Insulin-Stimulated Glucose Uptake Occurs in Specialized Cells within the Cumulus Oocyte Complex

Abstract

The oocyte exists within the mammalian follicle surrounded by somatic cumulus cells. These cumulus cells metabolize the majority of the glucose within the cumulus oocyte complex and provide energy substrates and intermediates such as pyruvate to the oocyte. The insulin receptor is present in cumulus cells and oocytes; however, it is unknown whether insulin-stimulated glucose uptake occurs in either cell type. Insulin-stimulated glucose uptake is thought to be unique to adipocytes, skeletal and cardiac muscle, and the blastocyst. Here, we show for the first time that many of the components required for insulin signaling are present in both cumulus cells and oocytes. We performed a set of experiments on mouse cumulus cells and oocytes and human cumulus cells using the nonmetabolizable glucose analog 2-deoxy-d-glucose to measure basal and insulin-stimulated glucose uptake. We show that insulin-stimulated glucose uptake occurs in both compact and expanded cumulus cells of mice, as well as in human cumulus cells. Oocytes, however, do not display insulin-stimulated glucose uptake. Insulin-stimulated glucose uptake in cumulus cells is mediated through phosphatidylinositol 3-kinase signaling as shown by inhibition of insulin-stimulated glucose uptake and Akt phosphorylation with the specific phosphatidylinositol 3-kinase inhibitor, LY294002. To test the effect of systemic in vivo insulin resistance on insulin sensitivity in the cumulus cell, cumulus cells from high fat-fed, insulin-resistant mice and women with polycystic ovary syndrome were examined. Both sets of cells displayed blunted insulin-stimulated glucose uptake. Our studies identify another tissue that, through a classical insulin-signaling pathway, demonstrates insulin-stimulated glucose uptake. Moreover, these findings suggest insulin resistance occurs in these cells under conditions of systemic insulin resistance.

Mammalian oocytes are surrounded by a layer of specialized granulosa cells called “cumulus cells” that differentiate at the time of antral follicle formation. Cumulus cells differ from the mural granulosa cells that line the follicle in function as well as gene and protein expression (1–3), including genes involved in glycolytic metabolism (4). The cumulus-oocyte-complex (COC) exists within the ovarian follicle from the antral follicle stage until after ovulation when, after the LH surge cumulus cells expand and secrete hyaluronic acid. Bidirectional communication through gap junctions, connexins, and paracrine signaling between the oocyte and surrounding cumulus cells is necessary for normal oocyte development, including oocyte meiotic maturation (5, 6). Additionally, the cumulus cells are particularly important for metabolism within the COC. The cumulus cells metabolize the majority of the glucose within the COC and provide metabolic intermediates to the oocyte, which has a poor capacity to metabolize glucose on its own, and preferentially metabolize pyruvate from the cumulus cells (7–10). It has been shown that oocytes denuded of their surrounding cumulus cells have low glycolytic activity (11) mediated, in part, by low phosphofructokinase activity (12).

Recently, the insulin receptor (IR) was identified in mouse oocytes and cumulus cells (13). Prolonged culture (10 d) of preantral mouse follicles with insulin resulted in phosphorylation of glycogen synthase kinase 3B in the oocyte, indicating that some activation of the insulin-signaling cascade occurs in oocytes (13). Previous work had identified IR mRNA in oocytes of the human, bovine, and rat (14–16); and IR protein in pig (17) and human oocytes (18). Insulin can affect cell growth, apoptosis, and metabolism in a variety of tissues (19); however, insulin-stimulated glucose uptake occurs predominantly in muscle, heart, and fat, as well as the blastocyst-stage embryo (20) and takes only minutes to detect, as opposed to hours as previously measured in the ovarian follicle. The primary metabolic function of insulin is to increase rapidly glucose uptake in the target tissues, primarily skeletal muscle and fat, in the postprandial state. A family of facilitative glucose transporters known as glucose transporters (GLUT) mediates glucose uptake into tissues. There are 14 members of the GLUT family that differ in their substrate specificity, kinetic characteristics, and subcellular distribution (21, 22). Insulin-stimulated glucose uptake into peripheral tissues is mediated through GLUT4 (23). However, the GLUT4 knockout mouse does not develop hyperglycemia (24), and soleus muscle from GLUT4 knockout mice can increase glucose uptake in response to insulin (25), indicating that other GLUT such as GLUT8 (20) and GLUT12 (26, 27) may also be insulin responsive in key target tissues.

