Responsiveness to a Physiological Regimen of GnRH Therapy and Relation to Genotype in Women With Isolated Hypogonadotropic Hypogonadism

Abstract

Context:

Isolated hypogonadotropic hypogonadism (IHH) is caused by defective GnRH secretion or action resulting in absent or incomplete pubertal development and infertility. Most women with IHH ovulate with physiological GnRH replacement, implicating GnRH deficiency as the etiology. However, a subset does not respond normally, suggesting the presence of defects at the pituitary or ovary.

Objectives:

The objective of the study was to unmask pituitary or ovarian defects in IHH women using a physiological regimen of GnRH replacement, relating these responses to genes known to cause IHH.

Design, Setting, and Subjects:

This study is a retrospective analysis of 37 IHH women treated with iv pulsatile GnRH (75 ng/kg per bolus).

Main Outcome Measures:

Serum gonadotropin and sex steroid levels were measured, and 14 genes implicated in IHH were sequenced.

Results:

During their first cycle of GnRH replacement, normal cycles were recreated in 60% (22 of 37) of IHH women. Thirty percent of women (12 of 37) demonstrated an attenuated gonadotropin response, indicating pituitary resistance, and 10% (3 of 37) exhibited an exaggerated FSH response, consistent with ovarian resistance. Mutations in CHD7, FGFR1, KAL1, TAC3, and TACR3 were documented in IHH women with normal cycles, whereas mutations were identified in GNRHR, PROKR2, and FGFR1 in those with pituitary resistance. Women with ovarian resistance were mutation negative.

Conclusions:

Although physiological replacement with GnRH recreates normal menstrual cycle dynamics in most IHH women, hypogonadotropic responses in the first week of treatment identify a subset of women with pituitary dysfunction, only some of whom have mutations in GNRHR. IHH women with hypergonadotropic responses to GnRH replacement, consistent with an additional ovarian defect, did not have mutations in genes known to cause IHH, similar to our findings in a subset of IHH men with evidence of an additional testicular defect.

Isolated hypogonadotropic hypogonadism (IHH) is a rare genetic disorder that results in failure to undergo normal pubertal development. Patients with IHH harbor mutations in genes involved in GnRH neuronal migration as well as GnRH synthesis and secretion (1). In most patients with IHH, physiological GnRH replacement restores normal levels of pituitary and gonadal hormones, allowing for testicular growth and spermatogenesis in men and ovulation in women (2, 3), thus implicating GnRH deficiency as the primary underlying etiology of the reproductive phenotype in IHH patients.

However, decreased pituitary responsiveness to GnRH has been described in IHH women in association with mutations in the GnRH receptor (4), indicating that in these patients, the defect lies at the level of the pituitary. Furthermore, GnRH replacement studies have unmasked functional defects at the level of the testes as well as the pituitary in subsets of IHH men (3). These findings are of particular interest because our growing understanding of the genetic architecture of IHH indicates that many of the genes that cause this syndrome are expressed in the mammalian pituitary (5–10) or gonad (11–19) in addition to the brain.

We therefore hypothesized that identification of women with abnormal responses to physiological GnRH replacement would provide greater insight into the pathogenesis of IHH. Hypo- or hypergonadotropic responses to GnRH would implicate pituitary or ovarian defects, respectively. Such findings would suggest that genes involved in gonadotrope or ovarian development and function that are also expressed in the hypothalamus should be given greater consideration in the search for new IHH genes.

Materials and Methods

IHH subjects

Thirty-seven women with IHH who received iv pulsatile GnRH treatment at a dose of 75 ng/kg per bolus during their first treatment cycle at the Reproductive Endocrine Unit of the Massachusetts General Hospital between 1980 and 2012 were included in this study. An additional 63-year-old IHH woman was treated with iv GnRH for 7 days to provide a basis for comparison of FSH levels in younger subjects found to have an ovarian defect. The diagnostic criteria for IHH have been previously described in detail (20). This study was approved by the Partners Human Research Committee and signed informed consent was obtained from each subject prior to participation.

