Follistatin-like 3 is a mediator of exercise-driven bone formation and strengthening.

Exercise is vital for maintaining bone strength and architecture. Follistatin-like 3 (FSTL3), a member of follistatin family, is a mechanosensitive protein upregulated in response to exercise and is involved in regulating musculoskeletal health. Here, we investigated the potential role of FSTL3 in exercise-driven bone remodeling. Exercise-dependent regulation of bone structure and functions was compared in mice with global Fstl3 gene deletion (Fstl3-/-) and their age-matched Fstl3+/+ littermates. Mice were exercised by low-intensity treadmill walking. The mechanical properties and mineralization were determined by µCT, three-point bending test and sequential incorporation of calcein and alizarin complexone. ELISA, Western-blot analysis and qRT-PCR were used to analyze the regulation of FSTL3 and associated molecules in the serum specimens and tissues. Daily exercise significantly increased circulating FSTL3 levels in mice, rats and humans. Compared to age-matched littermates, Fstl3-/- mice exhibited significantly lower fracture tolerance, having greater stiffness, but lower strain at fracture and yield energy. Furthermore, increased levels of circulating FSTL3 in young mice paralleled greater strain at fracture compared to the lower levels of FSTL3 in older mice. More significantly, Fstl3-/- mice exhibited loss of mechanosensitivity and irresponsiveness to exercise-dependent bone formation as compared to their Fstl3+/+ littermates. In addition, FSTL3 gene deletion resulted in loss of exercise-dependent sclerostin regulation in osteocytes and osteoblasts, as compared to Fstl3+/+ osteocytes and osteoblasts, in vivo and in vitro. The data identify FSTL3 as a critical mediator of exercise-dependent bone formation and strengthening and point to its potential role in bone health and in musculoskeletal diseases.

Inhibins and activins in human fetal abnormalities.

To date, the only routine clinical application of inhibin or activin measurement in testing for fetal abnormalities has been the use of inhibin A in prenatal screening for trisomy 21 (Down syndrome). Second trimester maternal serum levels of inhibin A are, on average, two-fold higher in Down syndrome than in unaffected pregnancies. Although the biology of altered second trimester maternal serum analyte levels in Down syndrome pregnancy cannot yet be explained, it seems that fetal products tend to be decreased, while placental products tend to be increased. This pattern holds true for inhibin A because maternal serum levels appear to be derived from placental rather than fetal sources. Therefore, the measurement of inhibins and activins in maternal fluids, although clinically useful and relatively easy to obtain, may not be helpful in studying their role in human fetal development. Studies in transgenic mice indicate a role for activin, follistatin, and activin receptor type IIA in development of the palate and craniofacial region. Cleft palate is a common birth defect and is associated with serious feeding and respiratory complications in newborns. We have begun to investigate the potential role of activin in human craniofacial development by examining the spatial and temporal expression of inhibin/activin subunits, follistatin and the activin receptors in the fetal palate. Palate tissues were collected at autopsy from fetuses ranging in gestational age from 9 to 42 weeks, and 8 week embryonic tissues were also examined. Tissues were either stored in paraffin for immunocytochemistry or were frozen for RT-PCR examination of the expression of inhibin/activin proteins or mRNAs, respectively. To date, betaA subunit, follistatin, and activin receptor, but not alpha and betaB subunit, mRNAs are present in palate tissues and inhibin/activin betaA immunoreactivity has been consistently observed in developing bone. Expression of the activin A subunit and its receptors in the human fetal palate are consistent with a developmental role. Studies are ongoing to determine whether altered activin biosynthesis is associated with cleft palate. Future studies of fetal tissues may help to elucidate other roles for the TGF-beta family in human development.

Dynamics of inhibin subunit and follistatin mRNA during development of normal and polycystic ovary syndrome follicles.

