Letrozole-Induced Polycystic Ovary Syndrome Attenuates Cystathionine-β Synthase mRNA and Protein Abundance in the Ovaries of Female Sprague Dawley Rats

ABSTRACT

Background

Polycystic ovary syndrome (PCOS) is an endocrine disorder that affects 10% of reproductive-aged women and leads to hyperandrogenism, anovulation, and infertility. PCOS has been associated with elevated serum homocysteine as well as altered methylation status; however, characterization of one-carbon metabolism (OCM) in PCOS remains incomplete.

Objectives

The aim of our research was to assess OCM in a letrozole-induced Sprague Dawley rat model of PCOS.

Methods

Five-week-old female rats (n = 36) were randomly assigned to letrozole [0.9 mg/kg body weight (BW)] treatment or vehicle (carboxymethylcellulose) control that was administered via subcutaneously implanted slow-release pellets every 30 d. For both treatment groups, 12 rats were randomly assigned to be euthanized during proestrus at one of the following time points: 8, 16, or 24 wk of age. Daily BW was measured and estrous cyclicity was monitored during the last 30 d of the experimental period. Ovaries were collected to assess mRNA and protein abundance of OCM enzymes.

Results

Letrozole-induced rats exhibited 1.9-fold higher cumulative BW gain compared with control rats across all age groups (P < 0.0001). Letrozole reduced the time spent at proestrus (P = 0.0001) and increased time in metestrus (P < 0.0001) of the estrous cycle. Cystathionine β-synthase (Cbs) mRNA abundance was reduced in the letrozole-induced rats at 16 (59%; P < 0.05) and 24 (77%; P < 0.01) wk of age. In addition, CBS protein abundance was 32% lower in 8-wk-old letrozole-induced rats (P = 0.02). Interestingly, betaine-homocysteine S-methyltransferase mRNA abundance increased as a function of age in letrozole-induced rats (P = 0.03).

Conclusion

These data demonstrate that letrozole-induced PCOS Sprague Dawley rats temporally decrease the ovarian abundance of Cbs mRNA and protein in the early stages of PCOS.

Introduction

Polycystic ovary syndrome (PCOS) is the most common endocrine disorder among reproductive-aged women, affecting 1 in 10 females (1). The etiology of PCOS is thought to be influenced by external and environmental factors, resulting in clinical abnormalities, such as anovulation, infertility, and hyperandrogenism. A hallmark of PCOS is the rapid hypothalamic pulsatility of gonadotropin release hormone (2), which stimulates the anterior pituitary hormones luteinizing hormone (LH) and follicle-stimulating hormone. Hypersecretion of LH is the central pathophysiologic driver of PCOS, which in turn leads to elevated circulating testosterone concentrations. Together, these imbalances lead to follicular arrest, amenorrhea, and infertility. Furthermore, a large concern for women with PCOS is the diminution in oocytes (3), as well as alterations to folliculogenesis (4).

A persistent imbalance of androgen production and the duration of PCOS diagnosis increases the risk for developing comorbidities, such as diabetes and cardiovascular disease (5). Aberrant methyl group metabolism and concomitant hyperhomocysteinemia is an independent risk factor for cardiovascular disease, as well as other chronic conditions (6). It is well established that women with PCOS have elevated blood and follicular homocysteine concentrations (7, 8), but the exact reason and its implications on ovarian function remain elusive. Furthermore, diabetic conditions compromise methyl group and homocysteine metabolism (9, 10). Given the common comorbidity of PCOS and diabetes (11), it is important to characterize ovarian one-carbon metabolism (OCM) and determine its relation in the pathogenesis of PCOS.