Baseline assessment

A detailed history was taken and physical examination performed on all subjects. Obesity was defined as a body mass index (BMI) greater than 30 kg/m². Subjects were classified as anosmic based on a 40-item University of Pennsylvania Smell Identification Test score of the fifth percentile or less for age (21) or on self-report. Hormone replacement therapy was discontinued for at least 1 month prior to the baseline biochemical assessment. Frequent blood sampling (every 10 minutes for 12 hours) was performed in all subjects to assess endogenous GnRH secretion as manifested by LH pulsatility. FSH, estradiol (E2), and progesterone (P4) were assayed from pools of the frequent sampling studies. Pulsatile LH secretion was analyzed as previously described (22, 23), and studies were categorized as pulsatile or apulsatile.

Response to physiological GnRH replacement

Pulsatile GnRH was administered using a closed iv system, as previously described (2). Daily blood samples for LH, FSH, E2, and P4 were drawn 40 minutes after a GnRH bolus at the same time of day. Beginning on day 7, serial pelvic ultrasounds were performed to monitor follicular growth and document ovulation. Women with IHH who did not ovulate in response to the standard dose of GnRH received additional treatment at the same or higher doses.

Genotyping

Genomic DNA was obtained from peripheral blood lymphocytes in 31 of 37 subjects. Exonic and proximal intronic (>15 bp from splice sites) DNA sequences of 14 genes implicated in the etiology of IHH were amplified by PCR and determined by direct sequencing. These genes include KAL1 (anosmin-1), GNRH1 (GnRH 1), GNRHR (GnRH receptor), KISS1 (kisspeptin), KISS1R (KISS1 receptor), NELF (nasal embryonic LHRH factor), FGF8 (fibroblast growth factor 8), FGFR1 (fibroblast growth factor receptor 1), PROK2 (prokineticin 2), PROKR2 (prokineticin receptor 2), TAC3 (tachykinin 3), TACR3 (tachykinin receptor 3), DAX1 (dosage-sensitive sex reversal, adrenal hypoplasia critical region, on chromosome X, gene 1), and CHD7 (chromodomain-helicase-DNA binding protein 7). PCR primers and amplification conditions for each gene have been described previously (24–33). Mutations were defined as rare sequence variants (RSVs) that were: 1) within coding regions ± 5 bp; 2) nonsynonymous; 3) present in less than 1% of alleles in 192 unaffected controls who had normal reproductive function by history and physical examination; and 4) not categorized as polymorphisms by the 1000 Genomes Browser (http://browser.1000genomes.org/).

Functional analyses

The in vitro functional data for the FGFR1, GNRHR, PROK2, and PROKR2 genes have been previously reported (25, 26, 32, 34–37). In instances in which in vitro data were not available, 5 different prediction algorithms were used to determine the potential significance of missense variants and 1 for intronic changes [PolyPhen (38), Mutation Taster (39), Panther (40), SIFT (41), pMUT (42), and Human Splicing Finder (43)].

Controls

One hundred three unaffected, regularly cycling (25–35 days) women, aged 19–42 years, provided daily blood samples across a complete menstrual cycle. All cycles were determined to be ovulatory based on a luteal phase P4 level of greater than 6 ng/mL (19 nmol/L). Eighty-one of these women served as controls in a previous study (44).

Assays

Serum LH and FSH were analyzed by RIA or using a 2-site monoclonal nonisotopic system (Axsym; Abbott Laboratories, Abbott Park, Illinois) calibrated with the same reference preparation used in the RIA (45) and expressed in international units per liter of the Pituitary Second International Standard 80/552. E2 was initially measured by RIA and then by using a 2-site monoclonal nonisotopic system (Axsym; Abbott Laboratories) with a functional sensitivity of 20 pg/mL (73.4 pmol/L), as previously described (46). P4 was measured by RIA or using a chemiluminescent enzyme immunoassay (ELISA; Immulite, Diagnostic Products Corp, Los Angeles, California) as previously described (46).

