CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

Developmental Biology 279 (2005) 20–30 www.elsevier.com/locate/ydbio

Oocyte control of metabolic cooperativity between oocytes and companion granulosa cells: energy

Koji Sugiura, Frank L. Pendola, John J. Eppig*

The Jackson Laboratory, 600 Main Street, Bar Harbor, Maine 04609, USA

Received for publication 24 September 2004, revised 4 November 2004, accepted 22 November 2004 Available online 15 December 2004

Abstract

Intercellular communication between oocytes and granulosa cells is essential for normal follicular differentiation and oocyte development. Subtraction hybridization was used to identify more highly expressed in cumulus cells than in mural granulosa cells of mouse antral follicles. This screen identified six genes involved in : Eno1, Pkm2, Tpi, Aldoa, Ldh1, and Pfkp. When oocytes were microsurgically removed from cumulus cell–oocyte complexes, the isolated cumulus cells exhibited decreased expression levels of genes encoding glycolytic , glycolysis and activity of the tricarboxylic acid (TCA) cycle. These decreases were prevented by culturing the cumulus cells with paracrine factors secreted by fully grown oocytes. Paracrine factors from fully grown oocytes exhibited greater ability than those from growing oocytes to promote expression of genes encoding glycolytic enzymes and glycolysis in the granulosa cells of preantral follicles. However, neither fully grown nor growing oocytes secreted paracrine factors affecting activity of the TCA cycle. These results indicate that oocytes regulate glycolysis and the TCA cycle in granulosa cells in a manner specific to the population of granulosa cells and to the stage of growth and development of the oocyte. Oocytes control glycolysis in granulosa cells by regulating expression levels of genes encoding glycolytic enzymes. Therefore, mouse oocytes control the intercellular metabolic cooperativity between cumulus cells and oocytes needed for energy production by granulosa cells and required for oocyte and follicular development. D 2004 Elsevier Inc. All rights reserved.

Keywords: Oocyte; Follicular development; Glycolysis; Tricarboxylic acid cycle; Granulosa cells; Phosphofructokinase; Oocyte-granulosa cell communication; Mouse

Introduction antral follicles than in cumulus cells (Joyce et al., 1999). Moreover, Lhcgr mRNA, encoding the luteinizing hormone Granulosa cells of preantral ovarian follicles (preantral receptor, is expressed at a higher level in mural granulosa granulosa cells) differentiate into two spatially and func- cells compared with those in cumulus or preantral granulosa tionally distinct populations during the preantral to antral cells (Camp et al., 1991; Peng et al., 1991). Differences in transition: cumulus cells, which are associated with the expression among the various granulosa cell popula- oocyte, and mural granulosa cells, which line the follicular tions likely reflect different roles of granulosa cell types on wall. While these populations are similar in many respects, the development of oocytes and follicles. Cumulus cells are there are differences in both transcripts and proteins probably specialized to support oocyte development and produced (Latham et al., 1999, 2004). For example, levels play an essential role in oocyte growth and metabolism by of Kitl mRNA, encoding the kit ligand, are higher in providing nutrients such as amino acids (Colonna and preantral granulosa cells and the mural granulosa cells of Mangia, 1983; Haghighat and Van Winkle, 1990) and substrates for energy production (Donahue and Stern, 1968; Leese and Barton, 1985). * Corresponding author. Fax: +1 207 288 6073. Communication between the granulosa cells and the E-mail address: [email protected] (J.J. Eppig). oocytes is bi-directional and essential for the development

0012-1606/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2004.11.027

YDBIO-01808; No. of pages: 11; 4C: 5 K. Sugiura et al. / Developmental Biology 279 (2005) 20–30 21 of both granulosa cells and oocytes (Eppig, 2001). Oocyte- oocyte complexes either from antral follicles of 22-day-old secreted factors affect the patterns of in the primed mice or from preantral follicles of 12-day-old mice various granulosa cell populations. For example, fully to collect either cumulus cells or preantral granulosa cells, grown oocytes suppress cumulus cell expression of Kitl respectively. These granulosa cell complexes are referred as and Lhcgr mRNA (Eppig et al., 1997; Joyce et al., 1999, oocytectomized complexes (OOXs) in this study. Mural 2000). The oocyte-derived factor suppressing both Lhcgr granulosa cells were isolated from antral follicles of 22-day- and Kitl mRNA levels is probably growth differentiation old mice as reported previously (Joyce et al., 1999). factor 9 (GDF9) (Elvin et al., 1999; Joyce et al., 1999, 2000). It is reasonable to hypothesize that other genes Construction and screening cumulus cell differentially expressed among cumulus cells, mural and transcript-enriched subtracted library preantral granulosa cells are regulated by oocyte-derived factors. Since cumulus cells are in closest contact with For construction of the library, mRNA was prepared from oocytes, and substantial evidence indicates their essential cumulus cells from 200 COCs, mural granulosa cells role during oocyte development (Eppig, 1991), genes up- roughly equivalent to the amount of cumulus cells or regulated in cumulus cells probably have an important role preantral granulosa cells from 350 preantral follicles, using in oocyte development. the Micro-FastTrack 2.0 Kit (Invitrogen, Carlsbad, CA) The overall objective of this project was to identify key modified as described previously (Eppig et al., 1997). The transcripts and processes occurring in cumulus cells that mRNA was converted into cDNA using the SMART PCR promote oocyte development. To this end, a library was cDNA Synthesis Kit (Clontech, Palo Alto, CA). Then, made using subtraction hybridization with cumulus cell subtraction was performed with the PCR-Select cDNA cDNA as the tester and a mixture of mural and preantral Subtraction kit (Clontech) according to manufacturer’s cDNA as the driver, thus enriching for cDNA derived from instructions. Cumulus cell transcript-enriched subtracted transcripts expressed more highly in cumulus cells than in cDNA and mural/preantral granulosa cell transcript- the other cell types. This revealed several transcripts enriched subtracted cDNA were generated by forward and encoding enzymes that drive the glycolytic pathway. reverse subtractions. Cumulus cell transcript-enriched sub- Oocytes carry out glycolysis poorly and they require that tracted cDNAs were cloned into a pCRII vector (Invitrogen) granulosa cells provide them with the products of glycolysis using TA Cloning Kit Dual Promoter (pCRII) (Invitrogen) for development (Biggers et al., 1967; Donahue and Stern, to generate a cumulus cells transcript-enriched subtracted 1968; Leese and Barton, 1985). The possible role of oocytes cDNA library, and differential screening was carried out in regulating the steady-state levels of these transcripts, and using the PCR-select differential screening kit (Clontech) also in controlling glycolytic and oxidative according to the manufacturer’s protocol. activity by granulosa cells was therefore evaluated. The results demonstrate that mouse oocytes regulate the inter- In situ hybridization cellular metabolic cooperativity between cumulus cells and oocytes needed for follicular energy production and required Clones from the library were digested with appropriate for oocyte development. restriction enzymes and were used for in vitro transcription to make 33P CTP-labeled riboprobes. In situ hybridization was carried out as described previously (Joyce et al., 2001) Materials and methods using 22-day-old primed mouse ovaries.

