Recent studies have found follistatin to be an important regulator of activin bioactivity. Whilst a number of assay formats have been described, all are of limited sensitivity and require the use of isotopes. Many use polyclonal antibodies. Furthermore, a wide range of follistatin preparations have been used as standards, complicating inter-laboratory comparison. We now describe an ultra-sensitive two-site enzyme immunoassay using a pair of mouse monoclonal antibodies raised against follistatin 288. The presence of sodium deoxycholate and Tween 20 in the diluent gave results for total (free and activin-dissociated) follistatin. The assay had a detection limit of <19 pg/ml and recovery of spiked follistatin 288 from amniotic fluid, serum seminal plasma, human follicular fluid and granulosa cell conditioned medium averaged 100.7 +/- 7.5%, 89.1 +/- 5.5%, 98 +/- 4.9%, 96 +/- 7.2% and 123.9 +/- 11% respectively. The intra- and interplate coefficients of variation were < 5%. An excess of activin-A (50 ng/ml) prior to assay did not affect follistatin recovery. Inhibin-A, inhibin-B, activin-A, activin-B and activin-AB had minimal cross-reactivity (<0.3%). However, follistatin 315 had a significant cross-reaction (9.9%). Serially diluted human samples gave dose-response curves parallel to the standard. Pooled human follicular fluid contained high concentrations of follistatin (approximately 242 ng/ml). Follistatin was also found in maternal serum during pregnancy (first trimester approximately 0.8 ng/ml, third trimester approximately 2.8 ng/ml), normal male serum (approximately 0.45 ng/ml), amniotic fluid (sixteen week approximately 3.63 ng/ml, term approximately 0.89 ng/ml), seminal plasma (2.4-30 ng/ml) and human granulosa cell conditioned media (approximately 0.44 ng/ml). Serial serum samples taken throughout the menstrual cycle of ten women showed fluctuating follistatin concentrations (approximately 0.62 ng/ml) with no apparent relationship to the stage of the cycle. Interestingly, pooled serum from postmenopausal women appeared to have higher follistatin levels than any of the normal women (approximately 1.4 ng/ml). The possible presence in certain samples of mixtures of follistatin isoforms with different immunoreactivities poses major problems of interpretation in this and all other current follistatin immunoassays. Further work is needed to identify the major immunoreactive forms in different tissues and fluids. Nevertheless, the new assay has a number of advantages over previous assays and should prove a useful tool for various clinical and physiological studies.
LW Evans, S Muttukrishna and NP Groome
TM Lovell, RT Gladwell, NP Groome and PG Knight
To study the potential involvement of inhibin A (inhA), inhibin B (inhB), activin A (actA) and follistatin (FS) in the recruitment of follicles into the preovulatory hierarchy, growing follicles (ranging from 1 mm to the largest designated F1) and the three most recent postovulatory follicles (POFs) were recovered from laying hens (n=11). With the exception of <4 mm follicles and POFs, follicle walls were dissected into separate granulosa (G) and theca (T) layers before extraction. Contents of inhA, inhB, actA and FS in tissue extracts were assayed using specific two-site ELISAs and results are expressed per mg DNA. InhB content of both G and T followed a similar developmental pattern, although the content was >4-fold higher in G than in T at all stages. InhB content was very low in follicles <4 mm but increased ~50-fold (P<0.0001) to peak in 7-9 mm follicles, before falling steadily as follicles entered and moved up the follicular hierarchy (40-fold; 8 mm vs F2). In stark contrast, inhA remained very low in prehierarchical follicles (< or =9 mm) but then increased progressively as follicles moved up the preovulatory hierarchy to peak in F1 (approximately 100-fold increase; P<0.0001); In F1 >97% of inhA was confined to the G layer whereas in 5-9 mm follicles inhA was only detected in the T layer. Both inhA and inhB contents of POFs were significantly reduced compared with F1. Follicular actA was mainly confined to the T layer although detectable levels were present in G from 9 mm; actA was low between 1 and 9 mm but increased sharply as follicles entered the preovulatory hierarchy (approximately 6-fold higher in F4; P<0.0001); levels then fell approximately 2-fold as the follicle progressed to F1. Like actA, FS predominated in the T although significant amounts were also present in the G of prehierarchical follicles (4-9 mm), in contrast to actA, which was absent from the G. The FS content of T rose approximately 3-fold from 6 mm to a plateau which was sustained until F1. In contrast, the FS content of G was greatest in prehierarchical follicles and fell approximately 4-fold in F4-F1 follicles. ActA and FS contents of POFs were reduced compared with F1. In vitro studies on follicle wall explants confirmed the striking divergence in the secretion of inhA and inhB during follicle development. These findings of marked stage-dependent differences in the expression of inhA, inhB, actA and FS proteins imply a significant functional role for these peptides in the recruitment and ordered progression of follicles within the avian ovary.