OCM is a ubiquitous system, comprising several essential nutrients functioning as coenzymes and methyl donors (12). Physiologic processes, such as maintaining serine and glycine homeostasis, DNA synthesis, and providing sufficient methyl groups for S-adenoyslmethioine–dependent transmethylation reactions, including gene expression, are reliant on a tightly regulated OCM system. Three central pathways are involved in the metabolism and balance of methyl groups and homocysteine: 1) folate-dependent remethylation of homocysteine to methionine, 2) folate-independent remethylation of homocysteine to methionine, and 3) irreversible catabolism of homocysteine via transsulfuration. The main enzymes involved in these reactions are represented in Supplemental Figure 1. The consequences of both hypo- and hypermethylation result in impaired gene expression, which has been closely linked to a myriad of metabolic diseases (13). The deleterious effects of hypermethylation on oocyte quality have been previously reported (14). Jia et al. (14) identified compromised mitochondrial DNA copy number and elevated follicular homocysteine concentrations in progressive polycystic gilt ovaries. Nevertheless, characterization of the key enzymes in the OCM pathway has not been examined. Therefore, the main objective of the present study was to induce the PCOS phenotype and characterize the OCM cycle as a function of disease progression. To achieve this objective, animals were maintained on a chemical treatment until 8, 16, or 24 wk of age, and differential mRNA, protein abundance, and enzyme activity between letrozole-induced and placebo control rats were measured.

Materials and Methods

Rats and diets

All animal studies were approved by the Institutional Animal Care and Use Committee at Iowa State University (IACUC #18-294) and were performed according to the Iowa State University Laboratory Animal Resources Guidelines. Female Sprague Dawley rats (n = 36) were purchased at 5 wk of age (Envigo) and were dually housed according to treatment group in a temperature-controlled room with a 12-h light-dark cycle. All rats were acclimated on a modified semipurified diet (AIN-93G) for 1 wk. Following acclimation, rats were randomly assigned to cage, experimental treatment, and age (8, 16, or 24 wk of age) of sacrifice. At the beginning of the experimental period, animals were divided into 2 groups: placebo (n = 18; n = 6/age) and letrozole (n = 18; n = 6/age) and subcutaneously implanted with a 30-d continuous, slow-release pellet (Innovative Research of American) to contain a 1-mg/kg body weight (BW) dose of letrozole (Sigma-Aldrich; no. 112,809-51-5) or a 1-mg/kg BW dose of the vehicle control, carboxymethylcellulose. Each 30-d dose was predetermined based on the average predicted change in BW as a function of age, providing us with a final consistent daily letrozole dosage of 0.9 mg/kg BW. Letrozole is an inhibitor of aromatase (CYP19A1), which is the key enzyme involved in the ovarian conversion of testosterone to 17β-estradiol (15). Rats were fed ad libitum a modified standard AIN-93G diet containing 50.4% kcal from carbohydrate, 17.3% kcal from protein, and 32.3% kcal from fat. For the last 30 d of the experimental period, vaginal cytology was monitored to determine the stage of the estrous cycle, as previously described (16). In addition, daily BW was recorded throughout the entire study. Animals were randomly assigned to be euthanized at proestrus in 1 of 3 groups (n = 12): 8, 16, and 24 wk of age. Letrozole-induced rats were euthanized in pseudodiestrus, because they failed to cycle into proestrus. Rats were anesthetized via a single intraperitoneal injection of ketamine/xylazine (90:10 mg/kg BW), and the epididymal fat pad, liver, kidneys, and ovaries were removed and weighed. Whole blood was collected via cardiac puncture for serum separation. Rats were euthanized via bilateral thoracotomy. One ovary was stored in RNAlater (Thermo-Fisher Scientific). Liver, kidney, and adipose samples were either snap frozen in liquid nitrogen or stored in RNAlater. All tissues were stored at −80°C until subsequent analysis.

Intraperitoneal glucose tolerance test

One week before euthanasia, rats underwent an intraperitoneal glucose tolerance test (IPGTT). Following a 14-h overnight fast with ab libitum water, animals were injected with a 1-g/kg BW dose of D-glucose in sterile 1× PBS after obtaining baseline blood glucose measurements via the lateral tail vein. Subsequent blood glucose measurements were obtained at 30, 60, 90, and 120 min postinjection using a standard glucometer (Bayer Healthcare). Data are reported as a change in blood glucose from baseline values.

