An Exploratory Study of Vitamin D and Matrix Metalloproteinase and Their Regulators in Polyendocrine Metabolic Ovarian Syndrome

Plasma MMP and TIMP levels were determined in 28 women with PMOS and 28 control women matched for age and BMI (<29 kg/m 2 ), recruited from the Hull in vitro fertilization (IVF) clinic [ 14 ]. A total of 36 women with PMOS who met those criteria were approached, of whom 28 agreed to participate. Sequential control women were then recruited within the age range of the PMOS: 31 were approached, of whom 28 agreed to participate. The study was approved by the Yorkshire and The Humber NRES ethical committee, UK (ethics number 02/03/043), and was performed in accordance with the approved ethical protocol. PMOS was diagnosed using Rotterdam consensus criteria, which included biochemical hyperandrogenemia, defined as free androgen index (FAI) >4, clinical hyperandrogenism, defined as a Ferriman-Gallwey score >8, amenorrhea or oligomenorrhea, and polycystic ovaries identified through transvaginal ultrasound [ 53 ]. All participants were otherwise healthy and had not taken prescription or over-the-counter medication for ≥9 months before enrollment. Endocrine causes that may mimic PMOS were excluded by appropriate testing, namely non-classical 21-hydroxylase deficiency, Cushing’s disease, androgen-secreting tumors, and hyperprolactinemia [ 3 ]. Table 1 summarizes the baseline demographic and biochemical profile of both groups. Subjects had a fasting blood sample that was taken in the luteal phase of the cycle (day 21) prior to initiating IVF treatment. Height, weight and waist circumference were recorded according to World Health Organization (WHO) recommendations [ 54 ], and BMI was calculated as kg/m 2 . Plasma was separated from fasting blood samples by centrifugation at 3500 g for 15 min, aliquoted and stored at −80 °C until analysis. Sex hormone-binding globulin (SHBG) and insulin were measured using a DPC Immulite 200 analyzer (Euro/DPC, Llanberis, UK). Plasma glucose, used for calculation of homeostasis model assessment-insulin resistance (HOMA-IR), and C-reactive protein (CRP) were measured using a Synchron LX20 analyzer (Beckman-Coulter, High Wycombe, UK). The FAI was derived as the ratio of total testosterone to SHBG multiplied by 100. Testosterone was quantified in serum by isotope-dilution liquid chromatography tandem mass spectrometry (LC-MS/MS) [ 14 ], with vitamin D and its metabolites assayed by the same isotope-dilution LC-MS/MS approach [ 55 ]. Plasma MMPs and TIMPs were quantified in EDTA plasma on the Slow Off-rate Modified Aptamer (SOMAscan) platform [ 56 ], following previously validated protocols [ 57 , 58 ]. An overview of the protocol involves the incubation of plasma samples with synthetic fluorophore-labeled SOMAmer reagents, linked to biotin through a photocleavable linker, to bind target analytes. The analyte-SOMAmer complexes were captured on a streptavidin-substituted matrix and released into solution by ultraviolet cleavage of the photocleavable linker. The complexes were then immobilized through analyte-associated biotinylation on a streptavidin matrix. After elution, the released SOMAmers were used as quantitative surrogates for the corresponding analytes, with quantification performed by hybridization to complementary oligonucleotides. Calibration relied on assay reference standards [ 59 ]. Raw intensities underwent standardization and normalization for hybridization, median signal and calibration signal [ 59 ]. All measurements were performed on SOMAscan Version 3.1, which covered 12 MMPs (MMP1, MMP2, MMP3, MMP7, MMP8, MMP9, MMP10, MMP12, MMP13, MMP14, MMP16 and MMP17) and 3 TIMPs (TIMP1,2,3), as listed in Table 2 .

