Evaluating the impact of the hexosamine biosynthesis pathway and O-GlcNAcylation on glucose metabolism in bovine granulosa cells

. Disruption of glucose


Introduction
Glucose is a major energy source for ovarian follicles, particularly the granulosa cells (GCs) (Armstrong et al., 1963;Boland et al., 1994;Boland et al., 1994a;Flint and Denton, 1969;Leroy et al., 2004a). Glycolytic metabolism is a major driver for follicles to successfully complete growth and steroidogenesis (Boland et al., 1993(Boland et al., , 1994a. Although energy metabolism throughout the stages of folliculogenesis has not been well characterized, ovarian follicles can develop to pre-ovulatory size via glycolysis alone, regardless of oxygen availability (Boland et al., 1994). Conversely, depletion of glucose availability hinders the growth of follicles and impairs their ability to develop an antrum or to attain pre-ovulatory size (Boland et al., 1994). Thus, while the preferred energy substrate for metabolism within follicles is clearly glucose, the requirement for oxygen, and thus the importance of oxidative phosphorylation, during folliculogenesis is less certain. Recent studies suggest that GCs cultured under low (5%) vs. high (21%) oxygen conditions exhibit increased rates of glycolysis, proliferation, and phosphofructokinase activity with low levels of ATP, mitochondrial mass, mtDNA, and mitochondrial membrane potential (Shiratsuki et al., 2016). As the follicle is known to be a hypoxic environment (Van Blerkom et al., 1997), the previously mentioned study indicates granulosa cells are adapting to low oxygen conditions by a means of increased glycolysis (Shiratsuki et al., 2016).
Gonadotropin stimulation of antral follicles promotes glucose uptake, while simultaneously enhancing lactate production, suggestive of an upregulation of glycolytic metabolism (Allen et al., 1981;Boland et al., 1994;Boland et al., 1993;Flint and Denton, 1969;Hillier et al., 1985). Indeed, high concentrations of lactate within the follicular fluid of antral follicles is further indication that the follicle, and thus the GCs therein, preferentially metabolize glucose via glycolysis, rather than via oxidative phosphorylation (Leroy et al., 2004a;Nandi et al., 2007Nandi et al., , 2008Orsi et al., 2005). In sheep, for instance, glycolytic metabolism is the preferred pathway for gonadotropin induced differentiation of GCs (Campbell et al., 2010). Similarly in pigs, GCs of small antral follicles are very proliferative and have a high rate of proliferation that is accompanied by a metabolic shift toward aerobic glycolysis rather than oxidative phosphorylation (Costermans et al., 2019). This metabolic shift, sometimes referred to as the Warburg effect (Warburg, 1925), is common to cancer and other rapidly proliferating cell types (Sun et al., 2019;Tekade and Sun, 2017;Warburg, 1925). In the cow, small antral follicles (3-5 mm) contain GCs undergoing immense proliferation, and higher concentrations of lactate are detected in the follicular fluid than large antral follicles (>8.5 mm) in which GC proliferation has begun to wane (Maucieri and Townson, 2021). These observations suggest a metabolic shift occurs from glycolysis toward oxidative phosphorylation as follicles grow in the cow, and as the GCs therein shift from proliferation toward differentiation (i.e., luteinization), but these concepts have yet to be explored experimentally. Further evidence supporting the perception that GCs thrive initially in a glycolytic environment comes from the observation that glucose utilization by GCs of preantral follicles is stimulated by paracrine factors from oocytes, which favor glycolysis (Sugiura et al., 2005).
There are a variety of pathways in which GCs metabolize glucose. These include glycolysis, the pentose phosphate pathway, the polyol pathway, oxidative phosphorylation and the hexosamine biosynthesis pathway (HBP) (Kim et al., 2010). In the current study, the HBP is of interest because it accounts for 2-5% of glucose metabolism in all cell types, yet becomes an important conduit for glucose processing when glucose is in abundance and glycolytic metabolism predominates (Buse et al., 2002;Copeland et al., 2008;Ma and Hart, 2013). Glucose metabolized via the HBP forms uridine diphosphate N-acetylglucosamine (UDP-GlcNAc), a substrate which is then used to modify serine and threonine residues of proteins that might otherwise be phosphorylated (Hart and Akimoto, 2009). The addition of these sugar moieties is referred to as O-linked N-acetylglucosaminylation (O-GlcNAcylation), and is regulated solely by two enzymes, O-GlcNAc transferase (OGT) and O-GlcNAcase (OGA) (Bond and Hanover, 2015). The process of O-GlcNAcylation is highly dependent on nutrient availability and thus is proposed as a nutrient sensing mechanism within cells, regulating a vast array of cellular processes including proliferation, signal transduction, transcription, and metabolism. O-GlcNAcylation is implicated in a variety of disease states including cancer, neurodegenerative disease, and diabetes (Lefebvre et al., 2010;Ma and Hart, 2013;Wani et al., 2017). Dysregulation of O-GlcNAcylation is a hallmark of insulin resistance (Copeland et al., 2008). Coincidentally, insulin resistance is a common, transient occurrence in postpartum dairy cows that can, in part, result in delayed ovarian cyclicity (De Koster and Opsomer, 2013;Lucy, 2016aLucy, , 2016b. Additionally, the enzymes of the glycolytic pathway, including phosphofructokinase, aldolase, glyceraldehyde-3-phosphate dehydrogenase, enolase, pyruvate kinase, and lactate dehydrogenase all undergo O-GlcNAc modification (Dehennaut et al., 2008). Considering that O-GlcNAcylation alters the functionality of target proteins (Hardivillé and Hart, 2014), it is conceivable that glycolysis may be regulated by O-GlcNAcylation. In the cow, not only is the HBP active within ovarian follicles, but the degree of O-GlcNAcylation differs between small and large antral follicles (Maucieri and Townson, 2021), impacts GC proliferation (Maucieri and Townson, 2021;Wang et al., 2022) and impacts oocyte competence and fertilization (Zhou et al., 2019). An important connection between O-GlcNAcylation and glucose metabolism in GCs of developing antral follicles may exist in the cow as they relate to follicular growth and ovarian function.
Based upon the above observations, we have formulated the working hypothesis that the degree of O-GlcNAcylation in GCs of bovine ovarian follicles varies according to relative follicle size, and that manipulating O-GlcNAcylation influences the function and fate of bovine GCs. The objectives in the current study were to characterize the basal metabolism of GCs from small and large antral follicles, and to then evaluate the effects of manipulating the HBP (and, thus, O-GlcNAcylation) on GC metabolism. Assessing glucose metabolism in GCs of antral follicles is expected to provide insight about how follicles utilize energy for growth and maturation that can ultimately lead to ovulation.