Isolation of oocytes and granulosa cells RNase protection assay

All mice were bred and raised in the research colony of Messenger RNAs were isolated as indicated above. the authors at The Jackson Laboratory. Cumulus cell–oocyte RNase protection assay was carried out using RPA-III kit complexes (COCs) were collected from antral follicles of (Ambion, Austin, TX) as described previously (Eppig et al., 22-day-old (C57BL/6J Â SJL/J) F1 mice that had been 1997). Riboprobes were transcribed from cDNA inserted primed with 5 IU equine chorionic gonadotropin (eCG) 44 into pCRII vector (Invitrogen), which were previously to 48 h earlier. Fully grown, meiotically competent oocytes digested with appropriate restriction enzymes. Sizes of were isolated by gentle pipetting of COCs. Growing, protected RNA fragments of Eno1, Rpl19, Pkm2, Tpi, meiotically incompetent oocytes were collected from the Aldoa, Ldh1, and Pfkp were 434 b, 386 b, 325 b, 302 b, 273 preantral follicles of 12-day-old mice by collagenase b, 248 b, and 215 b, respectively. The protected RNA digestion as described previously (Eppig, 1976; Eppig and fragments were analyzed by electrophoresis using a 6% Schroeder, 1989). urea-polyacrylamide gel. Then, the gels were dried and Cumulus cells and preantral granulosa cells were exposed to Fuji phosphor imaging plates and quantified collected as previously reported (Joyce et al., 1999). Briefly, using the Fuji Phosphor Imaging system (Fuji Medical oocytes were surgically removed from granulosa cells– Systems USA, Stamford, CT). Data are expressed as the 22 K. Sugiura et al. / Developmental Biology 279 (2005) 20–30 relative level of mRNA normalized with mRNA encoded by the granulosa cells were picked up in 2.5 Al of the final wash a housekeeping gene, Rpl19. medium and placed in the cap of a sterile 1.7 ml screw-cap microtube with 2.5 Al of the HEPES-WM containing one of Culturing oocytes and granulosa cells radiolabeled substrates (total culture volume of 5 Al). The cap was quickly and tightly fitted onto the tube containing The culture medium used was bicarbonate-buffered 1.5 ml of 25 mM NaHCO3, which had been equilibrated MEMa (Life Technologies, Inc., Grand Island, NY) with with a gas mixture of 5% O2,5%CO2, and 90% N2 at 378C. Earle’s salts, supplemented with 75 mg/l penicillin G, 50 For each experiment, three sham tubes were set up (5 Al mg/l streptomycin sulfate, 0.23 mm pyruvate, and 3 mg/ml drop of medium with no cells) to determine the non-specific crystallized lyophilized BSA. To maintain oocytes at the GV exchange of radiolabel between the incubation drop and the stage, 1, 6-dihydro-2-methyl-6-oxo-3, 4-bipyridine-5-car- NaHCO3. Three total counts were also set up per experiment bonitrile (milrinone) was added at the final concentration of to measure total radioactivity (2.5 AlofHEPES-WM 10 AM to maintain oocytes at the germinal vesicle stage. containing one of radiolabeled substrates was added directly Milrinone was added to all culture medium even if no to 1.5 ml of 25 mM NaHCO3 in the microcentrifuge tube). oocytes were present, or if the oocytes were incompetent to Twenty OOXs or 5 OOXs were used to estimate the ability undergo spontaneous maturation, for experimental balance. of granulosa cells to metabolize either glucose or pyruvate, All medium components were purchased from Sigma (St. respectively. After 3 h of culture at 378C, the NaHCO3 Louis, MO). In some experiments, granulosa cells were co- solution was removed from each tube and added to a 15-ml cultured with fully grown oocytes or growing oocytes at the scintillation vial and 10 ml of scintillation fluid was added concentration of either 2 oocytes/Al or 4 oocytes/Alof to each vial (ASC II, Amersham Biosciences). For detection 14 culture medium, respectively. These relative concentrations of CO2, 200 Al of 0.1 M NaOH was added to the vials, and were based on comparable oocyte volume (Chesnel and the vials were stored for at least 18 h at 48C allow Eppig, 1995). All cultures were performed under washed conversion of the dissolved CO2 and bicarbonate into mineral oil in 35-mm Petri dishes in 10–100 Al volume of carbonate before adding scintillation fluid. Disintegrations culture medium for 16 to 26 h, as indicated in the text, and per minute (dpm) of each vial were measured by a liquid were maintained at 378C in a modular incubation chamber scintillation counter. Because the present experiments were (Billups Rothenberg, Del Mar, CA) infused with 5% O2,5% performed only to compare the activities among the differ- CO2, and 90% N2. ent granulosa cell groups, controls to determine the efficiency of recovery were not conducted. Therefore, the Estimation of metabolic activity estimates of activity presented here are probably lower than actual values. A modified hanging-drop technique was used to estimate metabolic activity (O’Fallon and Wright, 1986; Rieger et al., Statistical analysis 1992). In this assay system, cells were cultured in a small volume as an inverted hanging-drop under the inside surface All experiments were preformed at least three times. For of a screw-cap of microtube, which contained bicarbonate paired comparison, Student’s t test was used. A P value 14 buffered solution. CO2 produced by oxidative phosphor- b0.05 was considered statistically significant. ylation derived from 14C-labeled substrates is rapidly taken up by the bicarbonate-buffered bottom solution. Because the hanging drop and the bottom solution are in very close Results proximity, a rapid exchange of H2O between two solutions also occurs. Thus, the system provides an effective way to Up-regulated expression of genes encoding glycolytic 3 3 collect and measure H2O produced by glycolysis when H- enzymes in cumulus cells labeled substrates were used. Medium used for the hanging drop was Whitten’s medium (Whitten, 1971) supplemented Six differentially expressed genes encoding glycolytic with 20 mM HEPES (HEPES-WM) containing one of the enzymes were identified from the screen of the cumulus radiolabeled substrates. Final concentrations of glucose and cells transcript-enriched subtracted library (Table 1). The pyruvate in the hanging drop were 5.6 mM and 0.23 mM, roles of these respective proteins in glycolysis are shown in respectively, and specific activities of [5-3H]-glucose Fig. 1. In situ hybridization showed that mRNA levels of (Amersham Biosciences, Piscataway, NJ), [U-14C]-glucose these genes were up-regulated in cumulus cells compared (Amersham Biosciences) and [2-14C]-pyruvate (Perkin with preantral or mural granulosa cells in ovaries of eCG Elmer, Wellesley, MA) were 18.0, 3.6, and 0.22 mCi/mmol, primed 22-day-old mice (Fig. 2). Transcripts encoded by respectively. Pfkp, Tpi, Eno1 and Pkm2 were also localized in theca cells, Before being cultured in the hanging-drop, all oocytes while those encoded by Aldoa and Ldh1 were not. Pfkp were removed from the groups that still had them and mRNA was detected at elevated levels in mural granulosa granulosa cells were washed twice with HEPES-WM. Then, cells closer to the antrum. The localization of mRNAs K. Sugiura et al. / Developmental Biology 279 (2005) 20–30 23

