| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Internal Medicine (L.-M.C., T.-Y.T.), National Taiwan University Hospital, Taipei, Taiwan; Graduate Institute of Clinical Medicine (L.-M.C.), National Taiwan University, Taipei, Taiwan; Division of Biostatistics and Bioinformatics (Y.-F.C., C.A.H.), National Health Research Institutes, Taipei, Taiwan; Department of Education and Research (W.H.-H.S.), Taichung Veterans General Hospital, Taichung, Taiwan; Division of Endocrinology and Metabolism (Y.-J.H.), Tri-Service General Hospital, Taipei, Taiwan; Department of Medical Research and Education (L.-T.H.), Taipei Veterans General Hospital, Taipei, Taiwan; University of Hawaii (J.G.), Honolulu, Hawaii; Hawaii Center for Health Research (B.R.), Honolulu, Hawaii; Division of Cardiovascular Medicine (T.Q.), Stanford University School of Medicine, Stanford, California; and Departments of Medicine and Obstetrics/Gynecology (Y.-D.I.C.), Cedars Sinai Medical Center and University of California at Los Angeles, Los Angeles, California
Address all correspondence and requests for reprints to: Lee-Ming Chuang, M.D., Ph.D., Department of Internal Medicine, National Taiwan University Hospital, 7 Chung-Shan South Road, Taipei, Taiwan. E-mail: leeming{at}ha.mc.ntu.edu.tw.
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
In animal models, administration of adiponectin has been shown to increase insulin sensitivity in liver, skeletal muscle, and adipose tissues (10, 11). Others and we have found that the plasma levels were also up-regulated under weight reduction and pharmacological treatment for diabetes (4, 12, 13). We found that the increase of the plasma levels of adiponectin correlated with the increase in insulin sensitivity measured with a modified insulin suppression test (12), consistent with a previous report that showed plasma levels of adiponectin were strongly correlated with insulin sensitivity evaluated by glucose disposal rate (14). Interestingly, adiponectin level might serve as a trait of insulin resistance linked to the future development of type 2 diabetes (15).
Plasma levels of adiponectin are under complex regulation. Adiponectin expression and/or secretion is increased by IGF-I and ionomycin and decreased by TNF-
, glucocorticoids, ß-adrenergic agonists, and cAMP (16, 17). Research on the limited number of candidate genes, such as PPAR
and APM1, has suggested a genetic background that might regulate adiponectin gene expression (18, 19). Further studies with whole genome scans have mapped chromosomal regions for adiponectin levels in two different populations (20, 21). These analyses revealed a quantitative trait locus (QTL) on chromosome 5p for a northern European population (20), and on chromosome 9p for a Pima Indian population living in Arizona (21). There are some smaller signals on chromosomes 2, 3, and 10 for Pima Indians and on chromosome 14 for a European population. These data suggest that there is a great ethnic heterogeneity to the genetic influence on plasma adiponectin levels. Therefore, we decided to conduct a genome-wide scan to identify QTLs influencing adiponectin concentrations in two populations recruited in the same Stanford Asia-Pacific Program of Hypertension and Insulin Resistance (SAPPHIRe), namely a Chinese population living in Taiwan and a Japanese population living in Hawaii.
| Subjects and Methods |
|---|
|
|
|---|
All participants were recruited for the Stanford Asia-Pacific Program of Hypertension and Insulin Resistance, the SAPPHIRe, a program designed to investigate the genetics of hypertension (22, 23) and insulin resistance as intermediate phenotypes (24, 25). The characteristics of the study population and the inclusion and exclusion criteria of the SAPPHIRe study were detailed in several previous publications (22, 23, 24, 25). Notably, diagnosed diabetic subjects based on the World Health Organization criteria were excluded. For this report, individuals were selected who: 1) had taken part in the genome-wide scan; and 2) had a determination of fasting plasma adiponectin level with a RIA (LINCO Research, Inc., St. Charles, MO). In this study, a total of 1359 subjects from 507 families were analyzed. The institutional review board of each participating center approved the protocols. All subjects received a clear description of the protocol and consented to participate in the study.
