Variations in PROKR2, but not PROK2, are associated with hypopituitarism and septo-optic dysplasia.

CONTEXT: Loss-of-function mutations in PROK2 and PROKR2 have been implicated in Kallmann syndrome (KS), characterized by hypogonadotropic hypogonadism and anosmia. Recent data suggest overlapping phenotypes/genotypes between KS and congenital hypopituitarism (CH), including septo-optic dysplasia (SOD).
OBJECTIVE: We screened a cohort of patients with complex forms of CH (n = 422) for mutations in PROK2 and PROKR2.
RESULTS: We detected 5 PROKR2 variants in 11 patients with SOD/CH: novel p.G371R and previously reported p.A51T, p.R85L, p.L173R, and p.R268C-the latter 3 being known functionally deleterious variants. Surprisingly, 1 patient with SOD was heterozygous for the p.L173R variant, whereas his phenotypically unaffected mother was homozygous for the variant. We sought to clarify the role of PROKR2 in hypothalamopituitary development through analysis of Prokr2(-/-) mice. Interestingly, these revealed predominantly normal hypothalamopituitary development and terminal cell differentiation, with the exception of reduced LH; this was inconsistent with patient phenotypes and more analogous to the healthy mother, although she did not have KS, unlike the Prokr2(-/-) mice.
CONCLUSIONS: The role of PROKR2 in the etiology of CH, SOD, and KS is uncertain, as demonstrated by no clear phenotype-genotype correlation; loss-of-function variants in heterozygosity or homozygosity can be associated with these disorders. However, we report a phenotypically normal parent, homozygous for p.L173R. Our data suggest that the variants identified herein are unlikely to be implicated in isolation in these disorders; other genetic or environmental modifiers may also impact on the etiology. Given the phenotypic variability, genetic counseling may presently be inappropriate.

Congenital hypopituitarism is a rare condition that may be associated with complex midline defects of the forebrain; these encompass a wide spectrum of phenotypes ranging from cleft palate to septo-optic dysplasia (SOD), holoprosencephaly (HPE), and incompatibility with life. SOD is defined as the combination of 2 of the following: 1) midline forebrain defects (eg, absent septum pellucidum, agenesis of the corpus callosum); 2) optic nerve hypoplasia; or 3) anterior pituitary hypoplasia and/or pituitary hormone deficiencies (1).

Both HPE and SOD are highly heterogeneous conditions, usually sporadic with a multifactorial etiology. However, an increasing number of early developmental transcription factors and associated pathway genes have been implicated in the etiology of HPE (eg, SHH, ZIC2, SIX3, TGIF) and SOD (eg, HESX1, SOX2, SOX3, OTX2) (2–7). These genes are expressed in regions that determine the formation of forebrain and related midline structures such as the hypothalamus and pituitary, and mutations in these genes are therefore associated with marked phenotypic heterogeneity (8).

Recently, we identified autosomal recessive and dominant mutations in fibroblast growth factor 8 (FGF8) in HPE and SOD (9). Although Fgf8 had previously been shown to maintain anterior pituitary cellular proliferation in mice through the regulation of Lhx3 (10), loss of function mutations in FGF8 in humans had so far only been associated with Kallmann syndrome (KS) and hypogonadotrophic hypogonadism (HH) (11). Our study was the first to provide a genetic link between these midline disorders and KS/HH, thus suggesting that other genes implicated in the pathogenesis of KS/HH, such as FGFR1 (receptor of FGF8), KAL1, prokineticin-2 (PROK2), or its receptor PROKR2, could also play a role in disorders such as HPE and SOD.

Heterozygous and homozygous loss of function mutations/variations in PROK2 and PROKR2 are implicated in KS/HH (Figure 1) (12–15). These proteins appear to be important for murine olfactory bulb development and subsequent GnRH neuronal migration from this region to the ventral forebrain (16, 17). Expression of Prok2 has been reported in the ependymal and subependymal layers of the olfactory bulbs, preoptic area, and median eminence in mice (18, 19).

Figure 1.

