The transcriptional activity of progestins used in contraception and menopausal hormone therapy via progesterone receptor A is dependent on the density of the receptor

Dose-response analysis ( Figure 1A ) was performed to determine the agonist efficacy ( Figure 1B ) and potency ( Figure 1C ) for P 4 and the selected progestins for transactivation via human PR-A in the same model system and the values are reported in Figure 1D . The MDA-MB-231 breast cancer cell line was considered a suitable model as it expresses negligible levels of endogenous steroid receptors, except the GR [ 18 ], to which some progestins can bind [ 19 , 20 ]. However, the transcriptional activity of the progestogens in the absence of transfected PR-A was negligible ( Supplementary Figure 1 ). While only statistically significant differences are discussed below, the possibility that the statistical power of the current experiments exclude other differences, cannot be dismissed. Results in Figure 1B show that the 2 nd generation progestin LNG displayed a similar agonist efficacy to R5020, which was set as 100% as it is considered a PR-specific ligand [ 21 ], while all other progestins and P 4 displayed a lower efficacy. Although this suggests that LNG is a full agonist for transactivation via PR-A, while P 4 and the other progestins are partial agonists, LNG, NET and GES had indistinguishable efficacies. No significant difference in efficacy was observed between P 4 , 1 st generation MPA and the 4 th generation progestins NES, NOMAC and DRSP, while the 1 st generation NET, 2 nd generation LNG and 3 rd generation GES were significantly more efficacious than P 4 via PR-A ( Figure 1B and 1D ). Notably, NET, LNG and GES are all structurally related to testosterone, while all other progestins, except DRSP which is derived from the mineralocorticoid receptor antagonist, spironolactone are structurally related to P 4 [ 8 ]. Our findings, together with our previous findings showing that P 4 as well as R5020, NET and LNG are full PR-B agonists, while only MPA is a partial agonist via PR-B [ 9 ], highlight the fact that progestins can display differential activity via PR-A and PR-B, underscoring the importance of investigating individual progestins via each PR isoform. Unlike our results showing that MPA displayed a lower efficacy than R5020, NET and LNG via PR-A ( Figure 1D ) and PR-B [ 9 ] overexpressed in the MDA-MB-231 cells, a study more than 20 years ago showed that MPA was more efficacious than all three these progestins in T47D breast cancer cells [ 22 ]. Differences between the former study and ours may be the fact that T47D cells not only co-expresses both PR-A and PR-B, but also other competing steroid receptors to which MPA can bind. Although the 4 th generation progestins were designed to be more PR-specific compared to the earlier generation progestins [ 23 ], the progestins within this generation were not the most efficacious and not always the most potent. For example, 4 th generation NOMAC was less potent than 3 rd generation GES via PR-A ( Figure 1A and C ) and displayed a similar potency to the 1 st generation progestins MPA and NET, as well as 2 nd generation LNG. On the other hand, the 4 th generation progestins NES and DRSP, were more potent than both 1 st generation progestins, but similar to the 2 nd and 3 rd generation LNG and GES, respectively. The 4 th generation progestins NES and DRSP were more potent than P 4 , whereas NOMAC was equipotent to P 4 . While EC 50 values higher than those determined in our study ( Figure 1C and 1D ) have been reported by others for R5020 [ 24 ], P 4 [ 24 – 26 ], MPA [ 26 ], NET-A [ 25 ], LNG [ 25 , 26 ], NES [ 25 ] and DRSP [ 26 ], these were from studies performed in T47D cells co-expressing PR-A and PR-B. Indeed, the potency determined for R5020 via PR-A in our study (81 pM) ( Figure 1C and 1D ), was consistent with a potency of 60.5 pM determined for PR-A overexpressed in a mouse mammary tumour cell line [ 27 ]. A recent study however reported much higher potencies for R5020, P 4 , MPA and NET, and a lower potency for LNG, in MDA-MB-231 cells stably expressing PR-A [ 10 ]. Similar to our findings, these authors also show that P 4 and MPA display lower efficacies than R5020. However, their results indicating similar