Genetics of Pubertal Timing

Abstract

Purpose of review

To summarize advances in the genetics underlying variation in normal pubertal timing, precocious puberty, and delayed puberty, and to discuss mechanisms by which genes may regulate pubertal timing.

Recent findings

Genome-wide association studies have identified hundreds of loci that affect pubertal timing in the general population in both sexes and across ethnic groups. Single genes have been implicated in both precocious and delayed puberty. Potential mechanisms for how these genetic loci influence pubertal timing may include effects on the development and function of the GnRH neuronal network and the responsiveness of end-organs.

Summary

There has been significant progress in identifying genetic loci that affect normal pubertal timing, and the first single-gene causes of precocious and delayed puberty are being described. How these genes influence pubertal timing remains to be determined.

INTRODUCTION

Puberty is a period of remarkable physical and psychological development that results in sexual maturation. Precocious and delayed puberty can be sources of distress, and there is compelling evidence that variation in pubertal timing may affect risks for conditions such as cardiovascular disease, breast cancer, and depression [1, 2**, 3]. Despite the strong heritability of pubertal timing, our understanding of the underlying genetics is limited. Recent advances have started to reveal the genetic background of normal pubertal timing and the first single-gene causes of precocious and delayed puberty. In this review, we discuss recent developments in the genetics of normal pubertal timing, central precocious puberty, and delayed puberty and discuss potential mechanisms for how implicated genes may affect pubertal timing.

VARIATION IN NORMAL PUBERTAL TIMING

The timing of normal pubertal onset varies substantially and is influenced by both genetic and environmental factors, with genetic factors accounting for an estimated 50–75% of the variation [4–6]. While numerous studies have examined associations between candidate genes and pubertal timing in small to moderate-sized cohorts [7–22], this section will focus on large-scale genome-wide association studies (GWAS), which provide the most definitive results to date, and on a series of studies on variants in FSHR, the gene encoding the FSH receptor.

Genome-Wide Association Studies

In the last decade, GWAS have been used to identify genetic loci that affect normal pubertal timing (Table 1). In 2009, four independent GWAS reported the first evidence that common genetic variants can influence pubertal timing [23–26]. Subsequent studies have used ever-larger cohorts, and a GWAS in 2017 used data in 368,888 women to identify 389 loci that influence pubertal timing [2**, 33]. These loci affect multiple biological pathways and sites of action, which are further discussed below.

Table 1.

(Original) Genome-Wide Association Studies on Pubertal Timing

Study [Reference]	Year	Ancestry of Study Population	Number of Individuals in Primary Analysis	Number of Individuals in Replication Analysis	Pubertal Marker Studied	Number of Genome-Wide Significant Loci Identified	% of Population Variance Explained
He et al. [23]	2009	European	17,438 women	—	Menarche	10	0.6
Ong et al. [24]	2009	European	4,714 women	16,373	Menarche	1	0.2
Perry et al. [25]	2009	European	17,510 women	—	Menarche	2	—
Sulem et al. [26]	2009	European	15,297 women	10,040	Menarche	5	—
Elks et al. [27]	2010	European	87,802 women	14,731	Menarche	32	1.3
Chen et al. [28]	2012	Hispanic	3,468 women	—	Menarche	0	—
Demerath et al. [29]	2013	African American	18,089 women	2,850	Menarche	0	—
Spencer et al. [30]	2013	African American	4,159 women	—	Menarche	0	—
Tanikawa et al. [31]	2013	Japanese	15,495 women	—	Menarche	0	—
Cousminer et al. [32]	2014	European	6,147 girls	—	Breast Tanner Stage	1	—
		European	3,769 boys	—	Genital Tanner Stage	1	—
Perry et al. [33]	2014	European	182,416 women	8,689	Menarche	106	2.7
Pyun et al. [34]	2014	Korean	3,437 women	—	Menarche	0	—
Day et al. [35]	2015	European	55,871 men	—	Voice Breaking	9	—
Shi et al. [36*]	2016	Chinese	8,073 women	8,322	Menarche	0	—
Day et al. [2**]	2017	European	329,345 women	39,543	Menarche	389	7.4

Most GWAS have used recalled age at menarche as a retrospective marker of pubertal timing in females. In males, age at voice breaking has been used as a similar marker. GWAS in men have largely revealed significant overlap between the sexes in the genetic factors that influence pubertal timing. However, some loci show different effect sizes or even effects in opposite directions [35, 37]. Potential mechanisms for these sex-specific effects on pubertal timing have been recently discussed by Cousminer et al. [37].