Structure of PROK2 and PROKR2. A, Ligand PROK2 consists of 4 exons with the alternative splicing event of exon 3 (orange) resulting in PROK1. The first 27 amino acids consist of the signaling peptide (green) followed by the AVITGA motif (brown), which is essential for PROK2 bioactivity. B, PROKR2 is a G protein-coupled receptor, thus spanning the membrane 7 times (blue; transmembrane domains) and encoded by 2 exons. Multiple variations have been detected in both proteins (indicated under figures by arrows). Asterisks indicate variants that have been detected herein, with the p.G371R variant of PROKR2 being the only novel variant (red). Purple type indicates variants that are known to be functionally deleterious in vitro and have been detected in KS (13, 15). Amino acid numbers are indicated above the figures.

In a recent collaborative study of 103 patients, including 68 patients with SOD, we showed that 4 patients with hypopituitarism harbored functionally significant mutations/variations in PROKR2 (20). Another recent report described 2 other patients with hypopituitarism who bore functionally significant PROKR2 variations (21). Therefore, to further investigate the role of PROK2/PROKR2 in hypothalamo-pituitary disorders, we expanded the screen to 422 patients (male:female ratio, 1.1:1) with hypopituitarism and SOD or HPE, which to our knowledge, is the largest cohort screened to date.

Patients and Methods

Patients

A total of 422 patients were recruited between 1998 and 2010; 168 were recruited at the London Centre of Pediatric Endocrinology, based at Great Ormond Street Hospital for Children and University College London (UCL) Hospitals in London; the remainder were referred from national (n = 157) and international (n = 97) centers. Ethical committee approval was obtained from the UCL Institute of Child Health/Great Ormond Street Hospital for Children Joint Research Ethics Committee, and informed written consent was obtained from patients and/or parents. Of the 422 patients screened, 375 (89%) had SOD and its variants, whereas 47 (11%) had HPE or midline clefts.

Mutation analysis

Primers for the PCR amplification (35 cycles) of the coding region of human PROK2 (NM_001126128) and PROKR2 (NM_144773) were designed using the online Primer3 program (http://frodo.wi.mit.edu/primer3). PCR parameters are available on request. Amplified DNA was then analyzed for mutations by direct sequencing, using BigDye version 1.1 sequencing chemistry (Applied Biosystems, Foster City, California) and analysis on a 3730X1 DNA Analyzer (Applied Biosystems/Hitachi, Tokyo, Japan). For any novel mutations/sequence variations detected in either gene, 480 ethnically matched controls (if available) were then screened at the corresponding residue. Changes were checked with reference to the dbSNP database (www.ncbi.nlm.nih.gov/SNP) and 1000 Genomes Project (www.1000genomes.org).

Functional analysis of novel variants

For cell surface quantification by ELISA, HEK-293 cells were cultivated in DMEM supplemented with 10% fetal calf serum and transfected by electroporation with a Gene Pulser Xcell eukaryotic system (Bio-Rad, Hercules, California) as described previously (13). Then, 107 cells were transfected with 2 μg of recombinant pRK5 plasmid vectors coding for the N-terminal hemagglutinin (HA)-tagged wild-type or mutant PROKR2 and made up to a total amount of 10 μg plasmid DNA with empty vector. Twenty hours after transfection, HEK-293 cells were washed with PBS and fixed with 4% paraformaldehyde in PBS for 5 min. The cells were then either permeabilized, using .05% Triton X-100 for 5 min, or not permeabilized as previously described (13). HA-tagged proteins were detected using monoclonal anti-HA peroxidase antibody 12CA5 (Roche Diagnostics, Mannheim, Germany) at 0.5 μg/ml. Because PROKR2 is a Gq-coupled receptor, we examined the signaling properties of novel variants by measuring intracellular calcium release and the accumulation of D-myo-inositol monophosphate (IP1) in HEK-293 cells in response to PROK2 ligand using IP-One HTRF assay kit (Cisbio Bioassays, Condolet, France), as previously described (13, 22). PROKR2 can also couple to Gs protein to generate cAMP, so we additionally evaluated this pathway to better characterize the functionality of novel PROKR2 variants using cAMP HTRF assay kit as previously described (13, 22).