efficacies for P 4 , MPA, NET and LNG contrast with ours. Differences between our study and that of Enfield et al . [ 10 ] may be attributed to differences in PR-A expression levels between the transiently transfected MDA-MB-231 cells and those stably transfected with PR-A. Next, we determined whether the density of PR-A would influence the efficacy and potency of a selected panel of progestogens, namely R5020, P 4 , MPA, NET and DRSP. The dose-response curves comparing the responses at the different PR-A densities are shown in Supplementary Figure 2 and the western blot analysis confirming increasing PR-A expression levels of 1x, 2x and 5x are shown in Figure 2A . Results show that the efficacy ( Figure 2B ) and potency ( Figure 2C ) of most, but not all, progestogens were modulated by increasing the expression levels of PR-A. While the efficacy of the natural ligand P 4 was not influenced by the density of PR-A, the efficacies of the progestins were influenced in a ligand- and PR-A density-dependent manner ( Figure 2B ). For instance, the efficacies for R5020, MPA and DRSP decreased at the 5x, but not 2x, concentration of PR-A. On the other hand, the efficacy for NET decreased at both 2x and 5x PR-A. When comparing progestin efficacies between 2x and 5x PR-A, significant differences were observed only for R5020 and DRSP. Notably, the increase in PR-A expression at both 2x and 5x resulted in a decrease in the basal luciferase activity in the absence of ligand ( Supplementary Figure 3 ). Indeed, unliganded effects of PR-A have previously been reported [ 28 ]. Thus, this decrease in basal activity may partly explain the observed decrease in efficacy for NET, at both 2x and 5x more PR-A. However, it is not that straightforward as the efficacy of R5020, MPA and DRSP did not decrease at both PR-A expression levels. While the efficacies of R5020 and DRSP were influenced by the density of PR-A, their potencies were unaffected ( Figure 2C ). In contrast, the potency of MPA increased at 2x PR-A although there was no significant difference in efficacies. While the efficacy of NET decreased at both 2x and 5x PR-A, the potencies increased to a similar extent at these densities. Surprisingly, even though the efficacy of P 4 did not change with increased PR-A expression, its potency increased at 2x PR-A and decreased at 5x PR-A ( Figure 2C ). Collectively, the data show that the density of PR-A influences the efficacy and potency in a progestogen-specific manner, which is independent of structural derivation or generation. Notably, the EC 50 values determined for P 4 , MPA and NET, at least for 2x PR-A, are similar to the values determined for these progestogens by Enfield et al ., [ 10 ] supporting our hypothesis that differences in EC 50 values between our studies are most likely due to the differing density of PR-A in the transiently transfected versus stably transfected MDA-MB-231 cells. Considering that the EC 50 values of all progestins for PR-A at all densities are either within or below the reported serum concentration ranges of the progestins in women [ 29 ], these activities may likely be mimicked in vivo . Our results are particularly relevant to tissues expressing different levels of PR-A as in the normal versus cancerous breast, and in tissues where PR-A is predominantly expressed such as the normal ovary [ 5 , 7 ]. In addition to transactivation, the PR can also repress the expression of target genes by tethering to other transcription factors such as NF κ B and activator protein (AP)-1 [ 3 ]. However, this mechanism of transrepression via the PR is relatively understudied. We thus investigated for the first time the comparative agonist activity of P 4 and the progestins R5020, MPA, NET and DRSP for transrepression via different densities of PR-A on a synthetic NF κ B-containing luciferase reporter construct, in the same model system. As attempts to perform dose-response experiments for the transrepression assay were thwarted by large variability between biological repeats, transfected cells were treated with 10 ng/mL PMA in the absence or presence of a single concentration (100 nM) of the progestogens for 24 hours. PMA was used to activate the NF κ B promoter prior to progestogen treatment ( Supplementary Figure 4 ). Results show that