Abstract
AbstractStudy questionDo very low levels of 1,25-(OH)2-D3 [<20nmol/L or <10ng/ml] alter ovarian cell function?Summary answerAnin vitroenvironment of severe vitamin D deficiency [<2nM] could contribute to impaired folliculogenesis/ovulation identified in women with PCOS, but levels above this [2-20nM] may have less impact; apart from in women who are also insulin-resistant.What is known alreadyVitamin D (VD) is known to regulate steroidogenic enzymes crucial for human granulosa and cumulus cell function (Merhi Zet al, 2014) and genes that play a critical role in folliculogenesis have a vitamin D response element on their promoters (Tiejun Set al, 1998); (Maestro Bet al, 2003); (Krishnan AVet al, 2007 & 2010). Women with PCOS have lower serum 25-(OH)-D3 compared to fertile controls but studies investigating the effects of vitamin D deficiency on reproductive outcomes are conflicting.Study design, size, durationA variety of cells and tissues were used for thein vitroexperiments of this study. Ovaries were used from mice (aged between 4-5 months) to initially establish expression and correlations of VDR and AMH protein. Human ovarian tissue was obtained from from women undergoing trans-abdominal hysterectomy with bilateral oophorectomy for benign gynaecological conditions, and was dissected to provide cortex, stroma, theca and granulosa cells from various-sized follicles. Luteinised granulosa cells (GLC) were harvested from follicular aspirates taken during oocyte retrieval from women undergoingin-vitrofertilisation (IVF) treatment at various clinics. A granulosa cell line (KGN-GC), that is established to correspond to immature granulosa cells from smaller antral follicles, was used to provide mechanistic insight. All tissue was obtained with appropriate ethical permission.Participants/materials, setting, methodsmRNA and protein expression of VDR/CYP27B1/AMH in various ovarian human and mice tissue compartments was assessed by qPCR, Westerns and IHC. GC, theca and GLC were cultured with a range of concentration of 1,25-(OH)2-D3 [0.02, 0.2, 2 and 20nM] with testosterone [5×10−7M] as an aromatase substrate in the presence or absence of either FSH/LH depending on the cell type. Steroid levels (E2, Androstenedione, 17-OH-P and P) were measured in the spent medium using RIA. KGN cells were also cultured in a similar manner with VD concentrations in the presence or absence of forskolin [25µM]. Aromatase mRNA expression was assessed by qPCR and aromatase promoter II activity by transfection withCYP19A1reporter construct. KGN cells were also cultured with insulin [10ng/ml and 100ng/ml] ± VD doses and the effect on mRNA expression of insulin receptor and AMH assessed by qPCR. KGN protein lysates were immunoprecipitated with anti-VDR antibodies to pull-down all existing forms of VDR (homodimers, heterodimers and solo). The immunoprecipitants were then immunoblotted with anti-RXR antibody to assess the heterodimer association between VDR-RXR in the presence of different concentrations of VD ± forskolin (to activate cAMP-dependent signalling pathways).Main results and the role of chanceVDR mRNA significantly increased in human theca from larger antral follicles (LAF) (7-12mm, n=11) compared to theca from smaller antral follicles (SAF) (5-6mm, n=4) (unpaired t-test *p=0.013). Additionally, VDR mRNA and protein was detected in both cortex and stroma from normal and PCO.CYP27B1(the enzyme that makes active 1,25-(OH)2-D3) mRNA expression was detected in stroma tissue, showing the human ovary can make local active 1,25-(OH)2-D3; however, the expression was extremely variable. Cell culture experiments demonstrated that 1,25-(OH)2-D3 was able to suppress androstenedione production in theca taken from LAF but not SAF (n=4-8; 2-way ANOVA, p=0.048 for treatment and p=0.0018 for follicle size). 