Assessment of estrous cyclicity

During the last 30 d of the experimental period, vaginal smears were obtained daily from all rats between 08:00 and 09:00. Smears were collected via a vaginal lavage of 15 μL using sterile PBS solution. Samples were mounted and stained with methylene blue as previously described (17, 18). The stage of the estrous cycle (i.e., proestrus, estrus, metestrus, and diestrus) was determined and classified by the relative proportion of leukocytes, epithelial cells, and cornified cells by light microscopy, as detailed previously (19).

Testosterone

Serum testosterone concentrations were determined using a commercially available ELISA (Crystal Chem).

Quantitative real-time PCR

Selection of the main methyl group metabolism enzymes was determined based on the critical enzymes involved in the folate-dependent remethylation, folate-independent remethylation, transmethylation, and transsulfuration reactions (Supplemental Figure 1). Total ovarian RNA was extracted from half of each ovary using a Qiagen RNAeasy Mini kit (no. 74,134) and measured using the LightCyler 96-well Real-Time PCR System (Roche) as previously described (20). Primers (Supplemental Table 1) were designed and obtained from Integrated DNA Technologies. Amplification efficiencies of target and reference gene assays were verified and data were analyzed using the Livak, ΔΔCt method for relative mRNA expression (21).

Western blotting

One ovary (5 mg) was homogenized in 200 μL lysis buffer and assessed for protein concentrations as previously described (20). In brief, ovarian lysates were diluted to 1.7 μg/μL in Laemmli loading buffer, and a total of 40 μg protein was loaded onto a 15% sodium dodecyl sulfate polyacrylamide gel for separation of proteins via electrophoresis (80 min; 200 V) in 1× Tris-Glycine SDS buffer. After separation, proteins were transferred to a nitrocellulose membrane via a fully wet transfer in 25 mmol/L Tris, 192 mmol/L glycine, 20% v/v methanol, pH 8.3 buffer via electrophoresis (120 min; 100 V). Membranes were washed with PBS and incubated with cystathionine β-synthase (CBS; 1:750 dilution) and α-tubulin (1:400 dilution) primary antibodies (CBS, cat. no MA517273; α-tubulin, cat. no sc-5286; ThermoFisher Scientific) in 5% nonfat dry milk in PBS-Tween buffer overnight at 4°C and incubated with a secondary antibody (IRDye 800CW Goat Anti-Mouse and 600CW Goat Anti-Rabbit) at a dilution of 1:5000 for 1 h at room temperature. The net intensity of each band was determined using Empiria Studio Software (Li-Cor) and normalized to α-tubulin.

Cystathionine β-synthase activity

Ovarian CBS enzyme activity was determined using a commercially available assay by Abcam (cat. no ab241043). Briefly, 1 ovary per rat (5.5 mg) was homogenized in CBS assay buffer using a mechanical homogenizer on ice at speed 3.5 for 5 s. Tissue lysates were then centrifuged at 10,000 × g for 15 min at 4°C. The remaining preparation of the assay was performed per the manufacturer's instructions using the Synergy H1 Hybrid Microplate Reader (BioTek Instruments). CBS activity was determined by performing the assay in kinetic mode with the gain setting on auto for 50 min, at 1-min intervals, resulting in a total of 50 reads. CBS enzyme activity is presented as nmol × min⁻¹ × mL⁻¹.

Serum cysteine and homocysteine measurements

Serum total cysteine and homocysteine were measured by HPLC using fluorescence detection, as previously described (22).

Statistical analysis

All data were analyzed with SAS 9.4 Statistical Software. Means were assessed for normality using Pearson residuals. Normally distributed data are presented as means ± SEMs and analyzed using unpaired t tests between treatment groups within each. When data were assessed as an overall effect across age and treatment groups, a linear mixed model with analysis of main effects of treatment and age in addition to simple effects of within-age and within-treatment groups was reported. Repeated-measures analysis of IPGTT and BW were assessed via a mixed-model analysis, and Satterthwaite approximations were used to estimate degrees of freedom for post hoc tests to compare pairwise treatment means within each age level for all indicators. Statistical significance was determined at a level of P < 0.05.