To further explore the developmental dynamics and possible roles of inhibin, activin, and follistatin in the development of human antral follicles as well as the relationship between mRNA and protein levels of these hormones within follicles, quantitative competitive RT-PCR assays were established to determine mRNA levels for the inhibin/activin subunits and both follistatin splice variants. Granulosa cell RNA was obtained by transvaginally aspirating follicles (6-23 mm) from carefully characterized normal women at different times of the follicular phase. alpha- and beta(A)-subunit mRNA levels increased significantly with follicle diameter (r = 0.56; P < 0.01 and r = 0.45; P < 0.05, respectively) and follicle maturity (r = 0.65; P < 0.001 and r = 0.58; P < 0.01, respectively), but beta(B) mRNA levels, which were at least 10-fold lower than levels of the other subunits, showed no relationship to size or maturity. Both follistatin 315 and 288 transcripts were detected in granulosa cells, but neither follistatin transcript varied significantly across the range of follicle sizes analyzed. In addition, granulosa cells contained three follistatin 315 mRNA transcripts for each follistatin 288 transcript, and the follistatin 315/288 ratio did not vary with follicle size. alpha-Subunit mRNA levels were positively associated with dimeric inhibin A protein in human follicular fluid from the same follicle aspirates (r = 0.71; P < 0.001). Similarly, beta(A)-subunit mRNA was associated with inhibin A (r = 0.59; P < 0.01), and beta(B) mRNA was associated with inhibin B (r = 0.67; P < 0.005) in these samples. Thus, the increase in inhibin subunit transcription and protein synthesis with follicle size suggests that inhibin biosynthesis might be important for continued development of the dominant follicle. To explore this hypothesis further, we compared mRNA levels for each of these transcripts in follicles obtained from six polycystic ovary syndrome patients (eight follicles) and compared the results to those from a group (n = 5) of normal follicles matched for mean diameter. Comparisons were also performed for a subset of polycystic ovary syndrome follicles (n = 5) matched for diameter and size range with the normal group. alpha-Subunit mRNA levels were 16-fold lower in both polycystic ovary syndrome follicle groups relative to size-matched normal follicles (P < 0.02), whereas beta(A)-subunit mRNA was significantly lower only when all polycystic ovary syndrome follicles were compared. beta(B)-Subunit and follistatin mRNA levels and the follistatin 315/288 ratio were not statistically different for any group. These results suggest that insufficient production of inhibin alpha and possibly beta(A)-subunits, but not follistatin, is associated with follicular arrest in polycystic ovary syndrome follicles.

Human follistatin-related protein: a structural homologue of follistatin with nuclear localization.

Follistatin-related protein is a recently discovered glycoprotein that is highly homologous in both primary sequence and exon/intron domain structure to the activin-binding protein, follistatin. We explored their potential for functional redundancy by investigating the relative affinities and kinetics of their interactions with activin, bone morphogenic protein-6, and bone morphogenic protein-7 and by exploring their expression and distribution in human tissues and cells. Follistatin and follistatin-related protein mRNA were ubiquitous by Northern analyses, although their sites of peak distribution differed, with follistatin-related protein and follistatin predominating in the placenta and ovary, respectively. Follistatin-related protein, like follistatin, preferentially bound activin with high affinity and in an essentially irreversible fashion. Although follistatin-related protein, like follistatin, possesses a signal sequence and no known nuclear localization signals, its secretion was undetectable in most cell lines by RIA. Intriguingly, follistatin-related protein was identified as a nuclear protein in human granulosa cells and all human cell lines tested. Furthermore, Western analyses of CHO cells transfected with human follistatin-related protein revealed this protein to reside within the insoluble nuclear protein fraction. We conclude that despite its remarkably high level of similarity to follistatin with regard to structure and activin binding kinetics, follistatin-related protein is a nuclear as well as a secretory protein that may perform distinct intracellular actions.

Follistatin-related protein (FSRP): a new member of the follistatin gene family.

The identification and characterization of follistatin related protein (FSRP) suggests that the follistatin (FS) gene family may actually contain two sub-families. The first includes FS and FSRP by virtue of their high degree of structural homology and comparable activin-binding activity, while the second sub-family contains extracellular matrix proteins that possess one or more 10-cysteine FS domains, but do not bind activin or related TGF-beta family members. Characterization of FSRP indicates that it binds activin with similar affinity and selectivity as FS, but does not bind heparin. Furthermore, although FSRP inhibits activin-mediated gene transcription in heterologous assays, FSRP is much less active than FS in the rat pituitary bioassay. When overexpressed in transgenic mice, FSRP may lead to interruption of follicular development and fertility in females but appears to have only a modest effect on males. These results suggest that FSRP is a structural, but not necessarily a functional homologue of FS.

Differential regulation of inhibin B and inhibin a by follicle-stimulating hormone and local growth factors in human granulosa cells from small antral follicles.