Isolation of granulosa cells
Bovine ovary pairs were collected from a slaughterhouse (Champlain Beef, Whitehall, NY) in 0.9% sterile saline with antibiotic-antimycotic (10,000 units/mL of penicillin, 10,000 μg/mL of streptomycin, and 25 μg/mL of Fungizone™; Gibco, Gaithersburg, MD). The ovaries were transported to the laboratory for processing at room temperature within 4-6 h of slaughter. Only ovaries morphologically staged to the mid-tolate estrous cycle were used (Ireland et al., 1980). Small (3-5 mm) and large (>8.5 mm) antral follicles were aspirated using a 21-gauge needle and luer lock syringe to collect granulosa cells (GCs). Samples were centrifuged at 84×g for 10 min at 4 • C to separate GCs from the follicular fluid. The pellet containing the GCs was separated from the follicular fluid and placed in cell culture as described below. The follicular fluid was discarded.

Granulosa cell culture
GCs were collected and cultured using a previously validated culture system (Maucieri and Townson, 2021). Briefly, the cells from individual ovary pairs were initially seeded into T25 flasks in DMEM/F12 medium (Gibco) containing 10% fetal bovine serum (Corning, Corning, NY) and antibiotic-antimycotic (10,000 units/mL of penicillin, 10,000 μg/mL of streptomycin, and 25 μg/mL of Fungizone™; Gibco) regardless of cell number/viability. The GCs were then incubated at 37 • C in 5% CO 2 and 95% air. After 72 h of culture (i.e., once the cultures had attained confluency), the GCs were switched to serum free conditions (described below) and were either treated for 24 h with the small molecule inhib-

Immunodetection of glycolytic enzymes
Cultures of GCs exposed to the treatments (Control, Thiamet-G, or OSMI-1) for 24 h were washed with HBSS (Gibco) and trypsinized with 1 mL TrypLE (Gibco) to detach the cells from the T25 flasks. The resulting suspension of cells was quenched with an additional 5 mL DMEM/F12 + 10% FBS media, removed from flasks, and pelleted via centrifugation at 84×g for 5 min. Supernatants were removed, GC pellets were set on ice, and 500 μL RIPA lysis buffer was added for 10 min. GCs were further lysed by aspirating through a 27-gauge needle. The entire sample was vortexed for 15 s and centrifuged at 15,295×g for 10 min. The resulting pellet was discarded but the protein lysate (supernatant) was subjected to quantification using a bicinchoninic acid (BCA) colorimetric assay (Thermo Scientific). Absorbance was measured at 562 nm by a Synergy HT Plate Reader (BioTek, Winooski, VT). The samples were diluted in RIPA buffer and Lamelli sample buffer to standardize protein concentration prior to boiling at 100 • C for 1-3 min.
For electrophoresis, 10 μg of total protein was loaded into each well of a precast, stain free 10% gel (BioRad, Hercules, CA). Following protein separation, the gels were activated and imaged for quantification of total protein using the BioRad ChemiDoc Imaging System. The proteins were then transferred to PVDF membranes (BioRad) using a semi-dry transfer system for 1 h at 45 mA (Hoefer, Holliston, MA). Post transfer, the gels were reimaged using the ChemiDoc Imaging System to verify complete transfer of the separated proteins. Following transfer, the membranes were incubated in tris-buffered saline (TBS) solution with 5% BSA and 0.2% Tween20 (TBST; Fisher Bioreagents, Pittsburgh, PA) as blocking agents, and agitated using a platform rocker for 3 h at room temperature. Following the blocking step, the membranes were incubated overnight at 4 • C on a platform rocker with rabbit antihexokinase (1:2500, C64G5), phosphofructokinase (1:2500, D4B2), pyruvate kinase (1:2500, C103A3), pyruvate dehydrogenase (1:2500, C54G1) (Cell Signaling Technology, Danvers, MA), lactate dehydrogenase (1:2500, NBP1-48336) (Novus Biologicals, Littleton, CO), O-GlcNAc transferase (1:1000 D1D8Q) (Cell Signaling Technology), or O-GlcNAcase (1:1000, SAB00311) (Sigma) primary antibodies. Following incubation with the primary antibody, the membranes were washed with TBST under agitation in the following order: 2X for 10-30 s, 1X for 15-min, and 3X for 5-min before incubating in secondary antibody for 1 h at room temperature on a platform rocker. Goat-anti-rabbit and goatanti-mouse antibodies conjugated to horse radish peroxidase (HRP) were used as the secondary antibodies (#7074, #7076 1:5000, Cell Signaling Technology). The membranes were washed as previously described and incubated in Clarity Western ECL Blotting Substrate (BioRad) for 5 min. Immunodetectable proteins were imaged on the ChemiDoc Imaging System. Protein values were standardized to total protein of the gel detected immediately after electrophoresis, prior to membrane transfer and immunodetection. Each protein of interest was quantified using Image J software (National Institute of Health).

