Mutations and polymorphisms in the gonadotrophin genes; clinical relevance

Recen_t studies on molecular pathogenesis of reproductive disorders have demonstrated an important genetic component in subfertility and infertility. The current methods of molecular biology, enabling detailed identification of mutations and polymorphisms, have revealed numerous mutations in the key genes involved in regulation of the reproductive functions. A good example are the common microdeletions that are detectable in the Y chromosome of azoospermic and oligozoospermic men (see, e.g. 28; 10; 20). When genes involved in hormone action are mutated, the phenotypic consequences are determined by the nature of the mutation. In a gain-of-function mutation, heterozygosity is sufficient to produce a phenotype, whereas in loss-of-function mutations usually only the homozygotes are affected. Both types of mutations are known with the receptor genes, but those of the ligands, i.e. hormones, are almost always of the loss-of-function type. When the genes involved in reproductive functions are affected, it is possible, but still rarely proven, that a given mutation affects the fertility of only one sex. Although the mutations impairing fertility should be self-limiting, they stay in the genetic pool due to their variable mode of inheritance and penetrance, and are in fact more common than assumed. Concerning gonadotrophins, the key hormones of the hypothalamic-pituitary-gonadal axis, microheterogeneity, due to variability in their carbohydrate structures, has been known for a long time. This phenomenon has been found to alter the intrinsic bioactivity of gonadotrophins, and in some cases to result in pathologies of gonadal function (7; 5; 46). The variability of gonadotrophin glycosylation has usually been considered to be hormonally regulated, and its eventual genetic basis remains totally unknown. Our recen_t studies on bioactivity/immunoreactivity ratios of gonadotrophins (especially of LH) in various physiological and pathophysiological conditions have shown that this parameter varies much less than previously demonstrated, and probably has less clinical value than proposed (18;16, 17). With regard to genuine mutations of the gonadotrophin genes, surprisingly few reports still exist. Since such mutations would severely compromise fertility, they are apparently sporadic and will be rapidly eliminated from the genetic pool. In contrast, a surprisingly common genetic variant of LH (v-LH) was recen_tly described. It alters the functional characteristics of LH, and is likely to be an important contributing factor in the individual physiological variation of LH action and in some pathologies of the pituitary-gonadal function. The purpose of this review is to summarize our current knowledge on mutations of gonadotrophin genes and the common genetic variant of LH. In addition, we review the data obtained from experimental studies on targeted disruption of the gonadotrophin subunit genes in mice, since these genetically modified animal models are very informative as regards the role of gonadotrophins in reproductive functions. The two gonadotrophins, follicle-stimulating hormone (FSH) and luteinizing hormone (LH) belong together with the thyroid-stimulating hormone (TSH) and chorionic gonadotrophin (CG) to the family of glycoprotein hormones. They are synthesized in gonadotroph cells of the anterior pituitary as heterodimers of the common α-subunit and the specific β-subunit which confers the hormonal specificity (Fig. 1). Both hormones are released to circulation in pulsatile fashion under the influence of the hypothalamic gonadotrophin-releasing hormone (GnRH). In addition, a complex feedback regulatory system through gonadal steroid (oestrogens, androgens, progesterone) and peptide (inhibins, activins) hormones modulates their secretion. Schematic presentation of molecular structures of the gonadotrophins and their receptors. a, LH and FSH are composed of a common α-subunit and a dissimilar β-subunit which are coupled by noncovalent interactions. The symbol ‘Y’ depicts the approximate locations of the N-linked carbohydrate side chains in the peptide subunits. b, The LH and FSH receptors are singular glycosylated proteins consisting of a long extracellular domain, the 7 times plasma membrane spanning transmembrane domain and a short intracellular tail. FSH and LH direct their actions to specific gonadal target cells through G-protein-coupled 7-times transmembrane receptors (Fig. 1). FSH receptors are present in ovarian granulosa cells and testicular Sertoli cells. Respectively, LH receptors can be found in the ovary in theca, granulosa and luteal cells, and in the testis in Leydig