While GnRH pulsatility provides one mechanism for differential FSH and LH regulation, other factors are clearly involved. A myriad of in vivo and in vitro studies clearly indicate a role for activins, inhibins, and follistatins in the selective regulation of FSH (3). As their names suggest, activins stimulate, while inhibins attenuate FSH synthesis and secretion. Both activins and inhibins are members of the transforming growth factor-p (TGFP) protein family. Activins are comprised of two closely related subunits (PA and PB) and form activin A and activin B by homodimerization (P A: p A and PB: PB) or activin AB by heterodimerization. Inhibins A and B are heterodimers (a:pA and a:PB) of a unique a-subunit and one of the two P-subunits they share with the activins. Follistatin is a single-chain peptide, structurally unrelated to the activins and inhibins. Follistatin inhibits FSH synthesis and secretion by binding activins irreversibly and blocking their actions.
Activins bind to one of two constitutively active serine/threonine kinase receptors, activin type II and type IIB receptors (ActRIIA and ActRIIB) (20). Ligand binding causes recruitment and transphosphorylation of an activin type I receptor, ActRIB (also known as ALK-4). ALK-4, like the type II receptors, is a serine/threonine kinase. A second type I receptor, ALK-2 or ActRI, is expressed in pituitaries (21), but recent data suggest that it does not play a role in activin signaling (22). Upon phosphorylation, ALK-4 is activated and phosphorylates one of two downstream intracellular signaling molecules, Smad2 or Smad3. Phosphorylated Smad2 or Smad3 multimers then hetero-oligomerize with another member of the Smad family, Smad4, in the cytoplasm, and the complex translocates to the nucleus. Once in the nucleus, the Smad2/Smad4 or Smad3/ Smad4 complex interacts with a DNA-binding partner to activate transcription by binding to cis-acting elements in the promoters of target genes. While it is assumed that the FSHp gene is a downstream target of activin signaling, this has not been shown directly. However, because activin stimulates increases in FSHp primary transcript levels within 30 min in perifused rat pituitaries, this rapid response suggests that FSHp may be a direct target (23).
How inhibins regulate FSH is less well understood. It is not known whether inhibins downregulate FSH by competing with activin for binding to activin receptors in the pituitary (24,25) or whether they act through their own receptor and signaling pathway. Support for the latter mechanism comes from two recent reports indicating the presence of high-affinity inhibin-binding proteins in murine ovarian tumors and in ovine pituitaries (26,27). Molecular characterization of these proteins will aid substantially in the identification and delineation of the inhibin signal-transduction cascade. More detailed analyses will identify how activin and inhibin signaling interact to affect FSH synthesis and secretion.
A variety of genetic models have been created to examine the roles of the activins, inhibins, and follistatins in reproduction and development (28-30). Next, we summarize how each of these models has contributed to our understanding of FSH regulation.
The inhibin a-subunit was deleted by homologous recombination in ES cells (31). Elimination of the a-subunit precludes the ability to produce both inhibin A and B, while leaving production of all forms of activin intact. A priori, one would expect FSH levels to be elevated in these mice because of the potent inhibitory effects of inhibin on FSH. In fact, in both male and female homozygous mice, FSH levels are significantly elevated. Inhibin-deficient mice also show significant increases in both serum activin A and B (32-34). Therefore, it is likely that the increases in serum FSH are attributable to decreased pituitary and gonadal inhibin and subsequent increases in serum and pituitary activin levels.
Inhibin-a also appears to be a tumor-suppressor gene, because virtually all inhibin-deficient mice develop gonadal sex-cord stromal tumors. Interestingly, elevated FSH levels may contribute to tumorigenesis in these mice. Inhibin- and GnRH-deficient mice have been created by interbreeding inhibin-a knockout mice with hpg mice. As reviewed above, hpg mice have very low levels of circulating gonadotropins. Mice homozygous for both the inhibin-a and hpg mutations do not develop gonadal or adrenal (33) tumors for at least 1 yr (35). More recent data indicate that FSH modifies tumor growth rather than causing tumorigenesis per se (36). That is, transgenic mice overexpressing human FSH at very high levels do not develop gonadal tumors, and inhibin-deficient/FSHp-deficient double-mutant mice still develop ovarian and testicular tumors, although at a slower rate than mice deficient in inhibin alone.
Mice carrying a transgene with 6 kb of the mouse inhibin-a promoter driving the expression of SV40 T-antigen (TAg) develop gonadal tumors of Leydig- and granulosa-
/theca-cell origin between 5 and 8 mo of age (37,38). When these mice are crossed with hpg mice to deplete endogenous gonadotropins, no tumors develop (39). These data suggest that the gonadotropins (particularly FSH) drive tumorigenesis through activation of the inhibin-a promoter, which in turn drives TAg expression, and/or that the gonadotropins have a direct effect on the inhibin-a-TAg-positive cells. This scenario is different than the case in inhibin-deficient mice in which tumor growth occurs when inhibin is removed and gonadotropin levels are elevated.