Immunohistochemistry and in situ hybridization of Prokr2 knockout mice

Prokr2 null and wild-type embryos were collected at embryonic day (E) 18.5 and fixed with 4% paraformaldehyde (Sigma, St Louis, Missouri) and dehydrated to 100% ethanol to be embedded in paraffin. Paraffin sections (7 μm) were used for both immunohistochemistry and in situ hybridization. In short, immunohistochemistry was performed by dewaxing sections with histoclear, followed by hydration and antigen retrieval using microwave with citric acid buffer (10 mm citric acid, .05% Tween 20 [pH 6.0]). Antibodies were obtained from hybridoma bank [Developmental studies hybridoma bank (University of Iowa) and National Hormone and Peptide Program (Harbor-University of California, Los Angeles Medical Center)] and used at 1:1000 concentration in 5% inactivated sheep serum. For immunofluorescence, secondary goat antirabbit biotinylated antibody 1:300 (Dako, Carpinteria, California) was used, followed by 1:500 streptavidin (Sigma). 3,3-Diaminobenzidine (DAB; Vector Laboratories, Burlingame, California) staining was used in accordance with the manufacturer's protocol. Slide in situ hybridization on paraffin sections was performed as described in Gaston-Massuet et al (23). To quantify the number of LH cells, 3 different sections at different axial levels were selected from 3 embryos per genotype. Quantification of arginine vasopressin (AVP), oxytocin (OT), and Ghrh neurons was performed using 3 sections of equivalent axial level between mutant and wild-type from the supraoptic to the tubular area of the hypothalamus. Data are presented as mean number of cells ± SEM, with Student's t test used for statistical analysis and a P < .05 value considered statistically significant.

Results

No mutations/variations were found in PROK2, whereas 11 unrelated patients exhibited mutations/variations within the coding region of PROKR2, 9 of whom have variations that have been previously described in KS and shown to be functionally significant (p.L173R [n = 4], p.R268C [n = 4], and p.R85L [n = 1]) (Supplemental Figure 1). p.L173R is known to disrupt cell-surface targeting of the receptor, whereas the latter 2 variants affect G protein coupling (13). The remaining 2 patients harbored the p.A51T and the novel p.G371R variations. None of the 11 patients with PROKR2 variations had changes in FGF8, FGFR1, KAL1, NELF, CHD7, WDR11, HESX1, SOX3, or SHH.

Patient with PROKR2 (p.R85L) variant

The c.254G>T, p.R85L variant was detected in heterozygosity in a male Caucasian patient (II) who presented at 6 years of age with combined pituitary hormone deficiency (CPHD; GH deficiency [GHD], TSH deficiency [TSHD], and ACTH deficiency [ACTHD]). He had normal visual acuity and normosmia, and magnetic resonance imaging (MRI) of the brain showed an absent anterior pituitary and an ectopic/undescended posterior pituitary. Puberty was induced with gonadotropins at the age of 23 years, and the patient has since remained on testosterone (Table 1 and Supplemental Table 1).

Table 1.

