| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Department of Clinical Sciences, Lund University, Malmö, Sweden (SE, ML, PW, BG, GB, EW, and OM); the Department of Internal Medicine, Malmö University Hospital, Malmö, Sweden (SE and OM); and the Department of Cardiology, Malmö University Hospital, Malmö, Sweden (ML).
2 Supported by grants from the Swedish Medical Research Council, the Swedish Heart and Lung Foundation, the Medical Faculty of Lund University (Malmö University Hospital), the Albert Påhlsson Research Foundation, the Crafoord Foundation, the Ernhold Lundströms Research Foundation, the Region Skane, the Hulda and Conrad Mossfelt Foundation, the King Gustaf V and Queen Victoria Foundation, and the Lennart Hanssons Memorial Fund. 3 Address reprint requests and correspondence to S Enhörning, Department of Clinical Sciences, Clinical Research Center, Malmö University Hospital, SE-205 02 Malmö, Sweden. E-mail: sofia.enhorning{at}med.lu.se.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: We investigated whether genetic variation of the human arginine vasopressin 1A (AVPR1A) gene is associated with phenotypic features resembling those of the V1aR–/– mouse.
Design: In a population-based cross-sectional study in southern Sweden, middle-aged individuals (n = 6055) were examined in 1991–1994. Associations between 4 AVPR1A tag single nucleotide polymorphisms (rs1042615, rs10784339, rs7308855, and rs10747983) and diabetes status, glucose and triglyceride concentrations, and BMI were analyzed. Furthermore, rs1042615 was related to diabetes status, glucose, and triglycerides within sex-specific quartiles of dietary fat intake (Q1Fat-Q4Fat) and BMI (Q1BMI-Q4BMI).
Results: Subjects carrying the T allele of rs1042615 had lower concentrations of triglycerides than did CC carriers (1.36 ± 0.77 compared with 1.42 ± 0.89 mmol/L; P = 0.014), especially in nondiabetic subjects (P = 0.001). Carriers of the rs1042615 T allele had higher fasting blood glucose (5.20 ± 1.44 mmol/L compared with 5.12 ± 1.22 mmol/L; P = 0.036) and a tendency toward an increased prevalence of diabetes (odds ratio: 1.22; 95% CI: 0.99, 1.51; P = 0.067) compared with CC carriers. The less common rs10784339, rs7308855, and rs10747983 were not consistently associated with metabolic variables. Among men, the rs1042615 T allele was associated with diabetes exclusively within Q4Fat (odds ratio: 2.22; 95% CI: 1.05, 4.71; P = 0.04) and Q4BMI (odds ratio: 1.81; 95% CI: 1.11, 2.93; P = 0.02).
Conclusion: The rs1042615 T allele is associated with features resembling the phenotype of the V1aR–/– mouse, including uncoupling of the usual direct relation between glucose and triglycerides and an increased prevalence of diabetes in subjects with a high fat intake or who are overweight.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The neurohypophyseal peptide arginine vasopressin (AVP) is involved in the inhibition of diuresis, modulation of ACTH release, stimulation of liver glycogenolysis, facilitation of thrombosis, and contraction of smooth muscle. These effects are mediated through 3 different vasopressin receptors: V1aR, V2R, and V1bR (also called V3) (3, 4). In patients with type 2 diabetes, the plasma AVP concentration has been reported to be elevated (5), and AVP infusion in humans leads to elevated circulating glucose concentrations (6). Thus, apart from an osmotically induced increase in AVP secretion due to hyperglycemia (7), it can be hypothesized that AVP or altered AVP sensitivity at the receptor or postreceptor level may be involved in regulating glucose homeostasis in humans.