DS Tannetta, SA Feist, EC Bleach, NP Groome, LW Evans and PG Knight
Active immunization of ewes against inhibin (IMM) consistently increases ovulation rate but this response is not always accompanied by the expected rise in plasma FSH. Inhibin-related molecules also have local auto/paracrine effects within the ovary and the ovulatory response to IMM could be due to neutralization of one of these effects, independent of changing FSH levels. To investigate this, ovaries were collected from long-term IMM (n = 6) and control (CON; n = 8) ewes killed 48 h after progestagen withdrawal (late follicular phase) and all follicles > or = 3 mm were recovered to determine intrafollicular levels of inhibin A, activin A and follistatin by specific two-site immunoassay and oestradiol and testosterone by radioimmunoassay. Blood samples were collected to assess plasma FSH, oestradiol and inhibin antibody titres. Although plasma FSH levels were similar in IMM and CON ewes, IMM ewes had approximately 3-fold more follicles > or = 3 mm (P < 0.0001) and approximately 3-fold more oestrogenic follicle (P < 0.001) than CON ewes. Compared with CON ewes, follicles from IMM ewes had much higher concentrations of activin A (approximately 6-fold; P < 0.001) and inhibin A (approximately 3-fold; P < 0.001) but only slightly more follistatin (approximately 1.4-fold; not significant). The activin A:follistatin ratio in follicles from IMM ewes (approximately 1:1) was significantly higher (P < 0.001) than in follicles from CON ewes (approximately 0.3:1). Levels of inhibin antibody measured in follicular fluid (FF) from IMM ewes were similar to plasma levels. Given that activin A has been shown previously to up-regulate FSH receptors and aromatase activity in rat granulosa cells, the increase in intrafollicular activin A, unaccompanied by a rise in the concentration of its binding protein (follistatin), could explain how long-term IMM enhances follicle development and ovulation rate without necessarily promoting a sustained increase in FSH secretion.
JA Keelan, KW Marvin, TA Sato, LM McCowan, M Coleman, LW Evans, NP Groome and MD Mitchell
To investigate labour-associated changes in production of activin and related hormones by gestational tissues we prepared extracts from amnion, choriodecidual and placental tissues delivered at term before labour (TNL; n=15), at term after spontaneous labour (TSL; n=15) or preterm (PTD; n=31) and measured concentrations of inhibin A, activin A and follistatin by ELISA. Activin concentrations in placental tissues were significantly (Mann-Whitney U-test; P<0.05) elevated with term labour (pg/mg protein, median; 1313 vs 2591), but in the PTD tissues concentrations were lower than those delivered spontaneously at term (3650 vs 2649). Inhibin concentrations also increased with term labour in the placenta (480 vs 686), but paradoxically decreased in amnion (188 vs 64) and choriodecidua (657 vs 358). Little or no significant changes in follistatin concentrations were observed. Concentrations of all three proteins were significantly correlated between amnion and choriodecidual tissues, and were significantly correlated with each other in most tissues (Spearman's ranked correlation; P<0.05). The activin:inhibin ratio in term amnion and choriodecidual tissues was increased 2 to 3-fold (P<0.0005 by Mann-Whitney U-test) after term labour, with similar trends also observed in the activin:follistatin ratio in placental tissue. These data suggest that a modest increase in placental activin and inhibin production may occur with labour at term. In addition, an increase in activin bioactivity may occur with labour, potentiating any paracrine effects of activin during parturition. The data, however, do not support an association between increased intrauterine activin biosynthesis and preterm delivery.
