Introduction

The term ovarian reserve (OR) indicates the number and the quality of ovarian follicles, providing a measure of female reproductive lifespan . The OR gradually declines with age, indicating a decrease in the primordial follicular pool and ovarian response to stimulation . Reproductive functions are regulated by gonadotropins, i.e., follicle-stimulating hormone (FSH), luteinizing hormone (LH), and chorionic gonadotropin (hCG), whose receptors (FSHR for the first one and LHCGR for the latter two) are expressed in the ovary in females. The action of these receptors relies on the simultaneous activation of different signaling pathways, such as the steroidogenic Gαs protein-mediated cAMP/protein kinase a (PKA)/cAMP responsive element-binding protein (CREB), the proliferation-associated ERK1/2, and the protein kinase B (AKT) pathways, and several other intracellular events . In the ovary of women of fertile age, gonadotropin-mediated intracellular signals trigger follicular recruitment, growth and selection, ovulation, luteinization of the corpus luteum, and steroid synthesis.

Several variants of gonadotropins and their receptors are widespread among human populations, resulting in genotype-specific intracellular signaling ; this suggests that some features of reproductive physiology are specific to individuals because of endocrine adaptation to specific environmental conditions occurring during human evolution . Therefore, a number of human reproductive phenotypes, in terms of menstrual cycle length, sensitivity to gonadotropin stimulation, OR, etc., are present that should be considered to optimize the present assisted reproductive techniques (ARTs).

To date, age, basal FSH levels , antral follicle counts (AFC), and anti-Müllerian hormone (AMH) levels have been the main parameters to guide clinicians in the estimation of OR to predict the number of oocytes retrieved and for ART “individualization” . To improve ART outcome, many controlled ovarian stimulation (COS) protocols have been developed to increase the yield of oocytes retrieval . According to the Italian national ART register, 9.9% of the cycles performed in 1 year are cancelled before oocyte retrieval because of poor ovarian response, which occurs in 9–24% of in vitro fertilization (IVF) cycles despite relatively high FSH dosage , or ovarian hyperstimulation syndrome (OHSS) . Indeed, several issues, such as the type of FSH preparation, starting dose, frequency of drug injections, and different gonadotropin combinations, should be considered to retrieve a proper number of oocytes and limit OHSS risk. Moreover, personalized COS regimens should be preferred for women affected by polycystic ovary syndrome (PCOS) because of individual-specific ovarian response to gonadotropins . However, a definitive consensus on the best gonadotropin combination in ART has not been achieved to date.

Despite the importance of OR evaluation, the identification of predictive biomarkers remains challenging. Single nucleotide polymorphisms (SNPs) are defined as amino acid changes of ≥1% frequency and could be predictive markers of ovarian response . Because FSH and LH regulate folliculogenesis, naturally occurring variants of genes encoding for these hormones and their receptors might be linked to OR and COS response, as previously suggested , but clear associations in this sense remain to be clarified.

Genetics of the FSHβ-subunit

The FSHβ-subunit confers hormone-specific biological properties and is encoded by the one-copy FSHB gene on chromosome 11p13 (4.2 kb) . The gene comprises one non-coding and two translated exons and was highly conserved during evolution because of its relevance in reproduction . Indeed, most of the few FSHB gene-inactivating mutations described in females ( Supplementary Table 1 ) are linked to infertility .

In accordance with the conservative nature of FSHB , about only 25 SNPs have been reported so far, mainly located in the non-coding region ( Supplementary Table 1 ). However, one of these SNPs was suggested as a candidate marker of hormone action . It was initially identified by studies enrolling infertile male patients, reporting an association between the SNP falling within the FSHB promoter region c.-211G>T (rs10835638 in the NCBI database; www.ncbi.nlm.nih.gov/pubmed ) and low FSH levels . The effect of this SNP in women was recently investigated in 365 normally cycling ART patients . Higher FSH levels in the follicular phase and decreased progesterone production in the luteal phase were reported in c.-211G>T T female carriers. Interestingly, a genome-wide association study found that c.-211G>T T was associated with longer menstrual cycles, later age at menopause, nulliparity, and lower endometriosis risk compared to the G variant, suggesting that differences in circulating FSH levels may affect the menstrual cycle length and female reproductive health . However, women enrolled in this study were approaching menopause; therefore, the results should be independently confirmed in cohorts of younger patients. In fact, this finding was challenged by a study that found an association between the c.-211G>T T genotype and lower FSH levels, which reasonably fit better with the longer menstrual cycle . These findings shed interesting cues on some evolutionary aspects: longer menstrual cycles and later menopausal age may be strategic during human evolution as they extend the reproductive lifespan and increase mating attempts, and therefore the reproductive success . Thus, the FSHB c.-211G>T SNP is proposed as a marker of COS outcome.