1,25-(OH)2-D3 had no effect on either 17-OH-P or P production from theca cells. Likewise, 1,25-(OH)2-D3 had no effect on either basal or FSH-mediated E2 production from unluteinised GC taken from AF. Surprisingly, 1,25-(OH)2-D3 significantly reduced both basal and LH-stimulated E2 production from luteinized GC (ANOVA p<0.0001; post-hoc t-test **p<0.005). Investigation of the mechanism behind this reduction in E2 production using the KGN-GC showed that extremely low doses of 1,25-(OH)2-D3 [0.02 and 0.2nM] significantly down-regulated forskolin-stimulated aromatase expression (n=6, One-way ANOVA **p=0.0018; post-hoc multiple t-test *p<0.05, **p<0.005). As higher doses [2 and 20nM] this attenuation was reversed. This pattern was also replicated using KGN cells transfected with aCYP19A1-PII-516reporter construct, showing that 1,25-(OH)2-D3 affected aromatase activity as well as expression (n=3; ANOVA****p<0.0001; post-hoc multiple t-tests for significance between treatments). Chronic exposure of KGN-GC to equivalent hyperinsulinemic doses [100ng/ml], significantly down-regulated insulin receptor mRNA expression, which was markedly potentiated by both lowest and highest doses of 1,25-(OH)2-D3 (n=5=8; One-way ANOVA*p=0.048; post-hoc t-tests *p<0.05, **p<0.005). Both doses of 1,25-(OH)2-D3 also reduced expression of insulin receptor in the presence of post-prandial levels of insulin; indicating the mechanism by which VD deficiency may reduce insulin sensitivity. Intriguingly, 1,25-(OH)2-D3 also further reduced the forskolin-induced downregulation of AMH mRNA expression (n=3; ANOVA *p=0.02; post-hoc t-test *p<0.05). Immunoprecipitation experiments showed a clear difference between levels of VDR associated with RXR in the presence of forskolin, compared with basal or VD ligands.Limitations, reasons for cautionIdeally, we would have used primary unluteinised GC and theca cells taken from various-sized follicles for the studies on the mechanism of action of 1,25-(OH)2-D3 and the VDR; but human ovaries are by necessity in short supply hence the use of the luteinized GLCs and KGN-GC line. By its nature human primary cells and tissues display considerable biological variability normally requiring large numbers for statistical analysis, which is not feasible due to the scarcity of such tissue. It is difficult to extrapolate from serum measurements of 25-(OH)-D3, the exact levels that the ovary would be exposed to in hypovitaminosis; hence the range chosen errered towards the lower end of VD deficiency.Wider implications of the findingsNumerous observational and intervention studies have been conducted on VD deficiency and various reproductive outcomes, with contradictory results. Whilst otherin vitrostudies (Merhiet al, 2014) have investigated effects of normal levels of vitamin D on steroidogenesis, by focusing on deficient levels our data has contributed significantly to understanding the mechanism of action of 1,25-(OH)2-D3 at a cellular level and its interaction with its receptor. We have shown for the first time an increase in VDR expression inthecaof larger follicles, which along with its ability of 1,25-(OH)2-D3 to decrease AMH expression, supports a role for 1,25-(OH)2-D3 in follicle progression. Uniquely we have shown a direct effect of deficient levels of 1,25-(OH)2-D3 on inhibiting androstenedione production from human thecal cells; which together with a reduction in LH-mediated E2 production could contribute to anovulation seen in women with PCOS and vitamin D deficiency. Human thecal and unluteinised GC are incredibly hard to obtain and in short supply for research purposes, highlighting the uniqueness of our data set. For the first time, we have also discovered a mechanism accounting for the numerous studies showing a correlation between low levels of 1,25-(OH)2-D3 and reduced insulin sensitivity i.e. by reducing insulin receptor expression. This would have implications for treating women with PCOS, insulin resistance and vitamin D deficiency. The ability of the ovary to make local bioactive 1,25-(OH)2-D3 via expression ofCYP27B1, could also explain the contradictory outcomes of clinical reproductive studies that rely on measurements of systemic VD levels. Encouragingly our data indicated that women (apart from those with PCOS and insulin resistance) must become severely deficient in VD before it impacts on their ovarian function.
Publisher
Cold Spring Harbor Laboratory