Table 1 RNase protection assays were used to assess relative Glycolytic genes found in the cumulus cell transcript-enriched subtracted steady-state levels of expression. As shown in Fig. 3, library mRNA levels were higher in cumulus cells compared with Name Symbol GenBank preantral granulosa cells collected from preantral follicles of accession no. commission no. 12-day-old mice, or mural granulosa cells, collected from Aldolase 1, Aldoa BC050896 4.1.2.13 antral follicles of either eCG-primed or unprimed 22-day- A isoform 1, alpha Eno1 BC024644 4.2.1.11 old mice. non-neuron Lactate Ldh1 NM_010699 1.1.1.27 Fully grown oocytes stimulate the expression of glycolytic dehydrogenase 1, genes in mural granulosa cells A chain Phosphofructokinase, Pfkp BC006926 2.7.1.11 platelet Next, the role of the oocyte in establishing these , Pkm2 BC016619 2.7.1.40 developmentally and spatially restricted patterns of gene muscle expression in granulosa cells was determined. Mural Triosephosphate Tpi BC046761 5.3.1.1 granulosa cells were co-cultured with fully grown GV-stage isomerase oocytes (2 oocytes/Al) for 16 h, and the steady-state mRNA levels encoded by these genes were assessed (Fig. 4). encoding Aldoa, Tpi, Eno1, Pkm2, and Ldh1 mRNAs in Controls were cultured without oocytes. Expression of the mural cells was greatest near the antrum and decreased in a genes encoding glycolytic enzymes was higher in mural centripetal gradient away from the antrum. granulosa cells co-cultured with fully grown oocytes than

Fig. 1. The metabolic pathway of glycolysis. The enzymes encoded by the genes studied here are enclosed by a square. Gene symbols are shown in parentheses. 24 K. Sugiura et al. / Developmental Biology 279 (2005) 20–30