Genotyping
Whole blood was obtained from all consenting family members for DNA extraction. DNA was prepared using commercial kits (Puregene, Gentra Systems, Minneapolis, MN). Genotyping was performed at the Marshfield Medical Research Foundation (Marshfield, WI) using the Weber screening set 9 (Research Genetics, Inc., Huntsville, AL). This procedure used 376 autosomal markers representing short tandem repeat polymorphisms (22) and yielded an average map density of 10 cM. Genotyping quality was monitored by typing 30 samples in duplicate. We estimated an error rate of approximately 1% based on these duplicate samples. All markers were inspected for mendelian errors, and genetic distances between markers were determined.
Genome-wide multipoint linkage analyses
The genome-wide scans on logarithm-transformed adiponectin, with adjustments for age and gender and with or without the additional adjustment for body mass index (BMI), were carried out for subjects of Chinese and Japanese origins separately, using a variance-components approach as implemented in the SOLAR (Sequential Oligogenic Linkage Analysis Routines) computer package. The variance-components model partitions the variability of a trait into components for a QTL, the residual polygenic component, and the random environmental component (26). Likelihood-ratio tests were used to test for the null hypothesis of no linkage (i.e. the variance component due to the QTL being 0). The likelihood ratio was derived by dividing the likelihood of the estimated variance component due to the QTL by the likelihood of this variance component being zero. Twice the logarithm of the likelihood ratio yields a test statistic that is asymptotically distributed as a 1/2:1/2 mixture of a
12 and a point mass at zero, denoted by
02. LOD scores were calculated as logarithm to base 10 of the likelihood ratios. A LOD score exceeding 3.3 was considered to be of genome-wide significance for evidence of linkage, whereas a LOD score greater than 1.9 was suggestive evidence for linkage (27). One-unit LOD support intervals (S.I.) were obtained by identifying the peak for the maximum LOD score on the plot of the linkage results, dropping down one LOD unit and finding the chromosomal region defined by the shoulders of the curve (28).
To verify the findings from the multipoint linkage analyses, we conducted a simulation study to examine the false-positive rates in our analyses using the Simulation module implemented in the MERLIN computer package (29). One hundred datasets that mimicked the original data in terms of marker informativeness, spacing, and missing data patterns were simulated. Marker data were simulated under the null hypothesis of no linkage to adiponectin levels. Phenotypic measurements, including age, sex, and adiponectin values, were preserved. The multipoint variance-components approach was then applied to these data sets; empirical false positive rates in a genome-wide scan were derived.
| Results |
|---|
|
|
|---|
There were a total of 1007 siblings (constituting 1205 sibpairs) from 360 nuclear families of Chinese, and a total of 352 siblings (constituting 333 sibpairs) from 147 nuclear families of Hawaiian Japanese recruited in this study (Table 1
). Nuclear families with at least two siblings were informative for linkage analysis. Among the 360 Chinese families, 24.7% have one parent genotyped, and 4.4% have both parents genotyped. In Japanese families, 60.5% of the families have one parent genotyped, and 10.9% of them have both parents genotyped. There was no significant difference in gender distribution, but the age and BMI were higher for the Hawaiian Japanese population (P < 0.0001, using the Generalized Estimating Equations method, results not shown). Mean levels of adiponectin were higher (P < 0.0001, results not displayed) in females in both Chinese and Hawaiian Japanese populations, yet no significant difference (P = 0.13, results not shown) in adiponectin concentrations was observed between ethnic groups (Table 2B
).