Data analysis

In subjects and controls, the day of ovulation was defined according to previously validated criteria (47). Data from the daily sampling protocol in controls were used to interpret the gonadotropin responses to GnRH observed in IHH women. Ovulation occurred no later than cycle day 20 in the controls. Therefore, GnRH treatment cycles were classified as normal if ovulation occurred on or prior to treatment day 20 and abnormal if ovulation occurred after day 20 or if the subject did not ovulate.

Gonadotropin and E2 levels during cycle days 1–7 in controls and during the first 7 days of pulsatile GnRH replacement (75 ng/kg per bolus) in IHH women with normal cycles were expressed as interquartile ranges [(IQRs) 25th to the 75th centile]. Hormonal data from IHH women with normal cycles were referenced against that from controls. Data from IHH women with abnormal responses to GnRH were then referenced against IHH women with normal cycles. Women with abnormal responses demonstrated either an inadequate gonadotropin response to GnRH consistent with pituitary resistance or an exaggerated FSH response to GnRH, indicating underlying ovarian resistance.

Phenotypic and genotypic characteristics of women in each group were compared using ANOVA and expressed as mean ± SD or using χ². BMI was not normally distributed and was expressed as median and IQR. Linear regression was used to determine the relationship between age and peak FSH level during the first 7 days of treatment in IHH women with normal cycles. Peak FSH was log transformed prior to analysis. A similar analysis was performed to determine whether there was an inverse relationship between BMI and peak LH in IHH women with normal cycles as has been demonstrated in both normal cycling women (48) and in women with polycystic ovarian syndrome (48–50). P < .05 was considered to be significant unless otherwise noted.

Results

Baseline characteristics

The clinical and biochemical features of the 37 women with IHH are summarized in Tables 1 and 2. All subjects were 16 years old or older at the time of treatment (mean age 29.8 ± 7.3 years), had low E2 levels in the face of low or inappropriately normal gonadotropins, no other pituitary hormone deficiencies, and no neuroanatomic cause of IHH. Forty-six percent of the women were anosmic.

Table 1.

Baseline Characteristics of the Cohort of Women With IHH Treated With iv GnRH

	Normal IHH	Abnormal IHH
		Pituitary Resistance	Ovarian Resistance
n	22	12	3
Anosmia, %	55%	42%	0%a,b
Age, y	28.5 ± 7.4 (16.8–41.9)	30.5 ± 6.9 (17.6–39.4)	36.0 ± 4.9a (31.2–41.0)
BMI, kg/m2, median (IQR)	22.4 (19.2–25.9)	28.0 (22.9–36.5)b	22.8 (22.6–25.9)
Partial puberty, %	46%	75%	0%
FSH, IU/L	0.87 (0.24–2.87)	0.44 (0.22–0.99)b	0.57 (0.50–0.70)b
LH, IU/L	0.28 (0.09–1.00)	0.20 (0.08–0.44)	0.34 (0.23–1.19)
Apulsatile baseline study, %	52%	83%	100%b
Ovarian volume, mLc	1.6 (1.4–2.7)	1.9 (1.0–2.2)	1.5 (1.1–2.4)

^aP < .05 vs pituitary resistance.

^bP < .05 vs normal response cycles.

^cMean ovarian volume in normal spontaneously cycling women is 9.5 mL (95% confidence interval 3.9–15.9 mL) (48).

Table 2.

Phenotypic and Genotypic Characteristics of Women With IHH Treated With iv GnRH

Subject no.	Age, y	BMI, kg/m2	Baseline Study	Partial Puberty	Olfactory Phenotype	Nonreproductive Phenotype	Gonadotropin Levels (d 1–7)a		Genetics
							FSH	LH	Gene	Variant	Functional Data (Reference)	Prediction Programsb
Normal response
1	20.0	22.7	A	N	nIHH	High arched palate, scoliosis			TACR3	p.W208X	Nonsense
2	22.8	21.3	A	N	nIHH				Neg
3	36.9	23.5	P	Y	nIHH	Cleft lip and palate, widely-spaced teeth			Neg
4	28.9	20.0	A	N	nIHH	Partial empty sella			N.A.
5	36.4	28.3	P	Y	nIHH	Marfanoid habitus			Neg
6	17.6	18.0	A	N	KS	Ocular defect			CHD7	c.3202–5 T>C (splice)		0c
7	26.6	33.6	A	Y	KS	Ocular defect, short fourth metacarpal			FGFR1	p.L630P		4
8	33.2	18.0	A	N	KS				Neg
9	38.6	25.4	P	Y	KS	Hearing loss			KAL1	p.K185N		2
10	31.5	18.8	A	N	KS				Neg
11	23.0	21.7	P	N	KS				N.A.
12	16.8	21.5	P	N	KS				Neg
13	18.5	18.9	A	N	KS	Clinodactyly			Neg
14	31.8	25.9	P	Y	KS				Neg
15	25.5	24.3	P	Y	nIHH				N.A.
16	41.9	35.6	P	Y	nIHH				Neg
17	28.5	22.1	P	N	nIHH				Neg
18	23.8	25.6	A	N	nIHH				TAC3	p.R80S		4
19	38.6	18.2	A	N	KS	Flat nasal bridge, PDA, hypoplastic fifth finger (CHAR syndrome)			N.A.
20	32.2	26.5	P	Y	KS				Neg
21	34.8	18.1	A	Y	nIHH				N.A.
22	19.9	26.9	A	Y	KS	Widely-spaced teeth			CHD7	p.L373F		1
Abnormal, pituitary resistance
23	23.4	20.9	A	Y	nIHH		↓	↓	Neg
24	27.9	21.7	P	N	KS	Short fourth metacarpal	NI	↓	FGFR1, GNRHR	p.R470L; p.Q106R, p.R262 (all partiald)	(26, 35, 37)
25	39.4	42.8	A	Y	KS	Clinodactyly, flat nasal bridge	↓	↓	Neg
26	30.0	23.3	A	N	nIHH	Partial empty sella	NI	↓	PROKR2	p.L173R (partial)	(25)
27	35.5	28.2	A	Y	nIHH		↓	↓	GNRHR	p.M1T + p.R139H (complete); p.Q106R (partial)	Start codon (35)
28	29.9	43.8	A	Y	nIHH		↓	↓	Neg
29	43.0	36.0	A	Y	nIHH	Widely-spaced teeth	↓	↓	Neg
30	27.1	37.7	A	Y	KS		NI	↓	Neg
31	33.5	27.7	A	Y	nIHH		↓	↓	N.A.
32	17.6	26.4	A	N	KS		↓	↓	Neg
33	32.3	32.4	A	Y	KS	Syndactyly, pes planus	↓	↓	Neg
34	26.7	20.1	P	Y	nIHH		NI	↓	GNRHR	p.N10K+p.Q11K (partial); p.P320L (complete)	(34, 36)
Abnormal, ovarian resistance
35	41.0	22.3	A	N	nIHH		↑	↑	Neg
36	35.9	22.8	A	N	nIHH	Mitral valve prolapse	↑	↑	Neg
37	31.2	28.9	A	N	nIHH	Cleft palate	↑	NI	Neg
63-year-old woman
38	63.0	23.5	A	N	nIHH		↑	↑	PROK2	p.155fsX1 (complete)	(32)

Abbreviations: A, apulsatile; KS, Kallmann's (anosmic or hyposmic IHH); N, no; N.A., DNA sample not available; Neg, negative; NI, normal; nIHH, normosmic IHH; P, pulsatile; Y, yes; ↑, high; ↓, low.

^aGonadotropin levels during the first week of GnRH treatment relative to normal IHH responders.

^bNumber of prediction programs (of 5) predicting this mutation to be deleterious.

^cA single prediction program, Human Splicing Finder, does not predict that this variant will alter splicing.

^dRefers to partial or complete loss of function.

Characterization of response to physiological pulsatile GnRH replacement

Normal response (n = 22)

Sixty percent of IHH women (22 of 37) ovulated during their first cycle of pulsatile GnRH at a dose of 75 ng/kg per bolus on or before day 20 (day 14.8 ± 2.2; range 11–18 days). FSH rose to the range seen in controls during the early follicular phase within 24–48 hours of GnRH treatment, whereas LH and E2 lagged slightly, reaching the normal range at 72 hours (Fig. 1). Seventy-one percent of women developed a dominant follicle (≥11 mm) by cycle day 10, in line with the follicular growth rate seen during normal menstrual cycles (51). The peak FSH response to GnRH treatment during week 1 increased with age (P = .03, r = 0.5). Baseline LH was inversely associated with BMI (P = .05, r = 0.4), but there was no relationship between BMI and baseline FSH (P = .1), peak LH (P = .2), or peak FSH (P = .9).

Figure 1.

In the 22 IHH women who ovulated within the normal time frame in response to initial exposure to a physiological dose of pulsatile GnRH (75 ng/kg, iv), gonadotropin and estradiol levels (mean ± 1 SD) rose to the normal range observed in the early follicular phase of regularly cycling women (shaded area, IQR) within the first few days of treatment.

Abnormal response (n = 15)

Fifteen IHH women did not ovulate during the first 20 days of their first GnRH treatment cycle (Fig. 2).

Figure 2.

A, Gonadotropin and estradiol levels (mean ± 1 SD) remained low in response to iv pulsatile GnRH (75 ng/kg) in 12 women, suggesting resistance at the level of the pituitary. B, A hypergonadotropic response (mean ± 1 SD) was observed in 3 other women who did not ovulate during the first 20 days of the GnRH treatment cycle. The FSH response in this group is similar to that seen during the first week of GnRH therapy in a 63-year-old IHH woman (squares). Shaded region represents the interquartile range for the 22 IHH women with normal responses to GnRH.

Pituitary resistance (n = 12)

Low LH and FSH levels (n = 8).

During the first week of treatment, 8 women demonstrated defects in both FSH and LH secretion in response to GnRH relative to IHH women with normal cycles (Fig. 2A). In 5 of the 8 women, gonadotropins eventually rose into the normal range after an additional week of treatment or with an increase in dose (100–150 ng/kg), allowing for normal dominant follicle development, ovarian E2 synthesis, and ovulation (in 4 of 5). In the remaining 3 women, prolonged treatment and/or dose increases to 100 ng/kg were unable to normalize gonadotropin levels. One of the 3 chose to undergo an additional cycle with high-dose (250 ng/kg) GnRH and successfully ovulated.

Preserved FSH secretion (n = 4).

Four women demonstrated an isolated defect in LH secretion during the first 7 days of treatment. Three of the 4 grew a dominant follicle (21–30 mm); however, none ovulated until receiving higher doses of GnRH (100–150 ng/kg) in subsequent treatment cycles.

Ovarian resistance (n = 3)

Three women with IHH demonstrated an exaggerated FSH response to GnRH during the first week of treatment, with FSH levels rising significantly above the 75th centile of levels observed in IHH women with normal cycles (ages 17–42 years) or controls (age 20–42 years) (Fig. 2B). The pattern of FSH during the first 4 days of treatment in these patients closely matched the response in a 63-year-old IHH woman treated with pulsatile GnRH (Fig. 2B). All 3 women grew a dominant follicle in association with elevated FSH levels. Subject no. 37 (aged 31 years) had delayed folliculogenesis but ovulated with prolonged GnRH treatment (75 ng/kg). Subject no. 36 (aged 36 years) had persistently low E2 levels, even in subsequent cycles with higher doses of GnRH, but eventually ovulated with high-dose gonadotropin treatment (450 IU Menopur). Subject no. 35 (aged 41 years) had an LH surge but an insufficient luteal phase [peak P4 <3.0 ng/dL (9.6 pmol/L)] in response to the 75-ng/kg replacement dose. Investigation into other causes of ovarian resistance in these women revealed normal karyotypes (n = 3), negative ovarian antibodies (n = 1), and the absence of autoimmune disease (n = 3). FMR1 premutation testing was not performed.

Comparison of baseline characteristics among different response groups

There was no difference in the age of presentation among the different response groups. BMI was highest among women with abnormal/pituitary resistant responses (42% obese) (P < .04) (Table 1). Nonreproductive phenotypes previously associated with IHH were present in all groups (Table 2). Women with partial puberty (spontaneous breast development) were no more likely to demonstrate a normal response to GnRH than women with absent puberty (Table 1).

Baseline FSH, but not LH, levels were higher in women with normal responses to GnRH compared with those with abnormal responses (P < .05). E2 levels were at the limit of detection of the assay and ovarian volume was diminished in all IHH groups compared with reproductive-aged women (48) (Table 1). The higher FSH levels observed at baseline in the IHH women with normal responses is consistent with the presence of pulsatile LH secretion in nearly half of these women as slow frequency GnRH stimulation favors FSH secretion (52). The higher incidence of apulsatile LH profiles in the women with abnormal responses to GnRH may reflect a more profound defect in GnRH secretion or pituitary responsiveness to GnRH. Anosmia was present in 46% of the cohort but was limited to women with normal or pituitary-resistant response cycles (Table 1).

Comparison of genotypes among different response groups

Genotyping was performed in 17 of 22 IHH women with normal cycles and in 14 of 15 IHH women with abnormal cycles. Six women with normal cycles had a monogenic, monoallelic variant in 1 of the 14 genes examined (Table 2). There were 4 missense RSVs (FGFR1, CHD7, TAC3, and KAL1) predicted to be deleterious by 1–4 prediction programs, 1 nonsense RSV in TACR3, and 1 splice-site RSV in CHD7 not predicted to disturb splicing by the 1 prediction program used. Among the IHH women with pituitary resistance, 1 had a monoallelic loss of function (LOF) mutation in PROKR2 (25), 2 harbored 3 LOF mutations in GNRHR each (34–36), and 1 had a biallelic LOF mutation in GNRHR (26, 35) combined with a monoallelic LOF mutation in FGFR1 (53). No RSVs were identified in the women with ovarian resistance. The 63-year-old woman with IHH had a homozygous deletion in PROK2, resulting in a nonfunctional, truncated protein (32).

Discussion

In most women with IHH, physiological replacement of GnRH activates the dormant reproductive axis, allowing women who have never before been exposed to GnRH to achieve normal follicular phase levels of gonadotropins and sex steroids in a matter of days and to ovulate within 2 weeks. This finding stands in contrast to men with IHH who typically require months to a year of GnRH therapy due to the much longer time required for testicular growth and spermatogenesis (54). The rapid response to GnRH replacement in IHH women not only has immediate fertility benefits but also provides an early opportunity to identify women with abnormal responses to physiological GnRH replacement. Demonstration of hypo- or hypergonadotropic responses to GnRH localizes defects to the pituitary or ovary, respectively, that may not be apparent in the absence of GnRH.

Pituitary and testicular, in addition to hypothalamic, defects were identified by Sykiotis et al (3) in a subset of IHH men (23 of 90 men) through detailed examination of gonadotropin, testosterone, and spermatogenic responses to long-term GnRH administration. Genotype-phenotype (ie, response to GnRH) correlation studies in this group of men were limited because the atypical responders to GnRH were predominantly mutation negative (75%) or had an isolated defect in KAL1 (25%). Furthermore, it is impossible to attribute gonadal defects in adult men with IHH entirely to genetic defects due to the potential for confounding by neonatal cryptorchidism, which influences gonadal function in the adult male (55). Systematic investigation of the response to GnRH in IHH women allowed us to confirm the presence of gonadal defects in patients with IHH. In part due to the larger number of genes (14 vs 8) sequenced, we were also able to relate phenotype to genotype in the small number of IHH women with abnormal responses to GnRH who demonstrated a greater variety of genetic defects than observed in our cohort of IHH men (3).

Most women with abnormal responses to initial exposure to pulsatile GnRH had low gonadotropin levels, indicating pituitary resistance. None of these women had signs of other pituitary hormone deficiencies, and only 1 had an abnormal pituitary on magnetic resonance imaging (partial empty sella). As expected, this group encompassed all women with mutations in GNRHR. Two women with mutations in GNRHR had preserved FSH but impaired LH secretion in response to GnRH replacement (4, 36), consistent with in vitro studies demonstrating that LHβ gene transcription is more sensitive to GnRH than FSHβ (36). The third woman with a mutation in GNRHR demonstrated a profound defect in both LH and FSH responsiveness (56), potentially due to a more severe genetic defect in GnRH receptor signaling, as suggested by in vitro studies (26, 34–36). Importantly, in all women with GNRHR mutation, impaired receptor signaling was overcome with higher doses of GnRH and ovulation was achieved. Our observations in IHH women with mutations in GNRHR differ from those of Sykiotis et al (3), who observed a normal response to GnRH in an IHH man with mutations in both GNRHR and FGFR1. The differential response between patients with GNRHR mutations in the 2 studies may well be explained by differences in genetic burden; the male patient had a monoallelic partial LOF mutation in GNRHR, whereas the women in the current study had biallelic partial or biallelic partial and complete LOF mutations in GNRHR (56).

One woman with pituitary resistance to pulsatile GnRH treatment harbored a heterozygous mutation in PROKR2 (L163R), previously shown to severely compromise PROKR2 signaling and expression in vitro (25), and had a partial empty sella on magnetic resonance imaging. These findings are consistent with studies implicating PROK2 and PROKR2 in pituitary gland development. PROKR2 is expressed in the human pituitary gland (8), and patients with hypopituitarism and abnormal pituitary anatomy (57, 58) have been shown to have mutations in PROK2 and PROKR2. The patient in the current study had an isolated defect in LH secretion and achieved ovulation with a higher dose of GnRH, suggesting that pituitary resistance associated with a prokineticin signaling defect is relative and, importantly for therapy, surmountable with increased GnRH dosage. The normal follicular growth rate, E2 levels, and successful ovulation in this patient after normalization of gonadotropin levels with a higher dose of GnRH also indicate the absence of an additional ovarian defect, consistent with the observation that PROK2-deficient female mice ovulate in response to gonadotropins (32). Interestingly, an IHH man in the cohort of Sykiotis et al (3) with a heterozygous PROKR2 mutation (R164Q) that also causes a severe disruption in PROKR2 signaling and expression nearly identical to that of L163R in head-to-head studies (25), responded normally to the standard male GnRH treatment regimen. The 63-year-old IHH woman in the current series with a homozygous deletion in PROK2 (32) also responded normally to GnRH with the hypergonadotropic response expected for her age. Thus, the functional role of the prokineticin pathway at the pituitary is incompletely understood, and it is possible that a second mutation in a gene relevant to pituitary function will be identified in the pituitary-resistant IHH woman with the PROKR2 mutation that can explain her abnormal response to GnRH.

Studies in a number of animal models have demonstrated that the fibroblast growth factor family also plays an important role in early pituitary gland development (5, 6). In situ hybridization studies in human embryos localizing FGFR1 to Rathke's pouch (9), the precursor of the anterior pituitary, and the observation that patients with abnormal pituitary anatomy or function have mutations in FGF8 or FGFR1 (9, 57) strongly suggest that fibroblast growth factor signaling plays a key role in pituitary development in humans. Interestingly, the 1 IHH woman with a monogenic RSV in FGFR1 (L630P) demonstrated a normal response to GnRH, suggesting that partial defects in FGFR1 signaling alone do not interfere with GnRH action at the pituitary. Although in silico prediction programs suggest that the L630P mutation is deleterious, its effects on receptor signaling have not been investigated in vitro. Therefore, we cannot exclude the possibility that this woman demonstrated a normal response to GnRH because this RSV has a minor effect on FGFR1 function.

CHD7 is also expressed in the human (10) and mouse (59) pituitary, yet the 2 women with heterozygous mutations in CHD7 responded normally to GnRH replacement. In the absence of an in vitro assay for CHD7, it is difficult to assess the functional impact of these RSVs; however, several prediction programs suggest that they do not significantly alter protein function. Thus, these women may have been able to respond normally to GnRH because CHD7 function is preserved, and a different mutation in an unidentified gene is responsible for their IHH. Alternatively, the normal pituitary response to GnRH observed in women with mutations in CHD7, in combination with reports that mice with heterozygous loss of CHD7 demonstrate normal gonadotropin responses to a GnRH agonist (60), may indicate that the CHD7 protein acts predominantly upstream of the pituitary in hypothalamic neurogenesis and/or GnRH neuronal migration.

Importantly, two thirds of the IHH women with pituitary resistance were mutation negative, providing further evidence that additional genes important for pituitary function will be identified in the IHH patient population. Studies of transcripts coexpressed in the pituitary and hypothalamus suggest that mutations in genes involved in cell differentiation, intracellular transduction, vesicle transport, or transcription may yet be found in patients with IHH (61).

Overall, IHH women with pituitary resistance tended to be heavier than the other groups. An inverse relationship between BMI and LH responsiveness to GnRH has previously been reported in regularly cycling women and women with polycystic ovarian syndrome (48–50), and in the latter group, further studies have localized the effect of BMI to the pituitary (50). However, because obesity cannot explain the pituitary resistance to GnRH in all the women in this subgroup, it is also possible that the obese IHH women with pituitary resistance share a unique genetic defect that causes both HH and obesity, as in patients with mutations in leptin (62) and its receptor (63). Of note, in the current study, the GnRH dose was weight-adjusted and was administered iv, making it unlikely that inadequate dosing accounts for the hypogonadotropic responses observed in a number of obese IHH women.

A small group of IHH women responded to GnRH with supraphysiological levels of FSH, implying a degree of ovarian resistance. Dual defects at the hypothalamus and ovary may also explain why none of these women had early signs of pubertal development in contrast to the normal responders and pituitary-resistant groups. The women with ovarian resistance were also normosmic, suggesting that the gene responsible for this IHH subtype is likely to act after GnRH neuronal migration along the olfactory tracts during early embryonic development, although these conclusions are limited by the small number of women with ovarian resistance.

The rate of rise in FSH in response to GnRH replacement in the hypergonadotropic women was identical to that in a 63-year-old IHH woman of postmenopausal age. Although older age may have contributed to the hypergonadotropic response observed in the 41-year-old woman in the ovarian resistance group, age is unlikely to account for the responses of the other 2 women in this group. Furthermore, the control group and the groups of IHH women with normal or hypogonadotropic responses to GnRH also included women in their 30s and early 40s. Because the hypergonadotropic women responded to treatment with higher-dose (or prolonged) GnRH or high-dose exogenous gonadotropins, the secretion of biologically inactive gonadotropins is an unlikely explanation for their results.

Because most genes associated with IHH are expressed in the human ovary (12–16, 18), we were surprised that none of the women with ovarian resistance had a mutation in any of the 14 genes tested. Similarly, genetic defects were absent in most IHH men with testicular resistance (3). Taken together, these studies indicate that there is an additional set of unidentified genes expressed in the hypothalamus that are also relevant for ovarian function. Potential candidates include genes encoding neurosteroidogenic enzymes or intracellular transduction signals.

Physiological replacement of GnRH recreates normal menstrual cycle dynamics in most women with IHH. However, hypogonadotropic responses in the first week of treatment identify a subset of women with pituitary dysfunction, only some of whom have mutations in GNRHR. In a further subset, hypergonadotropic responses to GnRH replacement indicate the presence of dual defects at the hypothalamus and ovary. Identification of many of the genes responsible for IHH has resulted in an explosion of information regarding the central control of reproduction. Investigation of genes with overlapping expression patterns in the hypothalamus and pituitary or ovary in this patient population provides an important strategy for identification of novel genes that are critical for both the central and peripheral control of reproduction.