Results

Letrozole-induced rats had higher cumulative BW

A main effect of letrozole treatment was observed for growth, resulting in higher cumulative BW gain as a function of time across all 3 age groups (P < 0.001; Figure 1). At 8 wk of age, letrozole-induced rats had ∼2-fold higher cumulative BW gain compared with their placebo counterparts (P < 0.01), and by 24 wk of age, letrozole-induced rats gained ∼1.8-fold more weight than the placebo group (P < 0.0001). There was a significant interaction between treatment and age (P < 0.0001). Moreover, pairwise comparisons determined differences in BW gain as a function of age in both the letrozole-induced and placebo rats (P < 0.0001).

FIGURE 1.

Cumulative body weight gain of rats on letrozole and placebo across 8, 16, and 24 wk of age. Vertical dotted lines indicate the end of the experimental period at 8 and 16 wk of age. Data are means ± SEMs; n = 6. Main effects of letrozole and age were tested in a linear mixed model of repeated measures at significance (P < 0.05). L, letrozole; P, placebo.

Letrozole did not impair glucose tolerance

There were no main effects of letrozole treatment on impaired glucose tolerance (P = 0.80) as measured by the change in blood glucose concentrations from baseline up to 120 min, indicating that the growth rate of the letrozole-induced rats was not concomitant with blood glucose intolerance (Figure 2). A main effect of age on a greater change in blood glucose concentrations was detected (P = 0.034), but there was no interaction between treatment and age (P = 0.13).

FIGURE 2.

Change in blood glucose concentrations (A) and area under the curve (B) in rats on letrozole compared with placebo following an intraperitoneal glucose tolerance test. Data are means ± SEMs; n = 6. Main effects of letrozole and age were tested in a linear mixed model of repeated measures for change in blood glucose, whereas AUC comparisons were made at each time point using an unpaired t test at significance (P < 0.05). L, letrozole; P, placebo.

Letrozole attenuated the frequency of proestrus occurrence

There are 4 stages of the estrous cycle: proestrus, estrus, metestrus, and diestrus; however, due to the effects of letrozole, a number of vaginal smears were unclassifiable potentially as a result of letrozole-induced acyclicity. The impact of letrozole exposure on the time spent at each stage of the estrous cycle was determined (Figure 3). Samples that only presented leukocytes and not able to be classified into 1 of the 4 estrous stages were termed pseudodiestrus and have been described in previous studies (23, 24). These samples were maintained in the analyses and indicated as pseudodiestrus (U). There was a main effect of treatment (P = 0.01) and age (P < 0.0001) on the percentage of samples that were in pseudodiestrus. Moreover, there was a main interaction between treatment and age on percentage in pseudodiestrus (P < 0.01). When simple effects were examined by extrapolating treatment effects within age groups, there was a higher prevalence of animals in pseudodiestrus (P = 0.0002) in letrozole-treated (22.7 ± 2.0%) compared with placebo-treated (12.7 ± 2.0%) rats at 24 wk of age. A shorter time (P < 0.0001) spent at proestrus was observed in the letrozole-induced rats compared with their placebo counterparts. Simple effects demonstrated this significance of shorter time in proestrus occurred at 16 (P = 0.003) and 24 (P < 0.01) wk of age. There were no main effects of treatment or age on the time spent in estrus or diestrus, but a prolonged occurrence of metestrus was observed in the letrozole-induced rats (P < 0.0001) and as a function of age (P = 0.03). When simple effects were assessed, letrozole-induced rats were in metestrus more frequently than placebo rats at 8, 16, and 24 wk of age (P < 0.001).

FIGURE 3.

Percent frequency of days spent in each stage of the estrous cycle for a total of 30 d. Data are means ± SEMs of the percent frequency; n = 6. Main effects of letrozole and age were tested in a linear mixed model, and simple effects between letrozole and placebo groups at each age are presented at significance (**P < 0.01, and ***P < 0.001). D, diestrus; E, estrus; L, letrozole; M, metestrus; P, placebo; Pr, proestrus; U, pseudodiestrus.

