|
|
||||||||
1 Unit of Endocrinology, Scientific Institute Casa Sollievo della Sofferenza, San Giovanni Rotondo, Italy
2 Research Division Joslin Diabetes Center, and Department of Medicine, Harvard Medical School, Boston, Massachusetts
3 Department of Clinical Sciences, University La Sapienza, Rome, Italy
| ABSTRACT |
|---|
|
|
|---|
2 loci and have also been linked to markers on 5p15 and 14q13. We investigated the role of these four loci in regulating serum adiponectin in a Caucasian population from Italy. Four haplotype-tagging single-nucleotide polymorphisms (ht-SNPs) (11377 C>G, 4041 A>C, +45 T>G, and +276 G>T) at the APM1 locus and the PPAR
2 Pro12Ala polymorphism were examined for association with serum adiponectin in 413 unrelated, nondiabetic individuals. Of the five SNPs tested, +276G>T was the only one to be associated with serum adiponectin ( P = 0.032), with "TT" individuals having higher adiponectin levels than other subjects. In a variance-components analysis of 737 nondiabetic members of 264 nuclear families, adiponectin heritability was 30%, with a small but significant proportion explained by the +276 genotype ( P = 0.0034). Suggestive evidence of linkage with adiponectin levels was observed on chromosome 14q13, with a LOD of 2.92 ( P = 0.000057) after including the APM1 +276 genotype in the model. No linkage was observed at 5p15. Our data indicate a strong genetic control of serum adiponectin. A small proportion of this can be attributed in our population to variability at the APM1 locus, but an as yet unidentified gene on 14q13 appears to play a much bigger role. insulin resistance; adipocytokines; heritability; haplotype-tagging single-nucleotide polymorphisms; linkage study
| INTRODUCTION |
|---|
|
|
|---|
Although adiponectin levels appear to be under genetic control (6, 10, 20), the specific genes that are involved are mostly unknown. Single-nucleotide polymorphisms (SNPs) as well as haplotypes at the APM1 locus have been reported to be associated with circulating adiponectin levels, but results have been controversial. We previously described an association between two SNPs in the adiponectin (APM1) gene (i.e., SNPs +45T>G and +276G>T, considered together as "TG" haplotype) and several features of the metabolic syndrome in 413 nondiabetic Caucasians from Italy (13). In a subset of 64 individuals, we found that the TG haplotype was also associated with low serum adiponectin levels (13). A different APM1 haplotype, defined by SNPs 11391G>A and 11377C>G in the promoter region ("GG" haplotype), was instead reported to be associated with low adiponectin levels in French Caucasians (20). Furthermore, an association with adiponectin levels has been reported for the PPAR
2 gene Pro12Ala polymorphism in Japanese (21), although this finding has not been replicated in other populations (18). Finally, serum adiponectin levels have been linked to regions on chromosomes 5p and 14q in a predominantly northern European population (6) and to 9p in Pima Indians (10).
To evaluate the role of APM1 gene variants in modulating adiponectin levels in Caucasians from Italy, we have expanded the population in which adiponectin levels were measured from 64 to 413 unrelated nondiabetic individuals and have extended the analysis to a total of four haplotype-tagging (ht) SNPs at the APM1 locus. In addition, the role of the PPAR
2 gene Pro12Ala polymorphism has been also evaluated in this population. Finally, we have measured heritability of serum adiponectin in 737 nondiabetic members of 264 Italian nuclear families. The same families were also utilized for linkage analysis to the two candidate regions (5p15 and 14q13) previously linked to adiponectin concentrations in Caucasians (6).
| SUBJECTS AND METHODS |
|---|
|
|
|---|
Unrelated individuals.
The study included 413 nondiabetic employees of the hospital Casa Sollievo della Sofferenza (San Giovanni Rotondo, Italy), who had fasting plasma glucose <7 mmol/l at screening and were not taking any medications. The recruitment criteria and the clinical characteristics of these subjects have been previously described (13).
Nuclear families.