cells. After ligand binding, the occupied receptor results in G-protein coupled activation of adenylyl cyclase and generation of cyclic adenosine monophosphate (caMP), the main second messenger of LH and FSH action. Also other second messenger systems, such as the inositol phosphates and intracellular free calcium, are involved in gonadotrophin action, but their physiological significance has not yet been delineated. Only sporadic cases on mutations of gonadotrophin genes have so far been reported, apparently because of their adverse effect on fertility (Table 1). The heterozygotes for such mutations are likely free of symptoms or only mildly affected, but the homozygotes are expected to have more severe failure of reproductive function. It is also possible that the severity of the symptoms depends on sex. More mutations in the gonadotrophin genes will apparently be discovered in the future, as their detection is now technically relatively easy and as the awareness of the possibility of inherited basis of reproductive disorders spreads. There are several earlier findings on hypogonadal males with low serum testosterone, elevated LH immunoreactivity but absent LH bioactivity (33; 3; 4). Since exogenous gonadotrophin administration was found to evoke normal rise in serum testosterone, it is possible that some of these cases are due to mutations in the LHβ gene, resulting in formation of immunoreactive but biologically inactive LH molecules. At the time of these reports, detailed genetic analysis of the conditions was not possible. Only one inactivating mutation of the LHβ gene has so far been described, where a single base substitution changed codon 54 from glutamine to arginine (47; Table 1). As a result, the LH formed maintained its immunoreactivity but was totally devoid of bioactivity, causing in the affected individual, a homozygous phenotypic male, absence of Leydig cells with lack of spontaneous puberty and infertility. Long-term treatment of the patient with hCG resulted in testicular enlargement, normal virilization, and a sperm count of 11 millions/ml with 50% motility. In the male heterozygotes found in the same family, impaired androgen production and infertility were observed. Female heterozygotes had normal sexual development and were fertile. Besides demonstrating the necessity of LH driven testosterone production for the postnatal sexual maturation and sexual functions in adult men, this genetic ‘model’ demonstrates that the androgen-dependent part of the male sexual differentiation occurring in foetal life proceeds without pituitary LH. It is possible that the onset of foetal testicular androgen production is autonomous, or it is regulated by hCG (15) or other, as yet unidentified, hormones or paracrine factors (14; 8a,9b). Two cases of inactivating mutation of the FSHβ gene have so far been characterized at the molecular level, by 25) and by 22) (Table 1). In the former, a 2-bp deletion in codon 61 (Val61X) was found in the FSHβ gene of a woman, giving rise to completely altered amino acid sequence between codons 61 to 86, followed by a premature stop codon, and lack of translation of amino acids 87 to 111. Consequently, the truncated FSHβ protein that was translated was unable to couple with the α-subunit to form dimeric bioactive FSH molecules. The affected woman had primary amenorrhoea with infertility. She had had apparently normal adrenarche, but no menarche or telarche. She could be treated with exogenous FSH that resulted in follicular maturation, ovulation and successful pregnancy. Her mother, a heterozygote for the mutation had suffered from menstrual irregularity and infertility, but it remains open whether her heterozygosity for the observed mutation was the cause of the symptoms. The second case of inactivating mutation of the FSHβ gene (22) was a female with similar phenotype, i.e. primary amenorrhoea and poorly developed secondary sex characteristics. She had undetectable serum FSH and oestradiol, high LH, and absent FSH response to GnRH stimulation test. Upon DNA sequencing she appeared to be a compound heterozygote for two mutations in the FSHβ subunit gene. One was the same as described by 25), and the other was a missense mutation that changed a cysteine (TGT) to glycine (GGT) at codon 51. No symptoms were found in the relatives heterozygous for either of the two FSHβ mutations. In addition, there exist several earlier reports on isolated FSH deficiency in men and women, but their genetic analysis was not possible at the time of most of these reports (38; 24; 13; 26; 27; 2; 6). In one patient, an FSH deficiency has been reported to be due to circulating antibodies to this hormone (39). Absence of bioactive FSH has also been