Relative uterine horn weights were atrophic as a result of letrozole exposure

A main effect of letrozole treatment resulted in a marked reduction in relative uterine horn weight (P < 0.0001), but there was no effect of age or the interaction of the two on relative uterine horn weight (Table 1). At 8 wk of age, letrozole-induced rats had a 53% decrease in relative uterine horn weight compared with the placebo group (P = 0.003), an 80% decrease at 16 wk (P = 0.0003), and a decrease of 81% by 24 wk (P = 0.003). Interestingly, there was a trend (P = 0.08) for higher relative ovarian weight as a result of letrozole exposure. There was a main effect of age (P < 0.0001) and the interaction of treatment and age (P = 0.017) on relative ovarian weight. When we examined the simple effects of letrozole treatment on relative ovarian weight within an age group, a difference was only observed at 8 wk of age (P = 0.003).

TABLE 1.

Relative organ weights of letrozole-induced or placebo Sprague Dawley rats at 8, 16, or 24 wk of age¹

	8 wk		16 wk		24 wk		P value
Characteristic	L	P	L	P	L	P	Trt	Age	Trt × age
BW, g	222 ± 4.18	186 ± 3.70	363 ± 5.82	239 ± 4.07	410 ± 9.16	275 ± 8.46	<0.0001	<0.0001	<0.0001
UH, g	1.14 ± .0138**	2.15 ± .0211	0.047 ± .00169***	2.42 ± .0354	0.052 ± .00211**	2.73 ± .0576	<0.0001	0.76	0.12
Ovary, mg	26.5 ± 1.81*	20.6 ± 0.922	16.4 ± 1.12	14.8 ± 1.31	16.8 ± 0.728	18.7 ± 1.46	0.08	<0.0001	0.02
Kidney, g	3.33 ± 1.02	3.13 ± 1.07	2.64 ± 1.04	2.92 ± 0.967	2.49 ± 1.06	2.76 ± 0.722	0.20	0.0002	0.06
Adipose, g	1.02 ± 0.167	0.880 ± 0.0274	2.92 ± 0.362	2.22 ± 0.246	3.69 ± 0.330	3.18 ± 0.343	0.14	<0.0001	0.73
Liver, g	3.75 ± 0.0814	3.74 ± 0.126	2.89 ± 0.081	2.96 ± 0.090	2.69 ± 0.0920	2.78 ± 0.141	0.61	<0.0001	0.87
BG, mg/dL	86.3 ± 4.54	95.0 ± 3.17	120 ± 8.86	108 ± 7.26	115 ± 3.63	116 ± 7.42	0.90	0.0005	0.26

¹Data are means ± SEMs; n = 6/group. Different from placebo group (*P < 0.05, **P < 0.01, ***P < 0.001). Data within age group determined by unpaired t test. Overall main (Trt and age) effects and their interaction (Trt × age) were determined via a linear mixed model. BG, blood glucose; BW, body weight; L, letrozole; P, placebo; Trt, treatment; UH, uterine horn.

Letrozole-induced rats had elevated circulating testosterone concentrations

To confirm the PCOS phenotype, circulating testosterone concentrations were measured in the rats. Letrozole-induced rats had higher circulating serum testosterone concentrations at 8 wk (P = 0.009), 16 wk (P = 0.005), and 24 wk (P = 0.02) of age (Figure 4). Concentrations in the letrozole-induced rats were elevated as much as 2.8-fold compared with the placebo rats.

FIGURE 4.

Serum testosterone concentrations for letrozole-induced and placebo control Sprague Dawley rats. Data are means ± SEMs; n = 6. Significant differences were assessed by unpaired t test. *P < 0.05, **P < 0.01. L, letrozole; P, placebo.

Circulating cysteine but not homocysteine concentrations were lower in the letrozole-induced rats

When comparing the letrozole-induced model to the placebo control rats, overall main effects indicated that rats exposed to letrozole exhibited significantly lower circulating cysteine concentrations (P = 0.02; data not shown). There was no main effect of age or interaction (treatment × age) on serum cysteine concentrations. In addition, there were no significant effects of treatment or age on circulating homocysteine concentrations (data not shown).