Serum inhibin B rises across the luteal-follicular transition, whereas inhibin A does not increase until the late follicular phase of the menstrual cycle. To test the hypothesis that inhibin B is secreted from preantral and small antral follicles and that FSH and local growth factors differentially regulate inhibin B and inhibin A from these developing follicles, human ovaries were obtained after oophorectomy. Basal secretion of inhibin B and inhibin A was examined in intact preantral follicles in culture (n = 6). Basal secretion and regulation of inhibin B and inhibin A secretion by gonadotropins, androstenedione, activin A, insulin, and IGF-I were examined in cultured granulosa cells from small antral follicles (n = 21). Inhibin B secretion from preantral follicle cultures was detectable at baseline (range, 17-96 pg/mL), whereas inhibin A was not detectable. In contrast, both inhibin B and inhibin A were detectable in granulosa cell cultures from small antral follicles. In granulosa cells from small antral follicles, FSH (30 ng/mL) stimulated inhibin A 3-fold (10.5 +/- 2.2 to 32.5 +/- 8.3 IU/mL; P < 0.001), but not inhibin B secretion (1730 +/- 354 to 2314 +/- 532 pg/mL; P = NS). Likewise, cAMP (1 mmol/L) stimulated inhibin A 4-fold (16.6 +/- 4.3 to 62.5 +/- 21.9 IU/mL; P < 0.002), but not inhibin B secretion (2327 +/- 546 to 1877 +/- 377 pg/mL; P = NS). hCG (30 ng/mL) did not stimulate inhibin A or inhibin B. Androstenedione (10(-)(7) mol/L), activin (30 ng/mL), insulin (30 ng/mL), and insulin-like growth factor I (IGF-I; 100 ng/mL) alone did not stimulate inhibin A or inhibin B secretion. Further, FSH-stimulated inhibin A secretion was not augmented by androstenedione, activin, insulin, or IGF-I. In contrast, the combination of IGF-I and FSH was the only treatment that stimulated inhibin B secretion (1742 +/- 380 to 2881 +/- 731 pg/mL; P < 0.03). However, FSH in combination with IGF-I resulted in greater stimulation of inhibin A (340%) than inhibin B (65%). These findings demonstrate that inhibin B is secreted from developing preantral and small antral follicles, but is not directly stimulated by FSH. However, the combination of FSH and IGF-I enhanced inhibin B secretion. In contrast, inhibin A is not secreted from preantral follicles, but in small antral follicles FSH and cAMP stimulate inhibin A secretion. Further, FSH in combination with IGF-I results in a greater degree of stimulation of inhibin A than of inhibin B. These findings suggest that FSH and IGF-I differentially regulate inhibin A and inhibin B secretion. However, additional growth factors or increasing granulosa cell number may contribute to the preferential serum inhibin B increase across the luteal-follicular transition in the menstrual cycle.

Follistatin: essential role for the N-terminal domain in activin binding and neutralization.

Follistatin is recognized to be an important regulator of cellular differentiation and secretion through its potent ability to bind and bioneutralize activin with which it is colocalized in many tissue systems. The 288-residue follistatin molecule is comprised of a 63-residue N-terminal segment followed by three repeating 10-cysteine "follistatin domains" also represented in several extracellular matrix proteins. We have used chemical modifications and mutational analyses to define structural requirements for follistatin bioactivity that previously have not been investigated systematically. Mutant follistatins were stably expressed from Chinese hamster ovary cell cultures and assayed for activin binding in a solid-phase competition assay. Biological activities were determined by inhibition of activin-mediated transcriptional activity and by suppression of follicle-stimulating hormone secretion by cultured anterior pituitary cells. Deletion of the entire N-terminal domain, disruption of N-terminal disulfides, and deletion of the first two residues each reduced activin binding to <5 % of expressed wild-type follistatin and abolished the ability of the respective mutants to suppress activin-mediated responses in both bioassay systems. Hence, the three follistatin domains inherently lack the ability to bind or neutralize activin. Activin binding was impaired after oxidation of at least one tryptophan, at position 4, in FS-288. Mutation of Trp to Ala or Asp at either positions 4 or 36 eliminated activin binding and bioactivity. Mutation of a third hydrophobic residue, Phe-52, reduced binding to 20%, whereas substitutions for the individual Lys and Arg residues in the N-terminal region were tolerated. These results establish that hydrophobic residues within the N-terminal domain constitute essential activin-binding determinants in the follistatin molecule. The correlation among the effects of mutation on activin binding, activin transcriptional responses, and follicle-stimulating hormone secretion substantiates the concept that, at least in the pituitary, the biological activity of follistatin is attributable to its ability to bind and bioneutralize activin.

Dynamic changes in the intrafollicular inhibin/activin/follistatin axis during human follicular development: relationship to circulating hormone concentrations.