the repression of the PMA-mediated induction on the NF κ B promoter by R5020 was similar at all PR-A densities ( Figure 3A ). P 4 and the selected progestins displayed similar agonist activity for transrepression via PR-A ( Figure 3B ), while none of the progestogens displayed transrepression activity in the absence of transfected PR-A ( Supplementary Figure 5 ). When comparing the agonist activity of the progestogens for transrepression via 1x and 2x PR-A, only NET displayed a greater activity at 2x PR-A ( Figure 3C ). Both MPA and NET displayed significantly greater transrepression activity via PR-A at the highest concentration of PR-A used in this study (5x), with the results showing that the transrepression activity of NET was enhanced in a PR-A density-dependent manner. Although the observed increase in transrepression with P 4 and DRSP at 5x PR-A was not statistically different from that obtained at 1x and 2x PR-A, the P 4 and DRSP-induced transrepression at 5x PR-A, was indistinguishable from that of MPA (5x PR-A) and NET (2x PR-A) ( Figure 3C ). The percentage repression obtained at each PR-A density is reported in Supplementary Table 1 . Given that transrepression is an important mechanism for the anti-inflammatory effects of steroid hormones, our results showing that all the progestogens are agonists for transrepression via PR-A may partly explain the mechanism underlying the current use progestins such as MPA, NET and LNG in inflammation-induced endometriosis [ 30 ]. Further studies are required to fully understand the physiological implications of these results particularly in tissues where PR-A is the predominant PR isoform expressed, and in inflammation-induced diseases such as breast cancer and endometriosis where PR-A plays a prominent role. Taken together, we show that selected progestins representative of all four generations displayed differential agonist efficacies and potencies for transactivation via PR-A. We show for the first time that P 4 and the selected progestins display similar agonist activity for transpression via PR-A. Moreover, we show that the progestogen agonist properties for transactivation and transrepression via PR-A are differentially influenced by the density of PR-A. Our findings indicating that the effects of the selected progestins via PR-A is progestin-specific and dependent on the density of the receptor, suggests differential progestin responses, particularly in tissues expressing varying PR-A levels.

The human MDA-MB-231 breast cancer cell line was gifted by Prof. Adrienne Edkins (Rhodes University, RSA) and was maintained as previously described in [ 12 ], with only mycoplasma negative cells used in experiments. R5020 was purchased from Perkin Elmer Life and Analytical Science, while P 4 , medroxyprogesterone acetate (MPA), norethisterone (NET), levonorgestrel (LNG), gestodene (GES), nestorone (NES), nomegestrol acetate (NOMAC), drospirenone (DRSP), and phorbol myristate acetate (PMA) were purchased from Sigma-Aldrich. The human PR-A expression vector (pSG5-hPR-A) [ 13 ] and the pSG5-empty expression vector containing no eukaryotic promoter or enhancer sequences [ 14 ] were received from Dr. Eric Kalkhoven (University Medical Centre, Utrecht, Netherlands) and Prof. Gunnar Mellgren (University of Bergen, Norway), respectively. The pTAT-2x-progesterone response element (PRE)-E1b-luciferase promoter-reporter construct [ 15 ] was received from Prof. Guido Jenster (Erasmus University of Rotterdam, Netherlands), while the 5x-nuclear factor kappa beta (NF κ B)-luciferase promoter-reporter construct was purchased from Stratagene (Houston, Texas). Promoter-reporter assays were performed as previously described [ 16 ], with some modifications. MDA-MB-231 breast cancer cells were transiently transfected with 900 ng (1x), 1 800 ng (2x) or 4 500 ng (5x) PR-A expression vector and 9 000 ng pTAT-2xPRE-E1b-luciferase reporter plasmid (transactivation), or 1 500 ng (1x), 3 000 ng (2x) or 7 500 ng (5x) PR-A expression vector and 3 000 ng 5xNF κ B-luciferase reporter plasmid (transrepression). The pSG5-empty vector was used to ensure a constant total DNA concentration in all experiments. For transactivation, cells were treated with increasing concentrations of the progestogens, while for transrepression cells