Glucocorticoids and medroxyprogesterone acetate synergize with inflammatory stimuli to selectively upregulate CCL20 transcription

Abstract

The pro-inflammatory cytokine, chemokine (C-C motif) ligand 20 (CCL20), is emerging as a therapeutic target for immune-based therapies. Cooperative regulation of CCL20 by glucocorticoids and progestins used in endocrine therapy and pro-inflammatory mediators could modulate immune function and affect disease outcomes. We show that glucocorticoids as well as medroxyprogesterone acetate (MPA), the progestin widely used in injectable contraception in sub-Saharan Africa, cooperate with pro-inflammatory mediators to upregulate CCL20 protein and/or mRNA in human peripheral blood mononuclear cells (PBMCs) and human cervical cell lines. Changes in CCL20 mRNA levels were shown to be synergistic, as assessed by Chou analysis, cell- and gene-specific and to involve transcriptional regulation, with a requirement for a nuclear factor kappa B (NF-κB) site and glucocorticoid receptor (GR) involvement. The novel results suggest a mechanism whereby MPA, like glucocorticoids, may impact inflammation both systemically and in the genital tract in patients using MPA and/or glucocorticoid therapy.

Keywords: CCL20; Glucocorticoid receptor; Inflammatory stimuli; Medroxyprogesterone acetate; Norethisterone; Synergism.

Fig. 1:. MPA and DEX, unlike NET, cross-talk with LPS to upregulate CCL20 but not IL6 or GILZ expression in PBMCs.

In (A-C), PBMCs were stimulated with 100 nM steroids or 0.1% (v/v) EtOH (vehicle, VEH) in the presence of 5 μg/mL LPS (+LPS) or sterile Milli-Q^® water (-LPS) for 24 hours. cDNA was synthesized and relative mRNA levels of CCL20 (A), IL6 (C), and GILZ (E) were determined by qRT-PCR and normalized to GAPDH mRNA levels. Secreted CCL20 (B) and IL6 (D) protein levels are shown. Fold changes were calculated relative to LPS only, which is set to 100%. The results show pooled data from (A, n=9), (B, n=11), (C, n=6), (D, n=4), and (E, n=5) donors, plotted as mean ± SEM. Non-parametric data (C & D) were analyzed using the Kruskal–Wallis test with Benjamini, Krieger, and Yekutieli posthoc test for multiple comparisons. Parametric data (A, B & E) were analyzed using one-way ANOVA with Benjamini, Krieger, and Yekutieli posthoc test for multiple comparisons. Statistical significance is denoted as *, **, ***and **** to indicate p<0.05, p<0.01, p<0.001 or p<0.0001, respectively.

Fig. 2:. MPA and GCs upregulate CCL20, but not IL6 mRNA expression in the absence and presence of TNFα or LPS in End1/E6E7 cells, via a GR-mediated mechanism.

End1/E6E7 cells were pre-treated with 1 μM RU486 or 0.1% (v/v) EtOH (vehicle, VEH) for 2 hours and thereafter treated with 100 nM steroids or 0.1% (v/v) EtOH in the presence of 1.15 nM TNFα or 0.0001% (v/v) BSA (A, C & E), or 5 μg/mL LPS or sterile Milli-Q^® water (B, D & F) for 24 hours. Relative mRNA levels of CCL20 (A & B), IL6 (C & D), and GILZ (E & F) were determined by qRT-PCR and normalized to GAPDH mRNA levels. Fold changes were determined relative to TNFα or LPS only, which were set to 100%. The data were plotted as mean ± SEM and include results from four independent experiments (n=4). Parametric data (C) were analyzed using one-way ANOVA with Benjamini, Krieger, and Yekutieli posthoc test for multiple comparisons. Non-parametric data (A, B, D, E & F) were analyzed using the Kruskal–Wallis test with Uncorrected Dunn’s multiple comparisons posthoc test Statistical significance is denoted as *, **, *** and **** to indicate p<0.05, p<0.01, p<0.001 or p<0.0001, respectively.