This set included 473 nondiabetic, first-degree relatives of 264 of the unrelated individuals described above. The clinical characteristics of subjects belonging to nuclear families are reported in Table 1. For 105 families no parents were available, 101 had one parent, and 58 had two parents. Eighty-five families included only the proband and one or two parents. In the remaining families, the sibship size ranged from 2 to 5.
|
Genotyping
SNPs at positions 11377, 4041, +45, and +276 of the APM1 gene were determined as described (13), and the PPAR
2 Pro12Ala polymorphism was determined by means of dideoxy single-base extension of an unlabeled primer (ABI PRISM SNaPshot Multiplex Kit). As previously shown (13), SNPs 11377, 4041, +45, and +276 are sufficient to tag all common (
5%) haplotypes at the APM1 locus, which account for 86% of all haplotypes. Microsatellite markers on chromosomes 5 (D5S807, D5S817, D5S2845) and 14 (D14S1280, D14S608, D14S551) were genotyped by means of electrophoresis of fluorescent-labeled PCR products using an ABI 310 Genetic Analyzer and standard protocols.
Data Analysis
Continuous variables were compared across genotype groups by ANOVA using the "PROC GLM" procedure of the SAS software package (SAS Institute, Cary, NC). Serum adiponectin, insulin, homeostasis model of insulin resistance index (HOMA IR), and triglycerides were analyzed in the logarithms. All analyses included sex and age as covariates. Distribution of haplotypes and their association with serum adiponectin were estimated using the score statistics proposed by Schaid et al. (16) and implemented in the HAPLO.SCORE software. This method allows adjustment for nongenetic covariates (age, sex) and provides a global test of association as well as haplotype-specific tests.
Adiponectin heritability and multipoint linkage to candidate regions were determined by variance-components analysis using the SOLAR software (1). SOLAR performs a variance-components analysis of family data that decomposes the total variance of the phenotype (serum adiponectin) into components that are due to genetic effects (i.e., polygenic, additive genetic variance), measured covariates, and random environmental effects. The relative contribution of genetic factors to serum adiponectin is then estimated by the heritability (h2), defined by the ratio of the genetic variance component to the residual (after removal of covariates) phenotypic variance. All variance-components analyses were performed with age and sex included as covariates in the model. To evaluate the contribution of the APM1 276 genotype to adiponectin variance, analyses were repeated after introducing a covariate coded as 1 for G/G individuals, 0 for G/T individuals, and 1 for T/T individuals. After verifying Mendelian inheritance by means of the PEDCHECK software and computing maximum likelihood estimates of allele frequencies, multipoint linkage analysis was performed for log-transformed adiponectin levels using the variance-components model implemented in SOLAR software package with age and sex as covariates. This model makes no assumptions about mode of transmission and uses all available data from the families. Linkage is assessed by comparing a polygenic model that does not incorporate genetic marker information with a model that does incorporate genotype data (i.e., identity by descent status). Empiric P values were determined by using the "lodadj" command within SOLAR (20,000 replicates). This approach samples the null distribution (the distribution of LOD scores obtained under the "no linkage" hypothesis) so that a sorted array of LOD scores is obtained and the location of the observed LOD provides the "empiric" P value.
| RESULTS |
|---|
|
|
|---|
Differences in serum adiponectin levels were observed across genotypes of the four APM1 ht-SNPs (Fig. 1). Statistical significance was reached at position +276 ( P = 0.032 adjusting for age and sex, P = 0.06 adjusting also for BMI), with allele "T" homozygotes having 23% higher adiponectin levels than allele "G" carriers [61.2 (SD 31) vs. 49.8 (SD 29) µg/ml]. A tendency to be associated with decreased adiponectin levels was observed for homozygotes for the 11377 "GG" haplotype, but this was not significant ( P = 0.149) (Fig. 1). When the four ht-SNPs were considered together, no significant association was observed between adiponectin levels and the overall haplotype distribution (global P = 0.18). Allele G at +276 was subdivided into five common haplotypes, four of which (nos. 1, 2, 3, and 5) were subtypes of the +45 to +276 "TG" haplotype that we previously described as being associated with insulin resistance (13). One of these haplotypes (no. 1 or "GATG") had a tendency to be associated with low adiponectin levels (haplotype-specific, P = 0.045) (Table 2). Based on the data of Vasseur et al. (20), this haplotype corresponded to the 11391 to 11377 GG combination that was associated with low adiponectin levels in French individuals. However, the other haplotype corresponding to such combination (no. 3 or "GCTG") did not show the same tendency in our population ( P = 0.75) (Table 2). No significant effect on serum adiponectin levels was observed for the PPAR
2 Pro12Ala polymorphism [46.9 (SD 29.6) µg/ml in Pro/Pro vs. 40.5 ± 25.4 µg/ml in X/Ala, P = 0.12], in agreement with previous data in healthy Europeans (18).