The most recent 2017 GWAS examined not only pubertal timing within the normal range but also early and late pubertal timing. This analysis suggested that, in girls, common genetic variants contribute more to early than late pubertal timing, and that in boys the opposite is true [2**]. The mechanisms underlying this sex difference are not clear [37], but this difference may explain why precocious puberty is more likely to be idiopathic in girls than in boys and may also explain, at least partially, why precocious puberty appears to be more common in girls and delayed puberty in boys [2**, 38, 39].

The above GWAS were conducted in cohorts of European ancestry. GWAS on age at menarche have also been conducted in non-European populations to determine the effect sizes of previously identified GWAS loci in different ethnic backgrounds as well as to potentially identify new loci. These GWAS, conducted in women of Hispanic [28], African American [29, 30], Japanese [31], Korean [34, 36], and Chinese [36*] origin, did not identify any additional genome-wide significant loci (Table 1), but statistical power was limited by the relatively smaller sample sizes in these studies (approximately 3,500 to 18,000) [28–31, 34, 36*]. These genome-wide studies, as well as additional studies examining select loci in Filipino [40], Chinese [41], American Indian [42], and Native Hawaiian [42] cohorts, have demonstrated possible differences in effect sizes between ethnic groups. Thus, it appears that there is considerable overlap across populations in genetic loci that influence pubertal timing, though the specific effects of each locus may vary.

Studies in FSHR

A series of studies have analyzed variants near FSHR, the gene encoding the FSH receptor, that appear to decrease signaling through the receptor. These variants have been associated with later age at testicular growth in boys and thelarche in girls [43, 44*, 45**]. However, recent GWAS have not associated these FSHR variants with age of menarche in girls or age at voice breaking in boys [2**]. A recent study investigating effects of several candidate gene variants on various hallmarks of puberty, found that a variant in LIN28B, the gene most strongly associated with variation in pubertal timing across multiple GWAS, affected timing of both thelarche and menarche, but variants in FSHR affected timing of thelarche but not menarche [45**]. At puberty, FSH stimulates ovarian production of estradiol, which stimulates growth of breast tissue, resulting in thelarche, one of the earliest hallmarks of female puberty. In contrast, menarche is a late event and likely involves maturation at multiple levels of the hypothalamic-pituitary-gonadal axis. Thus, it is possible that genes like LIN28B, which has been proposed to act in the hypothalamus, may impact all pubertal milestones, but genes like FSHR, which act peripherally, may preferentially affect thelarche with a smaller effect on menarche [45**].

Other Sources of Genetic Variation

Currently, the 389 loci identified through GWAS account for 7.4% of the variation in normal pubertal timing [2**]; GWAS for many other traits and conditions have similarly explained only a small proportion of variation [46]. It is possible that there are many more variants yet to be discovered in GWAS with even larger cohorts.

In addition, factors beyond single-nucleotide polymorphisms may contribute to variation in pubertal timing, factors including copy number variation or epigenetic changes [47, 48]. A duplication near the DBI (diazepam binding inhibitor) gene, believed to decrease testosterone and estrogen levels through effects on GABA signaling in mice, has been linked to age at menarche [47]. The latest GWAS associated lysine-specific demethylase genes with age at menarche, raising the possibility that epigenetic changes, including methylation, may also influence pubertal timing [2**]. Findings that global DNA methylation in adult women and epigenetic age (as determined by methylation of specific loci) in adolescent girls may be associated with age at menarche provide further support for this concept [49, 50].