|
|
|
A total of 376 microsatellite markers spread at every 10 cM across the genome were genotyped. The results of genome-wide multipoint linkage analyses for adiponectin concentrations for each population were shown in Fig. 1
. With adjustment for age and sex, a single peak with a maximum LOD of 3.19 at 39 cM of chromosome 15 was found in Chinese, whereas a smaller peak with a maximum LOD of 2.40 at 38 cM on chromosome 18 was noted in Hawaiian Japanese (solid line in Fig. 1
.). With further adjustment for age, sex, and BMI, the maximum LOD scores declined substantially in Chinese, but this phenomenon was not observed in Hawaiian Japanese (dotted line in Fig. 1
).
|
|
|
|
Computer simulations for the empirical P values from genome-wide scans
The false-positive rates of multipoint LOD scores from a genome-wide scan, using the variance-components approach based on the random autosomal chromosomes that were unlinked to the adjusted transformed adiponectin level, were derived by a simulation of 100 replicates. Significant linkage is the statistical evidence expected to occur 0.05 times in a genome scan (that is, with a probability of 0.05), whereas suggestive linkage would be expected to occur one time at random in a genome scan (27). For Chinese, under the null hypothesis of no linkage, the false positive rate of observing a maximum LOD score of 3.19 or greater was 0.01 in the genome scan. The chance of observing a maximum LOD score greater than 2 in the genome-scan in the absence of linkage was 0.17. For Japanese, the false-positive rates of observing maximum LOD scores greater than 3 or 2 were 0.02 and 0.26, respectively. These simulation results suggested that distributional irregularities in adiponectin that may result in violation of the assumption of the variance-components method have not resulted in a gross inflation of the type I error rate.
| Discussion |
|---|
|
|
|---|
The physiological and molecular control of adiponectin expression is not fully elucidated (16, 19). The adiponectin gene (APM1) promoter region contains consensus sequences for both peroxisome proliferator-activated receptor and glucocorticoid receptor (31, 32). Interestingly, we have demonstrated a significant interaction between the genotypes of APM1 and PPAR
in determining insulin sensitivity in these populations (25). Whether the genotypes of the two genes affect the circulating level of adiponectin remains to be seen. However, a number of previous studies have examined the association of polymorphisms of the PPAR
gene with circulating concentrations of adiponectin and metabolic diseases, with inconsistent results (33, 34). This suggests that other potentially causal genes might contribute to the adiponectin levels in the circulating blood.
In this study, we have examined linkage to QTLs of adiponectin in two populations. A significant linkage on chromosome 15 was found in Chinese; a smaller peak indicative of linkage on chromosome 18 was found in Hawaiian Japanese. After adjustment for BMI, in addition to age and sex, we found that in Chinese, the LOD score of the region on chromosome 15 declined remarkably from 3.19 to 1.31, suggesting that the effect of this locus is partially mediated by obesity. The LOD score for BMI on the same region is 0; the peak of LOD scores for BMI is 2.73 at 73 cM on chromosome 12 (our unpublished observations). Nevertheless, the same QTL on chromosome 15 showed evidence for other metabolic phenotypes, such as high-density lipoprotein cholesterol (our unpublished observations). How these putative QTLs for various traits interact needs further investigation. On the contrary, the LOD score of the region on chromosome 18 for Japanese did not decline significantly (from 2.40 to 2.23) when adjusted for age, sex, and BMI, indicating that the effect of chromosome 18 locus is independent of obesity. More interestingly, when further adjusted for age, sex, and BMI, peaks on chromosomes 3, 18, and 20 were identified in Hawaiian Japanese. These additional loci were not observed in the Chinese population. Our findings that different loci are mapped for adiponectin levels in Chinese and Japanese, together with previous genome scans for Caucasians and Pima Indians, indicate a great genetic heterogeneity in control of adiponectin levels among different ethnic groups.
The major peaks that we found on chromosome 15 and 18 were not seen in the previous analysis (20, 21), suggesting that the set of genes controlling the expression/secretion of adiponectin are different in each population. The only linkage found on chromosome 3 (112 and 119 cM) found in this study might be similar to the previous studies (20, 21). The adiponectin gene itself might be excluded because it lies at approximately 200 cM on chromosome 3.