Recent data show significant changes in the lipid and glucose metabolism of male mice lacking one of the vasopressin receptors—V1aR (V1aR–/–). On a normal Chow diet, these knockout mice had slightly higher fasting glucose concentrations and glucose intolerance than did wild-type (WT) mice (8). Interestingly, in contrast with what would be expected from the direct relation between circulating concentrations of glucose and triglycerides in human epidemiologic studies, V1aR–/– mice had lower concentrations of circulating plasma triglycerides than did WT mice (9), which suggested that the loss of V1aR function can uncouple the usual positive correlation between these 2 traits. The V1aR–/– mice weighed less than the WT mice after 16 wk of a normal Chow diet, but after being fed a high-fat diet the V1aR–/– mice were more prone to develop overt obesity and diabetes (8).
Given the phenotype of the V1aR–/– mice, we hypothesized that the common genetic variation of the human homolog to the V1aR gene, ie, the arginine vasopressin 1A (AVPR1A) gene, in a population-based sample would be associated with a similar phenotype of elevated glucose concentrations accompanied by reduced concentrations of triglycerides, although less pronounced than the complete knockout effect seen in V1aR–/–. Furthermore, the finding that a high-fat diet in the V1aR–/– mice resulted in exacerbation of overt obesity and diabetes prompted us to test the hypothesis that any association between genetic variation in the AVPR1A gene and impaired glucose homeostasis would be enhanced in subsets of our population with the highest intake of dietary fat.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Glucose and triglyceride analyses were carried out at the Department of Clinical Chemistry, Malmö University Hospital, which is attached to a recurrent standardization system. Glucose was measured in overnight fasting blood samples in whole blood glucose by a hexokinase-glucose-6-phosphate dehydrogenase method. Diabetes mellitus was defined by self report of a physician diagnosis or fasting blood glucose concentration 
6.1 mmol/L or use of antidiabetic medication. Overweight was defined as a BMI 25 and obesity as a BMI 30. Clinical characteristics of study subjects are shown in Table 1. The study protocols were approved by the ethics committee of Lund University, and all participants provided written informed consent.
|
The relative validity of the MDC method was evaluated in 1984–1985 in a sample of Malmö residents (105 women and 101 men) aged 50–69 y using 18 d of weighed records, 3 d every second month during a year, as the reference method. The Pearson correlation coefficients for total fat, adjusted for total energy, between the reference method and the MDC method were 0.69 in women and 0.64 in men (12). These values are generally higher than those found with the use of comparable dietary methods in other studies, which were performed in other populations (13).
Because dietary patterns and self-reported dietary intakes tend to differ according to sex (14, 15), the subgroup analyses of genotype-phenotype associations in different quartiles of fat intake were analyzed in male and female subjects separately. After logarithmic transformation, we regressed total fat intake on total energy intake in males and females separately. The residuals were saved and used to rank individuals. The study population was divided into quartiles (Q1Fat-Q4Fat), with Q1Fat representing those with the lowest and Q4Fat those with the highest relative intake of dietary fat. Using strata of residuals of fat intake in relation to total energy intake (Q1Fat-Q4Fat), instead of total fat intake, we reduced confounding from dietary over- and underreporting. Furthermore, as a diagnosis of diabetes commonly leads to a change in dietary pattern, subjects with known diabetes (ie, reporting a history of diabetes or being under treatment with antidiabetic agents) were excluded in the fat intake stratified analyses of genotype compared with diabetes and fasting blood glucose concentration (n = 5188). Thus, in these analyses, all diabetes cases were new onset diabetes cases defined solely by having fasting blood glucose 6.1 mmol/L on baseline screening.
Leisure-time physical activity was assessed on the basis of a list of activities (18 items) adapted from the Minnesota Leisure Time Physical Activity instrument (16). A score was obtained by multiplying the reported amount of minutes per week spent on a specific activity by an activity-specific factor. Based on this score, the population was ranked into quartiles.
Genotyping
DNA was extracted from frozen granulocyte or buffy coat samples collected from MDC-CVA with the use of QIAamp-96 spin blood kits (QIAGEN, Stockholm, Sweden) at the DNA extraction facility supported by SWEGENE. To analyze the AVPR1A polymorphism and capture the maximum of the genetic variance of the AVPR1A gene, data from HapMap were used (www.hapmap.org) to select 4 tag single nucleotide polymorphisms (SNPs): rs1042615, rs10747983, rs10784339, and rs7308855. Primers and probes were custom synthesized by Applied Biosystems (Foster City, CA) according to standard recommendations for the AB Prism 7900HT analysis system, and genotyped with polymerase chain reaction–based methods (17).
