1 Departments of Genetics
2 Physiology and Medicine, Southwest Foundation for Biomedical Research San Antonio, Texas 78245-0549
3 Axys Pharmaceuticals, La Jolla, California 92037
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
variance component linkage analysis; reproductive hormones; baboons
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Given that several proximate determinants of fertility exhibit significant additive genetic components, such as age at menarche (21) and age at menopause (29), direct determinants of fertility, i.e., reproductive hormones, may also exhibit a significant genetic influence in their variation in levels of expression. Indeed, twin studies have previously demonstrated a significant heritability for estrogen in males (3846%) (6, 24).
In females, estrogen levels vary throughout the menstrual cycle. This poses a major problem for the genetic analysis of estrogen in human females, for whom obtaining accurate data on menstrual cycle phase is difficult. Indeed, to accurately estimate phase of the menstrual cycle, sequential estrogen levels are required. Female baboons also exhibit menstrual cycles (12); however, in baboons, the perineal skin contains many estrogen receptors, therefore swelling of the perineum (turgescence) corresponds with relative precision to the phases in the menstrual cycle (19). With an accurate method to estimate menstrual cycle phase, the effects of the menstrual cycle on the expression of sex hormone levels can be modeled.
In addition to their physiological similarity to humans, a number of other factors make baboons an excellent model for genetic analyses of complex physiological phenotypes. DNA sequence data have demonstrated that homologous baboon and human protein-coding segments are generally more than 90% identical and noncoding regions greater than 85% identical (25). Chromosome painting studies have documented the homologies between specific human chromosomes and their baboon counterparts (5). For the present study, a critical advantage in using this species is the availability of a genetic linkage map. Using primarily microsatellite loci (n > 325) that have already been mapped in the human genome, researchers have generated an autosomal linkage map for Papio hamadryas (26). Construction of this linkage map utilized data from 690 baboons selected from a series of large, multigenerational pedigrees maintained at the Southwest Foundation for Biomedical Research (SFBR). The average marker distance in this baboon gene map is
7.2 cM. Each of the baboon autosomes has between 8 and 34 loci assigned to map locations.
Consistent with prior chromosome painting analyses (5), the baboon linkage map demonstrates the high degree of conservation between the human and baboon genome, with the differences being relatively straightforward. Specifically, human chromosome 2 is homologous to two separate baboon chromosomes, and the homologs of human chromosome 14 [Homo sapiens autosome (HSA14)] and HSA15 are fused in baboons. Similarly, the baboon equivalents of two other pairs of human chromosomes (HSA20 and HSA22; HSA7 and HSA21) are combined. Human-baboon comparisons of other chromosomes show that inversions and/or other rearrangements have created differences in the order of loci in the two linkage maps, but no interchromosomal translocations were observed.
Given the utility of the baboon model for the study of estrogen variation, we undertook a genome scan to identify genes that have significant effects on the expression of quantitative variation in serum estrogen levels using the SFBR pedigreed baboon colony.
![]() |
MATERIALS AND METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
Estradiol levels were assayed by radioimmunoassay using commercially available kits (Diagnostics Systems Laboratories) and following the human assay protocol. Previous research has demonstrated the appropriateness and effectiveness of using the human assay protocol for baboons (4). This assay does cross-react with estrone and estriol (0.86 and 0.57, respectively), but not with testosterone. To normalize the estrogen distribution, these data were natural-log transformed prior to analysis.
Turgescence Data
Given that animals were sampled irrespective to timing of menstrual cycle, the stage of the menstrual cycle had to be identified. Turgescence is the swelling of perineal (sex) skin, which serves as an indication of hormonal activity (19). This perineal skin inflation and deflation corresponds with relative precision to follicular and luteal phases, allowing for external determination of phase of menstrual cycle without the need for estrogen or progesterone serial measures (12). At SFBR, turgescence (perineal swelling) levels are monitored and recorded regularly (every 37 days) for adult female baboons, with an average cycle of 32 days. The scoring scheme assigns the complete deturgescence of the perineum as zero and maximum turgescence a score of four. Intermediate stages were scored as 1, 2, or 3 (corresponding to ascending order of level of turgescence). This method requires experience and knowledge of the individual animal, but it is a rapid and relatively accurate method of recording changes in the menstrual cycle (12).
Although the turgescence data has the potential to reduce the nongenetic variability by accounting for menstrual cycle variability, incorporation of this measure is not straightforward. Indeed, turgescence has not been used as a proxy measure for stage in menstrual cycle in quantitative genetic analyses. Therefore, there was no protocol on how to adjust data using turgescence observations. To determine which method would be the most beneficial in adjusting for menstrual stage, we created multiple covariates including turgescence, lag of turgescence, days plus or minus ovulation, and follicular phase. For turgescence level, we took the turgescence score at the time of the blood draw. However, the swelling of the perineum may not directly correspond to the estrogen level. Therefore, we also used lag of turgescence, the turgescence score 3 days prior to the blood draw. For days ± ovulation, the date of ovulation was identified by locating the first day of deturgescence and counting back 3 days; then the position relative to the blood draw was identified (19). For follicular phase, an animal was considered in the follicular phase if the blood draw occurred between menstruation and ovulation.
