A Quantitative Trait Locus on Chromosome 22 ... - Wiley Online Library

A Quantitative Trait Locus on Chromosome 22 for Serum Leptin Levels Adjusted for Serum Testosterone Lisa J. Martin,*† Michael C. Mahaney,† Laura Almasy,† James E. Hixson,‡ Shelley A. Cole,† Jean W. MacCluer,† Cashell E. Jaquish,§ John Blangero,† and Anthony G. Comuzzie†

Abstract MARTIN, LISA J., MICHAEL C. MAHANEY, LAURA ALMASY, JAMES E. HIXSON, SHELLEY A. COLE, JEAN W. MACCLUER, CASHELL E. JAQUISH, JOHN BLANGERO, AND ANTHONY G. COMUZZIE. A QTL on chromosome 22 for serum leptin levels adjusted for serum testosterone. Obes Res. 2002;10:602– 607. Objective: Studies have reported the existence of marked sexual dimorphism in serum leptin levels in humans with women having approximately two to three times the levels of men. We have shown that this sexual dimorphism has a strong genetic component arising from a genotype by sex interaction, but adjusting leptin levels for testosterone eliminates this interaction. Because interactions such as genotype ⫻ sex can confound the detection of quantitative trait loci (QTLs), we wanted to determine if there are QTLs associated with the expression of leptin adjusted for testosterone. Research Methods and Procedures: We performed a genome-wide scan using multipoint linkage analysis and implemented a general pedigree-based variance-component approach to identify genes with measurable effects on variation in leptin levels independent of testosterone in 318 Mexican Americans from the San Antonio Family Heart Study. Results: We detected significant evidence of linkage (log of the odds ratio ⫽ 3.44) for a QTL on chromosome 22.

Submitted for publication November 28, 2001. Accepted for publication in final form March 29, 2002. *Children’s Hospital Medical Center, Cincinnati, Ohio; †Southwest Foundation for Biomedical Research, San Antonio, Texas; ‡Human Genetics Center, University of Texas Health Science Center at Houston, Houston, Texas; and §National Heart Lung and Blood Institute, Bethesda, Maryland. Address correspondence to Dr. Lisa J. Martin, Center for Epidemiology and Biostatistics, Children’s Hospital Medical Center, 3333 Burnet Avenue, Cincinnati, OH 45229-3039. E-mail: [email protected] Copyright © 2002 NAASO

602

OBESITY RESEARCH Vol. 10 No. 7 July 2002

Discussion: Given these results, we hypothesize that a QTL on chromosome 22 may influence the level of leptin adjusted for testosterone. Key words: variance component linkage analysis, obesity, sex hormones

Introduction Leptin, the protein product of the LEP gene, is a key determinant of obesity, with increasing leptin levels associated with increased adiposity (1). As reported in previous studies (2– 8), there is pronounced sexual dimorphism in serum leptin levels. Typically these sexspecific differences have been considered a reflection of a higher percentage of total body fat and/or a different pattern of fat distribution in women than men (2– 4). However, serum leptin levels are higher in women compared with men regardless of body fatness. Recently, we identified a strong genetic component in sexual dimorphism in serum leptin levels and demonstrated that this genetic component to the sexual dimorphism persists even after conditioning on measures of body fat (9). Baumgartner et al. (10) have hypothesized that the sexual dimorphism present in leptin levels may be attributable to the sex hormones, specifically estrogen and testosterone. Therefore, we recently tested to determine if the genotype ⫻ sex interaction in leptin levels could be attributed to estrogen or testosterone (9). The genotype ⫻ sex interaction persists when adjusting for estrogen but is eliminated when adjusting for testosterone. This suggests that part of the sexual dimorphism seen in leptin levels may be attributed to sex differences in testosterone levels. Because interactions such as genotype ⫻ sex can confound the detection of quantitative trait loci (QTLs), we wanted to determine if there are QTLs associated with the

Linkage of Leptin Adjusted for Testosterone, Martin et al.

expression of leptin adjusted for testosterone. To accomplish this goal, we performed a genome-wide scan using multipoint linkage analysis implemented in a general pedigree-based variance-component approach to identify genes with measurable effects on variation in leptin levels independent of testosterone in 318 Mexican Americans from the San Antonio Family Heart Study.

