Introduction

Congenital or idiopathic hypogonadotropic hypogonadism (CHH or IHH) is a developmental disorder of sex maturation resulting from deficient GnRH release or action. CHH meets the criteria for a rare disease with an estimated prevalence of 1–10 cases per 100,000 births [1]. The condition manifests as absent or delayed puberty and infertility with low serum levels of gonadotropins: follicle-stimulating hormone (FSH) and luteinising hormone (LH). In about 50–60% cases, in addition to CHH, there is an impairment of the sense of smell [2]. The combination of anosmia and hypogonadotropic hypogonadism is referred to as Kallmann syndrome (KS). This unique phenotype is a result of shared embryonic developmental signalling pathways of GnRH and olfactory neurons. GnRH neurons originate in the nasal placode at an early stage of embryo development and then migrate along vomeronasal nerves towards the cribriform plate and into the forebrain to finally reach the hypothalamus by the time of birth [3]. Alterations in this migratory process or loss of GnRH neurons impair sexual development and reproductive competence. GnRH is a key factor in human reproduction, impacting both natural and assisted fertility [4, 5].

CHH is not only clinically heterogeneous, but is also even more complex in terms of genetic variability and inheritance patterns. High-throughput technologies have more recently allowed the rate of genetic discovery to greatly accelerate. To date, almost 40 genes have been found to play a role in Kallmann syndrome and normosmic congenital hypogonadotropic hypogonadism (nCHH) pathogenesis, but only up to 40% of cases can be explained by their mutations [6].

The list of CHH-associated genes is still expanding, with the latest report on AMH and AMHR2 in CHH and NDNF mutations exhibiting deleterious effects in Kallmann syndrome [7, 8]. While most CHH mutations are inherited as X-linked recessive, autosomal dominant, or autosomal recessive modes, some genes are shown to exert synergistic heterozygosity. Oligogenic, mostly digenic, interaction was identified for several CHH gene partners, among others: KAL1/PROKR2 [9], FGFR1/GNRHR [10], FGFR1/NELF [10, 11], FGFR1/FGF8 [12], KAL1/WDR11 [13], and SEMA3A/FGFR1 [14]. The overall prevalence of digenic mutations in CHH patients varies in different studies, but it is reported at about 10% [13–15]. Interestingly, in more than one-third of cases a single gene mutation is identified, thus it can be expected that the true frequency of oligogenicity is much higher [15].

Given that the complex architecture underlying idiopathic hypogonadotropic hypogonadism is still largely unknown, our study used targeted next-generation sequencing (NGS) to screen for DNA variants in a cohort of CHH/KS patients. Data on monogenic variants in ANOS1 and FGFR1, the classical CHH genes, were recently published by our group [16]. Here we report the identification of new oligogenic combinations of variants. All recognised genes are known to contribute to the phenotype of the Kallmann syndrome and normosmic congenital hypogonadotropic hypogonadism, but our results point to potential new “partners” underlying digenic and trigenic patterns.

Material and methods

Results

Oligogenic findings based on NGS custom panel screening

Patients were screened with the custom NGS panel with a mean read depth of 1142x and 90.3% coverage of the target region. Herein we report of patients with at least two heterozygous variants identified. A total of 12 variants were confirmed by Sanger sequencing in five individuals (Figure 1). Variants were found in SPRY4 (2 patients), SEMA7A (2 patients), POLR3B (2 variants in 1 patient), and SEMA3A, CHD7, PROKR2, CCDC141, NSMF, and SRA1 (each in a single patient). Gene coordinates, chromosomal localisation, and interactions are shown in Figure 2. Clinical phenotypes, characteristics of patients, and reported variants are detailed below.