Previous studies of normal human ovaries suggest that inhibins, activins, and follistatin (FS) are produced in a stage-specific pattern indicative of intraovarian, autocrine/paracrine roles in regulating follicle development. However, these studies relied largely on surgical specimens and thus include little information about the menstrual cycle stage or dominant follicle status at the time follicles or ovaries were obtained. The purpose of this study was to 1) determine the pattern of intrafollicular hormone biosynthesis across antral follicle development in normal women, 2) compare hormone concentrations in dominant and nondominant follicles from the same ovary, and 3) examine the relationship between dominant follicle hormone content and circulating hormone levels. Intrafollicular estradiol, progesterone, and inhibin A concentrations increased significantly with follicle size or maturity, whereas significant inverse relationships were observed for androstenedione and the androstenedione/estradiol (A:E) ratio. In contrast, neither inhibin B, activin A, nor free FS varied consistently with size or maturity. Estradiol, progesterone, and inhibin A levels and A:E ratio were significantly lower in nondominant follicles compared to the dominant follicle aspirated from the same ovary. Although intrafollicular and serum concentrations of each hormone followed the same general pattern as follicles develop, the human follicular fluid/serum gradients changed during the follicular phase and were different for estradiol and inhibin A, suggesting the presence of stage-specific differences in pharmacodynamics. These results are consistent with the hypothesis that the orderly transition from an activin-dominant to an inhibin A/FS-dominant microenvironment is critical for dominant follicle development.

Analysis of human follistatin structure: identification of two discontinuous N-terminal sequences coding for activin A binding and structural consequences of activin binding to native proteins.

A primary physiological function of follistatin is the binding and neutralization of activin, a transforming growth factor-beta family growth factor, and loss of function mutations are lethal. Despite the critical biological importance of follistatin's neutralization of activin, the structural basis of activin's binding to follistatin is poorly understood. The purposes of these studies were 1) to identify the primary sequence(s) within the N-terminal domain of the follistatin coding for activin binding, and 2) to determine whether activin binding to the native protein causes changes in other structural domains of follistatin. Synthetic peptide mimotopes identified within a 63-residue N-terminal domain two discontinuous sequences capable of binding labeled activin A. The first is located in a region (amino acids 3-26) of follistatin, a site previously identified by directed mutagenesis as important for activin binding. The second epitope, predicted to be located between amino acids 46 and 59, is newly identified. Although the sequences 3-26 and 46-59 code for activin binding, native follistatin only binds activin if disulfide bonding is intact. Furthermore, pyridylethylation of Cys residues followed by N-terminal sequencing and amino acid analysis revealed that all of the Cys residues in follistatin are involved in disulfide bonds and lack reactive free sulfhydryl groups. Specific ligands were used to probe the structural effects of activin binding on the other domains of the full-length molecule, comprised largely of the three 10-Cys follistatin module domains. No effects on ligand binding to follistatin-like module I or II were observed after the binding of activin A to native protein. In contrast, activin binding diminished recognition of domain III and enhanced that of the C domain by their respective monoclonal antibody probes, indicating an alteration of the antigenic structures of these regions. Thus, subsequent to activin binding, interactions are likely to occur between regions of follistatin located in different domains and separated by considerable lengths of linear sequence. Such interactions could have important functional significance with respect to the structural heterogeneity of naturally occurring follistatins.

Analysis of follicular fluid hormone concentrations and granulosa cell mRNA levels for the inhibin-activin-follistatin system: relation to oocyte and embryo characteristics.

OBJECTIVE: To explore the potential roles of inhibin, activin, and follistatin in human oocyte development by quantifying their intrafollicular biosynthesis. DESIGN: Prospective, nonrandomized study. SETTING: An IVF unit and academic research laboratory. PATIENT(S): Thirty one patients undergoing IVF. INTERVENTION(S): Human menopausal gonadotropins or human FSH (or both) were administered. Single-follicle aspirates (n = 110) were collected for analysis. MAIN OUTCOME MEASURE(S): Concentrations of dimeric, total and pro-alphaC inhibin forms; activin A; follistatin; estradiol; and progesterone were measured in follicular fluids. Granulosa-cell mRNA was analyzed for alpha, beta A, and beta B inhibin and activin subunits; follistatin; activin receptors; and beta-actin. Hormone concentrations and mRNA levels were correlated with oocytes or embryos from the same follicles. RESULT(S): Levels of progesterone and follistatin were significantly greater in follicles containing MI or MII oocytes than in those containing GV oocytes. Inhibin alpha-subunit mRNA levels were significantly higher in follicles containing maturing oocytes, the highest-quality oocytes, and oocytes that were subsequently fertilized. In contrast, inhibin alpha-subunit mRNA levels were significantly lower in follicles from which higher-quality embryos were obtained. CONCLUSION(S): Inhibin alpha-subunit biosynthesis is associated with normal oocyte and follicle maturation, but excessive alpha-inhibin is associated with poor embryo quality. None of the hormones analyzed were associated with oocyte or embryo quality.