As described above, the cumulus cells are responsible for the majority of glucose metabolism within the COC, and both cumulus cells and the oocyte express the IR. It is unknown however, whether either component of the COC is capable of insulin-stimulated glucose uptake. The measurement of lactate accumulation in media of primary cultures of human granulosa cells cultured with or without insulin has only provided indirect evidence that insulin may affect glucose metabolism in these cells (28, 29). Our aim was to determine whether the oocyte, cumulus cells, or both exhibit insulin-stimulated glucose uptake through classical insulin-signaling pathways, and if so, whether peripheral insulin resistance would impair this process. These studies are particularly relevant in light of the increasing incidence of obesity and insulin resistance in women of reproductive age, which appears to negatively impact fertility (30–34). Specifically, the impact of obesity on the oocyte, subsequent embryonic development, and pregnancy outcome is an active area of research (35–39).

Materials and Methods

Animal care and use

All mouse studies were approved by the Animal Studies Committee at Washington University School of Medicine and conform to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health. Female ICR (Harlan, Indianapolis, IN) mice, 3–4 wk, of age were used for all experiments. For high-fat experiments, mice (n = 40) were fed a high-fat diet containing 35.8% fat, 20.7% protein, and 35% carbohydrates (58% Fat energy-A1N-76A; TestDiet, Richmond, IN) from ages 3–7 wk before experiments; age-matched mice (n = 40) were fed standard mouse chow containing 4.8% fat, 73.9% carbohydrate, and 14.8% protein (12% Fat energy; TestDiet). After 4 wk on the diet measurements of fasting weight, blood glucose, and serum insulin were taken, and a glucose tolerance test was performed. Mice had access to food and water ad libitum.

Human cumulus cells

The Washington University Human Research Protection Office and Institutional Review Board approved all human studies. Cumulus cells were obtained from women undergoing infertility treatment at Washington University School of Medicine's Reproductive Endocrinology and Infertility Clinic using standard procedures in preparation of human oocytes for in vitro fertilization. Control cumulus cells were obtained from patients (n = 7) with a male factor-only infertility diagnosis with an average body mass index (BMI) of 27.4 and age of 29.6 ± 1.0 yr. Cumulus cells were also obtained from women with polycystic ovarian syndrome (PCOS) (n = 4) (based on Rotterdam criteria) with an average BMI of 34.8 and age of 28 ± 0.6 yr. Cumulus cells were removed from oocytes approximately 1–3 h after retrieval from the ovaries. One to three oocytes were placed into 200-μl drops of 80 U/ml hyaluronidase (Sigma Chemical Co., St. Louis, MO) in HEPES-buffered human tubal fluid (HEPES-HTF; Irvine Scientific, Santa Ana, CA) supplemented with 10% Synthetic Serum Substitute (SSS; Irvine Scientific). To effect removal, the oocytes underwent rapid pipetting through a 150-μm bore pipette tip (MidAtlantic Diagnostics, Mt. Laurel, NJ) for 10 sec before transfer to 100 μl wash drops of HEPES-HTF + 10% SSS. Rapid pipetting was continued in fresh wash drops until all cumulus cells were removed and the denuded oocytes were transferred to culture for clinical use. The spent drops of hyaluronidase and wash droplets containing the cumulus cells were collected and diluted 1:2 in HEPES-HTF + 10% SSS and pelleted. Samples were either used fresh or washed in PBS and stored at −80 C until use.

Isolation of mouse oocytes and cumulus cells

For compact cumulus cells and germinal vesicle (GV) stage oocytes, mice were injected with 10 IU PMSG (National Hormone and Peptide Program, Torrance, CA) and euthanized 48 h later. Ovaries were isolated and placed under M2 media (Sigma). Follicles were manually punctured and COC were separated and placed in a clean dish of M2 media. Cumulus cells were then removed from isolated COC by manual pipetting using pulled glass pipettes. For collection of expanded cumulus cells, mice were injected with 10 IU PMSG and 48 h later with 10 IU human chorionic gonadotropin (Sigma). Mice were euthanized 13 h after human chorionic gonadotropin injection, and oviducts were dissected out and placed in M2 media. Expanded COC were then isolated by dissecting oviducts and placing COC in a clean dish of M2 media containing 1 mg/ml hyaluronidase (Sigma) for 5 min. Oocytes were removed and cumulus cells were washed twice in PBS before use. Oocytes and cumulus cells were either used fresh or washed in PBS and stored at −80 C before use.

