| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
EDITORIAL |
Massachusetts General Hospital, Reproductive Endocrinology Unit, Boston, Massachusetts 02114
Address all correspondence and requests for reprints to: .Stephanie B. Seminara, Massachusetts General Hospital, Reproductive Endocrinology Unit, Boston, Massachusetts 02114. E-mail: sseminara{at}partners.org.
The past 2 yr have been witness to multiple successes in the field of genome-wide association studies (GWAS). These studies combed through the human genome in an unbiased fashion searching for links between clinical traits of interest and common (i.e. generally occurring in >5% of the population) genetic variants [also known as single nucleotide polymorphisms (SNPs)]. Because testing all of the millions of SNPs in a person’s genome is currently prohibitively expensive, the HapMap project has earmarked a subset of "tag" SNPs to facilitate the identification of haplotypes in a person’s DNA. SNPs that show positive association to the trait (or disease of interest) are then retested in an independent replication sample.
GWAS have been employed in several common diseases and clinical phenotypes. On the one hand, the past 2 yr have borne witness to several successful GWAS for quantitative traits. Collectively, however, most of these new genes only account for less than 5% of the genetic components of most complex diseases. In contrast, individuals with extreme or more discrete phenotypes are often employed in linkage studies. How much overlap exists between these two groups of alleles for a quantitative trait, i.e. the "common variants" that account for some of the disease vs. the "rare variants" or mutations that often occur in less than 1% of the population? Specifically, are the genes that have been implicated in the severe phenotype of GnRH deficiency with attendant hypogonadotropic hypogonadism (which is characterized by absent or markedly delayed puberty) the same genes that influence the timing of menarche in the general population? This is the question posed by Gajdos et al. (1) in this issue of JCEM.
Recent studies put this question into a larger perspective. Earlier this year, multiple GWAS identified more than 40 genetic loci associated with adult height variation (2, 3, 4). Like most GWAS of common diseases studied to date, despite the large number of loci identified, each variant appeared to have only a small impact on phenotype. For example, one group that had identified some 20 variants associated with adult height concluded that these SNPs only accounted for 3% of the variation in height in the population. Thus, GWAS appear powered to detect high-frequency but low-impact alleles. Moreover, in these studies, only two genes, HMGA2 (4) and GDF5-UQCC (5), had been previously implicated with abnormalities in stature. Therefore, two concepts are emerging: 1) although GWAS have had several recent successes, the degree of variation that can be attributed to the newly identified alleles is quite small; and 2) by and large, alleles identified from GWAS may be distinct from those identified in linkage studies and hence are valuable in surfacing new genes/proteins/pathways that were previously unsuspected. So a precedent in the literature does exist that genes identified at the phenotypic edges of a trait may play a role in population variation, but the proportion of genetic variation that can be attributed to many newly associated polymorphisms appears to be quite small.
This month in JCEM, Gajdos et al. (1) asked whether genes that have been implicated in hypogonadotropic hypogonadism also play a role in variation in the age of menarche. The authors did not perform a GWAS but used a derivative approach with candidate genes. These candidate genes are genes that have been identified in patients with hypogonadotropic hypogonadism, including GNRH, GNRHR, KISS1, GPR54, LEP, LEPR, FGFR1, PROK2, and PROKR. As such, they are involved in GnRH synthesis, secretion, responsivity, and GnRH neuronal migration.
Gajdos et al. (1) studied 1801 women from the Hawaii and Los Angeles Multiethnic Cohort who filled out questionnaires on age of menarche at their initial enrollment. Women were selected for this study from the two "tails" of the normal bell-shaped curve, i.e. if their menarche occurred before 11 yr of age or after 15–17 yr. Age of menarche is a gender-specific trait, and by focusing on this trait, the authors acknowledged that they would miss any role of the chosen candidate genes in male pubertal development. In addition, age at menarche should not be extrapolated to "age at puberty" because menarche is a late event in sexual maturation.
To tighten up the age data, the authors used information derived from a follow-up Multiethnic Cohort questionnaire in which age of menarche was assessed a second time. The authors noted that less that 1% of all women reported an age at menarche in the second questionnaire that was at the extreme opposite of their original category. However, the authors do not provide data on how many women answered the age of menarche question in an identical fashion on both assessments. In other words, how many women who initially reported menarche before 11 yr responded in the same manner when that question was asked decades later? Because the phenotypic assignment of this study was based on retrospective recall of an event that had occurred decades earlier, the details of its ascertainment are an important nuance of this study.
Tag SNPs selected to represent the common variants surrounding the 10 "candidate" genes that were nominally significant and suggestive of association were examined for in silico replication using data from a GWAS in the Nurses’ Health Study. The authors also genotyped these SNPs in 125 children with delayed puberty and their parents as a second replication population. In addition, some genes (GNRHR, GNRH1, FGFR1) were also screened in a delayed puberty subpopulation to detect additional variants to analyze in the menarche panel. The additional analyses revealed no association of any genetic variants to the age of menarche. Nominally significant and suggestive associations were examined in the data garnered from the Nurses’ Health Study, but these findings did not replicate. A very small panel of rare variants and mutations identified in hypogonadotropic hypogonadism patients also did not show association to age of menarche. However, correlations between ancestry and age at menarche were detected in Native Americans and Native Hawaiians.
Thus, common variations in 10 genes for idiopathic hypogonadotropic hypogonadism do not appear to play a major role in the timing of menarche in the general population. This should not be interpreted as a negative finding, but rather a demand bid for the investigative community to uncover other pathways contributing to the genetic variation of age of menarche. New candidates may come from newly identified genes for idiopathic hypogonadotropic hypogonadism, animal knockouts, or large-scale appropriately powered GWAS. As the authors point out, the identification of such genes will not only aid our understanding of the genetic architecture of this quantitative trait and expand our knowledge of the physiology underlying sexual maturation, but it will also pave the way for novel therapeutics for reproductive diseases. Thus, this study should be viewed as a first step in what will undoubtedly prove to be a long and fascinating biological journey to understand how the brain controls reproduction, one of science’s fundamental mysteries of life.
Footnotes
Abbreviations: GWAS, Genome-wide association studies; SNP, single nucleotide polymorphism.
Received September 11, 2008.
Accepted September 22, 2008.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