CENTRAL PRECOCIOUS PUBERTY

MKRN3

In 2013, mutations in MKRN3 emerged as an important genetic cause of central precocious puberty (CPP) in both boys and girls from whole-exome sequencing in multiple unrelated families. Loss-of-function mutations in MKRN3 both familial and sporadic cases of CPP have since been identified in various ethnic groups in the U.S., Brazil, Israel, Korea, Taiwan, and eight European nations [51–53, 54*, 55, 56]. Case series of families with multiple affected members have estimated the prevalence of MKRN3 mutations to be 33 to 46% in familial cases [57–59] and 0.4 to 3.8% in sporadic cases [52, 53, 60], and two small case series, including one in boys, have suggested even higher rates [53, 61*].

MKRN3 is a maternally imprinted gene; that is, only the paternal allele is expressed. Consistent with this imprinting pattern, only individuals who inherit an MKRN3 mutation from their father develop CPP [51]. In large familial cases, there is complete penetrance of MKRN3 mutations. However, two recent case reports suggested that both male and females with paternally inherited MKRN3 defects can sometimes be asymptomatic carriers [54*, 62].

Patients with CPP due to loss-of-function mutations in MKRN3 appear to be similar to CPP patients without MKRN3 mutations, with a median age of pubertal onset of 6.0 years in girls (range of 3.0 to 7.8 years) and 8.3 years in boys (range of 5.9 to 9.0 years) [51–54*, 56, 59, 61*]. MKRN3 mutations do not appear to cause any phenotypes other than CPP. The exact mechanism by which defects in MKRN3 lead to early activation of the hypothalamic-pituitary axis is unknown. MKRN3 encodes a protein that resembles E3 ubiquitin ligases, which facilitate ubiquitination of target proteins, leading to degradation by the proteasome or, in some cases, modification of protein function [63*]. Studies in mice show that Mkrn3 is expressed in the hypothalamus, and its expression decreases with sexual maturation [63*]. MKRN3 has therefore been postulated to have an inhibitory influence on reproductive endocrine activity in the hypothalamus in prepubertal children [64, 65].

DLK1

A defect in another imprinted gene, DLK1, was identified in 2017 through linkage analysis and whole-genome sequencing in a large family with multiple members with CPP [66**]. A ~14 kb deletion in DLK1was found to segregate with all four affected female members from a Brazilian family of African descent. Previously, deletions in DLK1 and surrounding genes had been identified as the cause of Temple Syndrome, a rare genetic disorder associated with CPP [67]. Subsequent screening for DLK1 mutations in an additional 19 individuals with CPP in the same study and in 60 girls with CPP in a separate study did not identify any rare variants or deletions, suggesting DLK1 mutations are a relatively rare cause of central precocious puberty [53, 66**].

The role of DLK1 in the regulation of pubertal timing is unknown. Studies in mice show that Dlk1 is expressed in the hypothalamus, and in vitro studies demonstrate that DLK1 may be a regulator of Notch signaling, which appears to be an important component of kisspeptin neuronal development in mice [66**, 68, 69]. Kisspeptin is stimulates secretion of GnRH in the hypothalamus; thus, DLK1 may affect GnRH secretion through the regulation of kisspeptin [66**].

The GWAS discussed above identified variants near the MKRN3 and DLK1 loci and specifically associated paternally-inherited variants with pubertal timing [2**, 33, 65]. These findings provide further evidence for the critical role of MKRN3 and DLK1 in the regulation of pubertal timing and demonstrate genetic links between the rare disease of central precocious puberty and normal variation in pubertal timing in the general population.

Kisspeptin and KISS1R

Gain-of-function mutations in the genes encoding kisspeptin and its receptor, KISS1 and KISS1R, respectively, have also been suggested to be causes of CPP [70*]. In 2008, a rare heterozygous mutation in KISS1R was identified in a Brazilian girl with CPP; this mutation led to prolonged intracellular signaling following kisspeptin binding [71]. A subsequent study in 2010 identified a rare heterozygous variant in KISS1 thought to confer resistance to degradation in a boy with CPP [72]. Studies have also investigated the role of common variants in KISS1 and KISS1R in individuals with CPP; these studies have suggested potential associations, but because of the small size of these studies, further replication studies are needed to confirm these findings [70*, 73]. Thus, variants in the kisspeptin pathway may be associated with central precocious puberty in some cases.