reported in carbohydrate-deficient glycoprotein syndrome (31). Although the latter two conditions are not due to mutations in gonadotrophin genes, they produce phenotypes very similar to those with the mutated FSHβ gene. In men, isolated FSH deficiency has been previously reported to cause variable degrees of oligo-astheno-teratozoospermia with normal testosterone production and virilization (38; 24; 13; 26; 27; 2), but also normal fertility has been reported (24; 13). Hence, these reports have not clarified the contentious question about the exact role of FSH in adult testicular function. The few cases of inactivating FSHβ mutations described above resemble phenotypically very much those caused by the inactivating mutation of the FSH receptor gene, reported recen_tly by us (1; 45). The females have hypergonadotropic amenorrhea with infertility, and complete arrest of follicular maturation. The men homozygous for this mutation display normal virilization with variable reduction of testicular size and suppression of spermatogenesis. As with isolated FSH deficiency (see above), some of the men have fully normal phenotype and fertility. Because of the mild phenotypic alteration, the men homozygous for the inactivating FSH receptor mutation could only be identified because of their siblingship with the women with the receptor defect and clear phenotype. In conclusion, both the FSH ligand and receptor defect indicate that FSH is more critical for female than male fertility. They also suggests that the male contraceptive strategies aiming at specific suppression of FSH secretion are likely to fail because of the wide variability of spermatogenic suppression in men devoid of this hormone or its action. Upon testing the suitability of various monoclonal antibody (Mab) combinations for LH measurement using the immunofluorometric assay (IFMA) principle (23; 34), we detected a healthy woman, with two children, whose LH was not detectable using one particular Mab combination (36). The Mab that did not detect her LH was directed against an antigenic epitope in the intact LH α/β dimer (assay 1). Interestingly, a number of other two-site assays combining two antiβ antibodies, or an antiβ and antiα antibody, detected normal LH levels in her serum. In fact, all combinations employing Mab's directed against a conformational α/β epitope underestimated her LH concen_tration. Her LH bioactivity and the ratio of bioactivity/immunoreactivity (using a subunit specific IFMA for immunoreactive LH measurement, assay 2) were normal, and in accordance with her normal fertility. Since her FSH and TSH levels were normal, we made the assumption that the structure of her LHβ subunit was abnormal. The LH levels of her family members were also analysed, and it was shown that her mother had similar nondetectable LH by assay 1, but her father, two sisters and brother, as well as her two children, displayed LH levels which by assay 1 were about 50% of those measured by the reference assay 2, hence the ratio of LH with assay 1/assay 2 was about 0.5. This lead us to the conclusion that the aberrant LH form was genetic in nature, with Mendelian fashion of inheritance. Our subject and her mother were apparently homozygous, and the father and children heterozygous for the variant LHβ allele. A scheme of the molecular alteration of v-LH and its influence on reactivity of the different epitopes with the Mab's is presented in Fig. 2. Schematic presentation of the reason of aberrant immunoreactivity in individuals with the variant form of LH in two-site immunometric assays. Either the changes in the amino acid sequence and/or an extra carbohydrate chain in the LH β-chain block or eliminate the antigenic epitope present in the α/β LH dimer. Combination of monoclonal antibody (MAB) α/β and MAB-β was used in Assay 1, and MAB-α and MAB-β in assay 2. , carbohydrate chain. When a larger material of serum samples from healthy Finnish volunteers were analysed (Fig. 3), and the ratio of LH with assay 1/assay 2 measured, the results fell clearly into three categories: (1) those between 1.0-2.0 (normal ratio individuals, i.e. homozygotes for the wild-type [wt] LHβ allele), (2) those at 0.5-0.75 (low ratio individuals; heterozygotes) and (3) those with the ratio near 0 (0 ratio individuals; homozygotes for the v-LHβ allele). The combined carrier frequency of the low and 0 ratio forms of LH in the Finnish population was 28%, and the distribution followed the Hardy-Weinberg equilibrium. The distribution of 249 normal subjects in the normal (▪), low (●) and zero (▴) ratio groups according to the results of the ratios of LH measured by assays 1 and 2. The LH level measured by assay 2 is shown on the abscissa. As