Statistics
SPSS statistical software (version 14.0; SPSS Inc, Chicago, IL) was used for all calculations. Data are expressed as means ± SDs. Because fasting concentrations of blood glucose and plasma triglycerides tended to be skewed to the right, all statistical analyses were performed after natural logarithmic transformation of the 2 traits. Significances of differences in continuous variables were tested with a t test or analysis of variance (ANOVA). Multivariate linear regression was used to test whether associations between genetic variants and continuous variables were independent of potential confounders. Significance of frequency differences in dichotomous variables was tested with a chi-square test. Multivariate logistic regression analysis was used to estimate odds ratios (ORs) and 95% CIs for dependent dichotomous variables in relation to genetic variance in crude and adjusted models. To test whether genetic associations differed according to fat intake and BMI, we stratified our population into quartiles of fat intake (Q1Fat, Q2Fat, Q3Fat, and Q4Fat) and BMI (Q1BMI, Q2BMI, Q3BMI, and Q4BMI). A 2-sided P value <0.05 was considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Genetic AVPR1A variance vs triglycerides
Carriers of at least one rs1042615 T allele had significantly lower concentrations of triglycerides (Table 2). After adjustment for age, sex, physical activity, and BMI, the association was still significant (P = 0.025). Further inclusion of glucose in the model strengthened the association even more (P = 0.001).
|
None of the other AVPR1A tag SNPs (rs10747983, rs10784339, and rs7308855) were significantly associated with triglyceride concentrations (Table 2).
Genetic AVPR1A variance vs glucose homeostasis
Individuals carrying at least one rs1042615 T allele had significantly higher fasting blood glucose concentrations than did carriers of the CC genotype (Table 2). The association between the T allele and elevated fasting blood glucose concentrations remained significant after adjustment for age, sex, physical activity, and BMI (P = 0.036) and was strengthened after additional inclusion of triglycerides in the model (P = 0.005). The OR for diabetes in carriers of the T allele was 1.22 (95% CI: 0.99, 1.51; P = 0.067). AVPR1A tag SNPs rs10747983, rs10784339, and rs7308855 were neither significantly associated with fasting blood glucose concentration nor with diabetes status.
rs1042615 In relation to diabetes after stratification for fat intake
The finding that the phenotype of human carriers of the rs1042615 T allele (increased glucose and decreased triglyceride concentrations) resembled that of the V1aR–/– mouse encouraged us to test for further phenotypic similarities. Because male V1aR–/– mice were more prone than WT mice to develop overt diabetes and obesity when fed a high-fat diet, we tested the rs1042615 T allele for phenotypic association within sex-specific strata of fat intake (Q1Fat-Q4Fat). Because diabetes commonly leads to changes in dietary patterns, we excluded patients with a previous diagnosis of diabetes (79 men and 76 women).
In men within Q4Fat, the rs1042615 T allele was significantly associated with diabetes, whereas there was no significant association in men belonging to Q1Fat-Q3Fat (Table 3). An interaction test was conducted between the genetic variants and fat intake. The interaction term (fat intake x genotype) was entered into a multivariate logistic regression together with genotype and fat intake, with diabetes as the outcome variable. The interaction was significantly related to diabetes in men (P = 0.044). Among women, there was no association between the T allele and diabetes in any of the quartiles of fat intake (Table 3). After pooling men and women in Q4Fat, we observed a borderline significantly higher prevalence of diabetes in T allele carriers than in carriers of the CC genotype (OR: 1.78; 95% CI: 0.99, 3.17; P = 0.052) that became significant after adjustment for sex, age, BMI, and physical activity (P = 0.013). In consequence, fasting blood glucose concentrations were slightly higher in men carrying the T allele than in carriers of the CC genotype in Q4Fat (5.13 ± 0.57 mmol/L compared with 5.27 ± 0.89 mmol/L) with P = 0.08 in crude analyses and P = 0.061 in analyses adjusted for BMI and physical activity, whereas fasting glucose concentrations did not differ between carriers and noncarriers of the T allele among Q1Fat-Q3Fat (Table 4).
|
|
|
|
rs1042615 In relation to triglycerides after stratification for fat intake and BMI
Finally, we tested whether the finding of relatively lower triglyceride concentrations in carriers of the rs1042615 T allele as compared with carriers of the CC genotype was dependent on strata of fat intake or BMI. Because the rs1042615 T allele was associated with diabetes in individuals within Q4Fat, and diabetes per se is strongly associated with elevated triglyceride concentrations, these analyses were performed in nondiabetic participants only.
Among nondiabetic men there was no consistent association between the T allele and low triglyceride concentrations in any of the fat quartiles (Q1Fat-Q4Fat) (Table 4). In contrast, nondiabetic women within Q4Fat carrying the T allele had significantly lower triglyceride concentrations than did carriers of the CC genotype. This association was no longer significant after adjustment for age, BMI, and physical activity (Table 4). No association was observed between the T allele and triglyceride concentrations in women in Q1Fat-Q3Fat (Table 4). Finally, no association was observed between the rs1042615 T allele and triglyceride concentrations among either men or women in the upper quartile of BMI (Table 6).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our hypothesis that common human genetic variation may lead to a phenotype similar to that of the V1aR–/– mouse was supported by many of our findings. Male V1aR–/– mice have higher fasting glucose concentrations than WT mice, but they do not develop diabetes, which is in accordance with our findings regarding the rs1042615 polymorphism in humans at the population level.
When fed a high-fat diet, male V1aR–/– mice were more prone to develop overt diabetes and obesity than the WT mice. We attempted to mimic the environmental background of a high-fat diet applied in the mouse model in an epidemiologic setting by testing the genetic association in strata of our population according to fat intake and BMI. The rs1042615 T allele was significantly associated with diabetes exclusively in the top quartile of fat intake and BMI, respectively. These results were also in accordance with previous findings in the mouse model.
To minimize the risk of over- and underreporting of fat intake, quartiles were based on the residuals of the regression line between fat intake and total energy intake. The median (interquartile range) of fat intake expressed as kcal/d in subjects without known diabetes in the top quartile of fat intake was 1161 (946–1439) for men and 871 (711–1059) for women.
Phenotype-dependent discrepancies between reported current dietary intakes and actual dietary patterns on a long-term retrospective basis may naturally be the case in patients with diabetes. Therefore, we excluded cases with known diabetes (positive self-report of diabetes or use of antidiabetic drugs) in the genotype-phenotype analyses in strata of reported fat intake, which left for analysis only subjects with diabetes diagnosed at the MDC screening event. Furthermore, in a subanalysis of males in Q4Fat, excluding all subjects who reported a history of substantial change in their dietary habits (n = 1666), the association between rs1042615 and diabetes was stronger (OR: 2.46; 95% CI: 1.05, 5.72; P = 0.037).
Given the fact that the knockout mice developed diabetes on the basis of either a high fat intake or obesity or both, we complemented the data derived from stratification of fat intake with that of BMI. The fact that we observed similar results in men in Q1BMI-Q4BMI as in those in Q1Fat-Q4Fat [ie, the rs1042615 T allele was associated with diabetes exclusively in cases of a high fat intake (Q4Fat) and high BMI (Q4BMI), respectively] supports our conclusion that the rs1042615 T allele results in a phenotype similar to that observed in male V1aR–/– mice.