Fig. 2. Localization of mRNA species found enriched by subtraction hybridization as detected by in situ hybridization using 22-day-old eCG-primed mouse ovaries. Upper and lower panels represent bright- and dark-field micrographs, respectively. Black and white arrowheads note antral follicles and preantral follicles, respectively. without oocytes. The level of Aldoa mRNA in mural co-culturing OOX cumulus cells with fully grown oocytes granulosa cells was not significantly increased by co-culture maintained high steady-state mRNA levels of these tran- with fully grown oocytes. scripts (Fig. 5, OOX/CC + FGO). Although there were no statistical differences, Tpi mRNA level in OOX cumulus Fully grown oocytes are required for the elevated expression cells also tended to decrease when compared with those in of glycolytic genes in cumulus cells cumulus cells cultured as intact COCs (P = 0.062) or OOX cumulus cells co-cultured with fully grown oocytes (P = The effect of oocytes on the expression of mRNAs 0.128). When OOX cumulus cells were cultured 16 h, encoding glycolytic enzymes in cumulus cells was deter- instead of 26 h, only Ldh1 and Pfkp mRNA were mined (Fig. 5). When OOX cumulus cells were cultured for significantly decreased compared with cumulus cells cul- 26 h (Fig. 5, OOX/CC), their levels of Eno1, Pkm2, Ldh1, tured as COCs or OOX + FGO (data not shown). Therefore, and Pfkp mRNA were decreased compared with cumulus paracrine factors secreted by fully grown oocytes are cells cultured as intact COCs (Fig. 5, COC). Furthermore, required for the higher steady-state levels of Eno1, Pkm2, Ldh1, and Pfkp mRNA in cumulus cells, and expression levels of Ldh1 and Pfkp are more sensitive to the factors.

Fully grown oocytes are required for elevation of both glycolytic and the TCA cycle activity by cumulus cells

Assays were conducted to assess whether these changes in steady-state mRNA levels affect glycolytic or the TCA cycle activity in cumulus cells. Glycolytic activity was estimated as the ability to metabolize [5-3H]-glucose to 3 14 H2O. The ability of cumulus cells to metabolize [2- C]- 14 14 pyruvate and [U- C]-glucose to CO2 was assessed as indicative of the levels of activity of both the TCA cycle and total glucose oxidation, respectively (Fig. 6B). Intact COCs, OOX cumulus cells (OOX/CC), and OOX cumulus cells co- Fig. 3. Steady-state mRNA levels of genes encoding glycolytic enzymes cultured with fully grown oocytes (OOX/CC + FGO) were during ovarian follicular development. Preantral granulosa cells (PG) were initially cultured for 16 h. Then, all oocytes were removed isolated from preantral follicles of ovaries of 12-day-old mice. Mural from the groups that still had them and the metabolism of granulosa cells (MG) were isolated from either early antral follicles of 22- day-old mice or the Graafian follicles of 22-day-old eCG-primed mice. the cumulus cells derived from these groups was estimated Cumulus cells (CC) were isolated from the Graafian follicles of 22-day-old over a 3-h incubation period in a hanging drop system (Fig. mice after eCG priming. 6A). As shown in Fig. 6C, cumulus cells initially cultured as K. Sugiura et al. / Developmental Biology 279 (2005) 20–30 25

Fig. 4. Effect of paracrine factors secreted by oocytes on the steady-state levels of mRNA encoding glycolytic enzymes in mural granulosa cells. Mural granulosa cells were cultured with (MG + FGO) or without (MG) fully grown oocytes and the levels of mRNA encoding glycolytic enzymes were detected by RNase protection assays. (Left) A representative experiment is shown. (Right) Quantified mRNA levels of genes encoding glycolytic enzymes in mural granulosa cells cultured alone (black bars) or co-cultured with fully grown oocytes (white bars) after normalization to Rpl19 transcripts. The values shown are the mean (FSEM) values from three experimental replicates. *P b 0.05; **P b 0.01. intact COCs (Fig. 6C, COC) exhibited greater activity to factors secreted by fully grown oocytes promote both 3 3 metabolize [5- H]-glucose to H2O versus cumulus cells glycolysis and the TCA cycle in cumulus cells. cultured in the absence of oocytes (Fig. 6C, OOX/CC). This decrease was prevented when OOX cumulus cells were co- Oocyte regulation of glycolysis and the TCA cycle is cultured with fully grown oocytes (Fig. 6C, OOX/CC+ developmentally controlled FGO). The same result was obtained upon estimation of the activity of the TCA cycle using the conversion of [2-14C]- As shown above, the steady-state levels of mRNAs 14 pyruvate to CO2 (Figs. 6D, E). Therefore, paracrine encoding glycolytic enzymes in cumulus cells are main-

Fig. 5. Effect of oocytectomy (OOX) on the steady-state mRNA levels of genes encoding glycolytic enzymes in cumulus cells. Cumulus cells were cultured either as intact COCs (COC), oocytectomized complexes (OOX/CC) or OOX cumulus cells co-cultured with fully grown oocytes (OOX/CC + FGO). The levels of mRNA encoding glycolytic enzymes were detected by RNase protection assays. (Left) A representative experiment is shown. (Right) Quantified mRNA levels of genes encoding glycolytic enzymes in cumulus cells cultured as intact COCs (black bars), OOX cumulus cells (hatched bars) or OOX cumulus cells co-cultured with fully grown oocytes (white bars) after normalization to Rpl19 transcripts. The values shown are the mean (FSEM) values from three experimental replicates. *P b 0.05. 26 K. Sugiura et al. / Developmental Biology 279 (2005) 20–30