no sex differences were detected, the male and female data are compiled. From 12) with permission. We then sequenced the LHβ subunit gene of our subject with the 0 ratio LH (37; 30), and found two point mutations, both resulting in amino acid change: Trp8 → Arg8 (TGG → CGG) and Ile15 → Thr15 (ATC → ACC) (Table 1). One of the mutations introduces an extra glycosylation signal (Asn-X-Thr/Ser) into the LHβ chain, which may introduce an oligosaccharide side chain into Asn13. The same structure is present in the hCG β-chain where Asn13 is glycosylated. The aberrant immunoreactivity of v-LH could be explained either by the two amino acid changes and/or the putative extra carbohydrate chain which could either alter or block, respectively, the antigenic epitope of the intact LH dimer (Fig. 2). It is worth noting in this context that the α/β epitopes which show restricted reactivity towards the LH variant are distributed over a fairly large area, in fact representing two separate independent epitope regions. This was shown by the fact that two α/β specific antibodies were able to form a sensitive sandwich assay (35). Recen_tly, two groups from Japan (11; 32) reported female patients with infertility, whose serum LH was unmeasurable with an immunoradiometric assay kit using two monoclonal antibodies, in the same fashion as in our measurements. The LHβ genes of these patients were subsequently sequenced, and the same two point mutations were found (11). Very recen_tly, one of the Japanese groups demonstrated with recombinant LH molecules, possessing either one or both of the two point mutations, that Asn13 in the mutated hormone carries an extra carbohydrate side chain, and that mainly the Trp8 → Arg8 mutation is responsible for the altered immunoreactivity of the hormone (43). However, all evidence so far available indicates that the two point mutations described above occur always together in the v-LH molecule (30). Moreover, the promoter region of the variant LHβ gene carries several mutations which alter the regulation of the v-LHβ allele as compared with wt-LH (M. Jiang, F.-P. Zhang, P. Pakarinen, K. Pettersson, & I. Huhtaniemi, unpublished observation). It was of particular interest that the same two mutations in the LHβ chain were found in both Japanese and Finnish subjects. We know now that this LH polymorphism is widely distributed in different ethnic groups, varying from 0 to 52% in frequency (12; 41; 29; 30). The highest frequencies are measured aboriginal Australians and Northern European populations, and the lowest in American Indians (Mayans and Hispanics from Texas) and Kotas (a tribe in the south of India). Judging from the structure of hCGβ, variant LHβ and wild-type LHβ, it is possible that the v-LHβ allele has appeared through gene conversion within the LHβ/CGβ gene cluster, which may be predisposed to genetic recombination (44). It is also possible that v-LH represents an ancestral LH form which is in the process of being replaced by the wild-type LH form. This process seems to be at different stage in different populations, as judged from the vast differences in frequency of the two LHβ alleles. The reason for this remains unclear. The homozygotes for v-LH in our material are apparently healthy, with no reported infertility or subfertility. In this respect our findings differ from those of the Japanese studies (11; 42) where v-LH was related to infertility and menstrual disorders. The B/I ratio of v-LH is higher than that of wt-LH, but its half-time in circulation is shorter (12; 43). Since the pulse-frequency of v-LH was normal (12), this results in overall LH action that is more potent at the receptor site, but shorter in duration. Such a change may influence the overall LH action in affected individuals, and there are data indicating that this indeed may be the case. Women homo- or heterozygous for the v-LH allele have higher levels of serum testosterone, oestradiol and sex hormone-binding globulin, and the frequency of the variant allele is significantly different between certain subtypes of polycystic ovarian syndrome PCOS, (41; J. Tapanainen, K. Pettersson, & I. Huhtaniemi, unpublished observation). We have also found that boys with v-LH have slower overall progression, gain of height and gain of testis weight at puberty (40). These findings suggest that the bioactivity of v-LH may be lower than that of the wt hormone. How exactly an LH form with higher biopotency at the receptor site, but with shorter half-time in circulation, differs from wt-LH in its overall in vivo action, needs further studies. The same applies to the heterozygous condition, where each LH pulse consists of a short and long-acting component. It is apparent on the