were treated with 10 ng/mL PMA in the absence or presence of 100 nM progestogen. Cells were lysed 24 hours later and the data analysed as previously described [ 16 ]. Protein samples were collected as described previously [ 17 ] and separated using 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE). The separated proteins were electroblotted to nitrocellulose membranes and blocked in 10% (w/v) fat-free milk powder. Membranes were probed using an anti-PR (Leica Biosystems, UK) or anti- glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Santa Cruz Biotechnology, USA) antibody, and the horseradish peroxidase (HRP)-conjugated goat anti-mouse secondary antibody (Santa Cruz Biotechnology, USA). Proteins were visualised by enhanced chemiluminescence (ECL) using the MyECL Imager (Pierce Thermo Scientific Inc., USA), and expression levels quantified using ImageJ ™ Software (v.1.8). The GraphPad Prism ® v9.00 software was used for data analysis. Non-linear regression and sigmoidal dose-response were used. A fixed Hill slope of 1 was chosen for transactivation. One-way ANOVA with Tukey’s multiple comparison and two-way ANOVA with Bonferroni’s multiple comparison post-tests were used for statistical analyses. The error bars indicate the standard error of the mean (SEM) of at least three independent experiments, each performed in triplicate.

Progestins or synthetic progestogens, are classified into four consecutive generations, and are widely used by women in contraception and menopausal hormone therapy (MHT) [ 1 ]. Progestins were designed to mimic the actions of the natural progestogen, progesterone (P 4 ) , by activating the progesterone receptor (PR), which exists as two functional isoforms, PR-A and PR-B (reviewed in [ 1 , 2 ]). Both isoforms can regulate transcription by the classical mechanism of either enhancing gene expression via the direct binding to progesterone response elements (PREs) in the promoter region of specific target genes (transactivation) or by repressing gene expression (transrepression) through tethering to DNA-bound transcription factors (reviewed in [ 3 ]). Evaluation of the relative efficacies (maximal response elicited by a progestin) and potencies (EC 50 values; the concentration of progestin eliciting half the maximal response) of progestins provides insight into their clinical relevance and side-effects. Surprisingly, studies evaluating these parameters for different progestins via the PR are limited, and most of the available studies do not compare the parameters for multiple progestins in parallel or specify the PR isoform involved. The importance of distinguishing between progestin activities via the PR isoforms is highlighted by the fact that PR-A and PR-B are expressed in several reproductive and nonreproductives tissues, and that the isoforms can have different activities [ 4 – 6 ]. A further level of complexity in understanding progestin action via the PR, is the varying expression levels of the PR isoforms found in some tissues [ 7 ], and the varying structures and functions of progestins [ 8 ]. While we recently determined for the first time the efficacies and potencies of selected progestins in parallel within the same model system for transactivation via PR-B [ 9 ], a similar study for PR-A has only very recently been reported [ 10 ]. We and others have also shown that the efficacy and potency of a progestin can be influenced by the density of PR-B [ 9 , 11 ]. For example, the potency of promegestone (R5020) for transactivation via PR-B increased with increasing PR-B levels [ 11 ]. Considering that the expression levels of PR-A vary in different tissues and disease states, it is surprising that similar studies have not been performed for PR-A. In this study, we performed a dose-response analysis for transactivation via overexpressed PR-A in the MDA-MB-231 breast cancer cell line using a panel of eight progestins, all structurally different and representative of all four progestin generations. We also investigated the agonist activity of a selection of these progestins for transrepression, as to our knowledge such studies are lacking. Moreover, we evaluated whether the density of PR-A would influence the agonist activity of selected progestins for both transactivation and transrepression via PR-A.