Fig. 6. Effect of paracrine factors secreted by fully grown oocytes on metabolic activities of cumulus cells. (A) Cumulus cells were cultured initially either as intact COCs (COC), oocytectomized complexes (OOX/CC) or OOX cumulus cells co-cultured with fully grown oocytes (OOX/CC + FGO), and their glycolytic activity (C), the TCA cycle activity (D) or amount of total glucose oxidation (E) were estimated as described in Materials and Methods. (B) 3 3 Metabolic routes for release of radiolabeled CO2 and H2O from radiolabeled substrates. H2O is released after metabolism of [5- H]-glucose through 14 14 14 14 glycolysis. Production of CO2 from [2- C]-pyruvate represents the activity of the TCA cycle whereas production of CO2 from [U- C]-glucose represents total glucose oxidation. The values shown are the mean (FSEM) values from at least three experimental replicates. **P b 0.01. tained by paracrine factors produced by fully grown Thus, either the growing oocytes of preantral follicles do not oocytes. Levels of these mRNAs expressed by preantral promote the expression of mRNAs encoding glycolytic granulosa cells are significantly lower, as demonstrated by enzymes above basal levels, or the preantral granulosa cells in situ hybridization, despite the presence of a growing are unable to respond to the paracrine factors produced by oocyte. The oocytes in the secondary (preantral) follicles these oocytes. To assess these issues of developmental and found in the ovaries of 12-day-old mice are about half the functional competence, cumulus cells or preantral granulosa volume of fully grown oocytes (Chesnel and Eppig, 1995). cells were co-cultured with either fully grown oocytes or K. Sugiura et al. / Developmental Biology 279 (2005) 20–30 27 growing oocytes, and glycolysis was estimated using the factors secreted from oocytes, but that only fully grown experimental design described above and in Fig. 7A. Both oocytes appear able to promote higher steady-state levels of cumulus cells and preantral granulosa cells initially co- transcripts. The absence of a stimulatory effect of growing cultured with fully grown oocytes (Fig. 7B, OOX/CC + oocytes demonstrates the specificity of the action of fully FGO and OOX/PG + FGO) showed greater metabolism of grown oocytes as well as the developmental control of this [5-3H]-glucose than those co-cultured with growing oocytes competence during oogenesis. (Fig. 7B, OOX/CC + GO and OOX/PG + GO). Further- The effect of fully grown oocytes and growing oocytes more, preantral granulosa cells co-cultured with fully grown on the TCA cycle activity by cumulus cells and preantral oocytes (Fig. 8, OOX/PG + FGO) exhibited higher steady- granulosa cells was estimated. As shown in Fig. 7C, OOX state mRNA levels of all the glycolytic genes assessed cumulus cells cultured with fully grown oocytes (Fig. 7C, except for Aldoa when compared with those in preantral OOX/CC + FGO) exhibited higher metabolism of [2-14C]- granulosa cells co-cultured with growing oocyte (Fig. 8, pyruvate when compared with OOX cumulus cells cultured OOX/PG + GO) or without any oocytes (Fig. 8, OOX/PG). without any oocytes or with growing oocytes (Fig. 7C, These results indicate that both cumulus cells and preantral OOX/CC and OOX/CC + GO). In sharp contrast, neither granulosa cells are capable of responding to the paracrine fully grown oocytes nor growing oocytes were able to

Fig. 7. Comparison of fully grown and growing oocytes on the metabolic activities of granulosa cells. Cumulus cells were isolated from 22-day-old eCG- primed mice. Preantral granulosa cells were isolated from 12-day-old mice. Oocytectomized complexes from either cumulus cells or preantral granulosa cells were cultured in the absence of oocytes (OOX/CC or OOX/PG), or co-cultured with either fully grown oocytes (OOX/CC + FGO or OOX/PG + FGO) or growing oocytes (OOX/CC + GO or OOX/PG + GO), and glycolytic activity (B), activity of the TCA cycle (C) or amount of total glucose oxidation (D) were estimated as described in Materials and methods. The values shown are the mean (FSEM) values from at least three experimental replicates. *P b 0.05; **P b 0.005; ***P b 0.0005. 28 K. Sugiura et al. / Developmental Biology 279 (2005) 20–30