Other Candidate Genes

Investigations of both common and rare variants in genes identified in GWAS (LIN28B, TACR3, LEPR, and ESR1 [2**, 33]) have not revealed any definitive associations with precocious puberty [70*, 74–79]. Though some small-scale candidate SNP or gene-association studies of additional candidate gene GNRH1, LHB, FSHB, TTF1, EAP1, NPVF, NPFFR1 (genes biologically linked to gonadotropin signaling) [80–82] and CYP19A1, CYP1A1, CYP17, and CYP1B1 (genes encoding steroidogenesis enzymes) [83, 84] have reported potential associations, the small size of these studies preclude any definitive conclusions.

SELF-LIMITED DELAYED PUBERTY

In self-limited delayed puberty (also called constitutional delay), pubertal onset is delayed but occurs spontaneously prior to age 18 years [85]. In contrast, in idiopathic hypogonadotropic hypogonadism (IHH), another condition that presents with delayed puberty, puberty does not start or is incomplete by adulthood [86]. In some cases, individuals with IHH undergo “reversal,” that is, activation of the HPG axis after age 18 years [87]. This section will focus on the recent developments in the genetics of self-limited delayed puberty and investigations into potential genetic links between these disorders.

IGSF10

IGSF10 was the first gene to be associated with self-limited delayed puberty through a whole-exome sequencing study of 76 affected individuals from 18 unrelated families [88]. Four heterozygous, potentially pathogenic missense variants in IGSF10 were identified in 28 individuals with delayed puberty from 10 families, and in 7 of these families these variants segregated fully with delayed puberty. Functional in vitro studies showed failed excretion and cytoplasmic retention of the mutant proteins [88]. Of note, a recent study did not identify significantly enrichment of IGSF10 mutations in individuals with delayed puberty compared to controls, but the high rate of variation in IGSF10 (a large gene) in control populations makes such enrichment difficult to demonstrate [89*].

The precise role of IGSF10 in human pubertal timing is yet to be elucidated. Studies in mouse and zebrafish models suggest that IGSF10 is required for the proper migration of GnRH neurons [88]. How defects in the GnRH neuronal network could result in delayed puberty is discussed further below.

Links to Idiopathic Hypogonadotropic Hypogonadism

Several studies in the last decade have investigated the hypothesis that self-limited delayed puberty and IHH share a genetic basis. Though initial studies examining individual IHH genes did not conclusively demonstrate such a genetic link [7, 90–92], a recent study using whole-exome sequencing to screen a panel of 21 IHH genes identified enrichment of potentially pathogenic variants in IHH genes in patients with delayed puberty compared to controls [93]. Of note, variants in the genes TAC3 and TACR3, which encode neurokinin B and its receptor, were identified in multiple individuals with delayed puberty. The neurokinin B pathway has also been implicated in IHH with reversal, as well as in the variation of normal pubertal timing in GWAS; thus, neurokinin B signaling appears to have a critical influence on pubertal timing [2**, 87].

However, another study in Finnish individuals with delayed puberty did not identify an enrichment of mutations in a panel of 24 IHH genes in individuals with delayed puberty compared to controls [89*]. The apparent discrepancy between these studies could be due to differences in how mutations were defined [89*, 93].

Other Candidate Gene Analyses

Candidate genes from disorders that include delayed puberty or IHH as a feature have been proposed, including LEP and LEPR [94, 95], IGSF1 [96], IGFALS [97], and GHSR [98]. Studies investigating the role of common variants in LEP, LEPR, and IGALS have not demonstrated an association with delayed puberty [94, 97]. Although rare-variant studies have identified two variants in GHSR (which encodes the ghrelin receptor) and one in LEP (which encodes leptin), whether rare variants in these genes are more common in individuals with delayed puberty compared to the general population is unclear [95–98]. An additional investigation in LIN28B, a gene identified in prior GWAS, did not identify any variants in 145 individuals with delayed puberty [2**, 99].

POTENTIAL MECHANISMS FOR REGULATING PUBERTAL TIMING

It has been proposed that puberty starts in the brain [100], yet the genes identified in the above studies have potential roles throughout the hypothalamic-pituitary-gonadal axis (Figure 1). How could these genes affect pubertal timing?