An inhibin-a overexpression model has also been generated by fusing the rat inhibin-a precursor downstream of the mouse metallothionein-I promoter (MT-a mice) (40). MT-a females have reduced serum FSH, but increased serum testosterone and LH levels. In addition, these females are fertile but produce smaller litters, in part because of a deficit in ovulation. A majority of MT-a females eventually develop unilateral or bilateral ovarian cysts. Male MT-a mice also have lower levels of FSH compared to controls, but are fertile and produce normal-sized litters. The testes decline in size as the animals age, and the seminiferous tubule volume and sperm counts are lower than in wild-type controls. No other testicular abnormalities were reported. These data indicate that, predictably, inhibin overexpression results in decreased FSH levels.
As discussed above, the inhibins and activins share p-subunits, and models of PA and PB deficiency as well as double pA/pB-knockouts have been produced. Because PB deletion results in deficiency in inhibin B and activin B and AB, it is difficult to predict a priori the effects on FSH synthesis and secretion. These mice are viable and are fertile, although some pB-knockout mice show defects in eyelid fusion depending on the strain background (41). 129/Sv/C57BL/6 hybrid background pB-deficient males have slightly elevated serum FSH levels relative to wild-type controls. A similar trend is observed in females, but the difference is not statistically significant. These data indicate that relatively normal FSH regulation can occur in the absence of the pB-subunit. This may arise because of compensatory changes in PA production. Indeed, PB-deficient mice show a significant upregulation of PA protein in ovarian tissue through a posttranscriptional mechanism (41).
Male PB-deficient mice breed normally, but females have some reproductive deficits. Ovarian morphology appears normal, but compared to heterozygotes or wild-type controls, homozygous-null females show a significant increase in gestation time and an impairment in the onset of labor. In addition, fetuses of PB-deficient females that survive birth, die shortly thereafter because of malnutrition caused by a nursing defect in the PB-deficient mothers.
Unlike the pB-null mutants, pA-deficient mice are not viable and die within the first 24 h after birth (42). These animals fail to develop whiskers and show craniofacial defects, including a lack of upper incisors and lower molars. Many animals also show cleft or incomplete palate formation, and therefore fail to suckle after birth. Mice completely devoid of activins and inhibins have also been generated by interbreeding PA- and PB-null heterozygotes and then interbreeding the compound heterozygous offspring to produce compound homozygotes lacking both PA and PB (42). Not surprisingly, the double mutant mice are not viable and die shortly after birth probably because of cran-iofacial defects similar to those observed in PA mutants. The defects appear to represent an additivity of the individual subunit mutants, including an eyelid-fusion defect, lack of whiskers, palate defects, and tooth defects. These data suggest that within the individual PA or PB mutant mice, there is little or no compensation by the preserved ligand for the deleted one. Unfortunately, because the mice die so close to parturition, the roles of activin A, activin AB, and inhibin A in FSH regulation cannot be assessed from these model systems.
As outlined above, activins act on target cells by first binding to one of two type II receptors. Both ActRIIA and ActRIIB are expressed in the pituitary gland and, therefore, both provide substrates for activin action in adulthood (21,43). At least in adult rats, pituitary expression of ActRIIA greatly exceeds that of ActRIIB (44); therefore, activin's the effects of activin on FSH in adult animals may be mediated principally through ActRIIA. Consistent with this hypothesis, targeted deletion of ActRIIA produces mice that are viable (although underrepresented at weaning, because of hypoplasia of the mandible in some newborns), but have reproductive defects (45). Mutant males are fertile, but FSH levels in pituitary gonadotropes and in serum are significantly reduced relative to controls. ActRIIA-deficient males also show delayed fertility, small testes, and seminiferous-tubule diameter, but normal spermatogenesis. In agreement with the data presented here for the FSHp mutants, FSH does not appear to be necessary for spermatogenesis in mice. In contrast to males, female ActRIIA-deficient mice are infertile. They show decreased levels of serum FSH, thin-walled uteri, small ovaries, higher than normal levels of follicular atresia, and a low incidence of corpora lutea. These data are consistent with the hypothesis that many of activin's actions on FSH synthesis and secretion in adult mice are mediated via ActRIIA. In addition, any compensatory changes in ActRIIB expression are insufficient to maintain normal wild-type FSH regulation.