Allelic variants of the follistatin gene in polycystic ovary syndrome.

In an earlier study of 37 candidate genes for polycystic ovary syndrome (PCOS), the strongest evidence for genetic linkage was found with the region of the follistatin gene. We have now carried out studies to detect variation in the follistatin gene and assess its relevance to PCOS. By sequencing the gene in 85 members of 19 families of PCOS patients, we found sequence variants at 17 sites. Of these, 16 sites have variants that are too rare to make a major contribution to susceptibility; the only common variant is a single base pair change in the last exon at a site that is not translated. In our sample of 249 families, the evidence for linkage between PCOS and this variant is weak. We also examined the expression of the follistatin gene; messenger RNA levels in cultured fibroblasts from PCOS and control women did not differ appreciably. We conclude that contributions to the etiology of PCOS from the follistatin gene, if any, are likely to be small.

Differential response to exogenous and endogenous activin in a human ovarian teratocarcinoma-derived cell line (PA-1): regulation by cell surface follistatin.

The activin/follistatin system is implicated in growth and differentiation of various cell types. Follistatin (FS), through binding and neutralizing activin, plays a major role in the regulation of activin bioavailability. We previously reported that ovarian PA1 cells constitutively secrete FS and show a decreased proliferation rate in response to exogenous activin only if cell surface associated FS is first removed by heparin treatment. These observations suggest that cell-associated FS prevents exogenous activin from accessing its receptor. We hypothesized that cell surface FS would differentially regulate the bioavailability of endogenous and exogenous activin in these cells. To examine the effect of endogenous activin, PA1 cells were stably transfected with an activin betaA-subunit complementary DNA (cDNA). The proliferation rate of five activin-secreting clones was measured by [3H]thymidine incorporation and compared with the proliferation rate of untransfected cells. In clones secreting levels of activin ranging from 22.6 +/- 7.1 to 42.4 +/- 9.9 ng/ml, proliferation was decreased by 31-72% at 96 h of culture, whereas one cell line secreting lower levels of activin (0.4 +/- 0.1 ng/ml) proliferated similarly to the untransfected cells, in which activin was not detectable. To further assess activin signaling, wild-type PA1 cells and activin-secreting clones were transiently transfected with an activin response element-luciferase reporter construct. Basal luciferase activity was 6-fold higher in activin-secreting clones than in wild-type PA1 cells. Exogenous activin (100 ng/ml) increased the transcriptional response of wild-type PA1 cells by 3-fold but did not increase reporter activity in activin secreting clones. Interestingly, the transcriptional response in activin secreting clones was always greater than the basal or activin-stimulated response in wild-type cells. Furthermore, we found that FS was removed from the cell surface by lipofectamine used for these transfections. Therefore, these results show that activation of the luciferase reporter gene occurs under conditions in which proliferation is affected, suggesting that the antiproliferative effect of activin could be due to a direct stimulation of activin signaling pathways. In summary, as opposed to exogenous activin, endogenous activin decreased proliferation of PA1 cells even in the presence of cell surface associated FS. These results are consistent with a model in which FS acts as a barrier for exogenous (endocrine-paracrine) but not for endogenous (autocrine) activin. In addition, the higher PA1 cell responsiveness to endogenous compared with exogenous activin, suggests that activin overexpression in PA1 cells may up-regulate an activin signaling component, or down-regulate an activin signaling inhibitor.

Characterization of inhibin/activin subunit, activin receptor, and follistatin messenger ribonucleic acid in human and mouse oocytes: evidence for activin's paracrine signaling from granulosa cells to oocytes.