MO Suescun, MO Suescun, L Lustig, RS Calandra, RS Calandra, NP Groome and S Campo
The aim of the present study was to evaluate inhibin secretion in rats with autoimmune orchitis. As we have previously described, experimental autoimmune orchitis (EAO) induced in rats by active immunization with testis homogenate and adjuvants is characterized by an interstitial mononuclear cell infiltrate and sloughing of the germinal epithelium. At 120 days after the first immunization 60% of the rats exhibited a severe orchitis with large areas of aspermatogenic seminiferous tubules in which only spermatogonia and Sertoli cells with cytoplasmic vacuolization remained attached to the tubular wall. None of the untreated (N) or control (C) rats revealed pathological alterations. Sixty percent decrease in testis weight was observed in rats with EAO compared with N or C groups. A 3-fold increase in serum FSH levels was observed in rats with EAO compared with N or C groups (19.8+/-3.7 vs 5.6+/-0.3 and 5.9+/-0.1 ng/ml respectively). A significant decrease in inhibin B levels was observed in rats with EAO when compared with N or C groups (40+/-4.6 vs 207+/-38.8 and 221.4+/-28.6 pg/ml respectively). An inverse correlation between inhibin B and FSH serum levels and a direct correlation between inhibin B and testis weight were found. Strong expression of the inhibin alpha-subunit in Sertoli cells of untreated and control rats was observed; this subunit was undetectable or poorly detectable in rats with orchitis. Positive staining for the inhibin alpha-subunit was also observed in Leydig cells of all groups studied. In conclusion, using a model of autoimmune orchitis our results show that circulating inhibin B levels and inhibin alpha-subunit expression in Sertoli cell cytoplasm closely correlate with the degree of damage of the germinal epithelium.
AE Drummond, M Dyson, E Thean, NP Groome, DM Robertson and JK Findlay
The contribution of specific follicle populations to dimeric inhibin production and inhibin subunit mRNA expression by the rat ovary has been investigated in two model systems, granulosa cells isolated from 25-day-old diethylstilboestrol (DES)-treated rats and post-natal rat ovaries, dispersed in culture or whole ovaries, using specific two-site immunoassays and 'real time' PCR. Media from FSH-stimulated granulosa cell cultures fractionated by gel filtration and RP-high performance liquid chromatography revealed two predominant peaks of alpha subunit activity which were attributed to alpha subunit and 31 k dimeric inhibin-A. The corresponding inhibin-B levels were low. FSH stimulation did not alter the ratio of inhibin-A:alpha subunit produced by granulosa cells. All three inhibin subunit mRNAs were expressed by granulosa cells, with eight-fold more alpha subunit mRNA relative to either of the beta subunits. Administration of DES to immature rats prior to the isolation of granulosa cells from the ovary led to beta(A) and beta(B) mRNA expression being down-regulated in the absence of any significant change in alpha subunit expression by the granulosa cells. Inhibin-A, -B and -alpha subunit were produced by basal and stimulated cultures of ovarian cells prepared from 4-, 8- and 12-day-old rats, indicating that primary, preantral and antral follicles contribute to total inhibin production. Consistent with these results, follicles within these ovaries expressed all three inhibin subunit mRNAs, with maximal expression observed in the ovaries of 8-day-old rats. The appearance of antral follicles in the ovary at day 12 led to a decline in the mRNA levels of each of the subunits but was most evident for the beta subunits. There was a profound influence of secondary preantral follicles on dimeric inhibin-A production, with FSH stimulation increasing inhibin-A relative to alpha subunit levels in cultures of ovarian cells prepared from 8-day-old rats. Thus, preantral follicles exposed to FSH contribute significantly to beta(A) subunit production by the ovary. In contrast, primary and preantral follicles did not produce inhibin-B in response to FSH stimulation. Transforming growth factor-beta (TGF-beta) enhanced, in a time-dependent manner, the production of the inhibin forms by ovarian cells in culture, although inhibin-B production was not responsive until day 8. The simultaneous treatment of ovarian cell cultures with FSH and TGF-beta elicited the greatest increases in production of all the inhibin forms. In summary, ovaries of 4-, 8- and 12-day-old rats expressed inhibin subunit mRNAs and produced dimeric inhibin-A and -B and free alpha subunit. Preantral follicles (day-8 ovarian cell cultures) were particularly sensitive to stimulation by FSH and TGF-beta and had a substantial capacity for inhibin production. The production of oestrogen by follicles may be instrumental in regulating inhibin production given that beta subunit mRNA expression was down-regulated by DES. The mechanisms by which inhibin-A and inhibin-B are individually regulated are likely to be similar during the post-natal period, when folliculogenesis is being established, and diverge thereafter, when inhibin-A becomes the predominant form in the fully differentiated ovary.