This was an exploratory study as there were no previous studies on the MMP and TIMP panel and their relationship to vitamin D to allow for a formal power calculation. Data were assessed for normality using the Shapiro–Wilk test and visual inspection of histograms and Q–Q plots. Continuous variables are presented as mean ± standard deviation (SD), while categorical variables are presented as frequencies and percentages. Comparisons between women with PMOS and healthy controls were performed using Welch’s independent samples t -test. Effect sizes were calculated using Cohen’s d with 95% confidence intervals (CI) and interpreted as small (0.2), medium (0.5) and large (0.8) according to conventional thresholds. To provide a distribution-free assessment of group differences, permutation testing (10,000 permutations) was performed for all primary comparisons. Given the large number of biomarkers evaluated, adjustment for multiple testing was undertaken using the false discovery rate (FDR) procedure, with adjusted q-values reported alongside nominal p -values. Associations between vitamin D metabolites and MMPs or TIMPs were assessed using Pearson correlation coefficients. Correlation analyses were conducted separately in women with PMOS, controls, and the combined cohort. To account for multiple correlation testing, FDR correction was applied. To determine whether observed associations were independent of adiposity, analyses of covariance (ANCOVA) were performed with BMI included as a covariate. Additional models incorporating insulin resistance, assessed by HOMA-IR, were explored where appropriate. Adjusted estimated marginal means and corresponding p -values are reported. All statistical tests were two-tailed, and statistical significance was defined as p < 0.05. Statistical analyses were performed using SPSS version 26 (IBM Corp., Armonk, NY, USA).

Polyendocrine Metabolic Ovarian Syndrome (PMOS; formerly polycystic ovary syndrome, PCOS) is the name recently agreed upon by international consensus, adopted to reflect the endocrine, metabolic and ovarian dysfunction of the condition and to avoid the implication of ovarian cysts conveyed by the former term [ 1 ]. PMOS is a common endocrine disorder of women of reproductive age and is the most prevalent endocrine abnormality in premenopausal women [ 2 ]. It is diagnosed using recognized international criteria after exclusion of other conditions [ 3 ], and an estimated 70% of affected women remain undiagnosed [ 4 ]. Beyond its reproductive features, PMOS carries substantial metabolic risk; affected women show higher rates of hypertension, dyslipidemia, type 2 diabetes and other cardiovascular risk factors [ 5 ]. Hyperandrogenemia, insulin resistance (IR), oxidative stress, chronic inflammation and vitamin D insufficiency are each implicated in the syndrome and are interrelated, so that their individual contributions are difficult to separate [ 6 , 7 ]. This complicates interpretation of circulating biomarkers, as markers such as heat shock proteins [ 8 ], complement and coagulation proteins [ 9 ] are altered in PMOS but are driven largely by coexisting obesity and IR [ 10 , 11 , 12 , 13 ]. Disturbances intrinsic to the syndrome are therefore difficult to distinguish from those secondary to obesity and IR. Vitamin D synthesis initiates in the skin, where ultraviolet B radiation (UV-B) converts 7-dehydrocholesterol to vitamin D 3 (cholecalciferol), with subsequent steps in the liver to 25-hydroxyvitamin D 3 (25(OH)D 3 ), and in the kidney to the active hormone 1,25(OH) 2 D 3 . 24,25-dihydroxyvitamin D 3 (24,25(OH) 2 D 3 ) is a measured metabolite in this pathway [ 14 ]. Levels of 1,25(OH) 2 D 3 are also contributed to by extra-renal macrophages through the type 2 interferon response, whose regulation differs from that of the kidney [ 15 ]. Vitamin D acts via binding to the vitamin D receptor (VDR) and heterodimerization with the retinoid X receptor [ 16 ], and through binding to the membrane VDR or the 1,25D-membrane-associated, rapid-response steroid-binding protein receptor [ 17 ]. The bound 1,25(OH) 2 D 3 –VDR complex regulates transcription of the insulin gene [ 18 ]. Low vitamin D status accompanies obesity, higher IR and raised testosterone, characteristics frequently found in PMOS [ 19 , 20 ], and vitamin D replacement