Fshb Knockout Mouse Model, Two Decades Later and Into the Future

Abstract

In 1997, nearly 20 years ago, we reported the phenotypes of follicle-stimulating hormone (FSH) β (Fshb) null mice. Since then, these mice have been useful for various physiological and genetic studies in reproductive and skeletal biology. In a 2009 review titled “FSHβ Knockout Mouse Model: A Decade Ago and Into the Future,” I summarized the need for and what led to the development of an FSH-deficient mouse model and its applications, including delineation of the emerging extragonadal roles of FSH in bone cells by using this genetic model. These studies opened up exciting avenues of research on osteoporosis and now extend into those on adiposity in postmenopausal women. Here, I summarize the progress made with this mouse model since 2009 with regard to FSH rerouting in vivo, deciphering the role of N-glycosylation on FSHβ, roles of FSH in somatic–germ cell interactions in gonads, and provide a road map that is anticipated to emerge in the near future. Undoubtedly, the next 10 years should be an even more exciting time to explore the fertile area of FSH biology and its implications for basic and clinical reproductive physiology research.

As a logical follow-up to my 2009 review on an Fshb KO model a decade ago and into the future, I summarize the applications of the mouse model in the second decade. Two tables summarize key data.

Follicle-stimulating hormone (FSH) is a heterodimeric glycoprotein that is synthesized by gonadotropes in the pituitary (1–3). FSH shares an α-subunit with luteinizing hormone (LH), and both LH and FSH contain a distinct hormone- and receptor-specific β-subunit (1–3). Of the three gonadotropin subunit encoding genes (Cga,Lhb, and Fshb), the Fshb is activated last in mouse gonadotrope development (4, 5). FSH binds to G protein–coupled, seven-helical-membrane-spanning FSH receptors that are expressed on ovarian granulosa and testicular Sertoli cells (1, 3). Together with LH, FSH plays critical roles in reproductive physiology and regulates gonadal growth, gametogenesis, and steroidogenesis (3). Recombinant human FSH expressed in mammalian cell expression systems is widely used in follicular induction and artificial reproductive technology protocols (6–8). Mutations in human FSHB and FSHR were previously considered rare. However, recently several patients have been identified to harbor inactivating or activating mutations in these genes that control FSH signaling pathways (9, 10).

There are no naturally occurring mouse mutants that harbor mutations in Fshb or Fshr genes, so assessing loss of FSH or FSH action (via FSH receptors) alone in reproductive physiology was difficult before production of ligand and receptor knockout (KO) mouse models via gene targeting approaches. A naturally occurring hypogonadal (hpg) mutant exists in which there is a functional deletion of the gonadotropin-releasing hormone (GnRH)–encoding gene (11). Consequently, both LH and FSH are severely suppressed in hpg mutants lacking GnRH peptide (11). One limitation of this model is the difficulty in selectively delineating the in vivo roles of FSH.

In contrast to GnRH, inhibin is a negative regulator of FSH. Inhibin is a gonad-derived heterodimeric peptide that consists of an α-subunit and an activin β subunit [activin A (βA) or activin B (βB)] (12). Activins were also originally purified as gonadal peptides and later identified in many tissues and exist as homodimers of either βA or βB. Mostly, locally produced activins (within the pituitary) stimulate FSH production (13, 14). Although activin AB (dimer consists of βA and βB subunits) also exists, its biological significance is unknown (13, 14). Inhibin antagonizes activin actions by preventing activin from binding to its receptor (12). In 1992, mice lacking inhibin α were reported to develop gonadal sex cord stromal tumors and demonstrate elevated levels of FSH in serum (15). At that time, it was unclear whether or how elevated levels of FSH regulate gonadal tumor development in the absence of inhibin. This necessitated the development of the Fshb null mouse model to directly assess the role of FSH in gonadal tumor development in the absence of inhibin.