Fig. 8. Effect of co-culture with either fully grown oocytes or growing oocytes on the steady-state mRNA levels of genes encoding glycolytic enzymes in preantral granulosa cells. Preantral granulosa cells were isolated from preantral follicles of 12-day-old mice. Oocytectomized preantral granulosa cells were cultured in the absence of oocytes (OOX/PG) or with either growing oocytes (OOX/PG + GO) or fully grown oocytes (OOX/PG + FGO) and mRNA levels of genes encoding glycolytic enzymes were estimated by RNase protection assays. (Left) A representative experiment is shown. (Right) Steady-state mRNA levels of glycolytic genes in preantral granulosa cells cultured in the absence of oocytes (black bars), or co-cultured with either growing oocytes (hatched bars) or fully grown oocytes (white bars) after normalization the level of Rpl19 transcripts. The values shown are the mean (FSEM) values from three experimental replicates. *P b 0.05; **P b 0.01. promote activity of the TCA cycle in preantral granulosa oocytes promote high steady-state levels of these mRNA cells (Fig. 7C, OOX/PG, OOX/PG + GO and OOX/PG + and glycolytic activity in cumulus cells. These results FGO). Nevertheless, fully grown oocytes, but not the indicate that fully grown oocytes secrete paracrine factors growing oocytes normally indigenous to preantral follicles, that stimulate glycolytic activity in granulosa cells at least in greatly stimulated the total utilization of glucose by part by promoting the expression of genes encoding preantral granulosa cells (Fig. 7D, OOX/PG + FGO). These glycolytic enzymes. Moreover, secretion of the factors results indicate that fully grown but not growing oocytes promoting glycolysis depends on the developmental stage secrete the paracrine factors that stimulate the glucose of the oocyte; only fully grown oocytes have this capacity. utilization in preantral granulosa cells, but probably only by Fully grown oocytes but not growing oocytes stimulated the 14 promoting glycolysis. Preantral granulosa cells are not CO2 release from pyruvate in cumulus cells, suggesting an capable of responding to the TCA cycle-stimulating factors increase in the TCA cycle in cumulus cells, thus secretion of produced by fully grown oocytes, even though cumulus the factor(s) is similarly dependent upon developmental cells can respond to these signals. stage of the oocyte. Furthermore, cumulus cells but not preantral granulosa cells were capable of responding to the TCA cycle-stimulating factors, indicating that preantral Discussion granulosa cells must undergo developmental changes to respond to the oocyte-derived factors. Thus, the capability The results presented here demonstrate that the steady- of oocytes to promote the TCA cycle activity appears state levels of mRNAs encoding enzymes involved in coordinated with the development of the ability of granulosa glycolysis are higher in cumulus cells and periantral mural cells to respond to it. Overall, the present study demon- granulosa cells than in the outermost mural granulosa of strated that energy production in antral follicles before the antral follicles or in preantral granulosa cell. Co-culture of preovulatory LH-surge is regulated by signals from the mural granulosa cells with fully grown oocytes isolated oocyte to the granulosa cells most closely associated with from antral follicles elevated levels of expression of these the oocyte, and both the production of these signals and the mRNAs in mural granulosa cells. Furthermore, fully grown ability to respond them are developmentally coordinated. K. Sugiura et al. / Developmental Biology 279 (2005) 20–30 29

The evidence of a granulosa cell role of nutritional cells co-cultured with fully grown oocytes. Further, when support for oocytes was first presented by Biggers et al. the OOX cumulus cells were cultured for a shorter period, (1967), who showed that oocyte maturation did not occur in 16 h, only Ldh1 and Pfkp mRNA levels were statistically glucose-containing medium unless cumulus cells or pyr- decreased (data not shown). These results indicate that uvate were also present. This was further supported by expression levels of Ldh1 and Pfkp are more sensitive to the reports that cumulus cells produce pyruvate from glucose or paracrine factors secreted by oocytes. Ldh1 encodes one of lactate in vitro (Donahue and Stern, 1968; Leese and Barton, the isoforms of , the critical enzyme 1985). Furthermore, oocytes require pyruvate as an energy regulating the amount of pyruvate by catalyzing both source for growth and resumption of meiosis (Biggers et al., pyruvate to lactate and the reverse. On the other hand, Pfkp 1967; Eppig, 1976). Thus, it is well accepted that granulosa encodes one of the isoforms of phosphofructokinase which cells play an important role in nutritional support of the is one of the key enzymes that regulates glycolysis (Uyeda, developing oocyte (Biggers et al., 1967; Buccione et al., 1979). It is important to note that genes