Theca interna: the other side of bovine follicular atresia

Currently, histological classifications of ovarian follicular atresia are almost exclusively based on the morphology of the membrana granulosa without reference to the theca interna. Atresia in the bovine small antral ovarian follicle has been redefined into antral or basal atresia where cell death commences initially within antral or basal regions of the membrana granulosa, respectively. To examine cell death in the theca interna in the two types of atretic follicles, bovine ovaries were collected and processed for immunohistochemistry and light microscopy. Follicles were classified as healthy, antral atretic, or basal atretic. Follicle diameter was recorded and sections stained with lectin from Bandeiraea simplicifolia to identify endothelial cells or with an antibody to cytochrome P450 cholesterol side-chain cleavage to identify steroidogenic cells and combined with TUNEL labeling to identify dead cells. The numerical density of steroidogenic cells within the theca interna was significantly reduced (P < 0.001) in basal atretic follicles in comparison with other follicles. Cell death was greater in both endothelial cells (P < 0.05) and steroidogenic cells (P < 0.01) of the theca interna of basal atretic follicles compared with healthy and antral atretic follicles. Thus, we conclude that the theca interna is susceptible to cell death early in atresia, particularly in basal atretic follicles.     

Sphingosine-1-phosphate and ceramide are associated with health and atresia of bovine ovarian antral follicles

The follicle destiny towards ovulation or atresia is multi-factorial in nature and involves outcries, paracrine and endocrine factors that promote cell proliferation and survival (development) or unchain apoptosis as part of the atresia process. In several types of cells, sphingosine-1-phospate (S1P) promotes cellular proliferation and survival, whereas ceramide (CER) triggers cell death, and the S1P/CER ratio may determine the fate of the cell. The aim of present study was to quantify S1P and CER concentrations and their ratio in bovine antral follicles of 8 to 17 mm classified as healthy and atretic antral follicles. Follicles were dissected from cow ovaries collected from a local abattoir. The theca cell layer, the granulosa cells and follicular fluid were separated, and 17β-estradiol (E2) and progesterone (P4) concentrations were measured in the follicular fluid by radioimmunoassay. Based on the E2/P4 ratio, the follicles were classified as healthy (2.2±0.3) or atretic (0.2±0.3). In both follicular compartments (granulosa and theca cell layer), sphingolipids were extracted and S1P and CER concentrations were quantified by HPLC (XTerra RP18; 5 µm, 3.0×150 mm column). Results showed that in both follicular compartments, S1P concentrations were higher in healthy antral follicles than in atretic antral follicles (P<0.05). The concentration of CER in the granulosa cells was higher in atretic antral follicles than in healthy antral follicles, but no differences were observed in the theca cell layer. The S1P/CER ratio in both follicular compartments was also higher in healthy antral follicles. Interestingly, in these follicles, there was a 45-fold greater concentration of S1P than CER in the granulosa cells (P<0.05), whereas in the theca cell layer, S1P had only a 14-fold greater concentration than CER when compared with atretic antral follicles. These results suggest that S1P plays a role in follicle health, increasing cellular proliferation and survival. In contrast, reduction of S1P and the S1P/CER in the antral follicle could trigger cellular death and atresia.     

mRNA expression pattern of gonadotropin receptors in bovine follicular cysts

Follicular growth and steroidogenesis are dependent on gonadotropin binding to their receptors in granulosa and theca cells of ovarian follicles. The aim of the present study was to evaluate the expression patterns of follicle-stimulating hormone receptor (FSHR) and luteinizing hormone receptor (LHCGR) in ovarian follicular structures from cows with cystic ovarian disease (COD) as compared with those of regularly cycling cows. Relative real-time RT-PCR analysis showed that the expression of FSHR mRNA in granulosa cells was highest in small antral follicles, then decreased significantly as follicles increased in size, and was lowest in cysts. FSHR mRNA was not detected in the theca cells of any follicular category, including cysts. LHCGR mRNA expression in granulosa cells was significantly higher in large antral follicles than in cysts, and not detected in granulosa cells of small and medium antral follicles. In theca cells, the expression level of LHCGR mRNA in medium antral follicles was higher than in small and large antral follicles, whereas that in follicular cysts it was similar to those in small and medium antral follicles, but higher than that in large antral follicles. Our findings provide evidence that there is an altered gonadotropin receptor expression in bovine cystic follicles, and suggest that in conditions characterized by altered ovulation, such as COD, changes in the signaling system of gonadotropins may play a fundamental role in their pathogenesis.     