|
|
2%). Thirty-eight percent of the residual variance could be attributed to other, as yet unidentified genetic factors. Evidence of linkage was observed on chromosome 14 at the same location where significant linkage with adiponectin levels was previously reported by Comuzzie et al. (6). The highest LOD score (LOD = 2.17, P = 0.0003, with age and sex as covariates) was detected in the vicinity of D14S551, 41 cM from pter (Table 3). The evidence of linkage at this locus increased, approaching genome-wide significance, when the APM1 +276 genotype effect on adiponectin variance was factored into the analysis (LOD = 2.92, P = 0.000057) (Table 3). Results were similar after adjustment for body weight or waist circumference, indicating that the linkage with adiponectin levels was not due to an effect of this locus on adiposity. No evidence of linkage was detected on chromosome 5 (LOD = 0.00), where another linkage peak had been previously reported (6).
|
| DISCUSSION |
|---|
|
|
|---|
The results of our study indicate that
30% of variance of circulating adiponectin can be accounted for by additive genetic factors, a proportion in agreement with that reported in other populations (6, 10, 20). An important aspect of our findings is the demonstration that adiponectin heritability is not affected by the inclusion of body weight in the genetic model. Thus the genes regulating adiponectin levels appear to be distinct from those affecting the adipose tissue mass.
A proportion of the genetic control of adiponectin levels can be attributed to genetic variability at the APM1 locus itself. SNP +276G>T appears to be the best SNP capturing such genetic effect, with allele G being associated with low and allele T with high adiponectin levels. We previously found allele G of SNP +276 to be associated with features of the metabolic syndrome, especially when occurring in conjunction with allele T at position +45 (13). Whether this is a manifestation of the same genetic effect underlying the association with low adiponectin levels is unclear at this time. SNP +276 is placed in intron 2 of the APM1 gene and does not have a known function. Thus it is probably a marker of some other variant affecting adiponectin expression. SNP +276 is in a linkage disequilibrium block encompassing most of the adiponectin gene, but whether and how far such block extends beyond the gene boundaries remain to be determined.
A tendency to be associated with low adiponectin levels was also observed for the APM1 haplotype 11391 to 11377, as previously described in French (20). However, this effect appeared to be secondary to the linkage disequilibrium between this haplotype and allele G at +276. Similarly, we could not confirm the association between PPAR
2 Pro12Ala SNP and adiponectin levels that was described in Japanese (21).
The proportion of adiponectin heritability that is accounted for by the APM1 variants that we studied, although significant, is rather small. Although one cannot exclude that as yet unidentified variants at the APM1 locus have a stronger effect, this finding suggests a role for other genes in the modulation of serum adiponectin levels. Supporting this hypothesis is the recent prospective study by Fumeron et al. (8), indicating that low adiponectin levels can predict the future development of hyperglycemia independently of APM1 genotypes. Linkage of adiponectin levels to chromosome 5p and 14q has been reported in a genome-wide screen of families of predominantly northern European ancestry (6). We have analyzed these two candidate regions in our population and have found evidence of linkage, although at the suggestive level, at the 14q locus. As expected, the LOD score at this location increases when the variance accounted for by the APM1 +276 polymorphism is subtracted from the total variance. The finding of linkage at chromosome 14q in two independent studies (Ref. 6 and this report) is highly suggestive for the presence in this region of a gene having a strong effect on adiponectin serum levels. By contrast, our data do not support linkage on chromosome 5p. This discrepancy could be due to differences in the combination of genetic factors regulating adiponectin levels in the two populations, one of northern European ancestry (6), the other from southern Europe. Population-specific mechanisms in the regulation of adiponectin levels are also suggested by the linkage of adiponectin levels to a locus on 9p in Pima Indians (10), a finding that is not observed in Caucasians (6).