Letrozole modulates the gene expression of 2 key enzymes in the OCM pathway

Of the 5 OCM enzymes examined with respect to transcript level, only 2 were altered (P < 0.05) in the letrozole-induced rats (Figure 5). An analysis within age groups determined a reduction in ovarian Cbs mRNA abundance in letrozole-treated rats of 59% (P = 0.05; Figure 5B) and 77% (P = 0.008; Figure 5C) at 16 and 24 wk of age, respectively. There was a trend for a 39% reduction of Cbs mRNA abundance in rats at 8 wk of age (P = 0.06; Figure 5A). In contrast, there was a trend for 85% higher hepatic Cbs transcript abundance in letrozole-induced rats compared with the placebo controls at 16 wk (P = 0.06) of (Supplemental Figure 2). When testing across both treatment and age, there was an overall effect of treatment as a function of age, as evidenced by increased ovarian betaine-homocysteine S-methyltransferase (Bhmt) mRNA abundance (P = 0.034) in the letrozole-treated rats (Figure 5A–C). No differences in ovarian Bhmt mRNA abundance were identified between treatment groups (P = 0.32). The abundance of mRNA encoding Cyp19A1 (cytochrome P450; aromatase) was significantly reduced by 75% in the letrozole compared with placebo control rats at 8 wk of age.

FIGURE 5.

Ovarian mRNA abundance of select enzymes in letrozole-induced and placebo control Sprague Dawley rats at 8 (A), 16 (B), and 24 (C) wk of age and cystathionine β-synthase (CBS) protein abundance (D). The abundance of mRNA was normalized to 18S ribosomal mRNA and expressed as relative to placebo rats, whereas protein abundance is expressed as arbitrary values when corrected for by α-tubulin. Data are means ± SEMs of the relative fold change of mRNA transcript abundance, protein abundance, and enzymatic activity, respectively; n = 6/group. Relative fold change, protein abundance, and enzyme activity were compared via an unpaired t test within in age group and deemed significant at P < 0.05. Bhmt, betaine-homocysteine S-methyltransferase; Cbs, cystathionine β-synthase; Cyp19A1, cytochrome P450 isoform 19a1, aromatase; Dnmt, DNA methyltransferase; Esr1, estrogen receptor 1; Gnmt, glycine N-methyltransferase; L, letrozole; Mtr, methionine synthase; P, proestrus.

CBS protein abundance, but not activity, is diminished in the early onset of PCOS

We measured the protein abundance and enzyme activity of CBS in rats with and without PCOS across 3 age groups. At 8 wk of age, letrozole-induced rats exhibited a 32% reduction in ovarian CBS protein abundance (P = 0.02), but there was no significant difference between treatment groups at 16 or 24 wk of age (Figure 5D). Furthermore, there was no significant effect of letrozole exposure on ovarian CBS enzymatic activity at any age (data not shown).

Discussion

Our previous work determined aberrations in methyl group and homocysteine metabolism, resulting in global hypermethylation (25) in an obese model of type 2 diabetes (T2D), suggesting potential alterations in the OCM pathway during the pathophysiologic progression of PCOS. Therefore, the objective in this study was to characterize OCM in an animal model of chemically induced PCOS and potentially one that displays an obese phenotype. In addition, similar to our studies in a hyperphagia-induced mouse model of obesity (26), the experimental premise included analysis of tissues at 3 ages to examine mechanisms involved during the progression of a PCOS phenotype.

Compromised CYP19A1 activity is one of the underlying conditions in the pathogenesis of PCOS, resulting in concomitant elevations in testosterone concentrations, thus making circulating testosterone an excellent biomarker for PCOS (27). Although polycystic ovaries are a consistent outcome of letrozole treatment (28), there are inconsistencies in the reported metabolic and phenotypic characteristics of letrozole exposure. For instance, several studies have reported almost full recapitulation of insulin-resistant PCOS that is observed in humans, as evidenced by observations of insulin sensitivity, increased adiposity, and obesity in the letrozole-induced mice (28, 29). In contrast, the letrozole-treated rats in this study had no indication of glucose intolerance compared with their control counterparts at any age. There were no differences in fasting blood glucose concentrations or glucose tolerance tests between treatment groups. There was a robust difference in absolute BW and cumulative BW gain across all age groups, but this was not concomitant with increased adiposity. The null impaired blood glucose findings in our study are consistent with other studies reporting a lack of glucose intolerance (15), insulin sensitivity (27, 30), and adiposity (29) in adult letrozole-induced rodents.