Research Methods and Procedures The Population Subjects in this study included 318 Mexican Americans (134 men and 184 women) distributed throughout 10 families (ranging in size from 7 to 56 individuals). These families are among those participating in the San Antonio Family Heart Study, a project designed to investigate the genetics of risk factors for atherosclerosis, noninsulin-dependent diabetes (type 2 diabetes), and obesity (11). Probands were 40- to 60-year-old Mexican American men and women enlisted (without regard to disease status) from a house-to-house survey of barrios that are at least 95% Mexican American. To assure large families, probands had to have a spouse and at least six age-eligible offspring and/or siblings living in San Antonio. All available first-, second-, and third-degree relatives of both the proband and the spouse were invited to participate. Ages of participants ranged from 18 to 92 years (average age, 37 years). All protocols were approved by the Institutional Review Board at the University of Texas Health Science Center, San Antonio, Texas. Phenotypes Leptin was assayed by radioimmunoassay (12), using a commercially available kit (Linco Research, St. Charles, MO) in serum samples collected after an overnight (⬃12 hour) fast. Serum leptin levels were checked for normality, and samples that fell outside of 3 SD units from the mean were removed to remain consistency with previous analyses (7,9). This resulted in serum leptin levels with a skew of 0.83 and a kurtosis of ⫺0.14. The analysis included the following covariates: sex, age, age2, age ⫻ sex, age2 ⫻ sex, diabetes status, and testosterone. Testosterone was assayed using commercially available radioimmunoassay kits (Diagnostics Systems Laboratory, Webster, TX). Genotypes DNA was prepared from lymphocytes and used for polymerase chain reaction with fluorescently labeled primers from the MapPairs Human Screening Sets versions 6 and 8 (Research Genetics, Huntsville, AL) containing 380 highly polymorphic microsatellite markers from 22 autosomes spaced at ⬃10 cM. PCR reactions were performed sepa-

rately according to the manufacturer’s specifications, and aliquots were pooled into multiplexed panels for typing on an automated DNA sequencer (model 377 with Genescan and Genotyper programs; Applied Biosystems, Foster City, CA). The distances between markers were computed from our data using the CRI-MAP software program (13) and verified for consistency with the genetic maps available from the Marshfield Medical Research Foundation (Marshfield, WI; www.mfldclin.edu/genetics) and University of Southampton (Southampton, UK; cedar.genetics.soton.ac. uk/public_ html/gene.html). The average spacing between markers was 10.0 centimorgans (cM), and the largest spacing was 26 cM (on chromosome 11; this was the only gap ⬎20 cM). Maximum likelihood estimates of allele frequencies were obtained using all pedigree information. Variance Components Linkage Analysis A variance-component model applied to extended family data was used to test for evidence of linkage of QTLs for phenotypes related to serum leptin levels with STR loci using the 10-cM genome-wide map. An extension of the strategy developed by Amos (14) was used to estimate the genetic variance attributable to a specific chromosomal location (15). 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. 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 (16), we tested the null hypothesis that states that the additive genetic variance due to a QTL (␴q2) 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 log of the odds ratio (LOD) score that is the equivalent of the classical LOD score of linkage analysis. Twice the difference in log likelihoods of these models yields a test statistic that is asymptotically distributed as a 1/2:1/2 mixture of a ␹2 variable and a point mass at zero (17). Extensive simulation suggests that the likelihood ratio test yields expected nominal p values for a wide variety of reasonable trait distributions (18). This quantitative trait linkage method has been implemented in the program package Statistical Oligogenic Linkage Analysis Routines (SOLAR) (15), which determines whether genetic variation at a specific chromosomal location can explain the variation in the phenotype (14,16,19). In variance-component linkage analysis, oligogenic effects can be incorporated, such that the genome is scanned for linkage and the chromosomal location that yields the largest marginal LOD score is retained for future conditional analyses (16). By fixing the position of the previous OBESITY RESEARCH Vol. 10 No. 7 July 2002

603


Figure 1: Maximum log of the odds ratio (LOD) scores by chromosome for the genome screen of serum leptin levels adjusted for testosterone. The chromosomal location is represented on the y axis and the LOD scores on the x axis. Marker density for each chromosome is represented by hash marks. LOD scales are provided for each chromosome with a maximum LOD score of 2 or greater.