Localization of an activin/activin receptor system in the porcine ovary.

The aim of this study was to locate a possible activin/activin receptor system within porcine ovaries containing functional corpora lutea. In situ hybridization was used to assess the gene expression of beta(A)- and beta(B)-activin subunits, and immunohistochemical studies were done to detect activin-A protein and activin receptor type II. mRNA expression of the beta(A)- and beta(B)-activin subunits was found in the granulosa from the unilaminar follicle stage onward, in the developing thecal layer of multilaminar and small antral follicles, in the theca interna of mid-sized antral follicles, in corpora lutea, and in the ovarian surface epithelium. Immunoreactive activin A protein could be detected at the same ovarian sites, but in thecal tissue of small antral follicles only. This protein was also demonstrated at the peripheral zone of oocytes from multilaminar and antral follicles. A positive immunoreaction for activin receptor was found in granulosa cells from multilaminar and older follicles and in oocytes from the earliest stages of follicular development onward. In late multilaminar follicles and in antral follicles, the oolemma was stained. Except for small antral follicles, a positive activin receptor immunoreaction was absent in the follicular theca. Activin receptor immunoreaction was furthermore present in corpora lutea and in the ovarian surface epithelium. It is concluded that, within porcine ovaries containing functional corpora lutea, an activin/activin receptor system is present in all intact follicles, the corpora lutea and the surface epithelium. Within follicles, granulosa and theca cells are the main sites of activin synthesis, while oocytes and granulosa cells are the main activin binding sites.     

The fine structure of the cumulus oophorus during follicular development in sheep

Summary The cumulus and membrana granulosa of non-atretic ovarian follicles from primordial up to a stage shortly before ovulation were studied by electron microscopy.The follicular cells of primordial follicles were undifferentiated and rested on a thick basal lamina. In secondary follicles the endoplasmic reticulum had proliferated forming an anastomosing network. In early antral and antral follicles (0.5–2.0 mm dia.) the ER was composed of short cisternae, the mitochondria had elongated and gap junctions were first observed. In late antral follicles (3.0–5.9 mm dia.) gap junctions were frequent. In the cumulus the glycogen was associated with electron lucent areas whereas in the granulosa it was invariably associated with membranes. In large antral follicles large membrane bound bodies were present in the basal cells of the cumulus. At early oestrus a distinctive mitochondrial morphology was noted in the granulosa but not elsewhere in the follicles. At mid oestrus numerous annular nexuses were present in the granulosa but not in the cumulus. At late oestrus numerous lipid droplets were formed in both cumulus and granulosa, the boundary with theca interna became indistinct and the basal lamina became incomplete.Deceased     

Insulin-like growth factor binding protein 4 expression parallels luteinizing hormone receptor expression and follicular luteinization in the primate ovary

It has been suggested that locally produced insulin-like growth factor binding protein 4 (IGFBP4) inhibits ovarian follicular growth and ovulation by interfering with IGF action. According to this hypothesis, IGFBP4-expressing follicles should demonstrate atresia, whereas healthy dominant follicles should be devoid of IGFBP4. Alternatively, according to this view, there could be constitutive expression of the inhibitory IGFBP4 but selective expression of an IGFBP4 protease in dominant follicles, allowing the follicle to mature and ovulate because of degradation of the binding protein. To examine these views concerning the role of IGFBP4 in primate follicular selection, we analyzed cellular patterns of IGFs 1 and 2, IGFBP4, and the IGFBP4 protease (pregnancy-associated plasma protein A [PAPP-A]) mRNA expression in ovaries from late follicular phase rhesus monkeys using in situ hybridization. The IGF1 mRNA was not detected, but the IGF2 mRNA was abundant in theca interna and externa of all antral follicles and was present in the granulosa of large preovulatory and ovulatory follicles. The IGFBP4 mRNA was selectively expressed by LH receptor (LHR) mRNA-positive theca interna cells of healthy antral follicles (defined by aromatase and gonadotropin receptor expression) and by LHR-expressing granulosa cells found only in large preovulatory and ovulatory follicles (defined by size and aromatase expression). The PAPP-A mRNA was abundant in granulosa cells of most follicles without obvious relation to IGFBP4 expression. Ovarian IGFBP4 mRNA levels were markedly increased after treatment with the LH analog, hCG, whereas IGF2 and PAPP-A mRNAs were not significantly altered. In summary, IGFBP4 expression appears to be associated with follicular selection, not with atresia, in the monkey ovary. The IGFBP4 is consistently expressed in healthy theca interna and in luteinized granulosa cells, likely under LH regulation. The IGFBP4 protease, PAPP-A, is widely expressed without apparent selectivity for IGFBP4-expressing follicles or for dominant follicles. These observations suggest that IGFBP4 or an IGFBP4 proteolytic product may be involved with LH-induced steroidogenesis and/or luteinization rather than with inhibition of follicular growth.     

Follicle-restricted compartmentalization of transforming growth factor beta superfamily ligands in the feline ovary

Ovarian follicular development, follicle selection, and the process of ovulation remain poorly understood in most species. Throughout reproductive life, follicle fate is balanced between growth and apoptosis. These opposing forces are controlled by numerous endocrine, paracrine, and autocrine factors, including the ligands represented by the transforming growth factor beta (TGFbeta) superfamily. TGFbeta, activin, inhibin, bone morphometric protein (BMP), and growth differentiation factor 9 (GDF-9) are present in the ovary of many animals; however, no comprehensive analysis of the localization of each ligand or its receptors and intracellular signaling molecules during folliculogenesis has been done. The domestic cat is an ideal model for studying ovarian follicle dynamics due to an abundance of all follicle populations, including primordial stage, and the amount of readily available tissue following routine animal spaying. Additionally, knowledge of the factors involved in feline follicular development could make an important impact on in vitro maturation/in vitro fertilization (IVM/IVF) success for endangered feline species. Thus, the presence and position of TGFbeta superfamily members within the feline ovary have been evaluated in all stages of follicular development by immunolocalization. The cat inhibin alpha subunit protein is present in all follicle stages but increases in intensity within the mural granulosa cells in large antral follicles. The inhibin betaA and betaB subunit proteins, in addition to the activin type I (ActRIB) and activin type II receptor (ActRIIB), are produced in primordial and primary follicle granulosa cells. Additionally, inhibin betaA subunit is detected in the theca cells from secondary through large antral follicle size classes. GDF-9 is restricted to the oocyte of preantral and antral follicles, whereas the type II BMP receptor (BMP-RII) protein is predominantly localized to primordial- and primary-stage follicles. TGFbeta1, 2, and 3 ligand immunoreactivity is observed in both small and large follicles, whereas the TGFbeta type II receptor (TGFbeta RII) is detected in the oocyte and granulosa cells of antral follicles. The intracellular signaling proteins Smad2 and Smad4 are present in the granulosa cell cytoplasm of all follicle size classes. Smad3 is detected in the granulosa cell nucleus, the oocyte, and the theca cell nucleus of all follicle size classes. These data suggest that the complete activin signal transduction pathway is present in small follicles and that large follicles primarily produce the inhibins. Our data also suggest that TGFbeta ligands and receptors are colocalized to large antral follicles. Taken together, the ligands, receptors, and signaling proteins for the TGFbeta superfamily are present at distinct points throughout feline folliculogenesis, suggesting discrete roles for each of these ligands during follicle maturation.     

The Expression Pattern of microRNAs in Granulosa Cells of Subordinate and Dominant Follicles during the Early Luteal Phase of the Bovine Estrous Cycle

This study aimed to investigate the miRNA expression patterns in granulosa cells of subordinate (SF) and dominant follicle (DF) during the early luteal phase of the bovine estrous cycle. For this, miRNA enriched total RNA isolated from granulosa cells of SF and DF obtained from heifers slaughtered at day 3 and day 7 of the estrous cycle was used for miRNAs deep sequencing. The results revealed that including 17 candidate novel miRNAs, several known miRNAs (n = 291–318) were detected in SF and DF at days 3 and 7 of the estrous cycle of which 244 miRNAs were common to all follicle groups. The let-7 families, bta-miR-10b, bta-miR-26a, bta-miR-99b and bta-miR-27b were among abundantly expressed miRNAs in both SF and DF at both days of the estrous cycle. Further analysis revealed that the expression patterns of 16 miRNAs including bta-miR-449a, bta-miR-449c and bta-miR-222 were differentially expressed between the granulosa cells of SF and DF at day 3 of the estrous cycle. However, at day 7 of the estrous cycle, 108 miRNAs including bta-miR-409a, bta-miR-383 and bta-miR-184 were differentially expressed between the two groups of granulosa cell revealing the presence of distinct miRNA expression profile changes between the two follicular stages at day 7 than day 3 of the estrous cycle. In addition, unlike the SF, marked temporal miRNA expression dynamics was observed in DF groups between day 3 and 7 of the estrous cycle. Target gene prediction and pathway analysis revealed that major signaling associated with follicular development including Wnt signaling, TGF-beta signaling, oocyte meiosis and GnRH signaling were affected by differentially expressed miRNAs. Thus, this study highlights the miRNA expression patterns of granulosa cells in subordinate and dominant follicles that could be associated with follicular recruitment, selection and dominance during the early luteal phase of the bovine estrous cycle.     