One unresolved and controversial issue at that time with regard to FSH action in the male is the need for FSH in spermatogenesis and to what extent FSH and testosterone contribute to spermatogenesis. The consequences of FSH immunoneutralization in male reproductive physiology appear to be species specific (16–21). Moreover, several missense mutations in FSHB gene have been identified in men and correlate to azoospermia and infertility (9, 10, 22). Patients with mutations in FSHR were reported to exhibit hypogonadism, subfertility, or infertility (9, 10, 22). Consistent with these reports, blocking endogenous FSH secretion in normal men causes significant reduction in sperm numbers. This phenotype was reversed by exogenous FSH but not by testosterone (3, 23). These data support those obtained with FSH or FSH receptor immunoneutralization in nonhuman primate models, where spermatogenesis arrest or various degrees of spermatogenic failure were consistently observed (19, 20). In several other studies, immunoneutralization of FSH in immature but not adult rats shows spermatogenic arrest. Studies with hamsters and sheep are in close agreement with those in rats and further confirm the critical need for FSH in maintaining spermatogenesis (19, 20). Similarly, germ cells were suppressed, and recombinant human FSH therapy partially restored spermatogenesis defects in GnRH-immunized rats (24). In hpg mice that lack a functional GnRH, testosterone alone restored spermatogenesis despite low intratesticular testosterone and undetectable serum FSH levels (25–27). A more refined and precise genetic approach was needed to study FSH action in the male.

To understand in vivo actions of exclusively FSH in a mouse model, to define how elevated levels of FSH modulate gonad tumor development in the absence of inhibin, and to settle the controversy of FSH action in the male in an in vivo setting as described earlier, we generated and reported phenotypic characteristics of Fshb null mice in 1997 (28). Fshb null male mice display normal fertility despite reduced testis size and sperm number and motility (28). Fshb null females are anovulatory, display hypoplastic ovaries, and undergo a block in folliculogenesis at the preantral stage (28). Superovulation of immature null females rescues the ovulation defect, indicating that FSH responsiveness is maintained in the absence of FSH (28). In a previous review, I described the significance and applicability of this model to a variety of studies that were published within the first decade of its generation (29). Here, I summarize the progress made in the second decade with this genetic model and describe some potential areas of future research to elucidate FSH actions.

Applications of the Fshb KO Mouse Model (2009–2017)

As mentioned earlier, the Fshb KO mouse model was published in 1997 (28). In the second decade since then, four major applications were reported with this mouse model, as highlighted in Table 1 and briefly described here.

Table 1.

Applications of Fshb Null Mouse Model Since 2009

Application	Approach	Ref.
Gonadotropin rerouting	Genetic rescue	30
In vivo role of N-glycosylation on FSHβ subunit	Genetic rescue	31
Bioactivity testing of recombinant human FSH glycoforms	Pharmacologic rescue	32
Granulosa-oocyte communication	In vitro culture and pharmacologic rescue	33–35
Sertoli cell and SSC maintenance	In vivo analysis and SSC transplantation	36

Abbreviation: SSC, spermatogonial stem cell.

An in vivo genetic strategy to redirect the trafficking of FSH into the LH pathway

In many vertebrate species, FSH is constitutively secreted, whereas LH is released from gonadotropes via a regulated pathway that involves dense core secretory granules (37, 38). The evolutionary significance of why LH and FSH have distinct modalities of secretion patterns from gonadotropes is not understood. Understanding gonadotropin secretion patterns could have several implications in basic and clinical research, including manipulation of gamete number and quality. Toward this end, in vitro transfection and mutagenesis studies of heterologous somatotrope-derived GH3 cells identified that the carboxy terminal heptapeptide in the human LHβ subunit is a critical determinant for trafficking the LH dimer via the regulated pathway (38–40). The LH dimer lacking this peptide sequence on the LHβ subunit was secreted from the constitutive pathway (38–40). Human FSH dimer containing a mutant FSHβ (the LHβ-specific heptapeptide sequence fused in frame) was secreted via the regulated pathway, similar to LH, and was released in response to a secretagogue (41).

