Oral Vitamin D supplementation impacts gene expression in granulosa cells in women undergoing IVF

Does oral Vitamin D supplementation alter the hormonal milieu of follicular fluid (FF) and the transcriptomic profile of luteinised granulosa cells (GCs) in women with Vitamin D deficiency undergoing IVF?

A transcriptomic signature relevant to oral Vitamin D supplementation in luteinised GCs was demonstrated, although Vitamin D supplementation did not alter hormone levels in FF.

Vitamin D deficiency is linked to lower live birth rates among women undergoing IVF. It is unclear whether Vitamin D elicits a targeted action in reproductive physiology or is a surrogate marker of overall well-being. Several in-vitro studies, but none in vivo, have examined the impact of Vitamin D on the periovulatory follicle, focusing on GCs as a proxy marker of oocyte competence.

We present a report of secondary outcomes from the SUNDRO clinical trial, which was launched in 2016 to determine whether Vitamin D supplementation can improve the IVF outcomes of women who are deficient in Vitamin D (<30 ng/ml). FF samples of 145 women who were randomised to receive Vitamin D or placebo from March 2017 to January 2019 were collected. All follicles that were aspirated in our study measured ≥11 mm on the day of hCG trigger. The first cohort of samples was collected from the dominant follicle of each participant and utilised for hormone profiling (n = 50 Vitamin D, n = 45 Placebo). For the second cohort, the follicle aspirates of each participant were pooled to create a single FF sample, which was used for the isolation of GCs for gene expression studies (n = 20 Vitamin D, n = 30 placebo). Six of the samples from the second cohort were used for RNA-sequencing analysis (n = 3 Vitamin D, n = 3 placebo).

Two academic infertility units were involved in the recruitment of the participants, who received a single dose of oral 25-hydroxyvitamin D (600 000 IU) or placebo, 2-12 weeks before oocyte retrieval. Women in both groups were deficient in Vitamin D, aged 18-39 years with a normal BMI (18-25 kg/m2) and <3 previous IVF cycles. The FF was aspirated at the time of oocyte retrieval and stored. Liquid chromatography tandem mass spectrometry was used to measure FF abundance of 25-hydroxyvitamin D, aldosterone, androstenedione, cortisol, cortisone, corticosterone, 11-deoxycorticosterone, 11-deoxycortisol, 21-deoxycortisol, dehydroepiandrosterone, dehydroepiandrosterone sulfate, dihydrotestosterone, oestradiol (E2), 17-OH-hydroxyprogesterone, progesterone (P4) and testosterone. GCs were isolated from pooled FFs and the transcriptome was evaluated by RNA-sequencing and RT-PCR. Ingenuity pathway analysis (IPA) was used to assess the top canonical pathways and upstream regulators mediating the action of Vitamin D.

At oocyte retrieval, FF concentration of 25-hydroxyvitamin D was 2.8-fold higher (P < 0.001) in the Vitamin D group (39.5 ng/ml; n = 50) compared to placebo (13.8 ng/ml; n = 45) but no other hormonal differences were detected. In the placebo group, but not the Vitamin D group, weak correlations of 25-hydroxyvitamin D concentration with P4 (r = 0.31, P = 0.03) and E2 (r = 0.45, P = 0.002) were observed. RNA-sequencing identified 44 differentially expressed genes in the GCs of patients who received Vitamin D (n = 3) compared to placebo (n = 3). RT-PCR demonstrated upregulation of VDR (vitamin D receptor), GSTA3 (glutathione S-transferase A3) and IL21R (interleukin 21 receptor), and downregulation of P T GS2 (prostaglandin-endoperoxide synthase 2), KLF4 (kruppel-like factor 4), T RP C4 (transient receptor potential cation channel subfamily C member 4), VEGF (vascular endothelial growth factor), RXRB (retinoid X receptor beta) and AGER (advanced glycosylation end-product specific receptor) genes in the Vitamin D (n = 17) versus placebo (n = 27) group. IPA suggested roles of Vitamin D in antioxidant defence.

Interpretation of the data is influenced by our intervention strategy (2-12 weeks prior to retrieval). As folliculogenesis may last 5-6 months, our protocol can only examine with confidence the impact of Vitamin D on the final stages of follicular growth. Furthermore, we examined the hormonal profile of the dominant follicle only, while the GC data reflect the transcriptome of all (pooled) follicles large enough to be used for IVF. Luteinised GCs from controlled ovarian stimulation were used in this study, which may be functionally distinct from the GCs of developing follicles. Moreover, the sample size for RNA-sequencing analysis was low (n = 3 per group), regardless of validation by RT-PCR that was performed on a larger cohort, introducing complexity to the IPA analysis, which required an input of data with P-adjusted <0.08 instead of <0.05 to be informative.

This is the first in-vivo study to show that Vitamin D supplementation alters gene expression in luteinised GCs. In contrast to some in-vitro evidence, no effect of the intervention on expression of genes encoding steroidogenic enzymes was observed. Unlike other studies, our results suggest that supplementation with Vitamin D is unlikely to directly influence hormone availability in FF. Our findings instead reinforce the hypothesis that Vitamin D could be considered one of the gatekeepers in protecting against an exaggerated response to ovarian stimulation.

The study has been funded by the Italian Ministry of Health (RF-2013-02358757) following peer review in the competitive 'Bando di Ricerca Finalizzata e Giovani Ricercatori 2013' for the clinical trial SUNDRO (EudraCT registration number 2015-004233-27). There are no competing interests.

EudraCT registration number 2015-004233-27.