Rerouting of follicle-stimulating hormone secretion and gonadal function

Abstract

Follicle-stimulating hormone (FSH) and luteinizing hormone (LH) are synthesized in the same pituitary cell, i.e., gonadotrope. They both consist of a common α-subunit that is noncovalently assembled with a hormone-specific β-subunit in gonadotropes. The heterodimers exit gonadotropes through distinct modes of trafficking and secretion. The FSH is constitutively secreted, whereas LH is secreted in pulses through the regulated pathway that involves dense core granules. Based on several in vitro mutagenesis studies, the carboxy terminus heptapeptide of human LH-β subunit is identified as a gonadotrope sorting determinant. When heptapeptide is genetically fused to human FSH-β subunit and the mutant transgene expressed on a Fshb null genetic background, the rerouted FSH mutant dimer enters the LH secretory pathway, stored in dense core granules, coreleased with LH on gonadotropin releasing hormone stimulation and rescues Fshb null mice as efficiently as the constitutively secreted wild-type FSH. The rerouted FSH markedly suppresses follicle atresia and significantly enhances ovulations per cycle and prolongs the female reproductive life span. Gonadotropin rerouting is emerging as a novel paradigm to treat ovarian dysfunction in women, and may explain the origins of ovarian cyclicity as well as provide clues to understand gene and protein networks that maintain optimal ovarian function throughout the female reproductive life span.

The gonadotropins, luteinizing hormone (LH) and follicle-stimulating hormone (FSH), are heterodimeric glycoproteins. They both share a common α-subunit that is noncovalently assembled with a hormone-specific β-subunit in gonadotropes of the anterior pituitary (1-3). The dimers are secreted through either a constitutive (FSH) or a regulated pathway (LH) from gonadotropes (1-3). The physiological regulators that are critical for gonadotropin production and secretion include the hypothalamic-derived gonadotropin releasing hormone (GnRH) and gonadal steroids (estrogen, progesterone, and testosterone) (1-3). Additionally, peptides that belong to the TGF-β superfamily that are locally produced within the pituitary or gonads, such as activins, inhibins, and follistatin, exclusively regulate FSH homeostasis (1-3). Activins promote and inhibins suppress FSH and follistatin prevents activin binding to its receptor, activin receptor-2 (Acvr2) which is predominantly expressed in gonadotropes. Mice lacking either one or both alleles of Acvr2 demonstrate reduced levels of FSH compared with age-matched controls (4).

Both FSH and LH bind to their cognate G-protein coupled receptors follicle -stimulating hormone receptors (FSHRs) and luteinizing hormone receptors, which are expressed on ovarian granulosa and thecal cells and regulate gonadal growth, gametogenesis, and steroidogenesis (2, 3). Low levels of FSH are mostly constitutively secreted and required for follicle growth. Moreover, LH is secreted in a pulsatile manner and a maximal amplitude LH peak triggers ovulation. A distinct secondary FSH surge also occurs at the time of ovulation in several species. Although gonadotropin actions in female reproductive physiology have been extensively studied in a number of species, mechanisms of gonadotropin trafficking and secretion are less understood. Furthermore, how gonadotropin secretion patterns affect gonadal function is not known.

GONADOTROPIN TRAFFICKING AND SECRETION—IN VITRO STUDIES

Mutagenesis of gonadotropin subunits and in vitro studies revealed their intracellular behavior in terms of subunit assembly and secretion of gonadotropin heterodimers containing these mutant subunits (5, 6). When polarized Madin-Darby canine kidney cells are co-transfected with cDNAs encoding human α-subunit and a wild-type (WT) or a mutant (Mut) LH-β subunit lacking the carboxy terminus heptapeptide (HP), LH dimer containing the WT HP exits through the basolateral side, whereas mutant LH dimer lacking the HP exits from the apical side (7). Moreover, the secretion rate of Mut LH-β subunit containing LH heterodimer was low compared with that of LH dimer containing the WT LH-β subunit from somatotrope-derived growth hormone-producing-3 (GH₃) cells (7).