Phenotypes in Patients with PROKR2 Variations

Patient No.	Mutation	Status	Sex	Ethnicity	Age, y	Endocrinopathy	MRI	Eyes	Other Features
I	p. A51T	HT	F	Chinese	0.9	GHD	APH	SOD, BL ONH	Pigmentary changes of right optic nerve
II	p. R85L	HT	M	Caucasian	6.0	GHD, TSHD, ACTHD, GnD	APA, EPP, hypoplastic infundibulum	Normal	Normal smell
III	p.L173R	HT	F	Caucasian	0.1	GHD, ACTHD	APH, partially descended PP	Normal	Hyperinsulinism treated with diazoxide and resolved; GI dysmotility
IV[a]	p.L173R	HT	M	Caucasian	0.6	GHD, ACTHD, TSHD, DI	Absent septum pellucidum, heterotopic gray matter, schizencephaly	SOD, BL ONH	Neonatal jaundice; spastic quadriplegia, epilepsy; sleep apnea and uncoordinated swallow diagnosed at age 18 yr
V	p.L173R	HT	M	Caucasian	0.7	GHD, ACTHD TSHD	APH, EPP, hypoplastic stalk	Normal	Hypopituitarism diagnosed in infancy; unilateral undescended testis, puberty induced age 12.5 yr; normal gonadotropins on retesting
VI	p.L173R	HT	M	Caucasian	1.0	GHD, ACTHD, TSHD	APH, EPP	SOD, BL ONH, retinal detachment, PHPV (L)	Autistic spectrum disorder and sleep disturbance
VII[b]	p.R268C	HT	M	Pakistani	1.3	Normal	APH, dysgenesis of CC, Dandy-Walker malformation, cerebellar hypoplasia	SOD, BL ONH, atypical visual evoked potential	Epilepsy, developmental delay, sleep disorder
VIII	p.R268C	HM	M	African[c]	1.2[d]	GHD, TSHD, ACTHD, GnD	APH, EPP, hypoplastic stalk	SOD, Rt ONH	Neonatal hypoglycemia, BL undescended testes
IX[b]	p.R268C	HT	M	Caucasian	0.8	GHD, TSHD, ACTHD, GnD	EPP, absent infundibulum	SOD, BL ONH	Micropenis, BL undescended testes, developmental delay
X	p.R268C	HT	M	Caucasian	1.4	GHD, TSHD	Agenesis of CC, focal malformation of Rt mesial frontal cortex	SOD, BL small ODs	Neonatal hypoglycemia, congenital heart disease, developmental delay, GORD
XI	p. G371R	HT	M	Caucasian	4.5	GHD	APH	SOD, BL ONH, astigmatism (L)	Mild aggression

Abbreviations: HT, heterozygous; HM, homozygous; M, male; F, female; Ht, height; Wt, weight; SDS, SD score; GnD, gonadotrophin deficiency; DI, diabetes insipidus; APA, anterior pituitary aplasia; APH, anterior pituitary hypoplasia; EPP, ectopic posterior pituitary; PP, posterior pituitary; CC, corpus callosum; ODs, optic discs; ONH, optic nerve hypoplasia; BL, bilateral; L, left; Rt, right; PHPV, persistent hyperplastic vitreous; GORD, gastroesophageal reflux disease. Age denotes age at presentation. The table shows endocrine deficits, ocular phenotypes, and results of MRI in patients with PROKR2 variations.

Patient deceased at age 24 years; cause of death unknown.

Patient phenotype has been recently reported (22).

Patient of mixed Black African and Caribbean origin.

Initial presentation with neonatal hypoglycemia; commenced hydrocortisone and T4 treatment on day 7 of life.

Patients with PROKR2 (p.L173R) variant

The 4 patients (III, IV, V, VI) carrying the heterozygous c.518T>G, p.L173R variant were all Caucasian and presented with variable phenotypes. Two patients (IV, VI) had SOD, and all four patients had multiple anterior pituitary hormone deficiencies, with the additional diagnosis of diabetes insipidus in patient IV (Table 1 and Supplemental Table 1). Patient III presented in the neonatal period with profound hypoglycemia and was diagnosed with cortisol deficiency (cortisol <30 nmol/L). Despite hydrocortisone treatment, she had increasing glucose requirements to maintain normoglycemia (up to 15 mg/kg/min) and had an inappropriately increased insulin concentration (11.4 mU/L) when hypoglycemic (blood glucose <3 mmol/L). A diagnosis of congenital hyperinsulinism was made, and diazoxide was commenced. Hyperinsulinism resolved by the age of 1 year, and the diagnosis of GHD was subsequently confirmed by glucagon provocation with commencement of recombinant human GH. She has since been diagnosed as having gastrointestinal (GI) dysmotility. In 3 of the 4 patients, the heterozygous variation was inherited from 1 of the parents. Only 1 of the parents manifested a phenotype; the heterozygous mother of patient V exhibited mild anosmia and delayed menarche with reported normal fertility. Surprisingly, the mother of patient VI was homozygous for the p.L173R variation and yet asymptomatic (Figure 2), with no evidence of abnormal gonadotropin secretion (peak LH, 30.3 IU/L; peak FSH, 8.6 IU/L, in response to GnRH; estradiol, 593 pmol/L). She had 3 other children (without fertility treatment) who were also heterozygous for the variation (data not shown). One male sibling of patient VI had a sleep disorder with behavioral problems, whereas another male sibling had epilepsy; their older sister was phenotypically normal.