Genotypes
DNA was extracted from white blood cells or liver cells using standard phenol/chloroform methods. Short tandem repeats (STRs) were amplified in PCR reactions using radioisotope and fluorescently labeled human primers. All microsatellite markers used were highly polymorphic in the baboon (average heterozygosity 0.71). PCR reactions (total volume of 25 µl for radioactive primers and 510 µl for fluorescently labeled primers) contained 50 ng DNA, 5 pmol of each primer, 0.20.3 U Taq polymerase, 0.25 mM dNTPs, 12 mM MgCl2, and additional buffer components. To increase specificity, "touchdown" PCR was used (for additional details, see Ref. 26), but annealing temperatures varied by marker. PCR reactions were performed separately, and aliquots were pooled according to multiplexed panels for typing with an automated DNA sequencer (Applied Biosystems model 377 with Gene Scan672 and Genotyper programs) (for additional details, see Ref. 26).
Data Analysis
Data preparation.
Before any quantitative analyses can be performed, data must be examined for outliers, as these could cause undue influence on quantitative genetic analyses. To eliminate outliers, observations that fell outside three standard deviations from the mean were removed. Additionally, because estrogen levels change during pregnancy, perimenopause, and menopause, animals that were pregnant, perimenopausal, or menopausal were removed from the analysis. Pregnancy was identified by the presence of a purple color on the perineal skin. If an animal displayed purple perineal skin within the period 2 wk before or after the blood draw, then the animal was considered pregnant. To identify perimenopausal and menopausal animals, cycle records for all females aged 20 yr or more were collected for 1 yr prior to the blood draw. If a female did not have menstrual flow or turgescence during the year, then the female was considered menopausal. If a female experienced sporadic menstrual flow or turgescence, then the female was considered perimenopausal.
Variance components linkage analysis.
A variance component model applied to extended family data was used to test for evidence of linkage of quantitative trait loci (QTLs) for estrogen variation with STR loci using the autosomal linkage map. An extension of the strategy developed by Amos (3) was used to estimate the genetic variance attributable to a specific chromosomal location (2). This approach is based on specifying the expected genetic covariances between arbitrary relatives as a function of the identity by descent (IBD) relationships at a given marker locus. For a simple model in which a major locus and residual additive genetic effects influence a trait, the covariance matrix () for a pedigree is given by
![]() |
The basic method of variance component linkage analysis also includes a QTL-specific component, which is used to test for linkage. Using a variance component model (8), we tested the null hypothesis that q2, the additive genetic variance due to a QTL, equals zero (no linkage) by comparing the likelihood of this restricted model with that of a model in which
q2 is estimated. The difference between the two log10 likelihoods produces a LOD score that is the equivalent of the classic LOD score of linkage analysis. Twice the difference in loge likelihoods of these models yields a test statistic, which is asymptotically distributed as a 1/2:1/2 mixture of a
2 variable and a point mass at zero (20). Extensive simulation suggests that the likelihood ratio test yields expected nominal P values for a wide variety of reasonable trait distributions (1). This quantitative trait linkage method has been implemented in the program package SOLAR (2), which determines whether genetic variation at a specific chromosomal location can explain the variation in the phenotype (3, 8, 16).
The use of the variance component approach requires an estimate of the IBD matrix. For the baboon pedigrees, a pair-wise maximum likelihood-based procedure was used to estimate IBD probabilities (2). To permit multipoint analysis for QTL mapping, an extension (20) of the technique of Fulker and colleagues (13) was employed. Estimates of the IBD probabilities were generated at any point on a chromosome using a constrained linear function of observed IBD probabilities of markers at known locations within the region. This multipoint procedure, which yields substantially greater power to localize QTLs than two-point methods, enabled direct localization of the QTL and construction of confidence intervals. For the current data set, a LOD score evaluation was performed every centimorgan along the chromosome, with the distances between markers having been determined using CRI-MAP (17).
Application to these data.
Using the variance component approach detailed above, heritabilities were estimated for estrogen. The initial analysis screened for general covariates including sex, age, age squared, age by sex, age squared by sex, turgescence, lag of turgescence, days plus or minus ovulation, follicular phase, and weight. All of the turgescence covariates were modeled as an interaction with sex, as males do not exhibit turgescence. Age and sex were included as covariates, because it is standard practice to screen these variables. Turgescence variables were included as covariates as an attempt to account for menstrual cycle variability. Weight was included as a covariate because previous research has suggested a relationship between estrogen and adiposity (30). Any covariate whose effect was significant at the P 0.10 level in the initial screening was retained for linkage analysis, and its effect was re-estimated simultaneously with genetic effects. To model the household effect, a matrix was made identifying animals that were housed in the same cage at the time of the blood draw.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
|
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In the quantitative genetic analyses, three covariates were significant: age, lag of turgescence by sex, and weight. Although significant, the effect of age was relatively small, possibly due to the fact the range of ages was limited (25 yr). The lag of turgescence was positively correlated with estrogen levels, as would be predicted by previous studies (12, 19). Weight and estrogen levels were negatively correlated, which is concordant with research in humans demonstrating a negative relationship between estrogen levels and adiposity (30).