QTL to the location that yields the largest LOD, it is held as a parameter, and other QTLs, which have lower genetic signals, may be detected (16). The use of the variance-component approach requires an estimate of the identity by descent (IBD) matrix. A pairwise maximum likelihood-based procedure was used to estimate IBD probabilities (15). To permit multipoint analysis for QTL mapping, an extension of the technique of Fulker et al. (20) was employed. Estimates of the IBD probabilities were generated at any point on a chromosome using a constrained linear function of the 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; the distances between markers were determined using CRI-MAP (13). 604


Results Figure 1 displays the results by chromosome from the variance-component linkage analysis for leptin levels adjusted for testosterone. We detected two chromosomal regions, one on chromosome 9 and one on chromosome 22, with LOD scores ⬎1.9 (Figure 1). The strongest signal detected was a maximum LOD score of 3.44 (p ⫽ 0.00003) on chromosome 22, near marker D22S1685 (Figure 2). Table 1 provides the parameter estimates for this linkage model. However, caution must be used when interpreting the QTL effect size, as potential bias exists (21). The one LOD unit support interval for the linkage signal on chromosome 22 spans a 25-cM region surrounding the peak LOD score (0 to 25 cM from p terminal; Figure 2). The only other chromosomal region to yield a LOD score ⬎1.9 was chromosome 9 at 20 cM, with a maximum LOD of 2.24. However, this fails to reach the generally accepted level for significant evidence of linkage (22), and in subsequent


Table 1. Parameter estimates from the linkage model with a QTL on chromosome 22

Figure 2: Estimated log of the odds ratio (LOD) functions obtained from multipoint quantitative trait linkage analysis of serum leptin levels adjusted for testosterone for chromosome 22.

oligogenic analysis, when the QTL on chromosome 22 was incorporated in the model, this signal was further reduced (data not shown).

Discussion As reported in previous studies (2– 8), there is pronounced sexual dimorphism in serum leptin levels. Previously, it has been suggested that these differences may reflect the fact that women generally have a higher percentage of total body fat than men (2– 4). However, serum leptin levels are higher in women compared with men regardless of body fatness (6). Recently, we identified a strong genetic component in sexual dimorphism in serum leptin levels and demonstrated that this genotype ⫻ sex interaction persists even after conditioning on body fat (9). Interestingly, when testosterone was included as a covariate, the genotype ⫻ sex interaction was eliminated, suggesting that part of the sexual dimorphism in leptin levels may be attributed to sex differences in testosterone levels. This observation is further supported by work of Elbers et al. (23), who demonstrated that in men, suppression of testosterone substantially increases serum leptin to levels similar to those of women. Given that interactions such as genotype ⫻ sex can confound the identification of QTLs, we wanted to determine if there are novel QTLs associated with the expression of leptin adjusted for testosterone. Although we still see a minor effect of the region of chromosome 2 previously reported to influence serum leptin

Parameter

Estimate ⴞ SE

Mean SD Sex Age Age ⫻ sex Age2 Age2 ⫻ sex Diabetes Testosterone e2 h2r h2q1

7.86 ⫾ 1.10 5.92 ⫾ 0.28 5.03 ⫾ 1.64 ⫺0.01 ⫾ 0.04 0.09 ⫾ 0.05 ⫺0.002 ⫾ 0.002 ⫺0.002 ⫾ 0.002 1.53 ⫾ 1.00 ⫺1.22 ⫾ 0.40 0.50 ⫾ 0.10 0 0.50 ⫾ 0.10

Values are maximum likelihood estimates ⫾ SE, the mean of the trait in men. QTL, quantitative trait locus; e2, random environmental effects; h2r, residual additive genetic effects; h2q1, QTL-specific effects.

levels (7), our results demonstrate that a substantial portion of the variability in serum leptin levels adjusted for testosterone can be attributed to a QTL on chromosome 22. This analysis suggests that a gene located near D22S1685 (CHLC.GCT10C10) is responsible for leptin variability independent of testosterone. This marker has been mapped to 22q11.2 through sequencing of chromosome 22 (24). There are several currently mapped genes in this region that could be hypothesized to have an effect on leptin levels. A potential candidate in this region is the gene for peroxisome proliferator-activated receptor-␣ (PPAR␣). Using Southern analysis of human rodent hybrids, PPAR␣ was assigned to 22q12-q13.1 (24) and is ⬃22 mb from the maximum LOD based on sequence data (25). PPAR␣ is a transcription factor that regulates fatty acid oxidation (26 –29). Recent research has demonstrated that PPAR␣ activation reduces serum leptin levels (30). Moreover, transgenic mice deficient in PPAR␣ have late-onset obesity with stable caloric intake (31). Interestingly, these mice also developed marked sexual dimorphism in accumulation of body fat and circulating lipid levels. Therefore, PPAR␣ may be involved in this sexual dimorphism. Another potential candidate is macrophage migration inhibitory factor (MIF). Using somatic cell hybrid panel PCR with human-specific primers, MIF has been localized to 22q11.2 (32). Based on sequence data produced by the Chromosome 22 Mapping Group at the Sanger Center and obtained from the World Wide Web at http:// www.sanger.ac.uk/HGP/Chr22, MIF is ⬃200 kb from OBESITY RESEARCH Vol. 10 No. 7 July 2002