encoding two 1990; Kim and Schuetz, 1991) and it has long been enzymes, phosphofructokinase and lactate dehydrogenase, believed, at least with regard to glucose metabolism, that which have critical roles in the glucose metabolic pathway, oocytes are passive recipients of the nutrients produced by are more sensitive to the paracrine factors secreted by cumulus cells. However, the present results show that oocytes. Phosphofructokinase activity is regulated by stimulation of nutritional support of granulosa cells is various metabolites acting as allosteric effectors (Kemp dependent upon paracrine factors secreted by the oocytes, and Foe, 1983; Reinhart and Lardy, 1980) and this enzyme at least before the induction of meiotic maturation by the activity might also be regulated by its phosphorylation LH-surge. Therefore, it is clear that the oocytes regulate the (Kagimoto and Uyeda, 1979; Sale and Denton, 1985). cooperativity between oocytes and cumulus cells for energy However, there are few reports about transcriptional metabolism. Since the oocytes themselves poorly utilize regulation of its activity. Vander Heiden et al. (2001) have glucose as an energy source, this regulation is probably shown that interleukin-3 (IL-3) is required for the expres- crucial for oocyte development. sion of phosphofructokinase in a murine pro-B-cell cell line. In contrast to the present indication that stimulation of In this report, levels of Slc2a1 (also known as Glut1) and glycolysis in cumulus cells is dependent upon the oocytes Hk2 mRNA encoding the glucose transporter 1 and before the LH-surge, it may not to be the case after the 2, which are also important for glucose surge, during the period of oocyte maturation. Sutton et al. metabolism, were decreased. Therefore, oocyte regulation (2003) showed that oocyte-secreted factors had no effect on of glycolysis in cumulus cells might involve expressional glucose uptake in cumulus cells during bovine oocyte controls of glycolytic enzymes in combination with those of maturation. Furthermore, although Zuelke and Brackett Glut1 and Hk2 and the posttranslational regulation of their 14 (1992) showed that glycolysis in bovine cumulus cells activity. Also, oocyte stimulation of CO2 release from was significantly decreased when oocytes were removed pyruvate in cumulus cells may result of the activation of the from COCs during in vitro oocyte maturation, the decrease TCA cycle by expressional and posttranslational controls of was only about 10% of intact COCs. This is in contrast to enzyme activities involved in the process as well as present result, where the difference between glycolysis in increased substrate availability via recruitment of mono- cumulus cells cultured as OOX or as COC was more than carboxylic acid transporters. 10-fold. These reports suggest that the ability of oocytes to The oocyte’s control of energy metabolism in granulosa regulate glycolysis after resumption of meiosis is decreased, cells may not be only to promote the supply of metabolites although the reason for this change is not clear. Since LH that the oocyte needs for its development. Oocytes regulate has been reported to activate glycolysis in cumulus cells the overall rate of follicular development. This was (Roy and Terada, 1999; Tsafriri et al., 1976; Zuelke and demonstrated by reaggregating growing oocytes isolated Brackett, 1992), it is possible that hormonal stimulation of from 12-day-old mice with the somatic cells of 0- to 2-day- glycolysis consequent to the LH-surge renders the stim- old mouse ovaries, which contain only primordial follicles, ulation by oocyte-derived paracrine factors superfluous. resulting in a doubling of the rate of follicular development, Alternatively, the glycolysis-stimulating action of oocytes thus correcting the asynchrony of the somatic and germ cell before the LH-surge persists in the cumulus cells during components of the follicles (Eppig et al., 2002). Thus, the oocyte maturation after the surge. Also, it might be possible bfolliculogenesis clockQ is set by the oocytes (Eppig et al., that this effect is unique in the mouse, as oocyte-secreted 2002; Matzuk et al., 2002). Moreover, growth and steroido- factors are required for FSH-stimulated cumulus cell genesis of cultured mouse follicles from preantral through expansion in mouse but not in bovine (Eppig et al., 1993; antral stages were dependent on the glucose concentration in Ralph et al., 1995). the culture media, indicating that follicular glycolysis is In the present results, the steady-state levels of mRNA required for growth and estradiol secretion (Boland et al., encoded by Eno1, Pkm2, Ldh1, and Pfkp were decreased 1994). Therefore, one mechanism by which the oocyte may after 26 h of culture in OOX cumulus cells compared with orchestrate follicular development is by the control of cumulus cells cultured as intact COCs or OOX cumulus granulosa cell energy metabolism. 30 K. Sugiura et al. / Developmental Biology 279 (2005) 20–30