Genetics of the FSHβ-subunit

The FSHβ-subunit confers hormone-specific biological properties and is encoded by the one-copy FSHB gene on chromosome 11p13 (4.2 kb) . The gene comprises one non-coding and two translated exons and was highly conserved during evolution because of its relevance in reproduction . Indeed, most of the few FSHB gene-inactivating mutations described in females ( Supplementary Table 1 ) are linked to infertility .

In accordance with the conservative nature of FSHB , about only 25 SNPs have been reported so far, mainly located in the non-coding region ( Supplementary Table 1 ). However, one of these SNPs was suggested as a candidate marker of hormone action . It was initially identified by studies enrolling infertile male patients, reporting an association between the SNP falling within the FSHB promoter region c.-211G>T (rs10835638 in the NCBI database; www.ncbi.nlm.nih.gov/pubmed ) and low FSH levels . The effect of this SNP in women was recently investigated in 365 normally cycling ART patients . Higher FSH levels in the follicular phase and decreased progesterone production in the luteal phase were reported in c.-211G>T T female carriers. Interestingly, a genome-wide association study found that c.-211G>T T was associated with longer menstrual cycles, later age at menopause, nulliparity, and lower endometriosis risk compared to the G variant, suggesting that differences in circulating FSH levels may affect the menstrual cycle length and female reproductive health . However, women enrolled in this study were approaching menopause; therefore, the results should be independently confirmed in cohorts of younger patients. In fact, this finding was challenged by a study that found an association between the c.-211G>T T genotype and lower FSH levels, which reasonably fit better with the longer menstrual cycle . These findings shed interesting cues on some evolutionary aspects: longer menstrual cycles and later menopausal age may be strategic during human evolution as they extend the reproductive lifespan and increase mating attempts, and therefore the reproductive success . Thus, the FSHB c.-211G>T SNP is proposed as a marker of COS outcome.

Genetics of the FSHR gene

FSHR is encoded by a 200-kb gene located on chromosome 2p21 and spans 10 exons and 9 introns . The mature receptor is a 696-amino acid protein that is organized in an extracellular domain encoded by exons 1–9, seven transmembrane domains, and a C-terminal intracellular domain encoded by exon 10. Alterations in the FSHR gene sequence may affect FSH binding, signal transduction, or receptor internalization, resulting in impaired FSH-dependent estradiol production and dominant follicle selection and growth . Most of the several naturally occurring FSHR mutations were classified as activating or inactivating and were associated with reproductive disorders ( Supplementary Table 2 ). Inactivation of homo- or heterozygous FSHR mutations was previously reviewed ; they were mainly found in exon 10 in women affected by ovarian dysgenesis and primary and secondary amenorrhea, e.g., the p.Ala189Val variant. Exogenous FSH administration failed to increase serum steroid levels in women carrying the mutated receptor, which affected the COS outcome , although contrasting results were also reported .

FSHR -activating mutations are relatively rare . In men they are associated with autonomous and gonadotropin-independent sustainment of spermatogenesis , while these variants were reported in very few women; these women were mainly affected by OHSS and promiscuous activation by hCG and thyroid-stimulating hormone. Because of the high constitutive activity of mutant FSHR, excessive number of mature follicles and high estradiol levels were obtained upon exogenous FSH administration in COS protocols, but few cases have been reported to date to draw solid conclusions.

The FSHR gene carries more than 2000 SNPs, although only one is considered a marker of ovarian response, as repeatedly observed in vivo and confirmed by in vitro experiments . In addition, p.Asn680Ser SNP in exon 10 is in linkage disequilibrium with p.Thr307Ala ( Supplementary Table 2 ). While too few studies have assessed p.Thr307Ala as a marker for ovarian response , serine at position 680 introduces a potential phosphorylation site in the FSHR intracellular domain , suggesting that SNP may affect the “sensitivity” to FSH stimulation . Indeed, FSHR p.Asn680Ser Ser mediates slower and weaker cAMP-dependent cell signaling activation and progesterone production than the Asn variant in vitro ( Figure 1 ) , confirming clinical data. Moreover, p.Asn680Ser Ser homozygous women have higher basal FSH levels, require higher FSH dose during COS, and display lower estradiol levels than heterozygous and p.Asn680Ser Asn homozygous women , thus providing p.Asn680Ser Ser as a variant that is “hyporesponsive” to FSH and can affect the response to COS . Interestingly, the mean duration of the reproductive lifespan is calculated to be approximately 39 years for p.Asn680Ser Ser homozygous women, which is 2 and 1 year more than that of Asn homozygous and heterozygous carriers, respectively . The haplotype p.680Asn/p.307Thr is the most frequent in Caucasians and has variable frequencies worldwide depending on ethnicity . Although the slight decrease in OR due to different sensibilities of FSHR variants per se may be practically negligible, the association between p.Asn680Ser and the FSHB c.-211G>T SNP may provide haplotypes featuring relevant endocrine differences linked to evolutionary consequences. Moreover, it may be relevant in ART because exaggerated basal FSH:LH ratio is linked to reduced OR, and the trend of this value might predict the response in infertile patients undergoing COS . Therefore, some studies attempted to correlate p.Asn680Ser with OHSS risk and severity , but the lack of statistically strong, independent confirmation makes this issue unresolved.