The estimates derived from the present analyses suggest that the loci on chromosomes 15 and 18 account for 35.1% and 50.7% of the variation in age/sex-adjusted adiponectin levels in the Chinese and Japanese population, respectively. It is worthy to note that the estimated heritability of adiponectin levels in Hawaiian Japanese is lower, though not significantly so, than in Chinese (0.48 vs. 0.70). But the mapped locus, which is independent of obesity, accounts for a larger portion of variation in adiponectin levels. This estimate must be interpreted with caution, however, because the multiple testing involved in a genome-wide scan will produce an upward bias, the magnitude of which is difficult to quantify (35).
Given the stochastic variability in linkage studies, our finding of potential heterogeneity across populations in the genetic control of adiponectin should be interpreted with caution. A more convincing case for heterogeneity might be made by exclusion mapping. On the other hand, there are some consistent findings in chromosome 3, with evidence of linkage to plasma adiponectin level in the previous two scans (20, 21) and ours in this study. However, the peaks of linkage are different among these studies [Comuzzie et al. (20), 201 cM; Lindsay et al. (21), 124 cM; this study, 112 and 119 cM]. Further fine mapping of this region might be of help to confirm and localize the QTL for adiponectin in the future.
There are many possible explanations for the failure of the current study to replicate the previous results, including differences in study design, sample size (in particular, the small sample size for Japanese in this study), or racial composition. However, with the same study design for SAPPHIRe, we here present evidence that novel regions on chromosomes 15, 18, and 20 may be important in influencing circulating levels of adiponectin in populations of Chinese and Japanese descent, in addition to a region on chromosome 3 that has been reported previously in other populations. Further fine mapping of these regions might help to identify the candidate genes that might control the expression and blood level of adiponectin in different populations.
| Acknowledgments |
|---|
| Footnotes |
|---|
Abbreviations: BMI, Body mass index; QTL, quantitative trait locus; SAPPHIRe, Stanford Asia-Pacific Program of Hypertension and Insulin Resistance; S.I., support interval(s).
Received April 5, 2004.
Accepted July 30, 2004.
| References |
|---|
|
|
|---|
agonist, rosiglitazone, increases plasma levels of adiponectin in type 2 diabetic patients. Diabetes Care 25:376380
ligands increase expression and plasma concentrations of adiponectin, an adipose-derived protein. Diabetes 50:20942099
2 genotypes in modulation of insulin sensitivity: a family-based association study. Diabetologia 46:977983[CrossRef][Medline]
) 2 gene Pro12Ala polymorphism may influence serum level of an adipocyte-derived protein, adiponectin, in the Japanese population. Metabolism 51:14071409[CrossRef][Medline]
2 Pro12Ala genotype on serum adiponectin concentrations in healthy Europeans. Metabolism 52:798[Medline]
This article has been cited by other articles:
![]() |
W.-S. Yang, Y.-C. Yang, C.-L. Chen, I-L. Wu, J.-Y. Lu, F.-H. Lu, T.-Y. Tai, and C.-J. Chang Adiponectin SNP276 is associated with obesity, the metabolic syndrome, and diabetes in the elderly Am. J. Clinical Nutrition, August 1, 2007; 86(2): 509 - 513. [Abstract] [Full Text] [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] |
||||
![]() |
X. Guo, M. F. Saad, C. D. Langefeld, A. H. Williams, J. Cui, K. D. Taylor, J. M. Norris, S. Jinagouda, C. H. Darwin, B. D. Mitchell, et al. Genome-Wide Linkage of Plasma Adiponectin Reveals a Major Locus on Chromosome 3q Distinct From the Adiponectin Structural Gene: The IRAS Family Study Diabetes, June 1, 2006; 55(6): 1723 - 1730. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |