Kallmann syndrome (KS) combines hypogonadotropic hypogonadism and anosmia or hyposmia, i.e., a deficiency of the sense of smell [1]. Anosmia/hyposmia is related to the absence or hypoplasia of the olfactory bulbs and tracts [2]. Hypogonadism is due to deficiency in gonadotropin-releasing hormone [3] and probably results from a failure of embryonic migration of gonadotropin-releasing hormone-synthesizing neurons [4]. These cells normally migrate from the olfactory epithelium to the forebrain along the olfactory nerve pathway [5]. In some KS patients other developmental anomalies can be present, which include renal agenesis, cleft lip and/or palate, selective tooth agenesis, and bimanual synkinesis [6].

This is a genetically heterogeneous disease, which affects 1:8000 males and approximately five times less females. Two different genes have so far been identified. Loss-of-function mutations in KAL1 (NCBI GeneID: 3730) [7–9] and FGFR1 (NCBI GeneID: 2260) [10] account for the X-chromosome linked form and an autosomal dominant form of the disease, respectively. KAL1 encodes anosmin-1, a locally restricted glycoprotein of embryonic extracellular matrices [11], which is likely to be involved in fibroblast growth factor-signaling [6,12]. Nearly 80% of the KS patients, however, do not carry a mutation in either of these genes [6].

Because the common infertility in affected individuals and, most importantly, the incomplete penetrance of the disease impede linkage analysis, the positional cloning strategies that have been taken to find causative genes were based on the analysis of rare KS individuals who carry chromosomal rearrangements detectable by cytogenetics techniques [7,8,10]. Here, we used a direct candidate gene approach and identified two novel genes underlying the disease.

We first considered GPR73L1/PROKR2 (NCBI GeneID: 128674), encoding the prokineticin receptor-2 (PROKR2) [13–15], a most relevant candidate because olfactory bulbs do not develop normally in mutant mice lacking this G protein-coupled transmembrane receptor, and these mice also have a severe atrophia of the reproductive system related to the absence of gonadotropin-releasing hormone-synthesizing neurons in the hypothalamus [16]. We thus sequenced the two coding exons of PROKR2 and flanking splice sites in 192 unrelated individuals (144 males and 48 females) affected by KS, including 38 familial cases. Ten different mutations (one frameshift and nine missense mutations) were detected in 14 patients (four familial and ten apparently sporadic cases) in the heterozygous (ten cases), homozygous (two cases), or compound heterozygous (two cases) state (Figure S1, Table 1, and Figure 1). Conservation of the mutated amino acid residues in bovine, murine, and rat orthologous sequences (Figure S2) argues in favor of a deleterious effect for all the missense mutations. However, two of these mutations, p.R268C and p.V331M, as well as a mutation (c.253C>T, p.R85C) affecting the same residue as the p.R85H mutation found in two KS cases and another missense mutation (c.1004C>G, p.T335M) not found in the cohort of KS patients, were detected, once each, in 500 alleles from ethnically matched (Caucasian) control individuals. No other nonsynonymous variant was found in the controls. In the absence of functional testing, one cannot be sure that each missense mutation found in KS individuals is causative of the disease. Nevertheless, together with the KS-like phenotype of Prokr2 knockout mice, the fact that the overall proportion of PROKR2 alleles carrying nonsynonymous mutations is significantly higher in KS patients (18 out of 384 alleles) than in controls (four out of 500 alleles; chi-square value = 13.5, p < 0.001) strongly argues in favor of the involvement of the gene in KS.

Prokineticin-2 (PROK2) [17] is the main ligand of PROKR2. We considered the possibility that mutations in PROK2 (NCBI GeneID: 60675), account for some KS cases, especially since mutant mice defective in Prok2 have a marked reduction in the size of olfactory bulbs and a loss of their normal architecture [18]. We sequenced the four PROK2 coding exons (including the alternative exon 3 [19]) and flanking splice sites in the same cohort of patients and found four different point mutations (two missense mutations, one frameshift mutation, and one single nucleotide substitution in the translation initiation sequence [20]), all in the heterozygous state, in two familial and two apparently sporadic cases (Figure S1, Table 1, and Figure 1). These mutations were not detected, or any other sequence variant, in 500 alleles from ethnically matched (Caucasian) control individuals. The p.G32R mutation affects the glycine residue of the N-terminal hexapeptide AVITGA (see Figure S2). This motif, which is conserved among prokineticins from mammalian and non-mammalian species, is critical for the bioactivities of these proteins [21]. The p.R73C mutation introduces a cysteine residue, which is expected to disrupt the formation of the disulfide bonds of the protein [19,21].