Figure 1

Targeted sequencing identifies oligogenic variants in CHH patients. A – Segregation of identified variants within families. Probands/affected individuals are identified by the arrow. Circles denote females; squares denote males. M: mutation; NT: not tested. B – Sequence chromatograms of affected patients. Family 1: SPRY4: c.722C>A and SEMA3A: c. 1303G>A; M2 chromatogram identifies both a rare variant and a common polymorphism rs7804122 (the dotted line arrow). Family 2: SEMA7A: c. 618_619delCC and CHD7: c. 2840G>A. Family 3: SEMA7A: c. 916G>A and SRA1: c. 377G>A. Family 4: CCDC141: c. 2299G>A and POLR3B: c. 1244T>C; 2570+1G>A. Family 5: PROKR2:c. 889G>T, SPRY4:c. 530A>G, and NSMF: c. 1261C>T. The site of mutation is identified by the arrow

Figure 2

Genes linked to oligogenic inheritance in the studied CHH cohort. A – List of genes with genomic coordinates for the GRCh38 reference assembly. B – Gene-gene interaction based on online tool https://genemania.org/. C – Chromosomal localisation of genes of interest based on https://www.ncbi.nlm.nih.gov/genome/tools/gdp

Discussion

In this study we report on new genetic findings supporting the importance of oligogenic inheritance in congenital hypogonadotropic hypogonadism. We have recognised five new oligogenic patterns/combinations: SPRY4/SEMA3A, SRA1/SEMA7A, CHD7/SEMA7A, CCDC141/POLR3B/POLR3B, and PROKR2/SPRY4/NSMF. All of the listed genes are known to contribute to phenotype of CHH, but our results point to potential new “partners” underlying digenic/trigenic inheritance. Herein we report both novel variants and those that were earlier recognised in the aetiology of CHH.

Both variants, SPRY4 c. 722C>A and SEMA3A c. 1303G>A, identified in patient 1 were previously reported in IHH/KS patients [14, 15]. SPRY4 protein acts as an inhibitor of the receptor-transduced mitogen-activated protein kinase (MAPK) and thus antagonises FGF signalling pathway [21]. Activity of this protein is important for proper regulation of neuron outgrowth, which has been proven in the animal model of Spry4^–/– mice [22]. Nucleotide variation in this gene has been associated with hypogonadotropic hypogonadism with or without anosmia [15]. Semaphorin 3A is a key player in GnRH neuron migration. It has been shown in mice Sema3a-null mutants that loss of Sema3a signalling alters the targeting of vomeronasal nerves and the migration of GnRH neurons into the brain, resulting in reduced gonadal size and infertility [23]. Later it was documented that a heterozygous mutation of SEMA3A leading to haploinsufficiency, accompanied by a second genetic hit in another gene affecting the GnRH pathway, produces the KS phenotype [14, 24]. Interestingly our patient presented symptoms of CHH with normosmia. Based on the function of SEMA3A gene and its involvement in guidance of vomeronasal axons, we would have expected to observe a deteriorated sense of smell in our patient. Thus, we speculate that heterozygous missense variant of SEMA3A may not be sufficient to modify the phenotype of the olfactory system.

SEMA7A gene encodes for semaphorin 7A, a membrane-anchored guidance protein known to enhance central and peripheral axon growth. It is required for proper axon tract formation during embryonic development [25]. Only one study reported on SEMA7A gene variants (p.R148W, p.R474Q) in 2 patients. SEMA7A variants have been shown to contribute to the oligogenic phenotype of both nCHH and KS, together with KISS1R and KAL1, respectively [14]. Identified in our cohort, SEMA7A: c.618_619delCC mutation is an out-of-frame deletion and thus predicted null variant due to a premature termination codon and consequently nonsense-mediated decay. In the same subject CHD7 mutation was identified. Chromodomain helicase DNA-binding protein 7 is a transcriptional regulator, and its mutations are responsible for up to 70% of CHARGE syndrome cases. However, several heterozygous CHD7 mutations have been recognised in nCHH/KS patients with no symptoms of CHARGE phenotype. Moreover, some variants were recognised in patients presenting with various phenotypes of nCHH, KS, or CHARGE syndrome [26–28]. Thus, a growing amount of evidence supports the hypothesis that nCHH/KS could be a milder allelic variant of CHARGE syndrome. Interestingly, in the patient’s family a case of choanal atresia was present. Association of Kallmann syndrome and choanal atresia was previously reported in a paper by Klein et al. [29]. The authors described several cases of both conditions affecting members of one family. Because the inheritance did not exactly match an autosomal dominant pattern it was called “irregular AD mode of inheritance”. We believe that what was observed and reported is now acknowledged as oligogenic inheritance, a phenomenon not known 30 years ago. Missense mutation CHD7 c.2840G>A revealed in our patient predicts a substitution Arg947Gln at the protein level and was previously reported in a single case of CHARGE syndrome [19]. Due to the lack of sufficient evidence the variant was previously classified as probably benign. In this case, reporting a second unrelated patient harbouring the same mutation seems to be a sufficient argument to revise variant and assort it from class 2 to class 3 (VUS). Our finding also provides further support to the allelic relationship of nCHH/KS, choanal atresia, and CHARGE syndromes.