Immunocytochemical study of cell proliferation in the cystic ovarian follicles in cows

We examined the frequency of proliferating cells in cystic, atretic and healthy antral follicles to determine whether a disorder of cell proliferation was responsible for the occurrence of bovine cystic follicles. Paraffin sections of healthy follicles and various stages of atretic and cystic follicles were immunostained with mouse monoclonal antibody to proliferating cell nuclear antigen (PCNA). The PCNA-positive cells were counted in 4 different regions of a follicle from the apical to the basal side. In the granulosa layer, a significantly higher frequency of PCNA-positive cells was observed in the healthy follicle in the basal region as compared with the apical region. A similar pattern of PCNA-positive cells population was observed in the granulosa layer of atretic follicles, although the frequency in the basal region was significantly lower in the atretic than the healthy follicle. The rate of cell proliferation in the granulosa layer of cystic follicles was markedly lower at the basal region than that of atretic follicles. In the theca interna, the frequency of PCNA-positive cells in atretic follicles at the early stages was higher than that in cystic follicles at the early stages. These results suggest that in the healthy follicle the proliferative activity of granulosa cells is higher in the basal than the apical region, and that the cell proliferation activity in the granulosa and theca interna may decrease in association with the induction of a follicular cyst.     

Localization of apoptotic cells in the cystic ovarian follicles of cows: a DNA-end labeling histochemical study

We examined the frequency of apoptosis in cystic follicular cells to investigate the cause of the delay in regression of cystic follicles. Paraffin sections of healthy antral follicles, early and late atretic ones, and early and late cystic ones were stained using the terminal deoxynucleotidyl transferase (Tdt)-mediated biotinylated deoxyuridine triphosphates (dUTP) nick end-labeling (TUNEL) method to detect apoptotic cells. In the granulosa layer of early cystic and atretic follicles, TUNEL-positive cells were evident. In the theca interna of both early and late atresia, high frequencies of TUNEL-positive cells were observed. In the theca interna, a high frequency of TUNEL-positive cells was noted in the early cystic follicles, whereas their frequency decreased in late cystic follicles. These results suggest that apoptosis occurs in the granulosa and theca interna cells of cystic as well as atretic follicles, but the frequency of apoptosis in theca interna cells decreases in late cystic follicles, which may be responsible for the delay of follicular regression.     

Distribution of extracellular matrix proteins type I collagen, type IV collagen, fibronectin, and laminin in mouse folliculogenesis

The extracellular matrix (ECM) plays a prominent role in ovarian function by participating in processes such as cell migration, proliferation, growth, and development. Although some of these signaling processes have been characterized in the mouse, the relative quantity and distribution of ECM proteins within developing follicles of the ovary have not been characterized. This study uses immunohistochemistry and real-time PCR to characterize the ECM components type I collagen, type IV collagen, fibronectin, and laminin in the mouse ovary according to follicle stage and cellular compartment. Collagen I was present throughout the ovary, with higher concentrations in the ovarian surface epithelium and follicular compartments. Collagen IV was abundant in the theca cell compartment with low-level expression in the stroma and granulosa cells. The distribution of collagen was consistent throughout follicle maturation. Fibronectin staining in the stroma and theca cell compartment increased throughout follicle development, while staining in the granulosa cell compartment decreased. Heavy staining was also observed in the follicular fluid of antral follicles. Laminin was localized primarily to the theca cell compartment, with a defined ring at the exterior of the follicular granulosa cells marking the basement membrane. Low levels of laminin were also apparent in the stroma and granulosa cell compartment. Taken together, the ECM content of the mouse ovary changes during follicular development and reveals a distinct spatial and temporal pattern. This understanding of ECM composition and distribution can be used in the basic studies of ECM function during follicle development, and could aid in the development of in vitro systems for follicle growth.     

BMPRIB and BMPRII mRNA expression levels in goat ovarian follicles and the in vitro effects of BMP-15 on preantral follicle development

This study evaluated the levels of bone morphogenetic protein receptors BMPRIB and BMPRII mRNA in goat follicles and the effects of bone morphogenetic protein-15 (BMP-15) on the in vitro development of cultured preantral follicles. Real-time polymerase chain reaction (PCR) was used to analyze the levels of BMPRIB and BMPRII mRNA in caprine preantral follicles and in small and large antral follicles. Preantral follicles (≥150 μm) were also isolated from goat ovaries and cultured for 18 days in α-MEM(+) supplemented with or without BMP-15 (10, 50, or 100 ng/ml). At the end of culture, some follicles were fixed for ultrastructural evaluation. Real-time PCR showed a reduction in BMPRII mRNA levels from the primary to secondary follicles. Higher levels of BMPRIB mRNA were observed in granulosa/theca cells from large antral follicles compared with small antral follicles. Moreover, BMPRII mRNA was expressed to a greater extent in cumulus-oocyte complexes from large antral follicles than in their respective granulosa/theca cells. In culture, 50 ng/ml BMP-15 positively influenced antral cavity formation and follicle growth after 18 days and also maintained follicular integrity. Thus, BMPRIB and BMPRII mRNAs are present in all follicular categories. BMP-15 (50 ng/ml) stimulates growth, antrum formation and the ultrastructural integrity of isolated caprine preantral follicles after 18 days of culture.     

Immunolocalization of 3 beta-hydroxysteroid dehydrogenase in human ovary

Immunohistochemical localization of 3 beta-hydroxysteroid dehydrogenase (3 beta-HSD) was performed in 55 cases of morphologically normal human ovaries by using a specific polyclonal antibody against purified human placental 3 beta-HSD. In small developing follicles, immunoreactivity was observed only in the theca interna but also became recognizable in the membrana granulosa with development of the follicle. At a late stage of folliculogenesis, the intensity of the 3 beta-HSD activity in the membrana granulosa was nearly equal to that of theca interna in 2 or 3 large follicles examined. One to several layers of theca interna cells just beneath membrana granulosa did not demonstrate any immunoreactivity of 3 beta-HSD or that of cytochrome P-450 17 alpha-hydroxylase. These unstained theca interna cells did not appear to be directly involved in ovarian steroidogenesis and might be designated as 'enzymically inactive theca interna cells.' Marked immunoreactivity was observed in luteinized theca and granulosa cells of the corpus luteum.     

Localization of preovulatory expression of plasminogen activator inhibitor type-1 and tissue inhibitor of metalloproteinase type-1 mRNAs in the rat ovary.

Proteinases and their inhibitors control follicular connective tissue remodeling associated with follicular rupture. We examined the regulation and cellular localization of plasminogen activator inhibitor type-1 (PAI-1) and tissue inhibitor of metalloproteinase type-1 (TIMP-1) mRNAs by in situ hybridization. [35S]UTP-labeled RNA probes were hybridized to ovarian sections of eCG-primed immature rats treated with hCG. Before hCG stimulation of ovulation, very low expression of PAI-1 mRNA was observed in theca cells. After hCG administration, expression of PAI-1 mRNA was increased in theca cells of most antral follicles, whereas expression in granulosa cells was limited to preovulatory follicles and only to areas where the basal membrane was dissociated. Before hCG treatment, low expression of TIMP-1 mRNA was observed in theca cells, but not in granulosa cells. After hCG treatment, TIMP-1 mRNA was greatly stimulated in theca cells irrespective of follicle size, while the expression in granulosa cells was limited to large antral follicles. The present study demonstrates cell-specific expression of PAI-1 and TIMP-1 mRNAs in the LH/hCG-stimulated ovary, thus confirming the localized control of preovulatory proteolysis by coexpression of both enzymes and their respective inhibitors.     

Gene expression of endothelin-1 in the porcine ovary: follicular development

We have investigated which follicular compartment and stage of follicular development are associated with endothelin-1 (ET-1) gene expression in the porcine ovary. The localization of mature ET-1 peptide and of its mRNA was determined by immunohistochemistry and by in situ hybridization. Stage of follicular development associated with ET-1 expression was investigated in terms of follicular class and occurrence of atresia. The latter was investigated by determining the occurrence of DNA fragmentation in apoptotic cells on adjacent sections to those used for ET-1 gene expression. Fifteen ovaries from 10 prepubertal pigs stimulated with gonadotropin were collected; a total of 1050 follicles were examined. Specific ET-1 immunoreactivity was restricted to the ovarian vasculature and to the granulosa cell compartment of antral follicles. The pattern of ET-1 mRNA expression was similar to that found for ET-1 immunoreactivity. Primordial, primary, and most secondary follicles did not express ET-1. The theca cell layer did not express ET-1 regardless of follicle developmental stage. ET-1 expression occurred with a significantly greater probability (P < 0.001 by the likelihood ratio test) in the granulosa cell compartment of antral follicles than in any other follicle class. Furthermore, in antral follicles, ET-1 expression occurred with a greater likelihood in large antral follicles than in small antral follicles (P < 0.001 by the likelihood ratio test). In small antral follicles, only 16.8% expressed ET-1; in contrast, 66.7% of large antral follicles exhibited ET-1 expression. It is interesting that in follicles in which ovulation had already occurred, intense ET-1 expression was found only in the prominent developing vasculature, the other cells present in the luteinized follicle did not display any ET-1 expression. The pattern of ET-1 gene expression observed in this study would be in agreement with our previous suggestion of a plausible physiological role for ET-1 in preventing premature progesterone production by granulosa cells of an antral follicle. The occurrence of atresia and expression of ET-1 in the same follicle was rare. Small and large antral follicles constituted 5.1% and 5.6%, respectively, of the examined follicles in this category. The majority of atretic follicles did not express ET-1 and, conversely, follicles that expressed ET-1 were not atretic. To the best of our knowledge, this is the first report in which large, nonatretic follicles are clearly identified as the population of follicles expressing ET-1. The results of this study delineate the follicular developmental stage and the compartment of when and where ET-1 may be physiologically meaningful.     

IGF-Ⅰ及其受体、IGF结合蛋白-2和LH受体mRNA在卵泡中的表达

利用原位杂交和原位DNA－3’末端标记的方法研究了胰岛素样生长因子河（IG－I）、IGF－I受体、IGF结合蛋白－2、和促性腺激素受体的信使核糖核酸（mRNA）在不同生长与闭锁阶段的大鼠卵巢卵泡中表达的变化。结果表明：IGF－I主要在正常生长的初级卵泡、窦前卵泡和小窦状卵泡中表达。在各生长与成熟阶段的卵泡中都检测到IGF－I受体mRNA,闭锁卵泡的IGF－I受体表达降低。窦前与窦状的生长和闭锁卵泡均表达IGFBP－2。促卵泡激素（FSH）受体在窦前和小窦状卵泡的表达水平比其在大卵泡中的高。窦前与小窦状卵泡仅在膜细胞中表达黄体生成素（LH）受体mRNA,大卵泡的膜细胞与颗粒细胞均表达LH受体,在闭锁卵泡中仅在膜细胞中观察到LH受体的信号。综上结果,提示IGF－I,IGF－I受体和FSH受体在窦前和小窦状卵泡中的协同表达对卵泡的早期发育有重要作用。LH受体mRNA特异地在大卵泡的颗粒细胞中表达可能与优势卵泡选择相关。     

Role of ovarian theca and granulosa cell interaction in hormone productionand cell growth during the bovine follicular maturation process

We have investigated the possible role of theca and granulosa cell interaction in the control of the hormone-producing activity and growth of granulosa and theca cells during bovine ovarian follicular development, using a coculture system in which granulosa and theca cells were grown on opposite sides of a collagen membrane. When follicular cells were isolated from small follicles (3-5 mm), theca cells reduced estradiol, progesterone, and inhibin production by granulosa cells to 14 +/- 5%, 64 +/- 6%, and 27 +/- 4%, respectively, of the production by granulosa cells cultured alone. On the other hand, when the cells were isolated from large follicles (15-18 mm), theca cells increased these levels to 253 +/- 34%, 156 +/- 24%, and 287 +/- 45%, respectively. Theca cells did not affect the growth of granulosa cells. Androstenedione production by theca cells was augmented by granulosa cells to 861 +/- 190% (in small follicles) and 1298 +/- 414% (in large follicles), respectively. The growth of theca cells was also augmented by granulosa cells (small follicle, 210 +/- 43%, and large follicle, 194 +/- 24%, respectively). These results indicate that theca cells secrete factor(s) inhibiting the differentiation of immature while promoting that of matured granulosa cells; they also suggest that granulosa cells secrete factor(s) promoting both the differentiation and growth of theca cells throughout the follicular maturation process.