Western blot analysis

Isolated cumulus cells were lysed in 30 μl protein lysis buffer, 50 mm Tris HCl (pH 7.5), 150 mm NaCl, 5 mm EDTA, 0.5% Nonidet P-40, and complete Mini protease inhibitor cocktail (Roche Applied Sciences, Indianapolis, IN). Other tissues used as positive controls (soleus muscle, fat, endometrial stromal cells, and testis) were lysed in 150 μl lysis buffer. Protein concentration was quantified using the BCA Protein Assay Kit (Pierce Chemical Co., Rockford, IL). Denuded GV-stage oocytes were used in all Western blot experiments in 100 oocyte aliquots per lane. After addition of 5% β-mercaptoethanol and Laemmli buffer, whole-cell lysates or oocytes were separated by SDS-PAGE and transferred to nitrocellulose. Nonspecific antibody binding was blocked in 5% nonfat dry milk powder in Tris-buffered saline for 1 h. All blots were then incubated overnight at 4 C in 5% nonfat dry milk or 5% BSA, according to the manufacturer's instructions with the following antibodies: IR (1:200; Santa Cruz Biotechnology, Inc., Santa Cruz, CA), IGF-1R (1:1000; Cell Signaling Technology, Danvers, MA), IR substrate-1 (IRS-1) (1:1000; Cell Signaling Technology), IRS-2 (1:1000; Cell Signaling Technology), Phosphatidyl-inositol-3 kinase (PI3K) p85 subunit (1:000; Cell Signaling Technology), PI3K p110α subunit (1:1000; Cell Signaling Technology), Akt (pan) (1:1000; Cell Signaling Technology), Phospho(p)-Akt (1:1000; Cell Signaling Technology), GLUT1 (1:3000; kindly provided by Dr. Michael Mueckler, Washington University) (40); GLUT4 (1:1000; kindly provided by Dr. Michael Mueckler, Washington University) (40); GLUT8 (1:1000; previously generated in our laboratory) (20); GLUT12 (1:1000; previously generated in our laboratory) (41); β-actin (1:5000; Millipore Corp., Bedford, MA). After primary antibody incubation, blots were incubated with goat-antirabbit IRDye 800 (1:10,000) or goat-antimouse IRDye 680 (1:10,000) (LI-COR Biosciences, Lincoln, NE) for 1 h at room temperature. Membranes were then scanned and analyzed using the Odyssey fluorescent imager (LI-COR Biosciences) and normalized to β-actin.

Insulin-stimulated 2-deoxyglucose (2-DG) uptake assay

For mice, isolated nonfrozen cumulus cells from approximately 10 mice were used per replicate. Human cumulus cells isolated from one patient were used per replicate. Cumulus cells were washed in PBS, centrifuged at 4500 rpm, and resuspended in 50 μl M2 media (Sigma) containing 5.6 mm d-glucose with or without 500 nm bovine insulin (Sigma) for 15 min at 37 C. Cells were then centrifuged and washed once in 50 μl 1 × Krebs-ringer solution (125 mm NaCl, 4.7 mm KCl, 2.5 mm CaCl₂, 2.4 mm MgSO₄, 25 mm NaHCO₃, 1.2 mm K₂HPO₄, and 1% BSA) without glucose, and then incubated in 50 μl 1 × Krebs-ringer solution without glucose for 15 min at 37 C in 5% CO₂. Cells were then centrifuged and resuspended for 1 min with 1 × Krebs-ringer containing 0.1 mm 2-deoxy-d-glucose and 1.2 μm (2 μCi) of [1,2-³H]-2-deoxy-d-glucose ([³H]-2-DG; MP Biomedicals, Solon, OH). The 2-DG uptake reaction was stopped by the addition of 50 μm ice-cold cytochalasin B. Samples were then washed twice in PBS and resuspended in 25 μl protein lysis buffer. A 10-μl sample was used to quantify protein concentration by BCA assay (Pierce, Thermo Scientific, Rockford, IL). The remaining sample was added to 20 ml scintillation fluid and [³H]-2-DG was counted and normalized to protein concentration of each sample. For calculation of nonradioactive 2-DG uptake in oocytes and blastocysts, fresh GV-stage denuded oocytes or expanded blastocysts were collected from 3-wk-old ICR mice, and basal or insulin-stimulated 2-DG uptake was measured in individual oocytes or blastocysts using enzymatic cycling assays previously developed and validated in our laboratory (20, 42). For all 2-DG uptake assays, uptake was calculated as the fold change in insulin-stimulated 2-DG uptake compared with the average basal 2-DG uptake.