Features of the FSHR p.N680S polymorphism. The S variant is linked to the slower kinetics of cAMP production, ERK1/2, and CREB phosphorylation, providing the in vitro evidence of a “FSH-resistant” receptor. This can explain the higher FSH levels and menstrual cycle length associated with the FSHR p.N680S S variant. Longer menstrual cycles likely result in increased reproductive lifespan of S homozygous women.

The c.-29G>A SNP located within the FSHR promoter region was found to be associated with the expression of the receptor , suggesting that relatively low FSHR number at the cell surface may be linked to the ovarian response to the hormone . In fact, the c.-29G>A was recently found to be associated with serum FSH levels in healthy, early pubertal Caucasian girls , indicating the SNP as one of the few markers of endocrine reproductive features in females.

Finally, gonadotropin receptor genes are known PCOS susceptibility loci . Women with PCOS display lower ovulation number than healthy females; thus, the OR of these women is preserved for a longer time, extending their reproductive lifespan . Generally, infertile PCOS patients are treated by increasing FSH levels, either by antagonizing the effect of estrogen by clomiphene citrate or by administering exogenous FSH. Recently, p.Asn680Ser Ser carriers were observed to be resistant to clomiphene citrate , highlighting the relevance of PCOS-dependent personalized COS regimens.

Taken together, the number of independent observations suggest that the FSHR c.-29G>A and p.Asn680Ser SNPs should be considered to predict the OR in ART.

Genetics of the LHB gene

The LHβ subunit confers biological functions to LH; these functions, in the ovary, include stimulation of steroidogenesis in theca cells, regulation of follicle maturation, and ovulation and maintenance of the corpus luteum. At the intracellular level, LH triggers anti-apoptotic and steroidogenic events, as demonstrated in granulosa cells in vitro . It is encoded by the LHB gene located in the chromosomal region 19q13.32 and comprises three exons and two introns.

Females carrying LHB mutations are rarely found and are always relatives of men harboring the same homozygous inactivating mutation that is usually associated with low LH levels, pubertal delay, and hypogonadism ( Supplementary Table 3 ) . An example is provided by two LHB mutations recently identified in a heterozygous 31-year-old man . He was affected by delayed puberty, azoospermia, and hypogonadism associated with LH absence. However, his sister, a carrier of the same mutations, underwent normal puberty and menarche despite LH deficiency; that was likely because of the presence of sufficient amount of estradiol required for triggering sexual development . This report suggests that LH activity in women may be replaced by the same exogenous hormone or hCG to recover fertility .

Few LHB gene SNPs are currently known ( Supplementary Table 3 ). A common variant of the LHB (p.Trp8Arg/Ile15Thr, called “V-LH”), which is relatively frequent in Northern European women, is related to FSH sensitivity and ovarian response . V-LH is more frequent in normogonadotrophic women exhibiting ovarian resistance to FSH stimulation in ART , suggesting that V-LH could be a predictive marker of COS in poor responder subgroups, which are characterized by low FSH sensitivity. In this case, they could benefit from LH supplementation rather than stimulation with high FSH doses . However, because the current findings do not provide sufficient evidence to include V-LH and, in general, LHB SNPs are among the markers of ovarian response, further investigations are needed to elucidate this issue.