Seahorse analyses
The GCs were sub-cultured to 96-well Seahorse XF microplates (Agilent Technologies, Santa Clara, CA) in serum-free DMEM/F12 medium containing a reduced concentration of ITS (insulin-10 ng/ml, transferrin-5.5 ng/ml, and sodium selenite-0.67 pg/ml) at a density of 75,000 cells/well. The GC cultures were exposed to 0.1% DMSO (Control), the HBP inhibitor, DON (6-diazo-5-oxo-L-norleucine, 50 μM; Cayman Chemical), the OGT inhibitor, OSMI-1 (50 μM, Cayman Chemical), or the OGA inhibitor, Thiamet-G (2.5 μM, Cayman Chemical), for 24 h. DON, OSMI-1, and Thiamet-G have all been previously validated as effective agents to disrupt global O-GlcNAcylation without adversely affecting the cells (Maucieri and Townson, 2021;Ortiz-Meoz et al., 2015;Walter et al., 2020;Yuzwa et al., 2008). The DMEM/F12 serum-free medium was exchanged with Seahorse XF DMEM medium (no glucose, glutamine, pyruvate, sodium bicarbonate, or phenol red, pH 7.5; Agilent Technologies) supplemented with 1 mM glucose. To prevent dilution of the treatments following the medium exchange, the treatments (DON, OSMI-1, Thiamet-G) were added again to the new medium at the appropriate concentration. The cultured GCs of small and large antral follicles were also subjected to glycolytic and mitochondrial stress tests as per the manufacturer's guidelines (Agilent Technologies). Briefly, these tests included three consecutive injections of treatments. For the glycolytic stress test, this entailed exposure of the cells to 10 mM glucose (glycolysis stimulant), followed by 1 μM Oligomycin (ATP synthase inhibitor), and then 50 mM 2-deoxy-glucose (an unmetabolizable form of glucose, used as a hexokinase and glycolysis inhibitor), separated by 18-min intervals. For the mitochondrial stress test, the cells were exposed to 1 μM Oligomycin (ATP synthase inhibitor), followed by 1 μM trifluoromethoxy carbonylcyanide phenylhydrazone (FCCP, uncoupling agent of the mitochondrial proton gradient), and then 0.5 μM Rotenone and Antimycin A(Complex I and III inhibitors), separated by 18-min intervals following an initial basal reading. Measured outputs were: extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) in the conditioned culture medium and were acquired every 3 min automatically by the Wave software. Seahorse measurements of ECAR and OCR are direct measurements of the relative rates of aerobic glycolysis and oxidative phosphorylation, respectively . Measurements of ECAR and OCR were automatically generated by the Seahorse XFe96 Wave software. All the inhibitors used in the above-described experiments were sourced from Cayman Chemical.

Statistics
In the instance that a sample from an individual pair of ovaries was run in either duplicate or triplicate, the mean of the two or three wells corresponding to that ovary pair was used as the value in the corresponding statistical analyses. In all instances, the individual cow or "ovary pair" is considered the experimental unit. A paired t-test was used to compare enzyme expression between untreated GCs obtained from small and large antral follicles within the same ovary pair. A repeated measures one-way ANOVA followed by Dunnett's multiple comparisons was used to evaluate enzyme expression relative to treatment. A repeated measures two-way ANOVA followed by Bonferroni's multiple comparisons was used to evaluate Seahorse analyses relative to treatment and time, and within follicle size classifications. A paired t-test was used to evaluate OCR and ECAR at individual time points between GCs obtained from small and large antral follicles. Significant differences were declared at P < 0.05. Tests were performed using GraphPad Prism 8 statistical software.

Granulosa cells of small antral follicles exhibit overall greater metabolic activity than those of large antral follicles
Overall, GCs cultured from small antral follicles exhibited overall greater metabolic activity than those from large follicles, as evidenced by greater ECAR (Fig. 1A, P < 0.05) and OCR (Fig. 1B, P < 0.05) under basal conditions.