Inhibition of PI3K signaling

A set of 2-DG uptake experiments was performed on a separate group of compact cumulus cells and human cumulus cells as described above with the following modifications: a 20-min preincubation period in M2 media containing 250 μm LY294002 (Sigma) or dimethylsulfoxide (DMSO) vehicle control was performed before 15 min treatment in M2 media containing 5.6 mm d-glucose with or without 500 nm insulin and 250 μm LY294002 or DMSO vehicle control. For Western blots of p-Akt and total Akt, compact cumulus cells, human cumulus cells, and GV-stage oocytes were collected and placed in glucose-free Krebs-Ringer solution with 250 μm LY294002 or DMSO vehicle control for 30 min. Samples were then treated for 15 min in Krebs-Ringer with 2.7 mm d-glucose, supplemented with or without 500 nm insulin and 250 μm LY294002 or DMSO vehicle control. Samples were then washed in PBS and frozen at −80 C before use.

Quantitative RT-PCR (qRT-PCR)

All qRT-PCR were performed on aliquots of compact cumulus cells collected from five to 10 mice, or from individual human patients. All oocyte samples were GV-stage oocytes collected in 200-oocyte aliquots. Total RNA was isolated from cumulus cell (n = 5–6) or oocyte samples (n = 5) using the Arcturus PicoPure RNA kit (Applied Biosystems, Foster City, CA). For positive control of GLUT4 mRNA expression, total RNA was isolated from soleus muscle or epigonadal fat (n = 4 each) from 3-wk-old ICR female mice using the RNeasy Mini kit (QIAGEN, Valencia, CA). cDNA was generated from 150–200 ng RNA using the Quantitect RT kit (QIAGEN). Quantitative RT-PCR was performed using the 7500 Fast Real-Time PCR System (Applied Biosystems). Each reaction was run in triplicate and included a control sample with no reverse transcription and consisted of 10 ng cDNA for cumulus cells, soleus muscle, or fat, or 7 ng cDNA for oocytes, 1 × Fast Power SYBR Green PCR System (Applied Biosystems), and 300 nm validated mouse or human primers for GLUT1, GLUT4, GLUT8, GLUT12, or β-actin mRNA (43). Quantifications were performed with standard curves generated with pCR 2.1 TOPO plasmids containing specific GLUT cDNA amplicons using the ΔΔ Ct method as previously described for these primer sets (43). Product specificity was confirmed by melt curve analysis and by electrophoresis and visualization of qRT-PCR products on a 2% agarose gel.

Statistical analysis

Experimental results are shown as means ± se. Results from insulin-stimulated glucose uptake assays were compared using Student's t test.

Results

Insulin-signaling pathway

Many of the insulin-signaling components necessary for insulin-stimulated glucose uptake are found at the protein level in both cumulus cells and oocytes (Fig. 1). All of the proteins found in mouse cumulus cells were also present in human cumulus cells. Most notably, whereas IRS-1 is present in cumulus cells of mice and humans, we did not detect IRS-1 in oocytes; and the inverse was true for IRS-2. The IRS-2 detected in the oocyte appeared to be heavily glycosylated (Fig. 1).

Fig. 1.

Representative Western blots showing components of the insulin-signaling pathway in mouse cumulus cells (cc) and denuded oocytes (do), and in human cumulus cells. Each lane contains a separate sample of either 5 μg of cumulus cell protein or 100 oocytes. Each protein was tested in four separate samples of cc or do and normalized to β-actin. Two samples of each tissue are shown. p85, PI3K-regulatory subunit; p110, PI3K catalytic subunit; Akt, murine thymoma viral oncogene homolog 1.