To determine the in vivo physiological consequences of rerouting FSH from the constitutive pathway into the regulated pathway, we achieved gonadotrope-specific expression of the HFSHB^WT and HFSHB^Mut transgenes encoding, respectively, the wild-type human FSHβ (hFSHβ) and a mutant FSHβ containing the LHβ carboxy-terminal heptapeptide sequence in pituitaries of independent lines of transgenic mice (30). These were then separately crossed onto and maintained on the Fshb null genetic background via a genetic rescue strategy (42). This rescue strategy permitted an evaluation of the intracellular behavior of the transgene-encoded human FSHβ containing interspecies FSH hybrid dimers in mouse pituitary gonadotropes and in the absence of the endogenous mouse FSH (30). A second advantage of this approach is that it allowed us to monitor the gonadal and fertility phenotypes in Fshb null mice expressing either a HFSHB^WT or HFSHB^Mut transgene (30).

Western blot analysis of pituitary proteins under nondenaturing conditions identified that the mutant hFSHβ subunit efficiently dimerized with the mouse α-subunit, and interspecies hybrid dimer was indeed formed (30). A variety of approaches were used to establish that the mutant FSH entered and was released from the regulated pathway. First, colocalization of mutant FSH and a dense core granule marker Rab27 (which normally colocalizes with LH in gonadotropes) were confirmed and visualized by immunofluorescence and confocal microscopy. Second, mutant but not the wild-type hFSHβ-containing FSH dimer was released in response to acute GnRH treatment. Finally, immunogold electron microscopy indicated that only the mutant FSHβ-containing dense core granule number decreased in response to acute GnRH. Rerouting FSH into the LH pathway did not affect LH secretion dynamics or expression of all other gonadotrope-specific markers, but net LH secreted was reduced, presumably as a result of mutant FSH competing for a limited number of dense core granules per gonadotrope (30).

Mutant FSH rescued Fshb null mice as efficiently as constitutively secreted wild-type FSH. These rescue females became cyclic, and corpora lutea were readily visible in ovarian sections, similar to those observed in the ovaries of mice rescued by wild-type FSH (30). Primordial and other growing follicle numbers did not change, nor were there differences in follicle activation. Moreover, premature ovarian failure as a result of rapid activation of follicles was not observed in mutant FSH-expressing Fshb null mice (30). Fertility per se in terms of litter size was not increased, but most interestingly, rerouted FSH increased follicle survival by preventing atresia (30). This increase resulted in a dramatically increased number of ovulations (six times more than in control mice) per estrous cycle for prolonged periods. Both the efficiency of blastocyst formation in vitro and the ability of fertilized embryos to produce viable offspring when transferred to oviducts of foster mothers were also increased (30). Our recent unpublished observations suggest an increased number of embryos implanted at embryonic day 10.5 in pregnant females expressing the rerouted FSH. Thus, one possible mechanism for normal litter size in these females could be the limited uterine capacity and consequently embryo loss.

Expression of several known FSH-responsive genes was significantly increased in Fshb null mice expressing the rerouted FSH. Serum estradiol levels were not elevated, but progesterone levels increased (30). Collectively, these studies indicated that ovarian cell responses could be fine-tuned and reproductive performance (fertility assessed by production of litters) extended by manipulating in vivo the FSH release pattern from pituitary. Thus, Fshb null mice provided a useful genetic platform to study the intracellular secretory behavior of engineered human FSHB transgenes and test the rescue by constitutively secreted wild-type FSH or mutant FSH released from the regulated pathway. Although FSH is not normally secreted from the regulated pathway, our studies suggest that the rerouted FSH concept could be tested in women undergoing in vitro fertilization protocols (e.g., by delivering FSH in pulses as opposed to a single bolus of injection each time, to have better yield and quality of eggs), where egg number and quality are both important criteria for success.

Identification of the in vivo effects of N-glycosylation on the FSHβ subunit

One of the characteristic features of FSH is the presence of four N-linked glycans (two on α-subunits and two on β-subunits) that are attached to asparagine (Asn) residues in a cotranslational process called N-glycosylation (1, 2, 43). The terminal and negatively charged sialic acid confers microheterogeneity to many glycoproteins, including FSH, and was shown to depend on the estrous cycle stage (1, 2, 43). More recently, macroheterogeneity caused by the presence or absence of N-glycans at the two Asn (N-glycosylation) sites was observed on the hFSHβ subunit. The mechanistic basis for microheterogeneity versus macroheterogeneity in FSH is not known. Purification and biochemical characterization studies indicated that distinct hypoglycosylated FSH glycoforms result from this macroheterogeneity (44–46). These glycoforms possess either one (on Asn7 or Asn24) or no N-linked glycans on the FSHβ subunit compared with the fully glycosylated FSHβ subunit, in which both Asn7 and Asn24 contain N-glycans (44–46). Depending on the molecular mass of the specific β-subunit in an FSH heterodimer, FSH glycoforms are referred to as FSH¹⁵ (no N-glycans on either Asn7 or Asn24), FSH¹⁸ (no N-glycan on Asn7), FSH²¹ (no N-glycan on Asn24), and FSH²⁴ (N-glycans are present on both Asn7 and Asn24) (44–46). It was also reported that these FSH glycoforms are expressed in an age-dependent manner, with the FSH²¹/FSH²⁴ ratio decreasing with age in human pituitaries (47, 48) (Table 2).

Table 2.

Summary of Different FSH Glycoforms

FSH Glycoform	Age Specificity	Target Cell of Action	Ref.
FSH15a	Young	Unknown	31, 45
FSH21/18	Young	Predominantly ovary	47, 49
FSH24	Old	Predominantly bone, adipose, and other tissues (?)	47, 50

^a In vivo studies indicate that FSH¹⁵, although detected in human pituitaries, is assembly and secretion incompetent in mice (31).

The in vivo functional significance of N-glycans on the hFSHβ subunit was directly tested in a set of genetic rescue experiments that used Fshb null mice (31). In the absence of the endogenous mouse FSHβ in this KO model, the behavior and rescue potential of the engineered mutant human FSHB transgene were easily monitored. First, gonadotrope-specific expression of a mutant complementary DNA (cDNA) encoding the human nonglycosylated FSHβ subunit (HFSHB^DGC) (i.e., both the Asn 7 and Asn 24 residues were mutated to Ala residues) was achieved. Second, independent lines of these transgenic mice harboring the mutant cDNA transgene were crossed onto the Fshb null genetic background (31). Finally, Fshb null mice expressing the well-characterized HFSHB^WT transgene served as the rescue positive controls, as described earlier with the FSH rerouting experiments (30, 42). The mutant FSHβ lacking the two Asn-linked N-glycans was appropriately expressed in gonadotropes and colocalized with the endogenous LHβ subunit. Although the nonglycosylated mutant FSHβ subunit was abundantly present when pituitary proteins were analyzed by Western blots under denaturing conditions, very little FSH dimer containing the nonglycosylated mutant FSHβ subunit was present under nondenaturing conditions (31). Consistent with these data, very low levels of FSH dimer containing the mutant nonglycosylated FSHβ subunit were detected in either short-term in vitro pituitary cultures or serum of Fshb null mice carrying the HFSHB^DGC transgene. This suggested that N-glycans on human FSHβ are important for efficient FSH dimer assembly and secretion in vivo (31). Even when secreted at low levels, this mutant FSH dimer consisting of no N-glycans (only two N-glycans as opposed to four N-glycans on wild-type FSH contributed by two each from the α and β subunits) on FSHβ subunit could be rapidly cleared from circulation.

The inefficient dimer assembly, secretion, and presumably rapid clearance of the mutant FSHβ-containing FSH dimer resulted in a failure to rescue Fshb null mice (31). Unlike the full rescue demonstrated by the wild-type hFSHβ containing FSH dimer, the nonglycosylated mutant FSHβ containing FSH dimer did not rescue the testis phenotypes or sperm parameters of Fshb null males (31). Thus, testis size remained small and sperm number and motility reduced, as in the case of Fshb null males. The nonglycosylated mutant FSHβ containing FSH dimer did not rescue Fshb null female mice, which were anovulatory, and their ovarian histology showed only immature follicles with no evidence of corpora lutea or estrous cyclicity (31). Real-time quantitative polymerase chain reaction (PCR) assays on total ovarian cDNA confirmed that the known FSH-responsive genes remain suppressed. Thus, even when low levels of nonglycosylated mutant FSHβ subunit (lacking the two N-glycans) containing FSH were present, they failed to rescue Fshb null mice (31). These studies demonstrate the applicability of Fshb null mice as a useful genetic model to uncover the in vivo roles of N-glycans on the FSHβ subunit.