has been reported to affect IR and steroidogenesis in affected obese women [ 21 ]. Severe deficiency is frequent, reported in 67–85% of affected women [ 21 , 22 ], and because vitamin D is sequestered in adipose tissue, its circulating concentration falls with increasing obesity, so the deficiency associated with IR is itself affected by adiposity [ 23 ]. Vitamin D has further roles in inflammation, oxidative-stress regulation and mitochondrial function, with antioxidant activity [ 24 ], and these actions may be independent of obesity [ 25 ]. Whether 25(OH)D 3 is anti-inflammatory is uncertain, with some studies reporting an effect and others none [ 26 ], although supplementation has been shown to reduce C-reactive protein (CRP) and stabilize interleukin-10 [ 27 ]; the anti-inflammatory actions of 1,25(OH) 2 D 3 , mediated through macrophage and T-cell responses, are more clearly established [ 28 , 29 , 30 ] ( Figure 1 ). In addition, vitamin D interacts with matrix remodeling pathways, including the MMP/TIMP axis, and it is reported that vitamin D suppresses MMP activity in chronic disorders, as well as in human uterine fibroid cells [ 31 , 32 ] ( Figure 1 ). Matrix metalloproteinases (MMPs) are the principal enzymes that degrade extracellular matrix (ECM) proteins during tissue remodeling [ 33 ], and their activity is regulated by the tissue inhibitors of metalloproteinases (TIMPs). Because remodeling reflects the balance between proteases and their inhibitors, the ratio of an MMP to its regulating TIMP may provide information on ECM turnover, and such ratios have been proposed as biomarkers of disease states [ 34 ]. MMP activity has been implicated in the pathogenesis of PMOS [ 35 ] and, within the ovary [ 36 ], both MMPs and TIMPs contribute to normal physiology, including follicular rupture and oocyte release, as well as to pathophysiology [ 37 ]. Reported MMP and TIMP findings in PMOS are inconsistent. An imbalance with reduced TIMP2 and increased MMP2/TIMP2 and MMP9/TIMP1 ratios has been described [ 35 ], yet circulating MMP2 and MMP9 have been reported as elevated [ 38 , 39 , 40 ], unchanged [ 35 ], or reduced [ 41 ]. Much of this variation appears to be related to body mass index (BMI), with which these proteases generally correlate, although increased MMP9 has also been reported independently of BMI [ 42 , 43 ]. As obesity and IR are closely associated with PMOS, it is unclear whether vitamin D metabolites relate to the MMP/TIMP system through the syndrome itself or through the accompanying obesity and insulin resistance. Studying women with PMOS who are neither obese nor insulin-resistant allows these two confounders to be removed rather than adjusted for statistically. We therefore correlated vitamin D 3 (25(OH)D 3 ) and its metabolites with circulating MMP and TIMP concentrations in non-obese, non-IR women with PMOS and in age- and BMI-matched controls without PMOS. The null hypothesis was that 25(OH)D 3 would not correlate with MMPs or TIMPs and that, after accounting for these parameters, MMP and TIMP concentrations would not differ between the non-obese PMOS and control groups.

Demographic, hormonal and metabolic data for the 28 PMOS subjects and 28 controls are shown in Table 1 . The two cohorts were weight-, IR- and age-matched with no evidence of increased systemic inflammation (CRP was not elevated and did not differ between the groups), but subjects with PMOS did have hyperandrogenemia and elevated anti-Müllerian hormone (AMH). 25(OH)D 3 , the active form 1,25(OH) 2 D 3 and the inactive form 24,25(OH) 2 D 3 did not differ between groups ( Table 1 ). Among the MMP and TIMP markers examined, only MMP16 was significantly lower in women with PMOS than controls (549.6 ± 58.3 vs. 678.0 ± 304.1 Relative Fluorescent Units (RFU), p = 0.037); this association did not remain significant following correction for multiple testing. No significant differences were observed for the remaining MMPs or TIMPs ( Table 2 ). Correlation analysis showed that there was an inverse association between 25(OH)D 3 and TIMP3 (r = −0.60, p = 0.005), and a positive association between 25(OH)D 3 and MMP16/TIMP3 r = 0.54, p = 0.017) and between 25(OH)D 3 and MMP17/TIMP3 (r = 0.6, p = 0.006) ( Table 3 ). There was a negative association between 24,25(OH) 2 D 3 and TIMP3 (r = −0.58, p = 0.007), but positive associations between 24,25(OH) 2 D 3 and MMP14/TIMP3 (r = 0.45, p = 0.049), 24,25(OH) 2 D 3 and MMP16/TIMP3 (r = 0.53, p = 0.015) and 24,25(OH) 2 D 3 and MMP17TIMP3 (r = 0.59, p = 0.006) ( Table 3 ). A trend to significance was seen for 1,25(OH) 2 D 3 and MMP14 (r = −0.464, p = 0.061); 25(OH)D 3 and MMP17 (r = 0.44, p = 0.052); 25(OH)D 3 and MMP16 (r = 0.397, p = 0.083); 24,25(OH) 2 D 3 and MMP17 (r = 0.419, p = 0.066) ( Supplementary Table S1 ). In multivariable regression analyses of the combined cohort ( n = 48), adjusting for age, BMI, and HOMA-IR, both 25(OH)D 3 and 24,25(OH) 2 D 3 were inversely associated with circulating TIMP3. Each 1-unit increase in 25(OH)D 3 was associated with a 51.1-unit reduction in TIMP3 (β = −51.1; 95% CI −97.2 to −5.0; p = 0.031). This association remained stable following bootstrap internal validation (bootstrap 95% CI −93.6 to −12.6; bootstrap p = 0.010). Each 1-unit increase in 24,25(OH) 2 D 3 was associated with an 858.7-unit reduction in TIMP3 (β = −858.7; 95% CI −1617.9 to −99.5; p = 0.028), with a bootstrap 95% CI of −1556.0 to −227.7 (bootstrap p = 0.005) ( Table 4 ). Age, BMI, and HOMA-IR were not independently associated with TIMP3 in either model. However, the overall explanatory power of the models was modest: the 25(OH)D 3 model had an R 2 of 0.116 and an adjusted R 2 of 0.034, while the 24,25(OH) 2 D 3 model had an R 2 of 0.120 and an adjusted R 2 of 0.038. These findings should therefore be interpreted as exploratory associations rather than evidence of a robust predictive model.

The PMOS patients recruited for this study were matched to controls for age, BMI, IR and CRP, as well as having equivalent levels of 25(OH)D 3 , 1,25(OH) 2 D 3 and 24,25(OH) 2 D 3 . Our hypothesis was that such matched groups would have no difference in systemic MMP and TIMP levels, nor differences in MMP/TIMP ratios, between cohorts. Only MMP16 was lower in PMOS compared to controls; however, this did not survive FDR testing. The absence of robust differences in circulating MMPs and TIMPs, following multiple-testing correction in PMOS compared to controls, suggests that there was no systemic dysregulation of the MMP/TIMP axis in this cohort. TIMP3 is unique among the tissue inhibitors of metalloproteinases because it is tightly bound to the extracellular matrix and regulates not only matrix metalloproteinases but also members of the A Disintegrin and Metalloproteinase (ADAM) family, particularly ADAM17/tumor necrosis factor-α converting enzyme (TACE) [ 44 , 45 ] Through these actions, TIMP3 influences ECM turnover, inflammatory signaling, epidermal growth factor receptor activation, angiogenesis, fibrosis and tissue remodeling [ 44 , 45 , 46 ]. Experimental studies suggest that TIMP3 plays an important role in ovarian physiology through regulation of follicular development, ovulation, and stromal remodeling, processes that are known to be altered in women with PMOS [ 37 , 47 ]. Furthermore, TIMP3 has been implicated in adipose tissue inflammation, insulin sensitivity, and hepatic fibrosis, linking extracellular matrix regulation to the broader metabolic disturbances associated with PMOS [ 44 ]. Circulating TIMP3 levels may not reflect tissue TIMP3 processes in the ovary; however, the observed inverse association between vitamin D metabolites and TIMP3 in this study raises the possibility that vitamin D may influence ovarian ECM homeostasis through mechanisms extending beyond classical MMP inhibition. While these findings require validation in larger cohorts, they suggest a potentially novel interaction between vitamin D metabolism and TIMP3-dependent tissue remodeling pathways in PMOS. Surprisingly, Vitamin D metabolites were found to independently and inversely associate with TIMP3 in the PMOS subjects, after adjustment for BMI and HOMA-IR. The strongest associations observed were inverse correlations between the vitamin D metabolites 25(OH)D 3 and 24,25(OH) 2 D 3 and TIMP3; moreover, vitamin D metabolites were positively correlated with MMP16/TIMP3. These findings were only present in the PMOS cohort; however, as circulating vitamin D concentrations themselves did not differ between PMOS and controls, this suggests that the observed relationships were not fully explained by adiposity or insulin resistance. Instead, it