Insulin-stimulated glucose uptake occurs in mouse cumulus cells but not oocytes

Based on previous evidence of greater glucose metabolism in cumulus cells compared with oocytes and the observed lack of IRS-1 in oocytes in these studies, we hypothesized that if insulin-stimulated glucose uptake was present, it may only occur in cumulus cells and be absent in the oocyte. We came to this conclusion, based on the fact that IRS-1 is usually associated with glucose metabolism and insulin-stimulated glucose uptake, whereas IRS-2 is linked to lipid metabolism (44). We observed consistent increases in glucose uptake as measured by 2-DG after insulin treatment in both compact and expanded cumulus cells (Fig. 2, A and B). This was somewhat surprising because only a handful of tissues display this metabolic property in response to insulin. We measured insulin-stimulated glucose uptake in individual mouse GV-stage oocytes and in the same experiment used mouse blastocyst-stage embryos as a positive control that is known to display insulin-stimulated glucose uptake using this assay (20, 42). We observed no significant change in glucose uptake in the oocytes (Fig. 2C) and an increase in insulin-stimulated glucose uptake in blastocysts (Fig. 2D), similar to the fold change in glucose uptake that we measure in cumulus cells (Fig. 2, A and B).

Fig. 2.

Metabolic assays showing insulin-stimulated glucose uptake occurs in cumulus cells, but not in denuded oocytes in mice. Insulin-stimulated uptake of 1,2-³H-radiolabeled 2-DG uptake measured in compact cumulus cells (A) and expanded cumulus cells (B) of mice. Insulin-stimulated nonradiolabeled 2-DG uptake was also measured in denuded oocytes (C) with blastocyst-stage embryos (D) measured in the same experiment as a positive control for the assay. Uptake for all experiments is expressed as fold change from basal in 2-DG uptake after 500 nm insulin treatment. For panels A and B, values are means ± se of four to five separate experiments; For B and C, values are means ± se of 30–35 individual oocytes or blastocysts. *, P < 0.05 compared with average basal uptake.

Inhibition of PI3K pathway or in vivo insulin resistance blunts insulin-stimulated glucose uptake in cumulus cells

We had observed that both the p85 and p110 subunits of PI3K are present in mouse and human cumulus cells, as well as mouse oocytes (Fig 1). The addition of a specific PI3K inhibitor, LY294002, was able to significantly decrease insulin-stimulated glucose uptake in mouse (Fig. 3A), and in human cumulus cells there was a trend (P < 0.1) for decreased glucose uptake (Fig. 3C), indicating that the canonical PI3K pathway is used in these cells. A downstream target of PI3K is Akt, which is phosphorylated in response to IR activation. We show that the use of LY294002 also prevents Akt phosphorylation by insulin in mouse (Fig. 3B) and human (Fig. 3D) cumulus cells. In contrast, insulin had no effect on Akt phosphorylation in oocytes, whereas LY294002 decreased Akt phosphorylation both with and without insulin treatment (Supplemental Fig. 1 published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org). Mice fed a high fat diet for 4 wk before measurement of insulin-stimulated glucose uptake displayed hyperinsulinemia and impaired glucose tolerance, as well as moderate weight gain but did not become hyperglycemic to the point of being considered diabetic (>250 mg/dl; Supplemental Table 1 and Supplemental Fig. 2). Insulin-stimulated glucose uptake in the cumulus cells of these mice was significantly blunted (Fig. 4A). Similarly, women with PCOS who also had a higher average BMI than control women without PCOS also had cumulus cells that displayed numerically, but not statistically, significant decreased insulin-stimulated glucose uptake (Fig. 4B). Despite hyperinsulinemia, the cumulus cells from mice fed a high fat diet did not display down-regulation of the IR or IGF-1R (Fig. 4C).

Fig. 3.

Insulin-stimulated glucose uptake and Akt phosphorylation are inhibited in mouse and human cumulus cells by LY294002, a specific inhibitor of PI3K activity. Mouse (A) or human (C) cumulus cells were preincubated with either 250 μm LY294002 or DMSO vehicle control before measurement of insulin (500 nm)-stimulated 2-DG uptake. For panels A and C, values are means ± se of four separate experiments. Western blots for phospho-Akt and total Akt in mouse (B) or human (D) cumulus cells preincubated with 250 μm LY294002 or DMSO control before treatment with or without 500 nm insulin. Western blot experiments were replicated three times. *, P < 0.05; †, P < 0.1 compared with fold change in DMSO-treated cells.

Fig. 4.

A high-fat diet in mice (A) or PCOS in women (B) blunts insulin-stimulated 2-DG uptake in cumulus cells. A, Mice were fed a high-fat or control diet for 4 wk before measuring insulin-stimulated 2-DG uptake in cumulus cells. B, Cumulus cells were obtained from women with PCOS or male factor-only infertility diagnosis, and 2-DG uptake was measured. C, Western blots of IR and IGF-1R in cumulus cells of mice fed a high-fat or control diet. Data for 2-DG uptake are means ± se of four to five separate experiments. *, P < 0.05 compared with fold change in control cells. cc, Cumulus cells; Con, control; HF, high fat.

Insulin-stimulated glucose uptake in cumulus cells occurs without GLUT4

Quantitative real-time PCR and Western blotting were used to determine which GLUT may be mediating glucose uptake in oocytes and cumulus cells. The ubiquitously present GLUT1, as well as the putative insulin-sensitive GLUT4, GLUT8, and GLUT12 were analyzed in cumulus cells as well as oocytes. Despite the observed increase in glucose uptake after insulin treatment (Fig. 2), GLUT4 mRNA was not detected in cumulus cells of mice (Fig. 5A) or humans (Fig. 5B) or in mouse oocytes (Fig. 5C). Due to this lack of detection, we used these GLUT4 primers on tissues known to have abundant GLUT4 mRNA expression, soleus muscle, and fat. In these tissues we observed robust GLUT4 mRNA expression and no GLUT4 mRNA in cumulus cells (Fig. 5D) when loading equal amounts of cDNA in the reaction and observing no difference in β-actin mRNA across these three tissues. Furthermore, we did not observe GLUT4 protein in mouse cumulus cells or oocytes and show soleus muscle as a positive control for our GLUT4 antibody (Fig. 5E). After GLUT1, the most abundantly expressed GLUT in cumulus cells were GLUT8 and GLUT12, and this pattern of gene expression was similar in both mice and human cumulus cells (Fig. 5, A and B). Oocytes also expressed high concentrations of GLUT1 mRNA; however, here GLUT12 mRNA was more abundant than GLUT8 mRNA, and total mRNA expression of any of these transcripts was more than 2-fold lower than expression in cumulus cells (Fig. 5C). Western blotting with a positive control for each antibody showed the presence of GLUT1, 8, and 12 in mouse cumulus cells, recapitulating what was observed from qRT-PCR (Fig. 5E). Oocyte Western blots showed GLUT1 and GLUT12 to be present, similar to the qRT-PCR data. However, a band around the size of GLUT8 was present in the oocyte at oversaturated concentrations and did not appear as a highly glycosylated smear typical of GLUT (41) and was not similar to the positive control. Based on this and the relatively low abundance of GLUT8 mRNA we do not believe this band in the oocyte lane to be GLUT8, but rather nonspecific binding.

Fig. 5.

qRT-PCR and Western blot analysis of GLUT1, 4, 8, and 12 in cumulus cells of mice and humans, and oocytes of mice. qRT-PCR was performed in mouse (A) and human (B) cumulus cells; as well as in mouse denuded oocytes (C). An additional GLUT4 qRT-PCR was conducted with cumulus using soleus muscle and fat as positive controls (D) for the GLUT4 primers. For qRT-PCR, values are means ± se for five separate samples. Western blots (E) of cumulus cells (cc) and denuded oocytes (do) with the following positive control samples (+): GLUT1, endometrial stromal cells (esc); GLUT4, soleus muscle; GLUT8, testes from transgenic mice that overexpress GLUT8; GLUT12, soleus muscle from transgenic mice that overexpress GLUT12. All samples were normalized to β-actin.

Discussion

These are the first studies to show that the cumulus cells that surround the mammalian oocyte display insulin-stimulated glucose uptake. This is a unique finding; although the mitogenic and antiapoptotic effects of insulin are seen in many cells, the metabolic effect of increased rapid glucose uptake is primarily observed only in muscle and adipose tissue. A notable exception is the blastocyst stage embryo that also displays insulin-stimulated glucose uptake (20). Insulin can signal through its own receptor as well as the IGF-1R, although at a lower affinity (45). We detected both the IR and the IGF-1R in cumulus cells and in oocytes. Previous work had also detected the IR in mouse cumulus cells and oocytes (13) and the IGF-1R in granulosa cells of mice and human follicles (46). In both these studies, however, IR and IGF-1R stimulation was only tested after chronic exposure, not acute activation. The activated IR can tyrosine phosphorylate adaptor proteins such as the IRS family, which recruit and phosphorylate downstream effectors. There are four IRS isoforms, but IRS1 and IRS2 are the most important to glucose metabolism (19) and were studied here, and interestingly showed cell-specific expression. We saw IRS1 only in cumulus cells of mice and humans, and IRS2 only in mouse oocytes. However, others have detected IRS-2 mRNA in cumulus cells of women (47) as well as IRS-1 mRNA and protein in rat granulosa cells (48). It is known that IRS1 and IRS2 can display distinct functions in different cell types (19, 49). Our data would indicate that IRS1 has a greater role in insulin-stimulated glucose uptake in the cumulus cells because IRS2 was not detected.

PI3K is a downstream target of IRS proteins, and signaling through this pathway is required for translocation of GLUT to the plasma membrane and insulin-stimulated glucose uptake to occur (23, 50). Treatment of adipose or muscle in vivo and in vitro with the selective inhibitor of PI3K, LY294002, blocks insulin-stimulated glucose uptake and GLUT4 translocation to the cell surface (50–53), whereas expression of constitutively active PI3K stimulates insulin-stimulated glucose uptake (54, 55). PI3K consists of the regulatory p85 and catalytic p110 subunits. We show that both the p85 and p110 subunits are also present in oocytes and cumulus cells, and that blocking this insulin-signaling pathway with LY294002 inhibits insulin-stimulated glucose uptake in cumulus cells of both mice and humans. Similar results are also reported in the blastocyst stage embryo (56, 57). Akt phosphorylation by PI3K is required for insulin-stimulated glucose uptake (23, 58). In our study, Akt phosphorylation by insulin in cumulus cells could be inhibited by the addition of LY294002, showing that insulin signaling through this pathway occurs. We also observed a decrease in Akt phosphorylation in denuded oocytes after treatment with LY294002; however, because insulin had no effect on glucose uptake in the oocyte, we did not measure the effect of PI3K inhibition on glucose uptake in the oocyte. The presence of Akt and glycogen synthase kinase 3A/B was previously reported in mouse oocytes (13); however, Akt phosphorylation in response to insulin was not assessed. During oocyte meiotic maturation Akt mRNA increases and is localized along the meiotic spindle (59). Treatment with LY294002 appears to alter localization of phosphorylated Akt during oocyte maturation and impairs polar body extrusion (60). The effects of LY294002 on Akt in the oocyte could be due to FSH signaling as opposed to insulin.

Both insulin and FSH use components of the PI3K pathway, and a number of studies have investigated the effect of FSH on COC metabolism. In mice, FSH-induced maturation of COC is also associated with increased glucose uptake and lactate production (61). Use of LY294002 inhibited FSH-induced glucose uptake and lactate production in culture media (61). The interaction of gonadotropins and cumulus cell metabolism has been noted by others that reported increased glucose consumption in cumulus cells (62, 63) and increased hexokinase activity in cumulus cells after FSH treatment (64). The impact of insulin on glucose metabolism has been noted in primary cultures of human granulosa cells from normal women or women with PCOS. Two studies reported that glucose uptake and lactate production in culture media were increased by insulin in a dose-dependent fashion (28, 29) and that granulosa cells from women with PCOS who were also anovulatory (28), or had peripheral insulin resistance (29), had decreased lactate production in response to insulin.

Differences between the above-mentioned studies and ours are noteworthy. First, these studies measured glucose consumption by COC or cultured granulosa cells over a 15- to 48-h time period, and second, they used d-glucose or assays for d-glucose or lactate production in the media (61–63). Glucose is continually metabolized as it enters the cell and thus cannot be used to accurately measure uptake or transport alone. In our studies we use the glucose analog 2-DG, which enters the cell, and then is phosphorylated by hexokinase, but not further metabolized, thus can accurately measure transport and not subsequent steps in glucose metabolism. However, glucose transport into the cell slows or stops once intracellular glucose reaches a certain concentration (65). In our study, we used a relatively low concentration of 2-DG in our studies (0.1 mm 2-DG and 1.2 μm [³H]2-DG) and a short time period (1 min) to measure uptake during the linear phase and avoid saturating the intracellular concentration of glucose. Similarly, in isolated adipocytes, insulin-stimulated 2-DG uptake is measured over 1 min (66, 67), and in blastocyst-stage embryos over 10 min (65).