Although microheterogeneity contributes to serum half-life and signal transduction properties of human FSH, the in vivo biological roles of hypoglycosylated human FSH glycoforms (designated FSH²¹ and FSH¹⁸) that occur as a result of macroheterogeneity were not clear. These age-dependent FSH glycoforms were expressed in a Chinese hamster ovary cell expression system, and the recombinant glycoforms were purified and extensively characterized in vitro (45, 46, 49, 51). Of particular note is the difference in receptor-binding potencies and in vitro activities reported between hypoglycosylated and fully glycosylated FSH glycoforms. Consistent with higher affinity and better binding characteristics in a radioreceptor assay, the hypoglycosylated FSH^21/18 (51) was more potent than the fully glycosylated FSH²⁴, when phosphorylated cyclic adenosine monophosphate response element–binding protein levels and induction of ovarian steroids were evaluated in vitro (49). To test the in vivo bioactivities of the recombinant human FSH glycoforms, Fshb null mice were used in a pharmacological rescue approach (32). Groups of immature female Fshb null mice (at 21 to 23 days of age) were intraperitoneally injected with either phosphate-buffered saline, hypoglycosylated FSH (FSH^21/18), or fully glycosylated FSH (FSH²⁴). First, ovarian gene expression changes were measured in a dose- and time-dependent manner by Taqman real-time PCR analysis. In addition, ovarian protein extracts were tested for levels of known phosphorylated proteins downstream of FSH-FSH receptor mediated signaling. Finally, the phosphorylated cyclic adenosine monophosphate response element–binding protein level was evaluated by immunofluorescence on formalin-fixed ovarian sections. In all the aforementioned assays, hypoglycosylated FSH was as effective as fully glycosylated FSH (32).

The recombinant FSH glycoforms were also tested in Fshb null male mice. Fshb null males at postnatal day 5 were intraperitoneally injected once daily with either phosphate-buffered saline, FSH^21/18, or FSH²⁴. The in vivo bioactivities of FSH glycoforms were evaluated on postnatal day 9 by expression analysis of selected FSH-responsive genes, testis weight gain response, tubule volume, germ cell number, and the number of proliferating Sertoli cells scored by in vivo bromodeoxyuridine labeling (32). In all these assays, the hypoglycosylated FSH^21/18 was more effective than the fully glycosylated FSH²⁴ (32). Contrary to the previous notion that hypoglycosylated forms of FSH^21/18 are not biologically active in vivo, recombinant hypoglycosylated FSH^21/18 elicited bioactivity in Fshb null mice. The bioactivity assays in the male seem to point out that FSH^21/18 is more potent than FSH²⁴. Whether in the female or male, the distinct signaling pathways regulated in vivo by different FSH glycoforms remain to be identified. Nevertheless, Fshb null mice (lacking FSH), as proposed originally in 1998 (42), proved to be useful as in vivo bioassay reagents for testing recombinant hypoglycosylated and fully glycosylated FSH glycoforms in a pharmacological rescue approach (32).

Establishing the intercellular communication between granulosa cells and the oocyte