K Ohshima, H Kishi, M Itoh, G Watanabe, K Arai, K Uehara, NP Groome and K Taya
Plasma concentrations of inhibin pro-alphaC, inhibin A and inhibin B were determined by enzyme-linked immunosorbent assay at 6 h intervals throughout the 4-day oestrous cycle of the golden hamster. Plasma concentrations of follicle-stimulating hormone (FSH) and oestradiol-17beta were also measured by radioimmunoassay during the oestrous cycle. Plasma concentrations of inhibin A increased from the early morning of day 1 (day 1=day of ovulation) and reached plateau levels at 0500 h on day 2. An abrupt increase in plasma concentrations of inhibin A was found at 1700 h on day 4, when the preovulatory FSH surge was observed. An increase in plasma concentrations of inhibin B occurred on day 1 and reached plateau levels at 1700 h on day 1. The levels remained elevated until 0500 h on day 4 and declined gradually by 2300 h on day 4. Plasma concentrations of inhibin pro-alphaC gradually increased with some fluctuation from day 1 to 1700 h on day 4 and then declined. Significant negative relationships were noted between plasma FSH and both dimeric forms of inhibin from day 1 to day 3. Significant positive relationships were found between plasma oestradiol-17beta and inhibin A or inhibin pro-alphaC throughout the oestrous cycle. In contrast, no significant relationship was found between plasma oestradiol-17beta and inhibin B. These findings suggest that both dimeric forms of inhibin play a role in the regulation of FSH secretion during follicular development. These findings also suggest that inhibin pro-alphaC could be secreted primarily by large follicles, and early atretic follicles could also be responsible for inhibin pro-alphaC secretion. On the other hand, the secretory pattern of dimeric inhibins might shift from inhibin B to inhibin A with follicular development.
K Ohshima, H Kishi, M Itoh, KY Arai, G Watanabe, K Arai, K Uehara, NP Groome and K Taya
The changes in plasma concentrations of inhibins A, B and pro-alpha C were determined in the cyclic golden hamster during follicular atresia induced with antiserum against luteinizing hormone releasing hormone (LHRH-AS) at 1100 h on day 4 (day 1=day of ovulation). Follicular status in the ovary was also studied by determining the number of follicles ovulating in response to human chorionic gonadotrophin (hCG) injection. The time-courses of changes in plasma concentrations of inhibins A, B and pro-alpha C were different from each other during induced follicular atresia and subsequent follicular development. Plasma concentrations of inhibin A decreased to 58.6% of initial values by 24 h after LHRH-AS treatment, and then remained relatively low until at least 60 h later. Plasma concentrations of inhibin B decreased to 64.2% of the initial values by 18 h after LHRH-AS treatment and remained at basal values for 36 h, but increased abruptly to greater than initial values at 42 h after the treatment. Plasma concentrations of inhibin pro-alpha C increased at 6 and 12 h, decreased suddenly to 21.9% of the initial values by 24 h after LHRH-AS treatment, and then gradually increased until 60 h after LHRH-AS. The number of follicles responding to hCG decreased gradually between 0 and 30 h after LHRH-AS, when no ovulations were observed, and then gradually increased until 60 h. The changes in follicular ovulatory responses to hCG correlated with the plasma profile of inhibin A throughout the experiment. These results suggest that inhibin A is mainly secreted by large antral follicles. In contrast, during the subsequent follicular development, the plasma concentration of inhibin B increased earlier than that of inhibin A. These results suggest that inhibin B is secreted by small and large antral follicles. Plasma concentrations of inhibin pro-alpha C were high at a time when plasma concentrations of oestradiol-17 beta had already decreased, indicating that inhibin pro-alpha C is secreted not only from healthy follicles but also from early atretic antral follicles.