Considering both the phenotypes of the Prokr2- or Prok2-deficient mice and the likely deleterious effect of the human frameshift mutations on protein synthesis, the various PROKR2 and PROK2 missense mutations underlying KS are predicted to be loss-of-function mutations too. Since the same developmental disease, namely KS, apparently results from insufficient signaling either through FGFR1 or through PROKR2, the two signaling pathways are expected to interfere at some level, possibly the activation of the mitogen-activated protein kinase pathway [13,14,22]. In addition, the KAL1 gene product anosmin-1, which has been reported to enhance fibroblast growth-factor-signaling through FGFR1 [12], could also play a role in prokineticin-signaling through PROKR2. Indeed, both anosmin-1 and PROK2 have binding affinities for heparan sulfate glycosaminoglycans, which are also well known co-receptors in fibroblast growth-factor-signaling [23–25]. Notably, only homozygous mouse mutants lacking Prokr2, Prok2, or Fgfr1 show abnormal olfactory bulbs [16,18,26], whereas it seems that heterozygous mutations in any of the orthologous genes can cause KS in humans ([10] and this study). A unifying explanation could be that the local amount of anosmin-1 exerts a critical dosage effect both on FGFR1 and PROKR2 receptor activation, hence a higher dose of anosmin-1 in the mouse (the murine Kal1 is expected to be pseudoautosomal [27,28], whereas the human KAL1 is X chromosome-linked) could protect heterozygous mouse mutants from the developmental failure.

It is noteworthy that KS patients who carry mutations in PROKR2 or PROK2 have variable degrees of olfactory and reproductive dysfunction (see Figure 1). In addition, they do not seem to have any of the occasional clinical anomalies that have been reported in the previously characterized genetic forms of the disease, namely bimanual synkinesis, renal agenesis, dental agenesis, and cleft lip or palate. However, the KS patient carrying the p.R73C mutation in PROK2 suffers from a severe sleep disorder and marked obesity, which might be related to the known circadian function of prokineticin-2 and its potential roles in sleep-wake regulation and ingestive behavior [29–31].

In human monogenic disorders, genuine dominance, where heterozygotes and homozygotes have the same phenotype, is unusual. In particular, most dominant developmental diseases are far more severe in the homozygous state. Therefore, the finding of both heterozygous and homozygous (or compound heterozygous) KS patients for a given mutation in PROKR2 (e.g., p.R85H or p.L173R, see Table 1) is quite remarkable and raises the question of a possible digenic mode of inheritance in heterozygous patients. In our cohort, none of the patients carrying a mutation in PROKR2 was carrying a mutation in PROK2 too. We did not find a mutation in FGFR1 in any of the individuals carrying a mutation in PROKR2 or PROK2, either. However, one of the patients heterozygous for the p.L173R mutation in PROKR2 (sporadic case) also carried a previously undescribed missense mutation, p.S396L, in KAL1 exon 8 (Figure S3), which was not detected in 500 alleles from control individuals. According to the predicted structure of anosmin-1 [32,33], the mutation modifies the first amino acid residue of the linker between the second and third fibronectin-like type III repeats [6], a residue that is conserved among orthologous proteins from vertebrates and invertebrates (Figure S3). This mutation thus can be regarded as causative of the disease. Likewise, the PROKR2 mutation carried by this patient, L173R, which was found in six unrelated KS individuals and none of the controls, is most likely pathogenic too. To date, this is the only case of possible digenic inheritance reported in KS. It is quite possible, however, that other patients heterozygous for mutations in PROKR2 or PROK2 also carry a mutation in another, still unknown KS gene. Indeed, mutations in KAL1, FGFR1, PROKR2, and PROK2 together account for approximately 30% of all the KS cases in our large series of patients, which indicates that still other genes underlying the disease remain undiscovered.

Written informed consent was obtained from all the individuals who participated in the study. Genomic DNA was obtained from peripheral blood samples or lymphoblastoid cell lines by using a standard phenol chloroform extraction procedure. The genomic DNA from 250 unrelated Caucasian individuals was used as a control (control individuals were not examined to look for Kallmann syndrome features). We used the ENSEMBL (http://www.ensembl.org) and UCSC (http://www.genome.ucsc.edu) genome databases to find the exon-intron structure and single nucleotide polymorphisms of PROK2 and PROKR2. We designed PCR primer sets for amplification of exons and exon-intron boundaries using PRIMER 3 software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi). Sequences of the primers used to amplify and sequence the PROK2 four-coding exons and PROKR2 two-coding exons in the patients and controls are listed below. The primers used for the KAL1 and FGFR1 coding exons have been reported [9,10]. All identified mutations were confirmed on a second PCR product. The alleles from control individuals were analyzed either by direct sequencing of the PCR products (PROKR2 exons) or by using a denaturing high-performance liquid chromatography standard procedure followed by sequencing of abnormal products (PROK2 exons, KAL1 exon 8).