Acknowledgments expression in mouse granulosa cells during ovulation. Endocrinology 142, 3187–3197. Kagimoto, T., Uyeda, K., 1979. Hormone-stimulated phosphorylation of We thank Drs. Ann M. Dorward, Mary Ann Handel, and liver phosphofructokinase in vivo. J. Biol. Chem. 254, 5584–5587. Edward Leiter, and Karen Wigglesworth for their helpful Kemp, R.G., Foe, L.G., 1983. Allosteric regulatory properties of muscle suggestions in the preparation of this paper. This research phosphofructokinase. Mol. Cell. Biochem. 57, 147–154. was supported by grant HD44416 from the NICHD to J.J.E. Kim, H., Schuetz, A.W., 1991. Regulation of nuclear membrane assembly and maintenance during in vitro maturation of mouse oocytes: role of pyruvate and protein synthesis. Cell Tissue Res. 265, 105–112. Latham, K.E., Bautista, D.M., Hirao, Y., O’Brien, M.J., Eppig, J.J., 1999. References Comparison of protein synthesis patterns in mouse cumulus cells and mural granulosa cells: effects of follicle-stimulating hormone and Biggers, J.D., Whittingham, D.G., Donahue, R.P., 1967. The pattern of insulin of granulosa cell differentiation in vitro. Biol. Reprod. 61, energy metabolism in the mouse oocyte and zygote. Proc. Natl. Acad. 482–492. Sci. U. S. A. 58, 560–567. Latham, K.E., Wigglesworth, K., McMenamin, M., Eppig, J.J., 2004. Boland, N.I., Humpherson, P.G., Leese, H.J., Gosden, R.G., 1994. The Stage-dependent effects of oocytes and growth differentiation factor 9 effect of glucose metabolism on murine follicle development and on mouse granulosa cell development: advance programming and steroidogenesis in vitro. Hum. Reprod. 9, 617–623. subsequent control of the transition from preantral secondary follicles to Buccione, R., Schroeder, A.C., Eppig, J.J., 1990. Interactions between early antral tertiary follicles. Biol. Reprod. 70, 1253–1262. somatic cells and germ cells throughout mammalian oogenesis. Biol. Leese, H.J., Barton, A.M., 1985. Production of pyruvate by isolated mouse Reprod. 43, 543–547. cumulus cells. J. Exp. Zool. 234, 231–236. Camp, T.A., Rahal, J.O., Mayo, K.E., 1991. Cellular localization and Matzuk, M.M., Burns, K.H., Viveiros, M.M., Eppig, J.J., 2002. Intercellular hormonal regulation of follicle-stimulating hormone and luteinizing communication in the mammalian ovary: oocytes carry the conversa- hormone receptor messenger RNAs in the rat ovary. Mol. Endocrinol. 5, tion. Science 296, 2178–2180. 1405–1417. O’Fallon, J.V., Wright Jr., R.W., 1986. Quantitative determination of the Chesnel, F., Eppig, J.J., 1995. Synthesis and accumulation of p34cdc2 and pentose phosphate pathway in preimplantation mouse embryos. Biol. cyclin B in mouse oocytes during acquisition of competence to resume Reprod. 34, 58–64. meiosis. Mol. Reprod. Dev. 40, 503–508. Peng, X.R., Hsueh, A.J.W., Lapolt, P.S., Bjersing, L., Ny, T., 1991. Colonna, R., Mangia, F., 1983. Mechanisms of uptake in Localization of luteinizing hormone receptor messenger ribonucleic cumulus-enclosed mouse oocytes. Biol. Reprod. 28, 797–803. acid expression in ovarian cell types during follicle development and Donahue, R.P., Stern, S., 1968. Follicular cell support of oocyte maturation: ovulation. Endocrinology 129, 3200–3207. production of pyruvate in vitro. J. Reprod. Fertil. 17, 395–398. Ralph, J.H., Telfer, E.E., Wilmut, I., 1995. Bovine cumulus cell expansion Elvin, J.A., Clark, A.T., Wang, P., Wolfman, N.M., Matzuk, M.M., 1999. does not depend on the presence of an oocyte secreted factor. Mol. Paracrine actions of growth differentiation factor-9 in the mammalian Reprod. Dev. 42, 248–253. ovary. Mol. Endocrinol. 13, 1035–1048. Reinhart, G.D., Lardy, H.A., 1980. Rat liver phosphofructokinase: Eppig, J.J., 1976. Analysis of mouse oogenesis in vitro. Oocyte isolation kinetic activity under near-physiological conditions. Biochemistry and the utilization of exogenous energy sources by growing oocytes. 19, 1477–1484. J. Exp. Zool. 198, 375–386. Rieger, D., Loskutoff, N.M., Betteridge, K.J., 1992. Developmentally related Eppig, J.J., 1991. Intercommunication between mammalian oocytes and changes in the metabolism of glucose and glutamine by cattle embryos companion somatic cells. BioEssays 13, 569–574. produced and co-cultured in vitro. J. Reprod. Fertil. 95, 585–595. Eppig, J.J., 2001. Oocyte control of ovarian follicular development and Roy, S.K., Terada, D.M., 1999. Activities of glucose metabolic enzymes in function in mammals. Reproduction 122, 829–838. human preantral follicles: in vitro modulation by follicle-stimulating Eppig, J.J., Schroeder, A.C., 1989. Capacity of mouse oocytes from hormone, luteinizing hormone, epidermal growth factor, insulin-like preantral follicles to undergo embryogenesis and development to live growth factor I, and transforming growth factor beta 1. Biol. Reprod. young after growth, maturation and fertilization in vitro. Biol. Reprod. 60, 763–768. 41, 268–276. Sale, E.M., Denton, R.M., 1985. Adipose-tissue phosphofructokinase. Eppig, J.J., Wigglesworth, K., Chesnel, F., 1993. Secretion of cumulus Rapid purification and regulation by phosphorylation in vitro. Biochem. expansion enabling factor by mouse oocytes: relationship to oocyte J. 232, 897–904. growth and competence to resume meiosis. Dev. Biol. 158, 400–409. Sutton, M.L., Cetica, P.D., Beconi, M.T., Kind, K.L., Gilchrist, R.B., Eppig, J.J., Wigglesworth, K., Pendola, F.L., Hirao, Y., 1997. Murine Thompson, J.G., 2003. Influence of oocyte-secreted factors and culture oocytes suppress expression of luteinizing hormone receptor messenger duration on the metabolic activity of bovine cumulus cell complexes. ribonucleic acid by granulosa cells. Biol. Reprod. 56, 976–984. Reproduction 126, 27–34. Eppig, J.J., Wigglesworth, K., Pendola, F.L., 2002. The mammalian oocyte Tsafriri, A., Lieberman, M.E., Ahren, K., Lindner, H.R., 1976. Dissociation orchestrates the rate of ovarian follicular development. Proc. Natl. between LH-induced aerobic glycolysis and oocyte maturation in Acad. Sci. U. S. A. 99, 2890–2894. cultured Graafian follicles of the rat. Acta Endocrinol. (Copenh.) 81, Haghighat, N., Van Winkle, L.J., 1990. Developmental change in 362–366. follicular cell-enhanced amino acid uptake into mouse oocytes that Uyeda, K., 1979. Phosphofructokinase. Adv. Enzymol. Relat. Areas Mol. depends on intact gap junctions and transport system Gly. J. Exp. Biol. 48, 193–244. Zool. 253, 71–82. Vander Heiden, M.G., Plas, D.R., Rathmell, J.C., Fox, C.J., Harris, M.H., Joyce, I.M., Pendola, F.L., Wigglesworth, K., Eppig, J.J., 1999. Oocyte Thompson, C.B., 2001. Growth factors can influence cell growth and regulation of kit ligand expression in mouse ovarian follicles. Dev. Biol. survival through effects on glucose metabolism. Mol. Cell. Biol. 21, 214, 342–353. 5899–58912. Joyce, I.M., Clark, A.T., Pendola, F.L., Eppig, J.J., 2000. Comparison of Whitten, W.K., 1971. Nutrient requirements for culture of preimplantation GDF-9 and oocyte regulation of kit ligand expression in mouse ovarian embryos in vitro. Adv. Biosci. 6, 129–139. follicles. Biol. Reprod. 63, 1669–1675. Zuelke, K.A., Brackett, B.G., 1992. Effects of luteinizing hormone on Joyce, I.M., Pendola, F.L., O’Brien, M., Eppig, J.J., 2001. Regulation of glucose metabolism in cumulus-enclosed bovine oocytes matured in prostaglandin-endoperoxide synthase 2 messenger ribonucleic acid vitro. Endocrinology 131, 2690–2696.