Inhibin, activin, and follistatin (FS) are gonadal proteins that appear to have a role in regulating folliculogenesis through possible paracrine and/or autocrine interactions. To further examine the potential role of activin in oocyte-granulosa cell communication, we developed a sensitive reverse transcription-polymerase chain reaction protocol to analyze mRNA for the alpha, betaA, and betaB inhibin/activin subunits, FS, and the four activin receptor subtypes in individual human and mouse oocytes. The resulting expression pattern was further compared to that in human cumulus granulosa cells. Our results indicate that neither ssA nor betaB mRNA was detectable in any human or mouse oocyte, that alpha subunit was marginally present in some of the human oocytes, and that FS mRNA was detectable in human but not mouse oocytes. On the other hand, inhibin/activin subunit and FS mRNAs were abundantly expressed in cumulus cells. In addition, mRNAs for all four activin receptor subtypes (ActRIA, ActRIB, ActRIIA, and ActRIIB) were easily detectable in both oocytes and granulosa cells and appeared to be differentially expressed in oocytes during nuclear maturation. Finally, RNAs for both zona pellucida 3 and growth-differentiation factor-9, which were originally used as oocyte-specific markers, were detected in human but not mouse cumulus cells, although at lower levels than observed in oocytes. Taken together with previous studies, these results indicate that oocytes may be capable of responding to, but not synthesizing, activin.

Imbalanced expression of inhibin and activin subunits in primary epithelial ovarian cancer.

OBJECTIVES: Inhibins and activins are related gonadal peptides with opposing biologic actions on gonadotropin regulation, cell differentiation, and proliferation. The previous study of activin in ovarian cancer cell lines suggests that activin may promote growth of ovarian cancer. Elevated serum inhibin levels were also found in ovarian cancer patients; however, the source of elevated inhibin is unknown. This study is designed to examine the expression of inhibin and activin subunits as well as activin receptor in primary ovarian epithelial tumors to explore their role in the process of ovarian epithelial tumorigenesis. METHODS: The protein and mRNA expression of alpha and betaA subunits of inhibin/activin as well as of activin receptor mRNA were examined with immunohistochemistry (IHC) and reverse transcription-polymerase chain reaction (RT-PCR) in 112 ovarian carcinomas. Cases included 59 serous, 23 endometrioid, 16 mucinous, 9 clear cell, and 5 undifferentiated carcinomas. We also tested normal ovary and benign and borderline ovarian tumors for comparison. These included 17 ovarian surface epithelial samples, 6 serous and 5 mucinous cystadenomas, and 9 serous and 7 mucinous tumors of low malignant potential. A total of 139 ovarian tumors were analyzed by IHC and a total of 63 ovarian tumor samples were tested by RT-PCR. RESULTS: Inhibin alpha subunit expression was found in 47% of ovarian surface epithelia and focal alpha immunoreactivity was seen in tumor stroma, but was not found in the epithelial component of ovarian cystadenomas, tumors of low malignant potential (LMP), or carcinomas. Activin betaA subunit was expressed in 93% of surface epithelia, in the epithelial component of all cystadenomas, in 81% of LMP tumors, and in 72% of carcinomas, but not in tumor stroma. Activin expression did not correlate with histologic grades, tumor types, and surgical stages. Activin receptor type I and II mRNA-amplified products were found in virtually all the surface epithelial samples and ovarian tumors. CONCLUSIONS: The data suggest that imbalanced expression of inhibin and activin subunits in ovarian surface epithelium may represent an early event which leads to epithelial proliferation. Unopposed betaA and activin receptor expression in epithelial compartment of ovarian tumors suggest that activin may be available as autocrine and/or paracrine factors in ovarian epithelial tumors. But exact roles of inhibin and activin in ovarian epithelial tumors remain to be defined.

Activin regulates betaA-subunit and activin receptor messenger ribonucleic acid and cellular proliferation in activin-responsive testicular tumor cells.

Activin, a member of the transforming growth factor-beta superfamily of growth and differentiation factors, has a number of actions in embryonic as well as adult tissues. These actions are mediated via a family of receptors containing two subtypes and at least two members of each subtype. Recent evidence demonstrates that activin-responsive cell lines containing different subsets of these receptors are valuable models for dissecting functional relationships among receptor subtype, signal transduced, and response obtained. TT cells, derived from a p53(-/-)/alpha-inhibin(-/-) mouse testicular tumor, respond to activin by proliferating, a response that can be inhibited by follistatin (FS) treatment. Using semiquantitative RT-PCR methods, we characterized steady state messenger RNA (mRNA) levels for the inhibin/activin subunits, FS, and activin receptor subtypes under basal conditions and in the presence of activin or FS. These cells produced ample immunoreactive activin A and FS, necessitating higher treatment doses to observe any modulation of cellular proliferation. Furthermore, in the presence of exogenous activin, mRNA levels for activin receptor type IIA (ACTRIIA) and betaA were significantly and profoundly suppressed. In addition, both ACTR1B and ACTRIIB were detectable and down-regulated by exogenous activin, although not to the degree observed for ACTRIIA and betaA. Finally, activin treatment at the higher doses, which decreased activin receptor mRNA levels, resulted in inhibition of cellular proliferation. Taken together with previous observations, our results support the model that these tumor cells respond to an autocrine activin signal by proliferating, whereas exogenous or excess activin results in down-regulation of activin receptor and activin biosynthesis, suggesting a potential autocrine/paracrine mechanism by which activin can modulate its own signal.