D D'Antona, FM Reis, C Benedetto, LW Evans, NP Groome, DM de Kretser, EM Wallace and F Petraglia
Activin A levels are elevated in maternal serum of pregnant women with hypertensive disturbances. Because follistatin is a circulating binding protein for activin A, the present study was designed to evaluate whether serum follistatin and activin A levels also change in patients with hypertensive disorders in the last gestational trimester. The study design was a controlled survey performed in the setting of an academic prenatal care unit. Healthy pregnant women (controls, n=38) were compared with patients suffering from pregnancy-induced hypertension (PIH, n=18) or pre-eclampsia (n=16). In addition, the study included a subset of patients with pre-eclampsia associated with intrauterine growth restriction (IUGR, n=5). Maternal blood samples were withdrawn at the time of diagnosis (patients) or in a random prenatal visit (controls), and serum was assayed for follistatin and activin A levels using specific enzyme immunoassays. Hormone concentrations were corrected for gestational age by conversion to multiples of median (MoM) of the healthy controls of the same gestational age. Follistatin levels were not different between controls and patients, while activin A levels were significantly increased in patients with PIH (1.8 MoM), pre-eclampsia (4.6 MoM), and pre-eclampsia+IUGR (3.2 MoM, P<0.01, ANOVA). The ratio between activin A and follistatin was significantly increased in patients with PIH (1.5 MoM) and was further increased in patients with pre-eclampsia (4.5 MoM) and in the group with pre-eclampsia+IUGR (2.6 MoM). Follistatin levels were positively correlated with gestational age in control subjects (r=0. 36, P<0.05) and in patients with PIH (r=0.46, P<0.05) or pre-eclampsia (r=0.61, P<0.01), while activin A correlated with gestational age only in the healthy control group (r=0.69, P<0.0001). The finding of apparently normal follistatin and high activin A levels in patients with PIH and pre-eclampsia suggests that unbound, biologically active, activin A is increased in women with these gestational diseases.
KJ Turner, S Macpherson, MR Millar, AS McNeilly, K Williams, M Cranfield, NP Groome, RM Sharpe, HM Fraser and PT Saunders
The biosynthesis of oestrogens from androgens is catalysed by the aromatase complex, an essential component of which is the aromatase cytochrome P450 (P450 arom) protein. Expression of a functional P450 arom is essential for normal fertility in males and females and the sequence of the protein is highly conserved. We have raised a new monoclonal antibody against a conserved peptide and validated it on fixed tissue sections of the rat, common marmoset (Callthrix jacchus) and human. The monoclonal antibody was used successfully for Western analysis and specifically reacted with a 55 kDa protein in microsomal extracts. On sections of ovaries in all three species, expression in follicles was specific to the mural granulosa cells of antral follicles and was present in corpora lutea. In the human and marmoset, staining of luteal cells was markedly heterogeneous and did not appear to vary consistently with the stage of the cycle. The intensity of immunostaining was elevated in corpora lutea from pregnant rats and following human chorionic gonadotropin rescue in the human. In the testis, the highest levels of expression were observed in the Leydig cells within the interstitium. In adult rat and marmoset, and possibly also in the human, some P450 arom was associated with the cytoplasm surrounding elongate spermatids but other germ cells were immunonegative. In conclusion, a new monoclonal antibody specific for P450 arom recognises the protein in rodent, primate and human. Its ability to work on fixed tissue sections will facilitate identification of individual cells expressing P450 arom within complex tissues.