microbiome and the brain, impacting brain functions, though the precise mechanisms remain largely elusive [ 7 – 9 ]. The vitamin D/vitamin D receptor (VDR) signaling pathway is involved in many neurodevelopmental disorders, including autism spectrum disorders and attention-deficit/hyperactivity disorder [ 10 – 12 ]. Vitamin D deficiency is reported to worsen maternal diabetes-mediated autism-like behaviors in mouse offspring [ 13 ]. Vitamin D and VDR regulate gene expression, gut barrier integrity, and modulate gut microbiota and inflammatory responses [ 14 , 15 ], indirectly modulating brain function through the gut-microbiota-brain axis [ 10 ]. Recent reports indicate that VDR is involved in claudin-1 (CLDN1) suppression and GI dysfunction [ 16 ] in progestin-mediated autism-like offspring, although the specific mechanism remains unclear. Additionally, it has been reported that estrogen receptor β (ERβ) regulates the basal expression of superoxide dismutase 2 (SOD2) and cellular redox balance [ 17 , 18 ], and its suppression in the brain contributes to autism spectrum disorder development and anxiety-like behaviors in offspring [ 19 – 22 ]. In this study, ERβ was overexpressed to investigate its potential effect on progestin-mediated oxidative stress, inflammation, and CLDN1 expression. We aim to investigate the potential mechanism for progestin-mediated GI dysfunction, hypothesizing that progestin exposure suppresses CLDN1 expression [ 16 ], subsequently triggering GI dysfunction. We explored the potential mechanism of progestin treatment-induced CLDN1 inhibition through the VDR in female mice. 17-hydroxyprogesterone caproate (17-OHPC) was selected as a representative of progestin for both in vitro and in vivo investigation, with the intestine-specific VDR knockdown mouse used as an animal model to mimic VDR deficiency and regulate CLDN1 expression [ 16 ]. The molecular consequences of progestin exposure in intestinal epithelial cells (IECs) were evaluated through gene expression, oxidative stress, and inflammatory responses. The potential role of altered expression of VDR [ 13 ] or ERβ through lentivirus infection [ 22 ] was investigated for progestin-mediated molecular effects. Additionally, progestin exposure-mediated changes in animal behaviors were evaluated through anxiety-like behaviors and autism-like behaviors, and the gene expression of SOD2, ERβ, and synaptophysin (SYP) in different brain regions, including the amygdala, hypothalamus, and hippocampus, was also evaluated as potential markers of autism/anxiety-like behaviors [ 21 , 22 ]. Finally, the possible role of VDR in progestin-induced GI dysfunction was assessed through analysis of intestinal permeability and gut microbiota.

No competing interest is declared.

This study was kindly supported by the National Natural Science Foundation of China, Project #: 82,060260; Futian Science Development Project #: FTWS2022055 and FTWS2021038; the Excellent Talent Team of Hainan Province Project #: QRCBT202121; and Hainan Province Clinical Medical Center Project #: QWYH202175.