Most importantly, Mut LH-β subunit containing LH dimer exits from GH₃ cells constitutively in contrast to that containing the WT LH-β subunit, which exited through the regulated pathway on stimulation by a hyperpolarizing agent such as barium chloride (8, 9). Interestingly, FSH-β containing this HP of LH-β subunit was redirected in GH₃ cells and also exited like LH through the regulated pathway. These in vitro studies identified that HP of LH-β subunit is a critical determinant for regulated sorting of LH in GH₃ cells (8, 9).

Several limitations are noted with the above in vitro transfection studies. First, a heterologous nongonadotrope cell line (GH₃) was used in these studies. Second, only the assembly and secretion into the cell culture medium were analyzed. Third, biologic consequences of retrafficked mutant dimers were not analyzed in a physiological context. Thus, it was necessary to achieve gonadotropin rerouting in vivo in a whole animal model to truly assess whether target cells, i.e., gonadal somatic cells, are able to decode constitutive vs. pulsatile patterns of gonadotropin signal inputs.

FSH REROUTING IN VIVO

Gonadotropin rerouting could be achieved in vivo in 2 distinct ways. In one approach, FSH could be redirected from its typical constitutive mode of secretion into the regulated pulsatile pathway by which LH is normally secreted. In a second approach, rerouting of LH from the pulsatile pathway into the constitutive pathway, by which FSH is normally secreted, could be achieved. Using the first approach, in vivo rerouting of FSH was achieved as described below.

A Genetic Approach to Redirect FSH Trafficking in Gonadotropes in vivo

To redirect FSH trafficking into the regulated pathway by which LH is normally secreted, a well-established genetic rescue strategy was employed. As a first step, independent lines of HFSHB ^WT and HFSHB ^Mut transgenic lines were generated (10). In a second step, these transgenic mice were separately intercrossed with already characterized Fshb null mice (11). Thus, Fshb null mice expressing either a HFSHB ^WT or HFSHB ^Mut transgene were generated. Because Fshb null mice continue to express common α-sub-unit, interspecies hybrid FSH heterodimers (mouse α + WT hFSHβ or mouse α + Mut hFSHβ) could be formed in gonadotropes (12). The advantage of this genetic rescue approach is that the intracellular behavior of only the hFSH WT or Mut FSH dimers could be evaluated in the absence of endogenous mouse FSH (10, 12). Indeed, Western blot analysis under nondenaturing conditions confirmed that interspecies hybrid dimers are assembled in pituitary extracts of rescue mice and the mice exhibited the expected sexually dimorphic pattern of expression, i.e., males have high basal levels and females have low levels of FSH-β protein (10). Expression of Mut FSH itself did not alter the expression of gonadotrope-specific marker genes (10). Within the gonadotropes, LH is stored in dense core granules (DCGs) and is released as a bolus on GnRH stimulation. Immunofluorescence analysis of pituitaries indicated that only Mut FSH but nor WT FSH was co-localized with Rab27, a DCG marker and also with chromogranin-A, a known gonadotrope chaperone. Quantification of Mut FSH with Rab27 or chromogranin-A co-labeled cells was nearly identical to that observed with LH with Rab27 or chromogranin-A co-labeled cells seen in control mouse pituitary sections (10). Acute GnRH stimulation is known to release LH stored in DCGs but not FSH (which is constitutively secreted) in mice. Indeed, Mut FSH but not WT FSH was released into blood within 1h in both male and female mice in response to acute GnRH stimulation as quantified by an FSH radioimmunoassay (10). Consistent with these results, immunogoldelectron microcopy of pituitary sections confirmed lower number of DCGs after GnRH treatment only in Mut, but not WT FSH, expressing mice. Thus, the HP is sufficient to redirect Mut FSH into the LH secretory pathway where it is stored in DCGs and released in response to acute GnRH, similar to LH in control mice.