We are one of the first groups to report using a slow, time-release pellet method for delivering letrozole, compared with the more common practice of daily intramuscular injections, a clearly stressful mode of delivery. Thus, it is evident that the phenotypic outcomes of letrozole treatment are variable and could be species, dose, mode of delivery, and/or developmental stage dependent (30–32). Furthermore, the letrozole-treated rats in this study were affected by the intervention, albeit to a more moderate level. We chose the 3 age groups of 8, 16, and 24 wk because they reflect earlier reproductive periods of approximately 9, 14, and 18 y of human age, respectively (33). Creating a model of PCOS that eliminates many of the confounding factors, such as insulin resistance and glucose intolerance, is important, as the model in this current study is therefore more suitable for elucidating the mechanisms that are involved in the pathogenesis of PCOS.

We observed significantly perturbed estrous cyclicity in our letrozole-induced rats. During metestrus, cornified epithelial cells along with leukocytes are present in the vaginal smears, whereas predominant presence of leukocytes indicates diestrus. Other studies employing letrozole-treated rodents have confirmed to be almost entirely acyclic, as assessed by exclusive presence of leukocytes, otherwise classified as pseudodiestrus (23, 24, 34). For the purpose of our model, acyclicity allows us to reduce the variability within the letrozole-treated group and examine differences in ovarian metabolism that are indicative of human PCOS conditions. In addition, we determined greater evidence of perturbations to ovarian function due to the observed uterine horn atrophy in the letrozole-induced rats across all stages. Uterine growth is responsive to circulating estrogen concentrations, and previous studies have reported dose-dependent suppression of the uterine weight upon letrozole exposure (32). Taken together, these observations provide us with evidence of a strong and robust model of PCOS.

A hallmark of the PCOS phenotype is elevated circulating testosterone concentrations (>70 ng/dL), leading to disruption of the hypothalamic-pituitary-ovarian axis (35). It was important for us to examine the influence of letrozole on circulating testosterone concentrations, because research has reported sex differences on OCM in animal models, but it has not been investigated whether conditions of PCOS affect OCM (36). Our letrozole-induced PCOS model resulted in an overall mean of 3.5 ng/mL of serum testosterone, which is comparable to the circulating testosterone concentrations of 5.5 ng/mL and 3.2 ng/mL that have been reported in 12-wk-old Sprague Dawley rats (37, 38). One distinction to make between the serum testosterone concentrations we observed in our letrozole-induced PCOS model compared with humans is that despite women with PCOS having elevated testosterone concentrations, they are not to the same magnitude of a healthy, male adult (240–950 ng/dL) (39). The effects of testosterone on OCM have predominantly been reported in prostate cancer models. For instance, high testosterone concentrations posttranscriptionally decreased Cbs gene expression specifically in a prostate cancer cell line (40). Furthermore, high testosterone exposure resulted in a compensatory decrease in the transsulfuration flux and glutathione production, partially explained by reduced CBS activity in the cell line. These findings are also corroborated by research examining the effects of estrogen replacement therapies on postmenopausal women (41, 42). For instance, estrogen was reported to have a dose-dependent effect on the transsulfuration of homocysteine to cysthathionine, whereby oral estradiol resulted in an increase in plasma glutathione concentrations and a reduction in circulating homocysteine (41, 42).

CBS is an enzyme encoded by the Cbs gene that is critical in the transsulfuration pathway by catalyzing the irreversible conversion of homocysteine to cystathionine (43). The primary role of CBS is its involvement in the canonical transsulfuration pathway that results in maintaining homocysteine concentrations, as well as the downstream production of cysteine and glutathione, the latter representing a major antioxidant (44). To date, when it comes to the female reproductive system, altered Cbs expression has been almost exclusively studied in ovarian cancer as a function of enhanced cellular proliferation, tumorigenic overexpression, and enhanced antioxidative capacity in cancer cells (43). Although it remains largely uncharacterized, because of the presence of increased circulating homocysteine concentrations (7) and elevated inflammatory stress (45) in the ovaries in women with PCOS, the OCM pathway is a logical candidate for the involvement in the pathogenesis of PCOS.