Fshb null female mice demonstrate a preantral stage developmental block in ovarian folliculogenesis, and therefore FSH is essential for antral follicle formation (28, 30). But these null mice ovulate in response to FSH activity provided via equine chorionic gonadotropin or other FSH analogs in a superovulation protocol and produce two-cell embryos (28, 52). However, the role of FSH in later stages of development or oocyte maturation per se is not clear. To address these questions, Fshb null female mice were used in a series of in vitro and in vivo experiments (33). First, oocyte growth was assessed in terms of oocyte size within the follicles on histological sections and freshly isolated oocytes from follicles of adult mice. These studies indicated that oocytes from Fshb null mouse follicles have nearly identical growth kinetics and reach full size in the absence of FSH (33). Second, consistent with these gross measurements, expression of an oocyte growth factor, Kit ligand-encoding messenger RNA splice forms (designated Kitl1 and Kitl2), was also not altered in granulosa-oocyte complexes isolated from control and Fshb null mice (33). Third, because accumulation of mitochondria in oocytes is necessary for supporting embryonic development, and defects in mitochondria are often linked to female fertility, mitochondrial DNA content was estimated as a surrogate for actual mitochondrial number. These data indicated that accumulation of mitochondrial DNA also occurred normally in the absence of FSH (33). Finally, meiotic competency of oocytes recovered from Fshb null females was tested in several ways, including quantifying the number of oocytes that underwent germinal vesicle breakdown and reached metaphase II after in vitro maturation, visualizing chromosomal alignment on the metaphase II plate of the meiotic spindle, and quantifying the expression of the cell cycle control proteins CDK1 and cyclin B by immunoblots (33). Collectively, these studies indicated that whereas early events of oocyte development (oocyte size and accumulation of mitochondria) were normal, a late event such as acquisition of meiotic competence is delayed in the absence of FSH (33). In another set of superovulation studies, maturation rates of oocytes recovered from Fshb null mice and subsequent blastocyst formation were lower than those in oocytes from control mice.

Because nearly all aspects of oocyte growth occur in the absence of FSH, but less efficiently, the aforementioned studies clearly indicate that FSH mediates somatic cell and oocyte communication that is necessary for optimal oocyte functions. FSH was postulated to mediate this communication between granulosa cells and oocytes via regulation of adhesion or gap junction proteins. Indeed, subsequent studies have found that FSH regulates connexins, the principal gap junction proteins, and cadherins, the cell-cell attachment proteins in both granulosa cells and oocytes (34). FSH also promoted oocyte development in vitro. When FSH activity was provided to Fshb null mice, the density of actin-rich transzonal projections increased between granulosa cells and oocytes and permitted active granulosa cell-oocyte communication within follicles (34).

Fshb null mice also served as a valuable model for establishing the dual role of regulating the LH receptor and epidermal growth factor receptor (EGFR) on granulosa cells and creating the proper follicular environment needed for ovulation (35). Cumulus-oocyte complexes obtained from ovaries of Fshb null females did not undergo expansion when incubated with epidermal growth factor (EGF) for 16 hours. Expression of typical cumulus expansion-specific genes including Has2,Ptgs2, and Tnfaip6 was not induced by EGF either. Oocyte-derived growth factors such as GDF9 and BMP15 were normally expressed, and their signaling in granulosa cells as measured by levels of phosphorylated SMAD proteins was also normal in the absence of FSH (35). Interestingly, loss of FSH led to significantly decreased expression of EGFR and its activity, measured by signaling via EGFR (decreased phosphorylated mitogen-activated protein kinase expression) in granulosa cells. Pharmacologic rescue of Fshb null females by equine chorionic gonadotropin supplementation restored both EGFR expression and activity after 48 hours. Furthermore, EGF-induced expansion genes were also restored in cumulus-oocyte complexes obtained from rescue mice. Together, these elegant in vitro and in vivo studies provide compelling evidence that FSH primes the follicles by regulating EGFR expression in granulosa cells for subsequent LH and EGF action to ultimately coordinate ovulation (35, 53–55).