The expanded section is accessible in Supplementary file Data S1 (for all online suppl. material, see https://doi.org/10.1159/000538692 ), and the primer details are provided in online supplementary Table S1. Antibodies against AP2 (sc-12726), RXRα (sc-515929), Sp1 (sc-17824), and SYP (sc-17750) were sourced from Santa Cruz Biotechnology. 17β-estradiol (E2, #E2758) and progesterone (P4, #P0130) were obtained from Sigma. Human VDR/ERβ expression and VDR knockdown lentivirus were previously prepared in our laboratory [ 13 , 22 , 23 ]. DNA isolated from Human Colon Stem Cells (CSC) was utilized, and the human CLDN1 promoter (2 kb upstream + first exon, identified by Ensembl ID: CLDN1-201 ENST00000295522.4) was amplified and inserted into the pGL3-basic reporter vector using the following primers with underlined restriction sites: CLDN1 forward: 5′-ttgg-ggtacc-cag tgg cac gat ctg ggc tca-3′ (Kpn1) and CLDN1 reverse: 5′-ttgg-ctcgag-tgc tca gat tca gca agg agt-3′ (Xho1). DNA methylation at the CLDN1 promoter was assessed using a methylation-specific PCR-based method, following established procedures with minor adjustments [ 24 – 26 ]. Genomic DNA was isolated from treated cells and subjected to bisulfite modification using the EpiJET Bisulfite Conversion Kit (#K1461, Fisher Scientific). The modified DNA was subsequently amplified using the following primers. Methylated primers: forward 5′-ttt ata gga gcg aga aga ttt acg a-3′ and reverse 5′-ccc taa cga ttt caa aac gac-3′; Unmethylated primers: forward 5′-ttt ata gga gtg aga aga ttt atg a-3′ and reverse 5′-ccc cta aca att tca aaa caa c-3′. The expected product sizes were 152 bp for methylated DNA and 153 bp for unmethylated DNA, with the CpG island spanning 172 bp. The melting temperatures (Tm) were determined to be 59.22°C for methylated DNA and 55.63°C for unmethylated DNA. Finally, the obtained DNA methylation data were normalized against the results from unmethylated DNA. mRNA levels were determined through real-time PCR using primers provided in online suppl. Table S1, and results were normalized using β-actin as the housekeeping gene. Protein expression was assessed via Western blotting or immunostaining using antibodies for 8-oxo-dG or CLDN1, and the scanned images were quantified using Image J. software [ 24 ]. Mapping of the 17-OHPC responsive element on the CLDN1 promoter was conducted using a luciferase assay. Epigenetic modifications on the CLDN1 promoter were evaluated using chromatin immunoprecipitation (ChIP) techniques and DNA methylation analysis as described above. Pro-inflammatory cytokines were assessed using the Bio-Plex Pro Mouse/Human Cytokine 23-plex Assay Kit as previously described [ 25 , 26 ]. The animal protocol adhered to US NIH guidelines (Guide for the Care and Use of Laboratory Animals, No. 85-23, revised 1996) and was reviewed and approved by the Institutional Animal Care and Use Committee at Hainan Women and Children’s Medical Center. In this study, nine female mice were allocated to each exposure group, following the protocol established in our previous report [ 16 ]. All experimental animals were housed under standard 12-h light/dark cycles and provided ad libitum access to mouse chow fortified with vitamin D3 (+VD; containing 0.81% calcium, 0.63% phosphorus, and 2.2 IU/g vitamin D3). Adult female mice, aged 3 months, were subjected to daily monitoring of estrous cycles via vaginal smears. Only mice exhibiting at least two regular 4- to 5-day estrous cycles were included in the study. VDR knockout mice at the intestine (shVDR) were generated and confirmed by genotyping as described previously [ 27 ]. Female mice were treated with either 17-OHPC (10 mg/kg body weight) to mimic medium doses of human progestin exposure as reported previously [ 16 ] or vehicle (VEH) for 1 month. They were then randomly assigned to four treatment groups: treatment (1): VDR wild-type (WT) mice receiving VEH (WT/VEH); treatment (2): VDR WT mice receiving 17-OHPC (WT/OHPC); treatment (3): VDR null mice receiving VEH (shVDR/VEH); treatment (4): VDR null mice receiving 17-OHPC (shVDR/OHPC). The treated mice were subjected to behavioral assessments during their light cycles, as detailed below. Fecal samples were collected for analysis of the fecal microbiome. Subsequently, mice were euthanized via cervical dislocation following the evaluation of GI symptoms. Serum was collected for analysis of the GSH/GSSG ratio and cytokine levels. Various brain tissues, including the amygdala, hypothalamus, and hippocampus, were isolated, flash-frozen in dry ice, and stored in a −80°C freezer for gene expression analysis. IECs were isolated using the method described below for further biological assays. Anxiety-like behavior was assessed using the marble-burying test (MBT) and elevated plus maze (EPM) test, while autism-like behavior was evaluated using the three-chambered social test. Details are provided in online supplementary Data S1. Treated mice underwent a fasting period before receiving FITC-dextran via gavage. Blood samples were collected, and serum FITC concentrations were measured using a fluorescence reader. Details of the assay are described in online supplementary Data S1. Data are presented as mean ± standard deviation, with each experiment independently repeated at least four times unless otherwise indicated. Statistical significance was determined using one-way analysis of variance (ANOVA) followed by the Tukey-Kramer test for multiple comparisons. Two-way ANOVA followed by the Bonferroni post hoc test was used to analyze interactions between factors. A p value of <0.05 was considered significant, and SPSS 22 software was used for analysis.

The animal protocol was approved by the Ethical Committee from Hainan Women and Children’s Medical Center with the reference #: 2020006.

Paul Yao prepared the manuscript. Liqin Cheng, Ling Li, and Paul Yao designed and supervised the entire study. Qingjun Shen, Xiaohan Liu, Qiaozhu Wu, and Shuangyun Zheng conducted part of the biological analysis. Li He, Xiangyue Zeng, and Irene Ma conducted part of the mouse experiments. Liqin Zeng and Xiaozhuang Zhang conducted the remaining assays. All authors read and approved this manuscript.