ActRIIB-deficient mice have also been generated by targeted deletion in ES cells (46). Animals homozygous for the mutation show a greater number and variety of deficits than the ActRIIA mutant mice, including cardiac and vertebral patterning defects. None of the homozygous mutant mice on an inbred 129/Sv background survive to weaning age. In contrast, 30% of the homozygous mice on a hybrid background (129/Sv/C57BL/6) are viable, and at least the males are fertile. There are no published data regarding serum FSH levels or gonadal function in these animals so it is not yet clear how ActRIIB-deficiency affects FSH regulation in adult mice.
Activin signaling is dependent upon transphosphorylation and activation of a type I receptor upon ligand binding to the type II receptor. Therefore, one would predict that mutations in the activin type I receptor may result in reproductive and other defects. The ActRIB (ALK-4) receptor has been deleted in ES cells, and homozygous mutants were generated by breeding of heterozygous mutants. ActRIB-deficient mice die during embryonic development, and the receptor appears to be required for gastrulation (47,48). As a result of the lethality, the role of ALK-4 in activin regulation of FSH has not yet been confirmed. In the future, genetic models in which ActRIB is deleted selectively in gonadotropes may help to clarify this issue.
Once activated, type I receptors phosphorylate and activate the intracellular signaling proteins known as Smads. Smad2 and Smad3 have been identified as downstream phosphorylation targets of activated type I receptors in the TGF^- and activin-signaling pathways. Homozygous Smad2-null mutant mice die during embryogenesis (49-51). At least two lines of Smad3-null mutant mice have been generated by homologous recombination. Mice in which exon 8 is deleted are viable, but die between 1 and 8 mo of age because of compromised immune function (52). Mice in which exon 2 of Smad3 is deleted are also viable, but develop metastatic colorectal adenocarcinomas between 4
and 6 mo of age (53). The cause of the difference in phenotypes between these two lines of Smad3-null mice is not clear, but it is possible that in at least one of the lines—a truncated, but functional Smad3—may still be produced (52). Once phosphorylated, Smad2 or Smad3 form complexes with a mediator Smad: Smad4. Similar to Smad2-null mutants, Smad4-deficient mice die early during embryonic development prior to gastru-lation (54,55).
The phenotypes of most of the Smad mutants preclude an assessment of Smad regulation of FSH synthesis in adult mice. However, Smad3-null mice lacking exon 2 are fertile (53), so some assessment of FSH synthesis and secretion prior to tumor formation should be possible in these animals. The fact that female homozygotes breed successfully suggests that FSH function is not severely compromised. The generation of gonadotrope-specific Smad2- and Smad4-deficient mice and double mutants also lacking Smad3 may be helpful in assessing the roles of these signal transducers in activin-induced FSH production in adult mice.
Finally, as described above, follistatins inhibit FSH production by binding and inactivating activins. Therefore, disruption of follistatin function may be predicted to increase FSH synthesis by increasing activin availability. Follistatin null-mutants, created by homologous recombination in ES cells, display a variety of developmental defects and die shortly after birth (56). Homozygous mutants have craniofacial defects, are growth-retarded, have shiny, taut skin, and show multiple skeletal and muscular defects. Most of these defects do not phenocopy those of activin-deficient mice, and therefore suggest that follistatins still have unidentified roles in development, perhaps interacting with other members of the TGFp family. Because of their early death, effects of follistatin-deficiency on FSH function cannot be assessed in these mice.
More recently, several lines of follistatin-overexpressing mice were created by expressing the mouse follistatin gene downstream of the mouse metallothionein-I promoter (MT-FS) (57). In the line with the highest and most widespread expression pattern (line 4), FSH levels are significantly decreased in both males and females relative to controls. Line 4 males also have the smallest testes and some are infertile. All females from line 4 are infertile, and have thin-walled uteri, small ovaries, and disrupted folliculogenesis. Taken alone, these data imply that overexpression of follistatin may lead to a sequestration of activin and concomitant decline in FSH, resulting in the small testes and disrupted estrous cyclicity in these mice. However, unlike other models in which FSH is decreased and males are still fertile (such as the ActRIIA- and FSH^-null mice), over one-half of the males in MT-FS line 4 are infertile. In addition, in some other lines of mice in which FSH levels are not affected, testes mass is decreased, and fertility is compromised. These data suggest that overexpression of follistatin within the gonads (confirmed by RNA blot analysis) and local abrogation of activin action (and possibly other TGF-P family members) is the primary cause of reproductive deficits in these mice.
Was this article helpful?
Once your pregnancy is over and done with, your baby is happily in your arms, and youre headed back home from the hospital, youll begin to realize that things have only just begun. Over the next few days, weeks, and months, youre going to increasingly notice that your entire life has changed in more ways than you could ever imagine.