Figure 2.

Parental screening of patient VI and pedigree mapping. A, Direct sequencing of PROKR2 in patient VI and his parents shows that the patient is heterozygous for the p.L173R variation (bottom panel, arrow), whereas the mother is homozygous (top panel) and the father has the wild-type sequence (middle panel). B, The genotypes were confirmed by restriction digest of PCR products using the BsaWI enzyme that does not cut in the WT sequence, thus leaving a band of 596 bp. Incubation with the patient's DNA results in 3 fragments, including the full-length product plus the cleaved 342- and 254-bp products, whereas the parent homozygous for the variant only shows the 342- and 254-bp products. C, Pedigree map of the patient's family presenting 2 generations. Black shading represents affected members, whereas unaffected carriers are marked by gray. The proband (II.4) was heterozygous for the p.L173R variation, whereas his mother (I.2) was the unaffected, homozygous carrier. She had 3 other children (II.1–3), all of whom were heterozygous for the variation but lacked a hypopituitary phenotype.

Patients with PROKR2 (p.R268C) variant

The PROKR2 c.802C>T, p.R268C variant was detected both in heterozygosity (VII, IX, X) and in homozygosity (VIII). All 4 patients were male and had SOD. Three patients had optic nerve hypoplasia and a hypoplastic anterior pituitary on MRI, whereas patient X had SOD with agenesis of the corpus callosum and small optic discs, but no anterior pituitary hypoplasia at the time of presentation (Table 1 and Supplemental Table 1). Patients VII and X had additional brain abnormalities (cerebellar hypoplasia, Dandy-Walker cyst, focal abnormality of mesial frontal cortex) with/without epilepsy and developmental delay. Three of the 4 patients had multiple pituitary hormone deficiencies. Although patient VII has not yet developed any pituitary hormone deficiencies, he is under regular clinical follow-up. Both parental DNA samples were only available for patient IX; the unaffected father was the heterozygous carrier. Only the maternal DNA sample for patient X was available, and she was not a carrier; parental DNA could not be obtained for patients VII or VIII.

Patients with other PROKR2 variations

Patient I is a female of Chinese origin with SOD who first presented at the age of 11 months with bilateral optic nerve hypoplasia and pigmentary changes of the right optic nerve. Pituitary MRI showed a hypoplastic anterior pituitary, with a eutopic posterior pituitary. She developed GHD (peak GH to stimulation, 3.3 μg/L) by the age of 6 years and commenced treatment with recombinant human GH. She entered puberty (breast stage 2) at the age of 10.3 years, and she is followed up regularly. To date, no other pituitary hormone deficiencies have been identified (Table 1 and Supplemental Table 1). Sequencing analysis revealed a heterozygous missense variation in PROKR2 (c.151G>A, p.A51T). Her unaffected mother was also heterozygous for this change. Although p.A51T occurs at a highly conserved residue, it has also been detected in 1 of our 480 controls and has recently been determined as functionally benign (21). Therefore, no further functional work was conducted.

The novel PROKR2 sequence variant (c.1111C>G, p.G371R) was detected in heterozygosity in a male patient (XI). He presented with SOD including unilateral optic nerve hypoplasia, anterior pituitary hypoplasia, and GHD (Table 1 and Supplemental Table 1). The sequence variant occurred at a highly conserved residue located within the intracellular C-terminal region of PROKR2 and was not detected in any of our 480 controls. The total amount and the amount at the cell surface of the variant are similar (Figure 3A), indicating that the trafficking properties of the p.G371R variant are not impaired compared to the wild-type receptor. Further functional analysis showed that the signaling activity of the variant was similar to that of the wild-type receptor for both Ca2+ release and IP1 accumulation (Figure 3, B and C). As shown in Figure 3D, the accumulation of cAMP from the variant after PROK2 stimulation was also comparable to that of the wild-type PROKR2. These results showed that the p.G371R variant does not alter either Gq or Gs signaling pathways, although we cannot exclude the possibility that this variation may cause defects in other aspects of PROKR2 signaling, such as Gi-protein coupling, which were not investigated in this study.