could be postulated that PMOS may not be characterized by major alterations in circulating vitamin D or extracellular matrix remodeling proteins individually, but rather by altered interactions between vitamin D metabolism and matrix remodeling pathways. The potential correlation of 24,25(OH) 2 D 3 with TIMP3 is interesting as 24,25(OH) 2 D 3 is regarded as an inactive metabolite in the vitamin D 3 metabolic pathway; however, 24,25(OH) 2 D 3 may not be an inactive metabolite [ 48 ], and it may have a physiological role in growth [ 49 ] and in dyslipidemia [ 50 ]. Previously, studies have linked vitamin D status with metabolic disturbances, including obesity and insulin resistance, in women with PMOS, but not necessarily with the condition itself [ 19 ]. It is reported that vitamin D interacts with matrix remodeling pathways, including the MMP/TIMP axis, and vitamin D suppresses MMP activity in chronic disorders, as well as in human uterine fibroid cells [ 31 , 32 ]. Balanced MMP/TIMP activity is fundamental for normo-ovarian physiology to enable follicular rupture and oocyte release [ 37 ]. Therefore, disturbances in the balance between MMP/TIMP activity may lead to the dysregulation of ovarian ECM remodeling and increased MMP activity, which has been reported for MMP9, which has been previously associated with the dysregulation of ECM remodeling in women with PMOS [ 51 ]. Additionally, higher MMP9/TIMP3 ratios have recently been identified in PMOS patients [ 43 ]. A preliminary hypothesis for the findings of this study would be that there is a novel Vitamin D-TIMP3 correlation in the context of PMOS. The inverse associations between vitamin D metabolites and circulating TIMP3 concentrations in the PMOS cohort reported, independent of BMI and HOMA-IR, may suggest that vitamin D interacts with ECM regulatory pathways in PMOS, a feature that has not been previously reported. No significant differences in circulating MMP and TIMP levels were observed between the control and PMOS cohorts, which suggests that systemic ECM remodeling is likely preserved in PMOS. Thus, it is possible that abnormalities in matrix remodeling are localized at the ovarian or endometrial tissue levels and are not reflected in peripheral blood measurements. TIMP3 is expressed in the liver, lungs, inflammatory cells, and many other tissues beyond reproductive tissue [ 44 ]; hence, it cannot be assumed that circulating TIMP3 levels originate from the ovary alone; however, the absence of differences in circulating TIMP3 does not exclude localized dysregulation within the ovarian microenvironment. A key strength of this study included the investigation of multiple vitamin D metabolites, identifying that circulating vitamin D and its metabolites did not differ between PMOS and matched controls, suggesting that global vitamin D metabolism was not disrupted in this cohort [ 52 ]. The cohort was closely matched to controls, allowing comparison of a panel of proteins without the bias of age, BMI, inflammation and IR. Blood samples were taken on day 21 of the menstrual cycle to limit any additional fluctuations due to changes in the menstrual cycle. Limitations include the modest number of subjects in each group, and there was a lack of ethnic diversity as the study was performed on an exclusively Caucasian population. The effect of hormonal stimulation prior to the 9-month washout period having a residual effect cannot be excluded. The exploratory study population was all undergoing IVF therapy, and a larger study in a broader population is needed to validate the results. Circulating measurements may not reflect ovarian or endometrial tissue concentrations; hence, there is a need for follicular fluid or endometrial tissue analyses. Further molecular-level analysis is required within a larger study to validate the potential role of vitamin D and its metabolites on tissue restructuring through TIMP regulation.

Exploratory analyses suggest potential inverse associations between vitamin D metabolites and TIMP3 together with weaker associations involving MMP14, MMP16 and MMP17. These findings suggest the novel hypothesis that vitamin D may influence ECM remodeling and ovarian stromal biology through regulation of TIMP3-dependent pathways independently of obesity and IR, rather than through MMPs directly.