Fig. 3:. Steroids and TNFα exhibit reciprocal regulation of CCL20, GILZ, and IL6 mRNA expression.

In (A-D), End1/E6E7 cells were stimulated with increasing concentrations of MPA (A & C) or CORT (B & D) in the presence of 1.15 nM TNFα or 0.0001% (v/v) BSA for 24 hours. In (E-H), End1/E6E7 cells were stimulated with increasing concentrations of TNFα or 0.0001% (v/v) BSA in the presence of MPA (E & G), CORT (F & H) or 0.01% (v/v) EtOH (vehicle, VEH) for 24 hours. Relative mRNA levels of CCL20 (A, B, E & F), GILZ (C & D), and IL6 (G & H) were determined by qRT-PCR and normalized to GAPDH mRNA levels. Dose-response curves were plotted as described in Materials and Methods. The data include three (n=3) independent experiments for (B, D, E, G & H) and four (n=4) for (A, C & F) pooled and plotted as mean ± SEM. Comparisons were done both within black or red curves and between black and red curves, denoting statistically significant differences by either stars or hashes, respectively. Non-parametric data (A, G & H) were analyzed using the Kruskal–Wallis test with Benjamini, Krieger, and Yekutieli posthoc test for multiple comparisons. Parametric data (B-F) were analyzed using one-way ANOVA with Benjamini, Krieger, and Yekutieli posthoc test for multiple comparisons. Statistical significance is denoted as *^/#, **^/##, ***^/### and ****^/#### to indicate p<0.05, p<0.01, p<0.001 or p<0.0001, respectively.

Fig. 4:. The cis-elements in the human CCL20 endogenous promoter and reporter construct.

The scheme is constructed using information from (Zhao et al., 2014, Zhao et al., 2017). The NUSBIScan Version 2.0 software (University of Basel, Switzerland) was used to predict putative GREs, with the minimum predictive score (Z score) set at 6. The red star (*) indicates the NF-κB site is mutated.

Fig. 5:. The synergism between DEX and inflammatory stimuli (TNFα and LPS) occurs at the transcriptional level on the CCL20 promoter via a mechanism involving the GR and NFκB p65.

In (A & B), HeLa cells transiently transfected with CCL20 promoter-reporter constructs, pGL2-MIP3α-LUC or pGL2-MIP3α/κBM-LUC were treated with 100 nM DEX or 0.1% (v/v) EtOH (vehicle) in the presence of 1.15 nM TNFα or 0.0001% (v/v) BSA (A), or 5 μg/mL LPS or sterile Milli-Q^® water (B) for 24 hours. In (C), HeLa cells transiently transfected with pGL2-MIP3α-LUC reporter constructs were pre-treated with 1 μM RU486 for 2 hours and then treated with 100 nM DEX or 0.1% (v/v) EtOH (vehicle) in the presence of 1.15 nM TNFα or 0.0001% (v/v) BSA, or 5 μg/mL LPS or sterile Milli-Q^® water for 24 hours. (D & E) show results of ChIP assays performed as described in Materials and Methods. Fold changes were calculated relative to TNFα (A) or LPS (B) only, which were set to 100%, or in (C -E), relative to vehicle only, which was set to 1. The data were plotted as mean ± SEM and include the following independent experiments: (A & B, n=3) and (C, n=4) each performed in triplicate and (D & E, n=5). Non-parametric data were analyzed using the Kruskal-Wallis test with Benjamini, Krieger, and Yekutieli (B) or Uncorrected Dunn’s (D & E) posthoc test for multiple comparisons. Statistical analysis for parametric data (B & C) was performed using one-way ANOVA with Benjamini, Krieger, and Yekutieli posthoc test for multiple comparisons. Statistical significance is denoted as *, **, and **** to indicate p<0.05, p<0.01, or p<0.0001, respectively. Letters of the English alphabet were used to show statistically significant differences whenever the Uncorrected Dunn’s postdoc test was used for multiple comparisons.