Figure 3.

Functional analysis of the HA-tagged novel variant G371R. A, Surface and total cell levels of the PROKR2 mutants. Amounts of HA-tagged receptors at the cell surface (nonpermeabilized cells, gray histograms) and in permeabilized cells (black histograms) were quantified by ELISA. B–D, Functional assays comparing wt PROKR2 and the PROKR2 p.G371R variant against negative control (mock). Induction of protein signaling through either the novel mutant or wt PROKR2 receptor by ligand PROK2 treatment resulted in similar levels (P > .05) of intracellular Ca2+ turnover (B) and IP1 (C) and cAMP (D) accumulation, indicating that p.G371R is not a pathogenic variant. Data are presented as mean ± SEM of 3 independent experiments.

Analysis of the hypothalamic-pituitary axis in the Prokr2-null embryos

Using immunohistochemistry against a variety of pituitary terminal differentiation markers (GH, ACTH, TSH, and α-glycoprotein subunit [gonadotropes and thyrotropes]), we aimed to investigate whether the phenotype in Prokr2 knockout mice at E18.5 would be comparable to the phenotypes we observed in our patients (Figure 4). Extensive Prokr2 expression has been reported in the developing preoptic region of the brain that contains the hypothalamic nuclei (24), whereas in humans, expression of PROKR2 has been shown by RT-PCR in the pituitary gland and central nervous system postnatally (25, 26). Expression of each of the markers was consistent between wild-type and mutant mice with a reduction in pituitary size, consistent with a reported reduction in global embryo size in the mutants (17). LH-immunoreactive cell number was significantly reduced in Prokr2−/− embryos (Figure 4, K, L, K′, L′, and M). In situ hybridization using specific probes to the endocrine hypothalamic neurons (median eminence and paraventricular/supraoptic nuclei) expressing Avp, OT, and Ghrh showed no overt differences between wild-type and Prokr2−/− embryos at E18.5 either morphologically (Figure 5, A–F) or quantitatively (Figure 5H). Our results indicate that mProkr2 is dispensable for proper formation of the hypothalamic-pituitary axis. The down-regulation in LH in the Prokr2-null embryos agrees with the role of mProkr2 in regulating GnRH neuronal migration previously reported by Matsumoto et al (16).

Figure 4.

mProkr2-deficient embryos exhibit normal terminal differentiation of hormone-producing cells. Coronal sections through the pituitary gland of the wild-type, Prokr2+/+ (A, C, E, G, I, K) and Prokr2−/− (B, D, F, H, J, L) embryos immunostained against GH (A, B), ACTH (C, D), TSH (E, F), α-glycoprotein subunit hormone (G, H), prolactin (I, J), and LH (K, L). Immunoreactivity of GH, ACTH, TSH, α-GSU, and PRL show no difference between wild-type and Prokr2-deficient embryos. Although the pituitary glands of Prokr2−/− embryos appear smaller, this is due to an overall reduction in head size of these mutant embryos. The number of LH-expressing cells appears reduced in Prokr2−/− embryos (K, L, and K′, L′, enlarged boxed areas in K and L, respectively), and quantification of LH cells demonstrates a statistically significant reduction in Prokr2−/− P < .05 (M). α-GSU, α-glycoprotein subunit hormone; PRL, prolactin. Scale bar in A represents 150 μm.

Figure 5.