We demonstrated a temporal decrease in ovarian Cbs mRNA abundance during the progression of PCOS. Moreover, we reported a significant transient reduction in CBS protein abundance, as it was only observed at 8 wk of age. In addition, we detected a trend (P = 0.07) in decreased circulating cysteine concentrations at 8 wk of age, with a significant overall decrease (P = 0.02) in our PCOS model, which is consistent with our CBS transcript and protein abundance. Due to the role of CBS in the transsulfuration pathway, this finding suggests reduced catabolism of homocysteine, which, in part, may explain the elevated follicular homocysteine concentrations that have been previously reported (7, 8). Although we did not report fasting blood insulin concentrations, previous research by Ratnam et al. (46) demonstrated the effects of insulin on decreasing CBS enzymatic activity. In their diabetic animal model, insulin treatment restored the elevated CBS activity to baseline. We previously established that diabetic rats exhibit elevated hepatic CBS protein abundance (25), and interestingly, we observed a trend (P = 0.08) in increased Cbs gene expression in the liver (data not shown). It is possible that our markedly lower ovarian Cbs expression and trend (P = 0.08) for elevated hepatic Cbs mRNA abundance is influenced by circulating insulin concentrations; however, we were not able to confirm this finding, since we did not observe a significant reduction in CBS activity.

Although letrozole treatment did not affect the abundance of Bhmt within each time point, we did observe increased transcript abundance in Bhmt as a function of age, exclusively in our letrozole-treated rats. To our knowledge, only one report indicates altered Bhmt expression in a model of PCOS. A study by Jia et al. (14) reported increased Bhmt gene expression in the ovaries of gilts with PCOS. Their findings were accompanied by hyperhomocysteinemia; therefore, elevated circulating concentrations of homocysteine resulted in perturbed ovarian OCM, particularly disrupting Bhmt and glycine N-methyltransferase activation in the oocytes of gilts. BHMT is the enzyme that functions in the transmethylation of homocysteine back to the amino acid methionine via the direct donation of one methyl group to homocysteine. Upregulation of Bhmt suggests prevention of hyperhomocysteinemia, as previous research has reported elevated BHMT activity at the expense of elevated homocysteine concentrations in models of progressive folate deficiency (47) and T2D (48). We found little to no effect of letrozole exposure on serum homocysteine concentrations. Due to the observed increase in Bhmt and decrease in Cbs transcript abundance, we suggest these opposing effects are maintaining serum homocysteine concentrations within the normal range similar to the placebo group.

In conclusion, our results demonstrate that progressive PCOS via letrozole exposure perturbs the enzymatic mRNA abundance and protein concentration of CBS at the early stage of PCOS (8 wk of age). The data and results provide a novel representation of the mechanistic OCM consequences in the ovaries during the progression of PCOS, one that is absent of metabolic characteristics such as increased adiposity and insulin resistance. In addition, the importance of determining the alterations in OCM enzymes allows us to apply our findings to understand how nutrition in the early stages of development may help mitigate alterations to the canonical OCM pathway and thereby limit the severity of PCOS. Future dietary intervention studies, such as intervening with a high dietary methyl diet, are warranted to examine the therapeutic role of nutrition in supporting OCM during progressive PCOS.

Notes

Supported by the National Institute of Child Health and Human Development (1 R03 HD095091-01A1).

Author disclosures: The authors report no conflicts of interest.

KLS is an editor of the Journal of Nutrition and played no role in the journal's evaluation of the manuscript.

Supplemental Figures 1 and 2 and Supplemental Table 1 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn/.

Abbreviations used: BHMT, betaine-homocysteine S-methyltransferase; BW, body weight; CBS, cystathionine β-synthase; CYP19A1, aromatase; IPGTT, intraperitoneal glucose tolerance test; LH, luteinizing hormone; OCM, one-carbon metabolism; PCOS, polycystic ovary syndrome; T2D, type 2 diabetes.