Genetics of the hCGβ-subunit genes

The hCGβ-subunit confers biological functions to hCG and is encoded by six genes ( CGBs ) located at 19q13.32 and belonging to a genomic cluster together with the LHB gene . Differently from LH, hCG has a highly glycosylated carboxyl terminal peptide (CTP), which confers a relatively long circulating half-life (several hours) compared to that of LH (60–90 min) . This CTP tail was used to produce an artificial chimeric gonadotropin, “corifollitropin alfa,” in which the CTP was attached to FSH. Because its half-life is longer than normal FSH, corifollitropin alfa is used for COS to reduce the number of hormone injections . CGB SNPs were associated with recurrent miscarriage , which was defined as three or more consecutive pregnancy losses before 22 gestational weeks or as spontaneous abortion of the embryo/fetus . In particular, the CGB5 and CGB8 genes seem to be fundamental for pregnancy progression because SNPs located within the CGB5 and CGB8 genes were associated with lower risk of recurrent pregnancy loss, while the rare non -synonymous substitutions (p.Val56Leu in CGB5 , p.Arg8Trp and p.Pro73Arg in CGB8 ) were associated with increased recurrent miscarriage risk . Although the role of these CGB transcript variants in miscarriage was argued , the molecular mechanisms underlying the link between these genes and pregnancy management remain unclear. However, there are no solid indications suggesting CGB SNPs as markers for OR or COS response.

Genetics of the LHCGR gene

LHCGR is encoded by a single 80-kb gene on chromosome 2p21, spanning 10 introns and 11 exons. Exons 1–10 encode the large extracellular region, while exon 11 encodes the serpentine transmembrane region, the intracellular tail, and part of the hinge region , although LHCGR exon skipping and alternative splicing provide several receptor isoforms . For instance, LHCGR lacking exon 10 was described in a male patient affected by Leydig cell hypoplasia type II, which resulted in hypogonadism due to impaired LH action . In this patient, eugonadism was restored by hCG injection, although endogenous LH failed, revealing that LH and hCG interact differently with the same receptor. Interestingly, exon 10 is primate specific and encodes the receptor hinge region, which is crucial in discriminating between LH- and hCG-induced signaling . Indeed, the Lhcgr of the monkey Callithrix jacchus lacks amino acids encoded by exon 10 and, interestingly, retains the pituitary production of a hCG-like molecule instead of LH.

The human species seems to feature a wide variety of LHCGR mRNA transcripts with barely known functions, such as the insertion of a 2.7-kb genomic region between exons 6 and 7. This results in an additional exon named “6A” , whose physiological importance is not yet fully understood, especially in women, while naturally occurring mutations within this exon cause type II Leydig cell hypoplasia .

Several LHCGR gene mutations have been described ( Supplementary Table 4 ) that result in noticeable phenotypes. LHCGR -inactivating mutations mainly result in truncated LHCGR protein. They were first observed in men affected by Leydig cell hypoplasia and pseudohermaphroditism and then reported in five females who were sisters of male patients carrying homozygous inactivating LHCGR mutations . Irregular menstrual cycles, recurrent ovarian cysts, increased LH:FSH ratio, normal androgens levels, and infertility were described despite normal female external genitalia and spontaneous puberty . Moreover, the p.Asn400Ser LHCGR loss-of-function mutation was identified in two sisters affected by the empty follicle syndrome, in whom oocytes could not be retrieved upon COS. No pre-ovulatory follicles, corpora lutea , or corpora albicantia were observed, suggesting LH independency of the early follicular phase . Moreover, p.Gly20Pro was observed in a woman affected by amenorrhea, recurrent ovarian cysts formation, and repeated chemical pregnancies after IVF. In her, FSH-induced ovarian response upon COS was excessive, resulting in the development of up to 35 follicles, although the number of oocytes retrieved was low . This report suggests that the presence of LHCGR mutations should be tested when oocytes cannot be recovered in IVF.

Moreover, only a few activating mutations were identified in females, and all were heterozygous and located within exon 11 ( Supplementary Table 4 ). They were linked to familial infertility, and patients presented normal LH and FSH levels despite precocious puberty .

The LHCGR gene harbors more than 300 SNPs, and most of them are located within the introns . The so-called “ins18LQ” is one of the most studied LHCGR polymorphic variants because of its relatively high frequency among Northern Europeans ( Supplementary Table 4 ). Some studies have suggested that the insertion may be associated with LH levels and play a role in the pathogenesis of PCOS, especially in women from the Mediterranean area ; however, because of the low number of associations, further independent confirmation is required to draw conclusions. Similarly, the p.Asn312Ser LHCGR variant was first observed in men with impaired spermatogenesis and then in women at increased PCOS risk .

Finally, new informative results may be provided by the analysis of the LHCGR expression levels, which may be linked to shorter ovarian stimulation period, better ART outcome , and premature luteinization in infertile women affected by diminished OR . However, more evidence is required before LHCGR mRNA can be proposed as a new biomarker for ovarian response.