Metabolic response to in vitro stress tests varies with antral follicle size
To evaluate further the metabolic capacity of the GCs from small and large antral follicles, cultured GCs were subjected to glycolytic and mitochondrial stress tests. As seen in Fig. 2, the glycolytic stress test revealed that GCs from both small and large antral follicles respond favorably to excess glucose by increased ECAR ( Fig. 2A; P < 0.05). Moreover, glycolysis (i.e., ECAR) remained elevated following inhibition of ATP Synthase via Oligomycin ( Fig. 2C; P < 0.05), but there was no concomitant effect on oxidative phosphorylation (i.e., OCR) ( Fig. 2D; P > 0.05). Inhibiting glycolysis with the glucose antagonist, 2-deoxy-Dglucose, decreased OCR in GCs from small antral follicles ( Fig. 2F; P < 0.05), but had no effect on GCs from large antral follicles ( Fig. 2F; P > 0.05) or on the ECAR of GCs from either follicle size ( Fig. 2E; P > 0.05).
The mitochondrial stress test (Fig. 3) revealed that GCs only from large antral follicles were sensitive to manipulation of oxidative phosphorylation through electron transport, mitochondrial membrane, and/ or ATP synthase mechanisms as evidenced by concomitant increases in ECAR and decreases in OCR. Specifically, oligomycin treatment, which prevents ATP synthase activity, increased ECAR (3A; P < 0.05) while simultaneously decreasing OCR (3B; P < 0.05). Uncoupling of the mitochondrial proton gradient via FCCP had no further effect on ECAR or OCR (3C and D, respectively; P > 0.05). However, inhibition of mitochondrial Complex I and III enzymes with rotenone and antimycin A, further increased ECAR (3E; P < 0.05), as it diminished OCR (3F; P < 0.05). There were no effects of these inhibitors on oxidative phosphorylation in GCs from small antral follicles (Fig. 3 A-F; P > 0.05).

Inhibition of the hexosamine biosynthesis pathway (HBP) and O-GlcNAcylation fails to impair granulosa cell metabolism, regardless of antral follicle size
Consistent with the above observations that overall metabolic activity is greater in GCs from small antral follicles compared to large antral follicles, evaluation of the effects of HBP manipulation revealed greater ECAR and OCR in GCs from small antral follicles compared to large antral follicles under control culture conditions (Fig. 4, compare open triangles vs. open circles, respectively; P < 0.05). Inhibition of the HBP via the GFAT inhibitor, DON (50 μM) had no effect on ECAR or OCR of GCs regardless of follicle size ( Fig. 4A and B, P > 0.05). Similarly, inhibition of OGA (via Thiamet-G) and OGT (via OSMI-1) had no effect on the ECAR or OCR of GCs regardless of follicle size ( Fig. 4C and D, P > 0.05). Additionally, and as expected (Maucieri and Townson, 2021), inhibition of the HBP decreased the expression of immunodetectable O-GlcNAcylated proteins in cultured GCs (results not shown).

Glycolytic enzyme expression varies by antral follicle size
As depicted in Fig. 5A and B, expression of immunodetectable phosphofructokinase was greater in GCs from small antral follicles compared to large antral follicles (P < 0.05). Expression of other key enzymes of glycolytic metabolism, including hexokinase, pyruvate kinase, pyruvate dehydrogenase, and lactate dehydrogenase, however, did not differ between the two antral follicle sizes ( Fig. 5C; P > 0.05).

Fig. 1. Extracellular Acidification Rate (ECAR) and Oxygen Consumption Rate
(OCR) of granulosa cells from bovine antral follicles according to follicle size. (A) ECAR (mpH/min) and (B) OCR (pMol/min) are depicted for small (SF; 3-5 mm) and large (LF; >8.5 mm) antral follicles. Results represent n = 10 ovary pairs, run in triplicate, with the mean±SEM of the respective measurement shown. Asterisks denote differences between SF and LF at P < 0.05.

Fig. 2. ECAR and OCR of granulosa cells of bovine small (SF) and large (LF) antral follicles using a glycolysis stress test. For the glycolysis stress test, ECAR (A) and OCR (B) after stimulation with glucose ECAR (C) and OCR (D) after inhibition with Oligomycin, and ECAR (E) and OCR (F) after inhibition with 2-deoxy-D-glucose.
Results represent n = 10 ovary pairs, run in triplicate, with the mean±SEM rate of the respective measurement shown. Asterisks denote a difference at P < 0.05.