Although the association of rs1042615 with glucose concentrations and diabetes in both nonstratified analyses and analyses stratified for fat intake and BMI, as well as the association with triglycerides in nonstratified analysis, were concordant with previous findings in the V1aR–/– mouse, no study has investigated whether triglyceride concentrations differ between V1aR–/– and WT mice in cases of a high fat intake. In humans, we found that the difference in triglycerides between nondiabetic carriers and noncarriers of the rs1042615 T allele was driven by subjects in the top quartile of fat intake, primarily women. However, there was only a borderline significant difference after multivariate adjustment including BMI, and in the top quartile of BMI we observed no difference in triglycerides according to genotype. Thus, although there was an association between rs1042615 and lower triglycerides in the sample as a whole as well as in nondiabetic females in the top quartile of fat intake, the lack of data from the V1aR–/– mice and nonsignificant findings in multivariate-adjusted and BMI-stratified analyses in humans makes it difficult to conclude whether the association between the rs1042615 T allele and lower triglyceride concentrations is enhanced by a high fat intake.
Studies of V1aR–/– and WT mice have shown that the relatively lower concentrations of triglycerides in V1aR–/– mice are caused by enhanced lipolysis that is at least partially a consequence of increased sensitivity of isoproterenol-mediated lipolysis and resistance to insulin-mediated inhibition of lipolysis. As a consequence, glycerol concentrations were increased in the V1aR–/– mouse; however, free fatty acids were reduced because of a simultaneous enhancement of β-oxidation (9). Because the V1aR–/– mouse is glucose intolerant and resistant to insulin-mediated glucose uptake, it can be speculated its hypermetabolism of fat is a compensatory mechanism for the reduced capacity to use glucose as a source of energy. In contrast with V1aR–/– mice, humans with insulin resistance and glucose intolerance usually have high concentrations of triglycerides. We speculate that the cause of relatively higher fasting glucose and lower fasting triglyceride concentrations seen in human carriers of the rs1042615 T allele may have an etiology similar to that of V1aR–/– mice, ie, preferential metabolism of fat instead of glucose as a consequence of reduced expression of the V1a receptor. However, it is important to remember that the regulation of glucose and fat metabolism has a multifactorial and polygenic nature. Thus, it is logical that the association between the rs1042615 T allele and elevations in glucose and reductions in triglycerides is small. Still, the uncoupling of the usual direct relation between glucose and triglycerides is interesting for the understanding of the complex pathophysiology of type 2 diabetes and its related alterations in blood lipids. In fact, similar to the AVPR1A polymorphism, the first type 2 diabetes susceptibility gene to be discovered—transcription factor 7-like 2 (TCF7L2)—is also associated with relatively lower circulating concentrations of triglycerides (18). A greater understanding of how multiple diabetes susceptibility genes interact, and their net effect on glucose and lipid metabolism, may contribute with clues that are important to the development of novel antidiabetic agents.
We do acknowledge a number of limitations of our study. The cross-sectional nature of our phenotypic and dietary data may lead to bias caused by the relation between diabetes and dietary patterns. Although we attempted to eliminate such bias by excluding patients with known diabetes in the analyses stratified by fat intake, these findings need to be confirmed in prospective studies. Furthermore, the significance levels for phenotypic associations with the rs1042615 T allele are relatively modest in relation to the number of tests that were performed. On the other hand, the phenotypes that we tested in relation to genetic AVPR1A variation were all prespecified and defined by previous findings in V1aR–/– mice (8, 9). Furthermore, the 4 SNPs studied were selected according to HapMap data (www.HapMap.org) to capture the maximum of common genetic variation in the AVPR1A gene. Although the rs1042615 T allele was associated with a range of metabolic features similar to those seen in V1aR–/– mice, we had no prior hypothesis concerning which of the 4 SNPs would be associated with such phenotypes, emphasizing the importance of replication of the results.
In conclusion, the rs1042615 T allele may be associated with features resembling the phenotype of V1aR–/– mice, including uncoupling of the usual direct relation between glucose and triglycerides and an increase in the prevalence of diabetes in subjects with a high fat intake or who are overweight. These findings may add to the understanding of the complex pathophysiological background of type 2 diabetes and related lipid disturbances.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