605


D22S1685. MIF has diverse functions including activity as a proinflammatory cytokine, a pituitary hormone, and a glucocorticoid-induced immunoregulator. Because tumor necrosis factor-␣ stimulates the expression of MIF in adipocytes, Hirokawa et al. (33) proposed that MIF may be involved in obesity. Moreover, recent research suggests that MIF may regulate glucose disposal and carbohydrate metabolism (34). Taken together, these results provide evidence that MIF may influence metabolism, and by extension, may influence serum leptin levels. In conclusion, this study reports a novel QTL for serum leptin levels corrected for testosterone. These results provide strong evidence for linkage with serum leptin adjusted for testosterone on chromosome 22 near D22S1685. Therefore, this chromosomal region likely contains an important and potentially novel gene influencing the expression of leptin corrected for testosterone. Because adjusting for testosterone level eliminates the genotype ⫻ sex interaction present in serum leptin levels, identification of genes that regulate serum leptin levels corrected for testosterone may provide insight into the possible sex differences in the complex regulation of body weight.

Acknowledgments The authors thank Edgar Rodriguez for technical assistance in conducting the leptin assays. This work was supported by National Institutes of Health grants GM15803, HL45522, DK44297, and MH59490. References 1. Zhang Y, Proenca R, Barone M, Leopold L, Freidman JM. Positional cloning of the mouse obese gene and its human homologue. Nature. 1994;372:425–32. 2. Considine RV, Sinha MK, Heiman ML, et al. Serum immunoreactive leptin concentrations in normal-weight and obese humans. N Engl J Med. 1996;334:292–5. 3. Haffner SM, Gingerich RL, Miettinen H, Stern MP. Leptin concentrations in relation to overall adiposity and regional body fat distribution in Mexican Americans. Int J Obes Relat Metab Disord. 1996;20:904 – 8. 4. Maffei M, Halaas J, Ravussin E, et al. Leptin levels in human and rodent: measurement of plasma leptin and ob RNA in obese and weight-reduced subjects. Nature Med. 1995;1: 1155– 61. 5. Ostlund RE Jr, Yang JW, Klein S, Gingerich R. Relation between plasma leptin concentration and body fat, gender, diet, age, and metabolic covariates. J Clin Endocrinol Metab. 1996;81:3909 –13. 6. Rosenbaum M, Nicolson M, Hirsch J, et al. Effects of gender, body composition, and menopause on plasma concentrations of leptin. J Clin Endocrinol Metab. 1996;81:3424 –7. 7. Comuzzie AG, Hixson JE, Almasy L, et al. A major quantitative trait locus determining serum leptin levels and fat mass 606


8.

9.

10.

11.

12.

13.

14.

15.

16. 17.

18.

19. 20.

21.

22.

23.

24. 25.