After covariates were accounted for, the heritability of serum estrogen levels was 55%. Although this estimate might appear high, the controlled environment in which these baboons live may reduce the phenotypic variation in estrogen levels, thus increasing the heritability estimate. Additionally, the cage effect explained 17% of the variation. How estrogen levels are influenced by a shared cage experience is unknown and should be explored in future analyses.
In the linkage analysis, we obtained evidence that 59% variation in estrogen levels could be attributed to baboon chromosome 10 in a region homologous to human chromosome 20, even though the serum samples were collected in both sexes and irrespective of menstrual cycle stage in females. We recognize that the random sampling utilized in this study is not optimal. Indeed, such a sampling strategy makes it more difficult to detect a genetic signal, because the variability due to sex and menstrual cycle confounds the familial inheritance pattern necessary to detect a genetic signal. Nonetheless, we detected a significant linkage signal, which may be an underestimate of the true effect, given the potential confounding factors listed above.
Additionally, we recognize that some individuals may criticize sampling males and females jointly, given the different expression patterns present in the sexes. However, estrogen has functions beyond reproduction, including a role in the cardiovascular system. Therefore, by analyzing both males and females, a better understanding of the regulation of estrogen, outside of the reproductive system, may be obtained.
Our results demonstrate that a substantial portion of the variability in estrogen levels can be attributed to baboon chromosome 10, a region homologous to human chromosome 20. This analysis suggests that a gene located near D20S100 and D20S171 is responsible for estrogen variability. Within the one LOD unit support interval for linkage with estrogen levels, there are several currently mapped genes, some of which could be hypothesized to have an effect on reproduction-related phenotypes in general and on levels of estrogen in particular. A potential candidate in this region is the gene for the melanocortin 3 receptor (MC3R), which has been localized in humans to 20q13.2 by fluorescence in situ hybridization (23). MC3R is expressed in brain, placenta, and intestinal tissues (14). Given that MC4R has been shown to mediate leptin stimulation of luteinizing hormone and prolactin (31), MC3R, which has a high degree of structural homology with MC4R, could also mediate reproductive hormone expression.
Another potential candidate is cytochrome P-450, subfamily XXIV (CYP24), which is localized to human 20q13.2-q13.3 by fluorescence in situ hybridization (18). CYP24 is a mitochondrial enzyme involved in the regulation of calcium homeostasis. Given that gonadotropin-releasing hormone utilizes calcium in the estrogen pathway, calcium appears to be a regulator for reproductive hormones (15). Therefore, polymorphisms in CYP24 might influence calcium homeostasis in the hypothalamus, such that production of follicle-stimulating hormone and therefore levels of estrogen could be affected.
A third potential candidate is the breast carcinoma amplified sequence (BCAS1), localized to 20q13 by comparative genomic hybridization studies (10). BCAS1 encodes a 585-amino acid protein of unknown function that is overexpressed in many breast cancer cell lines. Given that breast cancer is commonly associated with mutations in the estrogen receptor (9, 28, 32), the protein produced by BCAS1 could potentially influence the expression of estrogen through an interaction with the estrogen receptor.
In conclusion, this research presents several important findings. First, although proximate determinants of fertility exhibit significant genetic components, few studies have examined the direct determinants of fertility, the reproductive hormones. In this analysis, we have demonstrated that by accounting for menstrual cycle phase, estrogen levels are highly heritable. Second, this is the first study to report the localization of a QTL for a reproductive hormone in a primate species (including humans). These results provide strong evidence for linkage with baboon serum estrogen levels in a chromosomal segment homologous to the region of human chromosome 20 near D20S171. Therefore, this chromosomal region likely contains an important and potentially novel gene influencing the expression of estrogen levels in primates. Given that estrogen plays a major role in reproductive physiology and has been implicated in several disease processes, the identification of genes influencing estrogen levels could have far-reaching implications.
![]() |
ACKNOWLEDGMENTS |
---|
This work was supported in part by National Heart, Lung, and Blood Institute Grant HL-28972.
![]() |
FOOTNOTES |
---|
Address for reprint requests and other correspondence: L. J. Martin, Dept. of Genetics, Southwest Foundation for Biomedical Research, PO Box 760549, San Antonio, TX 78245-0549 (E-mail lmartin{at}darwin.sfbr.org).
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|