Prokr2 null embryos exhibit normal development of the neuroendocrine hypothalamus. Coronal sections through the brain of mouse embryos (represented in G) at E18.5, wild-type Prokr2+/+ (A, C, E), and mutant Prokr2−/− (B, D, F), hybridized with Avp (A, B), OT (C, D), and Ghrh (E, F). Avp- and OT-expressing neurons in the paraventricular and supraoptic nuclei appear similar between Prokr2+/+ and Prokr2−/− embryos, suggesting that Prokr2 is not required for the development of these structures. Expression of Ghrhr in the arcuate nucleus at the level of the median eminence appears unaltered between genotypes (E, F). H, Plotted graphs representing numbers of AVP, OT, and Ghrh neurons indicate no difference in the number of AVP, OT, and Ghrh neurons between wild-type (white columns) and Prokr2−/− embryos. arc, Arcuate nucleus; pvn, paraventricular nuclei; son, supraoptic nuclei; me, median eminence. Scale bar in A represents 300 μm.

Discussion

In this study, we have identified 11 patients with variable congenital hypopituitarism/SOD, who presented with sequence variations in PROKR2 (Table 1). Because the parental carriers included both maternal and paternal carriers, there is no suggestion of a parent of origin effect. Despite its established role in KS (14), we could not implicate the corresponding ligand, PROK2, in hypopituitarism. Our results of the largest cohort of patients with congenital hypopituitarism, including both CPHD and SOD, screened to date are consistent with our recently published data (20). Here, we report the identification of PROKR2 variants in patients with craniofacial/midline disorders and hypopituitarism, thus suggesting an overlap in genotypes/phenotypes between these conditions and KS (20). In our cohort, 9 of 11 patients were found to harbor previously described PROKR2 variations that had been shown to be functionally deleterious in vitro; the lack of dominant-negative effects of these variants suggests that their functional significance in vivo remains debatable (13). These variations are present in approximately 2% of our cohort with SOD; thus PROKR2 variations occur more frequently than any other genetic abnormalities identified in association with SOD to date (27). However, the extent to which these variations contribute to the phenotype is yet to be established. Of the 9 patients above, PROKR2 variations were detected both in homozygosity (n = 1) and in heterozygosity (n = 8). Interestingly, 1 patient (VI) with SOD/CPHD and structural pituitary abnormalities was heterozygous for the PROKR2 p.L173R variation, whereas his phenotypically normal mother was homozygous for the same variant; this variant has previously been identified in several patients with KS and was shown to disrupt cell-surface targeting of the receptor in vitro (12, 13, 28). We sought to better understand the phenotypes observed in our patients and the lack of a phenotype in the healthy homozygous mother by investigating a possible role for mProkr2 in the development and integrity of the hypothalamic-pituitary axis, using Prokr2 null embryos as a model. Our study of the homozygous knockout Prokr2−/− mice revealed a morphologically normal pituitary and hypothalamus with normal hormone-secreting cells except for LH+ve gonadotrophs, which were significantly reduced in the mutants and consistent with the previously reported KS-like phenotypes that these mice exhibit (16, 17). These observations suggest that in the absence of Prokr2, the murine hypothalamic-pituitary axis develops normally. Although this may not be a human model, the parallels between the normal hypothalamic-pituitary axis between our knockout mice and the healthy mother of patient VI are compelling and suggest that the variants identified in this report may not be sufficient to cause a hypothalamic-pituitary phenotype in isolation. Any contributory mechanisms of Prokr2/PROKR2 in the pathogenicity of hypopituitarism remain to be proven. In this cohort, the mother who was homozygous for the PROKR2 p.L173R variation clearly did not have KS (normosmic and fertile with 4 children that were conceived without assistance, and with normal gonadotropins and estrogen). In addition, 2 of the 4 patients who had reached pubertal age (II, VIII) required the induction of puberty and remained on testosterone treatment into adulthood, whereas the other 2 (IV, V) progressed through puberty spontaneously and remained off sex steroid treatment. Therefore, the possibility that Prokr2/PROKR2 is not, in isolation, causative of KS either must be considered; it may, however, contribute by modifying a phenotype caused by a defect in another gene(s) or environmental factor(s), as recently postulated by Raivio et al (20).

Comparison of our results with those already in the literature supports the above notion. First, we detected functionally significant variants at the p.R85 and p.R268 alleles in heterozygosity or homozygosity in patients with phenotypes ranging from CPHD to SOD. They had previously only been detected in KS and healthy controls (as had the p.L173R variant), albeit only in heterozygosity in the latter (12). Our data therefore strongly suggest that another gene or environmental factor is causing the more severe CPHD-SOD phenotypes. Indeed, although not proven in any of our hypopituitary patients, digenic cases of KS involving PROKR2 have been reported previously (12, 15, 28–30). In a recent study by Sarfati et al (28), male KS patients presenting with homozygous variations in PROKR2 were significantly more likely to exhibit cryptorchidism, microphallus, lower mean testicular volumes, and circulating gonadotropins than their heterozygous counterparts (28). This difference in the gene-dosage of PROKR2 supports a contributory role to the pathogenesis of KS. No differences were observed between homozygous or heterozygous females, although this was attributed to the very low number of cases of the former (n = 4). Thus, the extent to which PROKR2 variants contribute to either hypopituitarism or KS-associated phenotypes remains to be established. Care must be taken with the interpretation of results when patients presenting with such disorders also exhibit variations in the PROKR2 gene, particularly with respect to genetic counseling. Other genes known to be associated with these disorders should be screened and, should these be negative for mutations, then one cannot rule out the presence of an as yet unidentified mutated gene or possible environmental factors.

In addition to presenting with CPHD/SOD, some of our patients with PROKR2 variants had additional manifestations including epilepsy, sleep abnormalities, GI dysmotility, and diabetes insipidus. Although such variants have been detected for the first time in patients with the latter 2 disorders herein, a possible association is not altogether surprising given that prokineticin ligands and receptors are involved in various systems and processes including circadian rhythms in the brain, angiogenesis, neurogenesis, pain perception, immune responses, hematopoiesis, reproduction, and GI smooth muscle contraction (31–37). Epilepsy, sleep disorders, and diabetes insipidus are suggestive of a forebrain/hypothalamic phenotype (35, 36, 38). Avp and Prok2 mRNA both colocalize to a significant number of hypothalamic neurons in rats, and there is evidence of interaction between these pathways through AVP receptor-null mice (39, 40). However, Prokr2 knockout mice exhibited a morphologically normal hypothalamus with quantitatively normal expression of Avp, OT, and Ghrh. These data do not, however, exclude a role for PROKR2 in postnatal development of these hypothalamic disorders in our patients. However, considering that our patient with diabetes insipidus and epilepsy (IV) was heterozygous for the same variant that the unaffected mother of patient VI had in homozygosity, it appears unlikely that mutated PROKR2 is causative of these phenotypes. Additionally, we cannot definitively state that the GI dysmotility phenotype in patient III was caused by her heterozygous PROKR2 variation, this again being the same as the mother of patient VI. Any contributions of PROKR2 variants to these phenotypes would necessitate further studies, particularly with the aid of a postnatal murine model.

We have identified variations in PROKR2 at a higher frequency in SOD than any other previously described genetic factor; we also describe other clinical features in association with these variations, including GI and hypothalamic disorders (eg, diabetes insipidus). However, the role of PROKR2 is controversial; heterozygous and homozygous variants occurring across the protein induce comparable phenotypes or, as we have shown, more severe phenotypes in cases of the former than the latter. This is compounded by the incidence of homozygosity in the healthy mother of a heterozygous child with a severe form of hypopituitarism in the form of SOD. Our analyses of the pituitary and hypothalamus in Prokr2 knockout mice are largely inconsistent with the patient phenotypes, yet strongly support the normal presentation of the mother and her 3 unaffected (with respect to hypopituitarism) heterozygous children, although one needs to note the reduced LH in the murine null mutants. The number of genetically assigned causes of both KS and hypopituitarism is low, accounting for 30% of KS and 5–10% of SOD and related midline disorders, and given that none of our patients harbored mutations in any of the known genes in these disorders, there are clearly other genetic/environmental factors yet to be discovered. Subjects with sequence variants in PROKR2 represent a unique cohort; further careful genetic investigation is likely to aid in the identification of the missing genetic or epigenetic modifiers that interact with this pathway in humans to account for the phenotypic heterogeneity. Further research into uncovering these additional factors would help define the role, if any, of PROKR2 in these and other disorders discussed herein.