In summary, our data indicate that circulating levels of adiponectin are under multigenic control. Variability at the APM1 locus contributes to such regulation. However, an as yet unidentified gene on chromosome 14q appears to play a much bigger role. Identification of this gene carries with it the promise to provide novel insights into the molecular regulation of adiponectin production and/or metabolism and how this may predispose to insulin resistance.
| GRANTS |
|---|
|
|
|---|
| FOOTNOTES |
|---|
Address for reprint requests and other correspondence: C. Menzaghi, Unit of Endocrinology, Scientific Institute CSS, Poliambulatoria Giovanni Paolo II, Viale Padre Pio, 71013 San Giovanni Rotundo (FG), Italy (E-mail: c.menzaghi{at}operapadrepio.it).
| REFERENCES |
|---|
|
|
|---|
2 Pro12Ala genotype on serum adiponectin concentrations in healthy Europeans. Metabolism 57: 798, 2003.
This article has been cited by other articles:
![]() |
M. Tzoufi, C. Kanioglou, A. Dasoula, I. Asproudis, A. Tsatsoulis, C. Sismani, P. C. Patsalis, I. Georgiou, and M. Syrrou Mosaic Trisomy r(14) Associated With Epilepsy and Mental Retardation J Child Neurol, July 1, 2007; 22(7): 869 - 873. [Abstract] [PDF] |
||||
![]() |
C. Menzaghi, V. Trischitta, and A. Doria Genetic Influences of Adiponectin on Insulin Resistance, Type 2 Diabetes, and Cardiovascular Disease Diabetes, May 1, 2007; 56(5): 1198 - 1209. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Menzaghi, A. Coco, L. Salvemini, R. Thompson, S. De Cosmo, A. Doria, and V. Trischitta Heritability of Serum Resistin and Its Genetic Correlation with Insulin Resistance-Related Features in Nondiabetic Caucasians J. Clin. Endocrinol. Metab., July 1, 2006; 91(7): 2792 - 2795. [Abstract] [Full Text] [PDF] |
||||
![]() |
I. M. Heid, S. A. Wagner, H. Gohlke, B. Iglseder, J. C. Mueller, P. Cip, G. Ladurner, R. Reiter, A. Stadlmayr, V. Mackevics, et al. Genetic Architecture of the APM1 Gene and Its Influence on Adiponectin Plasma Levels and Parameters of the Metabolic Syndrome in 1,727 Healthy Caucasians Diabetes, February 1, 2006; 55(2): 375 - 384. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. G. Woo, L. M. Dolan, R. Deka, R. D. Kaushal, Y. Shen, P. Pal, S. R. Daniels, and L. J. Martin Interactions Between Noncontiguous Haplotypes in the Adiponectin Gene ACDC Are Associated With Plasma Adiponectin Diabetes, February 1, 2006; 55(2): 523 - 529. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. S. Kang, S. Y. Park, H. J. Kim, C. W. Ahn, M. Nam, B. S. Cha, S. K. Lim, K. R. Kim, and H. C. Lee The Influence of Adiponectin Gene Polymorphism on the Rosiglitazone Response in Patients With Type 2 Diabetes Diabetes Care, May 1, 2005; 28(5): 1139 - 1144. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Zacharova, J.-L. Chiasson, M. Laakso, and the STOP-NIDDM Study Group The Common Polymorphisms (Single Nucleotide Polymorphism [SNP] +45 and SNP +276) of the Adiponectin Gene Predict the Conversion From Impaired Glucose Tolerance to Type 2 Diabetes: The STOP-NIDDM Trial Diabetes, March 1, 2005; 54(3): 893 - 899. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |