GnRH-Deficient Phenotypes in Humans and Mice With Heterozygous Variants in KISS1/Kiss1

Abstract

Defects in a number of genes have been implicated in the pathophysiology of GnRH deficiency. Several lines of evidence suggest that KISS1 is a likely candidate gene for GnRH deficiency. These include identification of inactivating mutations in KISS1R (a.k.a. GPR54) in patients with GnRH deficiency, targeted deletions of either KISS1 or KISS1R in mice with hypogonadotropic phenotypes, and the demonstrated role of kisspeptin in stimulation of GnRH neurons and initiation of puberty. However, no inactivating mutations in KISS1 have been found in patients with GnRH deficiency. The aim of this study was to identify deleterious mutations in KISS1. After a systematic evaluation of genetic variation identified rare sequence variants (RSV) in KISS1 among 1025 probands with GnRH deficiency, the frequency and functional significance of RSV were studied. DNA sequencing was performed in RSV of all probands and controls with normal reproductive function (n = 535). In addition to human studies, reproductive phenotypes in heterozygous and double-heterozygous KISS1 and KISS1R mice were examined to test the hypothesis that heterozygous mutations in KISS1 can cause reproductive deficits either alone or in combination with mutations in other genes. Individual variants located within the mature kisspeptin peptide were added to CHO-KISS1R cells to measure the accumulation of inositol phosphate. Variants located outside the peptide coding region were also analyzed in vitro and in silico. Transcriptional efficiency of the g.1→3659C3T variant was examined in vitro. Among the 1025 probands, 15 carried heterozygous RSV in KISS1 seen in less than 1% of the controls. Of 5 variants located within the mature kisspeptin peptide, p.F117L (but not p.S77I, p.Q82K, p.H90D, or p.P110T) reduced inositol phosphate generation in vitro. Analysis in silico showed that 2 variants residing within the coding region but outside the mature peptide, p.G35S and p.C53R (but not p.A129V), were predicted to be inactivating. One variant residing outside the coding region (g.1→3659C3T impairs transcription in vitro and another (c.1→7C3T) lies within the consensus Kozak sequence. Four of 5 probands tested had abnormal endogenous LH pulse patterns at baseline. In the mouse model, serum testosterone concentration decreased with increasing heterozygous loss of Kiss1 and Kiss1r alleles [wild-type, 274 (99) vs double heterozygotes, 69 (16) ng/dL; r2 = 0.13; P = 0.03]. Compared with wild-type mice, Kiss1/Kiss1r double-heterozygote males had shorter anogenital distances [13.0 (0.2) vs 15.6 (0.2) mm at P34, P < 0.001] and females had longer estrous cycles [7.4 (0.2) vs 5.6 (0.2) days, P < 0.01]. In addition, mating pairs had decreased litter frequency [0.59 (0.09) vs 0.71 (0.06) litters/month, P < 0.04] and size [3.5 (0.2) vs 5.4 (0.3) pups/litter, P < 0.001]. These findings show that rare, deleterious, heterozygous variants in KISS1, which seem to be associated with GnRH deficiency and abnormal reproductive phenotypes in humans, exist. GnRH deficiency in humans is not caused by KISS1 RSV alone. As in the mouse model, it may result from heterozygous KISS1 variants acting in concert with other genetic factors and/or environmental factors.