LH Homeostasis in Rerouted FSH-Expressing Mice

Because rerouted FSH exits through the regulated secretory pathway, like LH, it was reasoned that it may interfere with endogenous LH homeostasis. Serum levels of LH both basally and after acute GnRH were significantly downregulated in mice expressing the Mut but not WT FSH (10). These levels nearly matched those seen in serum of Lhb ^+/− mice (10, 13). Further, Lhb mRNA and estrus stage-specific expression pattern of LH-β protein were unaffected, although the steady state levels of LH-β protein in pituitary extracts was reduced. As the number of DCGs is limited per cell, it is likely that rerouted FSH competed with LH and might have caused the overall reduction in LH levels in Mut–FSH-expressing mice (10). To test this, HFSHB ^Mut transgene was introduced onto Acvr2 ^+/− mice (4, 10) and as a result, the net synthesis of Mut FSH was reduced in pituitaries of these mice lacking one Acvr2 allele. On reduction of Mut FSH expression, LH levels were restored back to those seen in control mice and correspondingly acute GnRH simulation also caused significant increase in serum LH levels (10). Together, these in vivo genetic experiments corroboratively confirmed that Mut FSH indeed competes with a limited number of DCGs, occupying these LH-containing DCGs in gonadotropes.

Genetic Rescue of Fshb Null Female Mice by Rerouted FSH

Fshb ^−/− female mice are infertile as a result of a preantral stage block in ovarian folliculogenesis (11). In addition, WT FSH efficiently rescues Fshb ^−/− mice, and these mice show normal estrus cycles as well as become fertile and indistinguishable from the control mice (12). The rerouted Mut FSH also efficiently rescues Fshb −/− mice, and all the characteristics including estrus cycles, breeding performance, and fertility are indistinguishable from those observed in WT–FSH-expressing mice (10). Corpora lutea are readily apparent in ovaries, and while serum estrogen levels are unaltered, high progesterone levels are observed. Aromatase (Cyp19a1) mRNA, which is sensitive to FSH is suppressed in ovaries of Fshb null mice but was up-regulated equally in ovaries of either prepubertal or adult mice expressing WT or Mut FSH (10). Thus, rerouted FSH secreted from the LH secretory pathway efficiently and functionally rescues Fshb deficiency in mice.

Effects of Rerouted FSH on Ovarian Function

The most striking aspects of the ovarian phenotype of female mice expressing rerouted FSH include enhanced ovulations and reproductive longevity. The average number of natural ovulations increased by 6 to 7 times more in mice expressing rerouted FSH (10). The fertilized one-cell embryos derived from Mut FSH mice are also highly competent because when they were transferred to pseudo pregnant female mice, high number of pups were born compared with the transfers performed with WT–FSH-expressing mice (10). When cultured in vitro these fertilized embryos also developed into blastocysts with equal efficiency as those collected from WT–FSH-expressing mice. Ovarian follicle counts did not differ between WT and Mut–FSH-expressing mice nor did the size of the primordial follicle pool. Expression levels of the molecular machinery involved in primordial follicle activation were comparable to mice expressing WT FSH (10).

The observed increase in ovulations seen in Mut FSH mice are not due to the depletion of follicle pool because there was no evidence of premature ovarian failure and even at older ages (12 months and beyond), rerouted FSH-expressing mice remained fertile and continued to exhibit enhanced ovulations (10). It also does not appear to be related to new development of ovarian follicles as the neo-oogenesis concept has been challenged and disproven at least in mice by elegant lineage tracing experiments (14-16). Enhanced granulosa cell proliferation (BrdU+ granulosa cells), increase in number of antral follicles, and significantly suppressed atresia (decreased cleaved caspase-3 expression) were evident in ovaries and may explain the basis for enhanced ovulations and reproductive longevity in rerouted FSH-expressing mice (10).

Ovarian Gene Responses to Rerouted FSH Signal Inputs

Ovarian granulosa cells express FSHRs and respond to constitutively secreted FSH to regulate genes important for proliferation and differentiation. To determine how granulosa cells respond to rerouted FSH secreted from the LH pathway presumably in defined pulses, candidate gene expression analysis was performed on ovaries obtained from WT and Mut–FSH-expressing mice.