basis of our observations that, although the subjects homo- or heterozygous for the v-LHβ allele are largely healthy, there are subtle differences in the regulation of their LH secretion (mutations in the LHβ promoter) and in LH action (shorter T1/2 in circulation and increased bio/immuno ratio at the receptor site). Whether the fertility of the affected individuals is altered is not yet known. It is tempting to speculate that the v-LH allele could offer a reproductive advantage to the affected individuals in some environmental conditions, thereby explaining the high frequency of the polymorphism in some populations, as well as its wide variation world-wide. Although the pathophysiological significance of the LH variant remains open, it is important for the clinician to be aware of its occurrence. It may be the reason for unexpected results of LH measurements. Several widely applied commercial immunoassay systems still employ antibodies, similar to the α/β-dimer specific antibodies we have studied, showing a more or less restricted ability to recognize this commonly occurring, biologically active LH variant. Two targeted gene disruption (knock-out) studies have been reported with the gonadotrophin subunit genes, those of the common α-subunit (19) and of FSHβ subunit (21). With regard to the common α-subunit gene, the null mutant mice were hypogonadal in adult age, as expected. However, they displayed normal male sexual differentiation in utero, despite the total lack of gonadotrophins and TSH, which further strengthens the contention that the early foetal testicular androgen production, a prerequisite for male-type sexual differentiation, occurs independent of pituitary LH. The same seemed to be true with the human male with inactivating LHβ mutation (47); he had normal sexual differentiation in utero in the absence of functional LH. The intrauterine stimulus of foetal testicular steroidogenesis may be in the human hCG (15), or it may be a nongonadotropic paracrine or endocrine factor. We have recen_tly shown that many regulatory peptides, active elsewhere in the adult body, are potent stimuli of foetal testicular steroidogenesis (8a,9). This knock-out model shows that only the postnatal sexual development, in both sexes, is dependent on pituitary gonadotrophin secretion. The FSHβ knock-out (21) produced in female mice a phenotype where the FSH-deficient animals were infertile due to a block in folliculogenesis prior to antral follicular formation. The null-mutant males had normal sexual differentiation, were fertile, but had somewhat decreased testicular size. Interestingly, the phenotypes of these mice were practically identical with those reported by us in humans with inactivating mutation of the FSH receptor gene (1; 45). These two ‘models’ demonstrate clearly that FSH is needed in the female for follicular maturation and is mandatory for maintenance of fertility. In the male, in contrast, the role of FSH is less compelling. Whereas the finite testicular size is determined by FSH, through stimulation of postnatal Sertoli cell proliferation, the pubertal onset of spermatogenesis does not need FSH. In adult age, FSH improves spermatogenesis qualitatively and quantitatively but is not absolutely necessary for this process. This review describes the currently known and mutations in the gonadotrophin subunit These conditions have us a about the and of gonadotrophin secretion and as well as on of the gonadotrophin genes and mutations of the gonadotrophin subunit genes to be but there is a surprisingly common genetic variant of the LHβ gene. The mutations produce specific deficiency of either and the phenotypes of the affected individuals are largely as expected on the basis of knowledge on actions of the two In addition, some unexpected findings have been These the of foetal testicular steroidogenesis on pituitary LH also on and the apparent of spermatogenesis on FSH. Female fertility seems to be absolutely dependent on FSH, whereas LH seems to be more critical for the male sexual functions. The phenotypes caused by the human mutations are by similar findings in mice targeted of the gonadotrophin subunit The polymorphism detected in the LHβ gene is surprisingly with to 50% frequency in some It also variability with total absence on some The variant form of LH differs clearly from wild-type hormone in its intrinsic bioactivity and half-time in circulation, for which it is apparent that the overall LH action of the variant differs from that of the wild-type hormone. There are in the pituitary-gonadal function of individuals the variant LHβ but its clinical significance and possible with and of LH action to be The studies in this were by from The of and The