Role of FSH in Sertoli cell and male germline stem cell maintenance

In the male, FSH acts on Sertoli cells by binding to FSH receptors (56, 57). Sertoli cells constitute the microenvironment for the spermatogonial stem cells (SSCs) and physically support germ cells (58, 59). FSH regulates Sertoli cell number during prepubertal testis development, and Fshb null male mice have reduced Sertoli cell number (60). The self-renewal of SSCs was previously thought to be mediated by FSH-dependent expression of glial cell line–derived neurotrophic factor (GDNF) in Sertoli cells based on in vitro studies (61, 62). To address this question more thoroughly, Fshb null male mice were used in combination with germ cell marker expression and transplantation studies (36). First, immunolocalization studies indicated that whereas the number of GDNF receptor α1⁺ germ cells was the same, the number of cadherin 1⁺ and KIT proto-oncogene receptor tyrosine kinase⁺ germ cells was reduced in the testis in the absence of FSH. These data are consistent with an overall reduction in Sertoli cell and germ cell numbers in the absence of FSH (60). Second, real-time quantitative PCR analysis of somatic cell marker gene expression in the germ cell–depleted testis indicated there was no difference between control and Fshb null groups. Most importantly, Gdnf expression was not affected in the absence of FSH (36). Finally, when GFP reporter-tagged SSCs were transplanted into germ cell–depleted control or Fshb null juvenile or adult male mice (recipients), no differences were found the in number of GFP+ colonies in the testes of control or Fshb null male recipients (36). Thus, these studies of Fshb null mice unequivocally established that FSH does not regulate GDNF expression in Sertoli cells in vivo, and there were no abnormalities in SSCs and their microenvironment in testes in the absence of FSH.

Final Remarks and Looking Into the Future

Fshb null mice have been useful for studies at the pituitary and gonad levels. In the second decade after they were originally described, Fshb null mice were extensively used in genetic and pharmacologic rescue-based experiments (30, 31, 32, 34, 35). Although Fshb null mice were not directly used, recent FSH immunoneutralization (which creates FSH deficiency similar to the loss of FSH in Fshb null mice) studies found that FSH plays a critical role in adipose tissue homeostasis (50). These studies, coupled with previous reports on FSH actions in bone osteoclasts (63), have tremendous clinical significance, particularly in the context of postmenopausal transition in aging women (64). These studies on extragonadal FSH actions particularly in bone and adipocytes are important in the context of estrogen-mediated effects on these tissues. Strikingly, research on rodent models suggested a direct effect of FSH on bone in the presence of estrogen (63, 64). Furthermore, in ovariectomized mouse models, a neutralizing FSH polyclonal antibody prevented adiposity and increased mitochondrial biogenesis and conversion of white to brown adipose tissue (50). Thus, one potential future application of these studies could be to block FSH action or supplement with low-dose estrogen as a therapeutic option for treatment of postmenopausal bone loss and visceral adiposity.

Several exciting lines of research on Fshb null mice are anticipated in the next decade. First, Fshb null mice will continue to be useful to address several issues unresolved in reproductive biology, such as whether the FSH dependency of spermatogenesis is truly species-specific or there are technological refinements needed with regard to timing of blocking FSH action specifically in the adult male. Another issue that could be addressed is the temporal sequence of FSH action on ovarian granulosa cells. Second, more refinements to blocking or enhancing FSH action at desired times will become feasible with the development of mouse models with temporally controlled deletion or expression of Fshb. These refined models should allow us to identify phenotypes of activating or inactivating FSH action in desired tissues or cell types at desired times. These refined genetic models may also reveal whether any compensation occurred in Fshb global KO mice in which FSH signaling is absent from birth. Third, these in vivo models will help elucidate the extragonadal actions of FSH and identify signaling pathways in nongonadal cells. Fourth, FSH-dependent signaling cascades in gonads or extragonadal tissues can be broadly identified via recent proteomic and phosphoproteomic approaches (65–68). Fifth, the in vivo roles of recently characterized various FSH glycoforms (32, 47, 51) can be delineated with Fshb null mouse models. Finally, the future goal will be to extrapolate FSH-based mouse genetic models to the clinic to better diagnose and treat FSH-dependent disorders in patients. Similarly, when personalized medicine becomes more routine and provides insights into variations in FSH actions in individual patients, it may be feasible to make patient-specific mouse models by using the emerging genome editing tools (69–72).

Glossary

Abbreviations:

Asn

asparagine

cDNA

complementary DNA

EGF

epidermal growth factor

EGFR

epidermal growth factor receptor

FSH

follicle-stimulating hormone

Fshb

follicle-stimulating hormone β–encoding gene

GDNF

glial cell line–derived neurotrophic factor

GnRH

gonadotropin-releasing hormone

hFSHβ

human follicle-stimulating hormone β

hpg

hypogonadal

KO

knockout

LH

luteinizing hormone

PCR

polymerase chain reaction

SSC

spermatogonial stem cell

βA

activin A

βB

activin B