A two-site chemiluminescent assay for activin-free follistatin reveals that most follistatin circulating in men and normal cycling women is in an activin-bound state.

Follistatin (FS) is a monomeric protein that binds and regulates the bioavailability of activin. Previously, we found circulating levels of total FS to be similar in men and cycling women. Because relative amounts of activin-bound and free FS are important considerations in determining activin bioavailability, we asked here whether the relative proportions of these two changed during different physiologic states. For this, we developed a two-site, solid-phase, immunochemiluminescent assay for free FS. The assay recognizes the 288 or 315 amino acid variants of human FS and has a detectable limit of 1 ng/mL. Inhibin, transforming growth factor-beta, or alpha-2-macroglobulin do not cross-react or interfere in this assay. Preincubation of FS with activin results in dose-dependent loss of immunoreactivity, confirming specificity of the assay for free FS. Human follicular fluid, pituitary extract, and serum with added FS dilute parallel with the recombinant human FS-288 standard. Recovery of recombinant human FS-288 from serum is quantitative. Using this assay, we found circulating concentrations of free FS to be at or below the detection limit of the assay throughout the menstrual cycle. Comparison of circulating total and free FS levels in postmenopausal or cycling women and normal men suggested that at least 90% is activin-bound. In contrast, measurable quantities of free FS were found in follicular fluid and pituitary extracts. The results of this study, showing that most circulating FS is normally activin-bound, argue against an endocrine role for FS and suggest that a major role of circulating FS is to bind and neutralize the bioactivity of circulating activin. The roles of FS as a local autocrine or paracrine regulator of activin in target tissues, where FS exists in free form, or as an endocrine regulator in human pathophysiology, warrants further investigation.

Characterization of intrafollicular steroid hormones, inhibin, and follistatin in women with and without polycystic ovarian syndrome following gonadotropin hyperstimulation.

The etiology of polycystic ovary syndrome (PCOS) is unexplained. Since no major deficiencies are reported in serum FSH or inhibin, we hypothesized that abnormal levels of a paracrine modulator of FSH action within the ovary may be associated with the arrest of follicular growth seen in the PCOS ovary. Follicular fluid aspirates were collected from women with (n = 7) or without (n = 17) PCOS during oocyte retrieval for in vitro fertilization. Aspirates were assayed for total inhibin, inhibin A (InhA), inhibin B (InhB), and follistatin (FS), as well as for estradiol, progesterone (P4), androstenedione, and total protein. Hormone levels were compared between women with and without PCOS using all aspirates (some of which were collected from multiple follicles at once) and also between aspirates containing fluid from a single follicle only (PCOS, n = 30; non-PCOS, n = 107). P4 levels were significantly (p < 0.01) reduced in PCOS versus non-PCOS women as evidenced by analysis of all follicles as well as in single-follicle aspirates only. In addition, InhA, P4, and FS increased with follicle volume, and InhB decreased significantly in non-PCOS, but not in PCOS, follicles. Therefore, although follicular development can be induced in PCOS patients with gonadotropins, hormonal responses within the ovary appear inappropriate in terms of concentrations or patterns of secretion. These data support the concept that PCOS is associated with a deficit in the paracrine control of folliculogenesis.

Presence of activin, inhibin, and follistatin in epithelial ovarian carcinoma.