SRA1 is a transcriptional coactivator of steroid nuclear receptors. To date, four missense mutations in the SRA1 gene have been reported in CHH: p.P20L, p.Q32E, p.T103A, and p.I79T [30]. Mutations were revealed in families presenting homozygous, compound heterozygous, and digenic pattern of inheritance. In the digenic case, SRA1 mutation was accompanied by PNPLA6 missense variant. Herein we report a patient harbouring heterozygous SRA1 mutation c.377G>A that predicts substitution of arginine for histidine at residue 126 (p.Arg126His) and SEMA7A gene mutation c.916G>A (p.Val306Ile). Heterozygous SEMA7A mutations were noted in both normosmic HH and KS, but in the case of anosmia the second reported variant was a deleterious mutation of KAL1 gene, which is crucial for the development of both GnRH and olfactory neurons. Therefore, SRA1 and SEMA7A seem to have no impact on olfactory function, which is concordant with the phenotype of our patient.

CCDC141 encoding for coiled-coil domain containing protein was recently shown to play an important role in GnRH neuron migration [31]. Its heterozygous mutations were identified in patients with Kallmann syndrome and normosmic CHH [32, 33]. Herein we report a patient with a novel variant of CCDC141, a missense mutation (c.2299G>A) that leads to Asp to Asn substitution at residue 767. The variant was predicted as deleterious by several programs used for in silico analysis. In the same patient we have also identified two biallelic mutations in POLR3B gene, which encodes for the polymerase III subunit. The T-to-C transition c.1244T>C leading to an amino acid change p.(Met415Thr) was previously reported in IHH and 4H leukodystrophy [20, 34, 35]. The mutation was identified in multiple individuals, but all of them were carrying a second missense mutation in the POLR3B gene. These observations led to the question of whether the p.(Met415Thr) variant is harmful at all, although prediction programs are congruent in terms of variant pathogenicity. Alternatively, a combination of two mutations impacting both alleles is necessary to evoke the effect [20]. In addition to the p.(Met415Thr) variant, we have identified an intronic mutation c.2570+1G>A in POLR3B. According to NNSPLICE, the mutation creates a splice defect. Notably, the proband’s parents were heterozygous for POLR3B mutations, the splice site variant was inherited from the mother, while the missense POLR3B variant was of paternal origin. Thus, our finding supports the hypothesis that a severe effect on phenotype can be observed when the Met415The variant is combined with a mutation affecting the transcription from the second allele.

Patient 5 carries a heterozygous mutation in SPRY4 c.530A>G (p. Lys177Arg), which was implicated in KS [15]; moreover, he harbours two novel missense mutations c.889G>T (p.Val297Ile) in PROKR2 and c.1261C>T (p.Leu241Phe) in NSMF gene. Notably, the SPRY4 and NSMF variants were of maternal origin, while the PROKR2 variant was paternal. PPROKR2 coding for prokineticin receptor 2 is a gene implicated in both IHH and KS via defective GnRH neuron development [9]. NMDA receptor synaptonuclear signalling and neuronal migration factor encoded by NSMF gene, previously known as NELF, has been shown to contribute to CHH pathogenesis [36].

Herein we present several sequence variants in nCHH and KS patients. According to the ACMG–AMP criteria [37, 38], we classified them as variants of uncertain significance (VUS, class 3), probably pathogenic (class 4) or pathogenic (class 5). Oligogenic inheritance does not show regular Mendelian segregation patterns, nor does it follow Mendelian mechanisms. However, in the absence of adequate tailored criteria for variant classification we used the one meant for interpretation of sequence variants in monogenic diseases. We thus believe that the variant categories assigned in our study might be underestimated.

In conclusion, with accumulating data, it is becoming increasingly clear that oligogenicity may account for the phenotype in many complex diseases. Hopefully, with the growing evidence of oligogenic patterns, it will be possible to fine-tune guidelines and provide a tool for precise sequence variant classification in non-Mendelian conditions.