is located on human chromosome 2. Nature Genet. 1997;15: 273– 6. Saad MF, Damani S, Gingerich RL, et al. Sexual dimorphism in plasma leptin concentration. J Clin Endocrinol Metab. 1997;82:579 – 84. Martin LJ, Mahaney MC, Almasy L, et al. The sexual dimorphism in serum leptin levels is a result of genotype by sex interactions. Obes Res. 2002;10:14 –21. Baumgartner RN, Waters DL, Morley JE, Patrick P, Montoya G, Garry PJ. Age-related changes in sex hormones affect the sex difference in serum leptin levels independently of changes in body fat. Metabolism. 1999;48:378 – 84. Mitchell BD, Kammerer CM, Blangero J, et al. Genetic and environmental contributions to cardiovascular risk factors in Mexican Americans—the San Antonio Family Heart Study. Circulation. 1996;94:2159 –70. Ma ZM, Gingerich RL, Santiago JV, Klein S, Smith CH, Landt M. Radioimmunoassay of leptin in human plasma. Clin Chem. 1996;42:942– 6. Green P, Falls K, Crooks S. Documentation for CRI-MAP, version 2.4. St. Louis, MO: Department of Genetics, Washington University School of Medicine; 1990. Amos CI. Robust variance-components approach for assessing genetic linkage in pedigrees. Am J Hum Genet. 1994;54: 535– 43. Almasy L, Blangero J. Multipoint quantitative trait-linkage analysis in general pedigrees. Am J Hum Genet. 1998;62: 1198 –211. Blangero J, Almasy L. Multipoint oligogenic linkage analysis of quantitative traits. Genet Epidemiol. 1997;14:959 – 64. Hopper JL, Matthews JD. Extensions to multivariate normal models for pedigree analysis. Ann Hum Genet. 1982;46: 373– 83. Allison DB, Neale MC, Zannolli R, Schork NJ, Amos CI, Blangero J. Testing the robustness of the likelihood-ratio test in a variance-component quantitative-trait loci-mapping procedure. Am J Hum Genet. 1999;65:531– 44. Goldgar DE. Multipoint analysis of human quantitative genetic variation. Am J Hum Genet. 1990;47:957– 67. Fulker DW, Cherny SS, Cardon LR. Multipoint interval mapping of quantitative trait loci, using sib pairs. Am J Hum Genet. 1995;56:1224 –33. Goring HH, Terwilliger JD, Blangero J. Large upward bias in estimation of locus-specific effects from genomewide scans. Am J Hum Genet. 2001;69:1357– 69. Lander E, Kruglyak L. Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nat Genet. 1995;11:241–7. Elbers JM, Asscheman H, Seidell JC, Frolich M, Meinders AE, Gooren LJ. Reversal of the sex differences in serum leptin levels upon cross-sex hormone administration in transsexuals. J Clin Endocrinol Metab. 1997;82: 3267–70. Dunham I, Shimizu N, Roe BA, et al. The DNA sequence of human chromosome 22. Nature. 1999;402:489 –95. Sher T, Yi HF, McBride OW, Gonzalez FJ. CDNA cloning, chromosomal mapping, and functional characterization of the


26.

27.

28.

29.

30.

human peroxisome proliferator activated receptor. Biochemistry. 1993;32:5598 –5604. Issemann I, Green S. Activation of a member of the steroid hormone receptor superfamily by peroxisome proliferators. Nature. 1990;347:645–50. Gottlicher M, Widmark E, Li Q, Gustafsson JA. Fatty acids activate a chimera of the clofibric acid-activated receptor and the glucocorticoid receptor. PNAS. 1992;89:4653–7. Keller H, Dreyer C, Medin J, Mahfoudi A, Ozato K, Wahli W. Fatty acids and retinoids control lipid metabolism through activation of peroxisome proliferator-activated receptor-retinoid X receptor heterodimers. PNAS. 1993;90:2160 – 4. Kersten S, Seydoux J, Peters JM, Gonzalez FJ, Desvergne B, Wahli W. Peroxisome proliferator-activated receptor alpha mediates the adaptive response to fasting. J Clin Invest. 1999; 103:1489 –98. Ye JM, Doyle PJ, Iglesias MA, Watson DG, Cooney GJ, Kraegen EW. Peroxisome proliferator-activated receptor (PPAR)-alpha activation lowers muscle lipids and improves

31.

32.

33.

34.

insulin sensitivity in high fat-fed rats: comparison with PPARgamma activation. Diabetes. 2001;50:411–7. Costet P, Legendre C, More J, Edgar A, Galtier P, Pineau T. Peroxisome proliferator-activated receptor alpha-isoform deficiency leads to progressive dyslipidemia with sexually dimorphic obesity and steatosis. J Biol Chem. 1998;273:29577– 85. Budarf M, McDonald T, Sellinger B, Kozak C, Graham C, Wistow G. Localization of the human gene for macrophage migration inhibitory factor (MIF) to chromosome 22q11.2. Genomics. 1997;39:235– 6. Hirokawa J, Sakaue S, Furuya Y, et al. Tumor necrosis factor-alpha regulates the gene expression of macrophage migration inhibitory factor through tyrosine kinase-dependent pathway in 3T3–L1 adipocytes. J Biochem. 1998;123: 733–9. Benigni F, Atsumi T, Calandra T, et al. The proinflammatory mediator macrophage migration inhibitory factor induces glucose catabolism in muscle. J Clin Invest. 2000; 106:1291–300.


607