Inhibition of O-GlcNAcylation impairs expression of glycolytic enzymes
Observing the similarity in expression of glycolytic enzymes between GCs of small and large antral follicles, we then examined the effect of manipulating O-GlcNAcylation on these same enzymes. As seen in Fig. 6, augmenting O-GlcNAcylation (via Thiamet-G) had no effect on any of these enzymes, regardless of follicle size ( Fig. 6A and B; P > 0.05). Conversely, inhibition of O-GlcNAcylation (via OSMI-1) impaired the expression of hexokinase in GCs of small antral follicles only ( Fig. 6A and B; P < 0.05) and inhibited the expression of phosphofructokinase, pyruvate kinase, and pyruvate dehydrogenase in GCs of both small and large antral follicles ( Fig. 6A and B; P < 0.05). There was no effect of inhibition of O-GlcNAcylation on the expression of lactate dehydrogenase, regardless of follicle size (Fig. 6A and B; P > 0.05).

OGT inhibition via OSMI-1 does not affect OGT or OGA protein expression
As OSMI-1 impaired expression of glycolytic enzymes and has been shown to impair global O-GlcNAcylation in GCs (Maucieri and Townson, 2021), its effects on OGT and OGA expression were evaluated. No effects were observed for either OGT or OGA expression in GCs from small and large follicles (Fig. 7, P > 0.05). (A and B, respectively), uncoupling with FCCP (C and D respectively), or after inhibition with rotenone and antimycin A (R + A) (E and F, respectively). Results represent n = 10 ovary pairs, run in triplicate, with the mean±SEM of the respective measurement shown. Asterisks denote a difference at P < 0.05.