Figure 1. (Original) Putative sites of action for genes implicated in precocious puberty, normal variation in pubertal timing, and delayed puberty.

Select genes are listed next to their likely sites of action within the hypothalamic-pituitary-gonadal axis. Of note, some genes implicated in precocious puberty (DLK1, MKRN3, and KISS1) and in delayed puberty (TACR3) are also associated with variation in normal pubertal timing, indicated by underlining.

The earliest known physiologic event in puberty is an increase in pulsatile GnRH secretion from the hypothalamus; the mechanisms that determine the timing of this event is determined are largely unknown [101]. The first physical signs of puberty, such as testicular growth in boys and thelarche in girls, occur months to years after this initial increase in GnRH secretion [102], and voice breaking in boys and menarche in girls occur even later [2**].

Given this physiology, there are at least three potential mechanisms for variation in the timing of appearance of physical hallmarks of puberty (Figure 2). First, there may be differences in the timing of the initial emergence of reproductive endocrine activity (Figure 2B). The genes KISS1, KISS1R, TAC3 and TACR3 encode signaling molecules and receptors that promote GnRH secretion and may function as part of this putative pubertal “switch” [2**, 86].

Figure 2. (Original) Potential mechanisms underlying variation in pubertal timing.

The graph schematizes the hormonal and physical events that mark pubertal onset and shows potential mechanisms for variation in pubertal timing. The solid black curve in panel A represents typical hormone production over time, with a period of robust activity in infancy (“minipuberty”), a relative quiescence in childhood, and reemergence of activity at puberty. The dashed line represents the “threshold” concentration of gonadotropins and/or sex-steroids needed for the appearance of physical signs of puberty, such as testicular growth in males and thelarche in females. These physical signs appear when hormonal activity reaches the threshold, indicated by the circle.

Three potential mechanisms for variation in pubertal timing are illustrated in Panels B–D. The solid black curve depicts normal pubertal timing. Variations resulting in differences in pubertal timing are depicted in gray; these include variation in the timing of activation of the GnRH neuronal network (B), in the rate of rise of gonadotropins/sex-steroids (C), and in end organ responsiveness to gonadotropins and/or sex steroids (D). These mechanisms are not mutually exclusive.

Second, differences in GnRH neuronal activity or in downstream signaling events could alter the rate at which reproductive hormones rise after this initial activation (Figure 2C). This mechanism could explain how genes involved in the development and migration of GnRH neurons, such as the ANOS1 gene (also known as KAL1) affect pubertal timing [2**, 86].

Third, end-organ responsiveness to reproductive hormones may vary, thus changing the “threshold” for the degree of reproductive endocrine activity needed to induce physical characteristics of puberty (Figure 2D). This mechanism could explain how variants near the FSH receptor gene, FSHR, affect the timing of testicular growth in boys and breast development in girls [43, 44*, 45**].

These various mechanisms, which are not mutually exclusive, emphasize that puberty is not a single event but rather a process that unfolds over time, and that there has already been significant reproductive endocrine activity by the time physical hallmarks of puberty become apparent. Future studies may reveal additional mechanisms for variation in pubertal onset.

CONCLUSION

The last decade has seen significant progress in the identification of genes that affect pubertal timing, with the discovery of hundreds of common genetic variants that affect normal pubertal timing and the first single-gene causes of precocious and delayed puberty. Future studies with larger and more diverse cohorts will undoubtedly identify additional genes that affect pubertal timing. The next challenge will be to determine precisely how these genes and pathways regulate pubertal timing so we can better understand the physiology underlying growth and development and the links to adult disease risks.

KEY POINTS.

• Genome-wide association studies have identified hundreds of common genetic loci that affect normal pubertal timing in both sexes and across ethnic groups. Several of these loci have also been implicated in precocious and delayed puberty. Some loci may preferentially impact certain pubertal milestones but not others.

• The first single-gene causes of precocious puberty (MKRN3, DLK1) and delayed puberty (IGSF10) are being characterized.

• Mechanisms for how these genetic loci influence pubertal timing may include effects on the development and function of GnRH neurons, their activation at the time of puberty, and the responsiveness of end-organs.