Several distinct ovarian gene expression patterns emerged depending on the pattern of FSH signal input (constitutive vs. pulsatile) in vivo (10). The first set included preferential up-regulation of prosurvival genes (Bcl2 and Igf1) in rerouted FSH-expressing mice. The second set contained several known FSH-responsive genes (Esr1, Kcnj8, S100a6, S100g, and Tagln), also up-regulated in rerouted FSH-expressing mice. Interestingly, the third set of genes involved genes normally responsive to LH signaling (Cap1, Fas, Ptgs2, Star, and Tsg6). This group of genes may probably sense the pulsatile nature of gonadotropin signal input by rerouted FSH. However, rerouted FSH did not bind ovarian luteinizing hormone receptors and transduce signals, because when introduced onto Lhb ^−/− genetic background, the HFSHB ^Mut transgene did not rescue LH deficiency (10). There was also no up-regulation of phospho-CREB protein levels in rerouted FSH-expressing mice. Finally, the last set of genes did not show any preference for either constitutive or rerouted FSH signal input (10). Thus, FSH rerouting in vivo clearly indicated that the target granulosa cells interpret and decode the FSH signal inputs (pulsatile vs. constitutive) and regulate distinct gene networks.

Effects of Rerouted FSH on Male Reproductive Function

Follicle-stimulating hormone binds to FSHRs expressed on Sertoli cells within testes and regulates their proliferation and differentiation (2, 3). Fshb null mice displayed reduced testis size, a decrease in sperm number and motility, and qualitatively normal spermatogenesis but normal male fertility (11). The WT FSH rescues Fshb null male mice and restores the testis size, sperm number, and motility back to normal levels (10, 12). Similarly, rerouted Mut FSH also completely rescues Fshb null male mice (10). When adult testis phenotypes including sperm characteristics were compared, no differences were noted between WT and Mut–FSH-expressing rescue mice (10). Hence, unlike adult ovarian granulosa cells, adult Sertoli cells do not appear to discriminate the pattern of FSH signal input in vivo. Whether gene networks in proliferating Sertoli cells are altered prepubertally in WT and rerouted FSH-expressing mice remains to be identified.

FSH REROUTING IN VIVO—BASIC AND CLINICAL IMPLICATIONS

A short 7-amino-acid heptapeptide uniquely present in the C’-terminus of LH-β, when genetically fused to FSH-β is able to redirect FSH trafficking in vivo into the LH secretory pathway in pituitary gonadotropes. This rerouting of FSH causes enhanced ovulations and increased female reproductive life span in mice. These FSH rerouting mice offer an excellent model to study the basic and clinical aspects of gonadotropin action in the female. First, dynamic changes in FSHRs could be studied using constitutive and pulsatile FSH signal inputs in target cells with regard to receptor occupancy, internalization, endosomal signaling events, and receptor recycling (17). Second, the distinct patterns of gonadotropin trafficking and secretion in vivo may explain the origin of estrus cycles in vertebrates because primitive vertebrate species contained a mono-gonadotropic hormone ancestor with dual FSH- and LH-like activities, and there was no evidence of regulated mode of gonadotropic hormone secretion (10). Third, based on successful in vivo rerouting of FSH, it would now be possible to achieve the rerouting of LH by which an LH mutant lacking the HP could be engineered to exit through the constitutive secretory pathway in Lhb ^−/− mice. Fourth, the FSH rerouting principle may be applicable to several other pulsatile hormone signaling systems to explain the basis of hormone action in target cells. Fifth, although FSH is normally secreted through the constitutive route, the ability to mobilize a pulsatile mode of FSH administration could provide novel therapeutic options in women seeking fertility treatments using ART protocols. Sixth, obese women often have defects in hormone secretion (18, 19) and FSH-rerouted mice could provide an in vivo model to delineate how obesity causes changes in hormone secretion pattern, which in turn affects ovarian vs. pituitary function (20). Finally, FSH rerouted mice may provide a novel genetic model to understand ovarian aging and identify gene or protein networks that maintain ovarian function and fertility for prolonged periods of time.