Glucose uptake is mediated through GLUT, and insulin-stimulated glucose uptake in muscle and fat is mediated through GLUT4 (23). At the transcript level, GLUT4 mRNA has been detected in granulosa cells of the human (47), rat (48), sheep, and cattle (68, 69). Whereas GLUT4 has been detected in mouse granulosa cells by immunohistochemistry (61) and in rat granulosa cells by Western blot (48), others failed to detect GLUT4 in rat granulosa cells by Western blot (70). Here, we used qRT-PCR, with validated primers that run at a high efficiency in the qRT-PCR (43) with a positive control tissue for these primers, as well as Western blotting with an established antibody (71) and positive control for that antibody. We did not detect GLUT4 mRNA or GLUT4 protein in cumulus cells or oocytes. There are differences in function and gene expression between cumulus and granulosa cells (1–3), including differences in genes involved in glucose metabolism (4), which may explain some of the differences compared with studies that used only granulosa cells. Additionally, the two studies that did detect GLUT4 protein (48, 61) used an antibody different from the one used in our study. One study that did use qRT-PCR to detect GLUT4 mRNA failed to span an intron (48) and thus could have detecting genomic DNA, not mRNA. Based on these observations, insulin-stimulated glucose uptake in cumulus cells is likely mediated through a GLUT other than GLUT4. Despite a lack of GLUT4 in cumulus cells, earlier studies from the GLUT4 knockout mouse indicated that other insulin-sensitive GLUT may exist (24, 25). Indeed, previous work in our laboratory has shown that GLUT8 can mediate insulin-stimulated glucose uptake in blastocyst-stage embryos (20), and recent work from our laboratory and others indicates that GLUT12 may also be an insulin sensitive glucose transporter (26, 27). We show that both GLUT8 and GLUT12 are present in cumulus cells of mice and humans and may be responsible for insulin-stimulated glucose uptake in these cells.

Insulin-stimulated glucose uptake into peripheral tissues is impacted by obesity and insulin resistance. Maternal obesity is increasing and affects almost a quarter of US women (31). A recent study examining 43,163 embryo transfers from infertility clinics reported that obesity was associated with a significant decrease in clinical pregnancy rate with the use of autologous, but not donor, oocytes, indicating that influence of the oocyte alone can impact pregnancy outcome (34). Our studies show that peripheral insulin sensitivity also has metabolic consequences for the cumulus cells surrounding the oocyte. We fed mice a high fat diet for 4 wk, which resulted in moderate weight gain, impaired glucose tolerance, and hyperinsulinemia. This diet caused a reduction in insulin-stimulated glucose uptake in the cumulus cells from these mice. In women, PCOS is a common metabolic disorder associated with insulin resistance, and anovulation (35, 72). Women with PCOS are not always obese; however, obesity is more common (73), and the BMI of women with PCOS was greater than the BMI of control women in our study (34.8 vs. 27.4, respectively) and consistent with obesity. We did not observe significantly decreased insulin-stimulated glucose uptake in cumulus cells of women with PCOS compared with controls; however, our number of women with PCOS used for this study (n = 4) was relatively low. We also did not observe down-regulation of the IR or the IGF-1R in mice fed a high fat diet; others have observed IR expression is down-regulated in granulosa cells of women with PCOS who were also insulin resistant (29). These results from our high fat-fed mice indicate that the same conditions that lead to peripheral insulin resistance can also impair insulin-stimulated glucose utilization in the cumulus cells. Alternatively, the blunted response to insulin in the cumulus cells of high fat-fed mice could be an adaptation to protect the germ line from elevated insulin concentrations.

In conclusion, these studies identify another tissue that through a classical insulin-signaling pathway demonstrates insulin-stimulated glucose uptake, albeit through a nonclassical insulin-sensitive glucose transporter. This metabolic action of insulin was shown in both a mouse model and in human cells. Moreover, under in vivo conditions of insulin resistance and/or obesity, this insulin sensitivity in the cumulus cells is impaired. The cumulus cells are critically important to oocyte growth, maturation, and metabolism, and thus the oocyte's ability to develop into a viable embryo. Possibly, the insulin responsiveness of cumulus cells could be used as a biomarker of oocyte competence in clinical assisted reproduction. Obesity-associated oocyte abnormalities may be due, in part, to this impaired insulin action in the cumulus cells. Future studies on the mechanisms of insulin resistance in cumulus cells will be beneficial.