Activin induces proliferation in epithelial ovarian carcinoma cell lines, whereas follistatin (FS), an activin binding protein, inhibits this action. To test the hypothesis that activin production, in excess of inhibin and FS, results in cell proliferation in epithelial ovarian tumors, messenger RNA (mRNA) expression of the activin family of proteins, FS, and activin type I and II receptors was examined in 25 primary epithelial ovarian tumors and tumor epithelium in culture (n = 7) using RT-PCR. Activin A was measured in the serum of ovarian cancer patients, and activin A, total inhibin, and FS protein secretion was measured from primary epithelial tumors in vitro. The effect of activin and FS on cell proliferation was assessed by measuring [3H]thymidine incorporation. All results were compared with normal ovarian epithelium. All epithelial ovarian tumors expressed mRNA for the alpha, beta A, and beta B subunits; FS 288 and 315; and the activin type IA, IB, II, and IIB receptors. beta A mRNA expression, as assessed using semiquantitative RT-PCR, was 3-fold greater in cultured tumor epithelium than in primary tumors (band density 0.86 +/- 0.17 vs. 0.28 +/- 0.09; P < 0.01). In addition, beta A mRNA was abundantly expressed in normal epithelium in culture (n = 2), whereas only trace amounts were seen in 2/9 primary epithelial samples. Activin protein was secreted by 24/25 primary epithelial ovarian tumors (range 0.2-155.8 ng/mL). In contrast, total inhibin was secreted by only 2/25 (range 0.01-0.92 ng/mL), whereas free FS was not detectable in the medium of any tumor (< 0.5 ng/mL). Treatment with activin or FS did not consistently affect cell growth. Measurement of serum activin A in a subset of subjects and in 27 additional subjects with epithelial ovarian carcinoma (n = 33) revealed preoperative activin A levels > 3 SD above the mean for pre- and postmenopausal women in 13/33 (39%) subjects. We conclude that in epithelial ovarian cancer: 1) beta A subunit mRNA is expressed, 2) activin protein is secreted more frequently than inhibin and in greater quantities than FS, 3) beta A subunit mRNA expression is greater in neoplastic and normal epithelium in culture than in the primary tissue, 4) the majority of tumors in culture do not respond to activin or FS treatment with proliferation, and 5) serum activin levels may reflect tumor secretion in some patients. Thus, activin A appears to be available as an autocrine/paracrine factor in epithelial ovarian tumors and may contribute to circulating levels, but its role in tumorigenesis has yet to be defined.

Multiple immunoreactive inhibin proteins in serum from postmenopausal women with epithelial ovarian cancer.

Inhibin is an ovarian protein previously shown, using a nonspecific assay, to be elevated in serum of women with ovarian cancer. However, inhibin is secreted in multiple biochemical forms, including dimeric inhibin A and B and alpha inhibin precursors (pro-alphaC), each of which can now be specifically measured. We have examined the secretion of inhibin B and pro-alphaC inhibin in serum from women with epithelial ovarian cancer (EOC) for the first time, and have compared these analytes to inhibin A and total inhibin (inhibin A + B + pro-alphaC) as potential serum markers for EOC in postmenopausal women. Of all the immunoreactive inhibin proteins studied, the best serum marker was pro-alphaC, with 22% of women with EOC having levels that exceeded the range of values in women without EOC. Since CA 125 and pro-alphaC levels were not significantly correlated, combination of these markers resulted in 87% of EOC cases having elevated preoperative serum levels, a 9% increase over CA 125 alone. These data suggest that alpha inhibin secretion, especially pro-alphaC, may be useful in addition to CA 125 as a serum marker for EOC in postmenopausal women.

Activin inhibits basal and androgen-stimulated proliferation and induces apoptosis in the human prostatic cancer cell line, LNCaP.

LNCaP cells, derived from an androgen-sensitive cell line widely employed as an in vitro model of human prostate cancer, have been shown to express activin receptors. Activin is a local regulator of cellular growth, appears to play a key role in mesoderm induction and differentiation during development, and has been implicated in gonadal tumorigenesis. Follistatin, a monomeric glycoprotein that specifically binds and neutralizes activin, is often coexpressed with activin and, thus, modulates the autocrine/paracrine biological activity of this potent growth factor. We tested the hypothesis that LNCaP growth is modulated by the activin/follistatin system. Recombinant human activin A inhibited [3H]thymidine incorporation in a dose-dependent fashion with an ED50 of approximately 0.43 +/- 0.3 nM. Activin (0.1-3 nM) also inhibited dihydrotestosterone (DHT)-stimulated [3H]thymidine incorporation in LNCaP cells. Similarly, recombinant human inhibin A inhibited LNCaP proliferation, but was only 1/100th as potent as activin. Furthermore, activin (3 nM) induced a 3-fold increase in the extent of labeling of low mol wt DNA fragments typical of apoptosis. Activin-induced apoptosis was also indicated by an increase in the number of cells with reduced DNA content, as measured by flow cytometry of activin-treated cells. Both activin-mediated inhibition of cell proliferation and induction of apoptosis could be completely blocked by recombinant human follistatin. Based upon these results using an in vitro model, we speculate that activin functions locally to oppose androgen-driven cell proliferation and, thus, is a key factor controlling prostate growth. Reduced activin biosynthesis, increased follistatin secretion, or signaling defects in the activin receptor system should be further investigated in future studies as potential mechanisms underlying enhanced androgen-independent growth of human prostate cancer cells.