Discussion
Glucose metabolism in preantral and early antral follicles is considered primarily glycolytic (Boland et al., 1993(Boland et al., , 1994aHarris et al., 2007;Leese and Lenton, 1990). Glycolytic metabolism diminishes the need for oxygen by granulosa and theca cells, allowing it to be conserved  for use by the oocyte exclusively (Biggers et al., 1967;Harris et al., 2007). Indeed, lactate is a major glycolytic product of follicular metabolism and this is supported by the observation that cultured GCs produce large quantities of lactate, even in the presence of oxygen (Billig et al., 1983;Boland et al., 1993;Leese and Barton, 1984). The current thinking is that glycolysis affords a rapid and efficient avenue for cells to convert glucose to ATP during rapid cell division. In the bovine ovary, small antral follicles consisting of highly proliferative GCs have less glucose, but greater lactate concentrations in the follicular fluid than large antral follicles in which GC proliferation has begun to wane (Maucieri and Townson, 2021). This suggests there may be a point during follicle maturation wherein GCs become less reliant on glycolysis for rapid growth, and transition to oxygen consuming pathways (i.e., oxidative phosphorylation) to support differentiation (e.g. luteinization) and steroidogenesis (e.g., estradiol production). Experiments in the current study are among the first, to our knowledge, to investigate these aspects of glycolytic and oxidative metabolism in GCs at the cellular level, using highly sensitive methods that can measure potential changes in metabolism in real time. Additionally, we evaluated the influence of the hexosamine biosynthesis pathway (HBP), and the novel process of O-GlcNAcylation, on glucose metabolism in GCs considering potential changes to glycolytic enzymes.
Consistent with previous work suggesting GCs transition from glycolytic to oxidative metabolism as follicles grow (Boland et al., 1993(Boland et al., , 1994a(Boland et al., , 1994bLeroy et al., 2004bLeroy et al., , 2008Maucieri and Townson, 2021;Nandi et al., 2007;Orsi et al., 2005), we discovered direct evidence of this transition at the cellular level by measuring metabolism in cultured GCs using the Seahorse XFe96 analyzer. This machine detects extracellular proton secretion and oxygen consumption as indices of glycolysis and oxidative phosphorylation, respectively, in real time. Using an established glycolytic stress test, we determined that cultures of GCs from small antral follicles were acutely sensitive to glucose availability and relied upon glycolysis (with or without oxidative phosphorylation) to provide cellular energy. The increase in ECAR, without an accompanying effect on OCR, following an influx of glucose is indicative of this action (Fig. 2 A and B). Additionally, inhibiting oxidative phosphorylation via oligomycin had no effect on OCR, but again increased ECAR (Fig. 2 C and D), inferring that GCs from small antral follicles do not rely on oxidative phosphorylation as part of their metabolism. Conversely, cultures of GCs from large antral follicles ostensibly have the capacity to utilize a variety of energy substrates (i.e., glutamine, fatty acids, etc.), and readily metabolize them via oxidative phosphorylation. This was evident during the portion of the glycolytic stress test in which 2-deoxy-D-glucose (2-DG, a non-metabolizable form of glucose) was introduced. Briefly, oxidative phosphorylation (as measured by OCR) continued unimpeded, but only in GCs from large antral follicles (Fig. 2F). The decline in OCR in response to 2-DG in GCs from small antral follicles (Fig. 2F) suggests these cells rely upon glucose processed directly through glycolysis to provide cellular energy. In similar work conducted by Shiratsuki and coworkers, cultured bovine GCs in the presence of low oxygen levels (5%) exhibited upregulated glycolysis and increased cellular proliferation (Shiratsuki et al., 2016). Such dependence upon glycolysis versus oxygen consumption is likely cell type specific. This thinking is supported by studies from others who found that the OCR of cancer cells, for instance, remains unaffected by treatment with 2-DG (Wu et al., 2007). In contrast, neuronal cells exhibit a remarkable decline in OCR when exposed to 2-DG (Dodson et al., 2022). Collectively, these observations support the concept that glycolysis shifts toward oxidative phosphorylation as GCs mature within growing antral follicles. They also complement the results of previous studies in which metabolic endpoints were measured primarily using whole follicle cultures and mouse models (Boland et al., 1993(Boland et al., , 1994a(Boland et al., , 1994bGook et al., 2014).
Conceptually, the transition of glucose metabolism as GCs mature and further differentiate in growing antral follicles is becoming clearer, but the rationale as to why this transition occurs is less understandable. Perhaps GCs extensively metabolize glucose glycolytically during early follicle development to support the oocyte that, alone, executes glycolysis inefficiently (Sugiura et al., 2005). Unlike oocytes, GCs produce large quantities of pyruvate and lactate, which in turn serve as ideal metabolic substrates for oocytes (Donahue and Stern, 1968;Leese and Barton, 1985). The metabolic transition in GCs may also be controlled by the demands of the oocyte, wherein direct (Eppig, 1979(Eppig, , 1991 and indirect communication (Sugiura et al., 2005) occurs. In this manner the advancement of follicular growth might be intrinsically regulated, influenced by granulosa cell-oocyte communication as well as the transition in glucose metabolism. Oxidative metabolism favors the capacity of GCs of large antral follicles to produce increased amounts of steroid. Gonadotropins stimulate oxygen uptake in mitochondria of ovarian cells, and this effect accompanies an increase in steroidogenesis (Dimino et al., 1980). Regardless of whether or not the transition in GCs occurs to support the oocyte, is controlled by the oocyte, and/or simply reflects greater capacity of the GCs to differentiate and produce steroid, a similar pivot in glucose metabolism also occurs during in vitro development of murine oocytes and in vitro derived bovine embryos (Harris et al., 2009;Thompson et al., 1996).
In this study, although inhibition of the HBP (via DON) had no adverse effect on glucose metabolism by GCs (i.e., ECAR and OCR), the downstream disruption of O-GlcNAcylation directly (via OSMI-1) provided evidence that GCs from small antral follicles remain reliant on glucose as a primary substrate. An absence of glucose channeled through the HBP to facilitate O-GlcNAcylation had no effect on metabolism, but it did impair the expression of enzymes involved with glycolysis. In other studies, disruption of O-GlcNAc cycling also reduces mitochondrial membrane potential and ATP production (Wang et al., 2022). In the current work, direct inhibition of O-GlcNAcylation (via OSMI-1) impaired the expression of phosphofructokinase, pyruvate kinase, and pyruvate dehydrogenase (Fig. 6) in GCs of both small and large antral follicles. This outcome is consistent with the Wang and coworkers study in which GCs were treated with BADGP, another OGT inhibitor, that inhibited the activity of pyruvate kinase (Wang et al., 2022). Hexokinase expression, a key enzyme for both glucose utilization (via the HBP) and metabolism (via glycolysis), was impaired only in GCs of small antral follicles (Fig. 6). Thus, a connection likely exists between O-GlcNAcylation and the regulation of glycolysis, particularly in GCs of small antral follicles, but it is unclear why inhibition of the HBP and O-GlcNAcylation failed to impact glucose metabolism in the current study. Perhaps stimulating steroidogenesis and stressing metabolic capacity of the GCs of small follicles with gonadotropin (i.e., FSH) and/or growth factor (IGF-1) driven conditions is needed. LH and FSH reportedly increase glucose uptake and glycolysis in granulosa cells (Billig et al., 1983;Boland et al., 1993;Hillier et al., 1985) and IGF-1 mobilizes GLUT 4 (Dupont and Scaramuzzi, 2016). In any event, this area offers fertile ground for future investigation, and at least one study has determined that the glycolytic enzymes, phosphofructokinase, pyruvate kinase, and lactate dehydrogenase, are all modified by O-GlcNAcylation (Wang et al., 2017). Here we provide evidence that some of these same enzymes might be modified by O-GlcNAc and regulated by the degree of O-GlcNAcylation in bovine GCs.

Conclusion
In conclusion, granulosa cells (GCs) from bovine small antral follicles exhibit greater overall glucose metabolism than GCs of large antral follicles and are more reliant on aerobic glycolysis than oxidative phosphorylation. During folliculogenesis, GCs of antral follicles apparently undergo a transition in metabolism primarily from aerobic glycolysis to oxidative phosphorylation as the follicle grows. The reasons for this require further investigation, but it is likely that the transition facilitates the rapid growth and maturation of the follicle. Although inhibition of the hexosamine biosynthesis pathway in this study failed to adversely affect GC metabolism, it appears there are effects downstream (via direct inhibition of O-GlcNAcylation) in which expression of glycolytic enzymes is impaired. Thus, a relationship between O-GlcNAcylation and glycolytic metabolism in bovine GCs is supported (Fig. 8). Future work will help determine whether direct O-GlcNAcylation of glycolytic enzymes impacts the metabolism, function, and fate of the GCs.

CRediT authorship contribution statement
Abigail M. Maucieri: Conceptualization, the study, experimental Colored arrows and text denote pathways and effects of O-GlcNAcylation on glycolytic enzymes, respectively, corresponding to GCs of small antral follicles (GREEN), large antral follicles (RED), or both types of antral follicles (YELLOW). Glucose uptake by both types of follicles is processed similarly through the hexosamine biosynthesis pathway for O-GlcNAcylation (upper LEFT) and the glycolytic pathway (MIDDLE). However, whereas GCs of small follicles metabolize glucose glycolytically (resulting in lactate production), GCs of large follicles further process glucose through oxidative phosphorylation in the mitochondria. Notably, disruption of O-GlcNAcylation affects specific glycolytic enzymes, in a follicle size dependent manner. Glucose 6-phosphate (Glucose-6P), fructose 6-phosphate (Fructose-6P), fructose 1-6, bisphosphate (Fructose-1-6, biP), Glucosamine 6-phosphate (Glucosamine-6P), O-GlcNAc transferase (OGT), O-GlcNAcase (OGA) hexokinase (HK), phosphofructokinase (PFK) pyruvate kinase (PK), lactate dehydrogenase (LDH), pyruvate dehydrogenase (PDH). Created with biorender .com. design, and wrote and revised the manuscript, performed the experiments, Data curation, and conducted the statistical, Formal analysis. David H. Townson: Conceptualization, the study, experimental design, and wrote and revised the manuscript, acquired funding and provided the resources needed to complete the experiments and manuscript. 17.

Declarations of competing interest
None.

Data availability
Data will be made available on request.