Advertisement
Research Article

Polyunsaturated Fatty Acids Modulate the Association between PIK3CA-KCNMB3 Genetic Variants and Insulin Resistance

  • Ju-Sheng Zheng,

    Affiliations: Department of Food Science and Nutrition, Zhejiang University, Hangzhou, China, Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, United States of America

    X
  • Donna K. Arnett,

    Affiliation: Department of Epidemiology, University of Alabama at Birmingham, Birmingham, Alabama, United States of America

    X
  • Laurence D. Parnell,

    Affiliation: Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, United States of America

    X
  • Yu-Chi Lee,

    Affiliation: Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, United States of America

    X
  • Yiyi Ma,

    Affiliation: Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, United States of America

    X
  • Caren E. Smith,

    Affiliation: Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, United States of America

    X
  • Kris Richardson,

    Affiliation: Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, United States of America

    X
  • Duo Li mail,

    chaoqiang.lai@ars.usda.gov (CQL); duoli@zju.edu.cn (DL)

    Affiliation: Department of Food Science and Nutrition, Zhejiang University, Hangzhou, China

    X
  • Ingrid B. Borecki,

    Affiliation: Department of Genetics, Washington University School of Medicine, Saint Louis, Missouri, United States of America

    X
  • Katherine L. Tucker,

    Affiliation: Department of Health Sciences, Northeastern University, Boston, Massachusetts, United States of America

    X
  • José M. Ordovás,

    Affiliation: Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, United States of America

    X
  • Chao-Qiang Lai mail

    chaoqiang.lai@ars.usda.gov (CQL); duoli@zju.edu.cn (DL)

    Affiliation: Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, United States of America

    X
  • Published: June 27, 2013
  • DOI: 10.1371/journal.pone.0067394

Abstract

Background

Neighboring genes PIK3CA and KCNMB3 are both important for insulin signaling and β-cell function, but their associations with glucose-related traits are unclear.

Objective

The objective was to examine associations of PIK3CA-KCNMB3 variants with glucose-related traits and potential interaction with dietary fat.

Design

We first investigated genetic associations and their modulation by dietary fat in the Genetics of Lipid Lowering Drugs and Diet Network (GOLDN) study (n = 820). Nine single-nucleotide polymorphisms (SNPs) were selected for analysis, covering more than 80% of the SNPs in the region. We then sought to replicate the findings in the Boston Puerto Rican Health Study (BPRHS) (n = 844).

Results

For KCNMB3 missense mutation rs7645550, meta-analysis indicated that homeostasis model assessment of insulin resistance (HOMA-IR) was significantly lower in minor allele T homozygotes compared with major allele C carriers (pooled P-value = 0.004); for another SNP rs1183319, which is in moderate LD with rs7645550, minor allele G carriers had higher HOMA-IR compared with non-carriers in both populations (pooled P-value = 0.028). In GOLDN, rs7645550 T allele homozygotes had lower HOMA-IR only when dietary n-3: n-6 PUFA ratio was low (≤0.11, P = 0.001), but not when it was high (>0.11, P-interaction = 0.033). Similar interaction was observed between rs1183319 and n-3: n-6 PUFA ratio on HOMA-IR (P-interaction = 0.001) in GOLDN. Variance contribution analyses in GOLDN confirmed the genetic association and gene-diet interaction. In BPRHS, dietary n-3: n-6 PUFA ratio significantly modulated the association between rs1183319 and HbA1c (P-interaction = 0.034).

Conclusion

PIK3CA-KCNMB3 variants are associated with insulin resistance in populations of different ancestries, and are modified by dietary PUFA.

Introduction

Type 2 diabetes (T2D), one of the most common chronic diseases in the world, is predicted to affect 285 million adults in 2010 and 439 million adults by 2030 [1]. Insulin resistance (IR) is central to the development of T2D, and usually precedes diagnosis of T2D by many years. Observational and intervention studies suggest that IR could be prevented by dietary and lifestyle modifications [2], including quantity and quality of dietary fat [3]. However, the beneficial effect of n-3 polyunsaturated fatty acid (n-3 PUFA) on the prevention of IR and T2D is still controversial [4][13]. A possible reason for this discrepancy is the fact that genetic factors alone or through interactions with diets and lifestyle influence the pathogenesis of IR and T2D [14][16]. Although numbers of genetic variants associated with IR continue to grow, relatively few studies have examined the interactions of genetic variants with fatty acid intake for IR. One of the novel candidate genes is PIK3CA, encoding phosphatidylinositol 3-kinase (PI3K) Class-IA p110α catalytic subunit. PI3K is a key mediator of insulin signaling and implicated in human diseases including diabetes and cancer [17]. PI3K is activated by the tyrosine-phosphorylated insulin receptor substrate following insulin stimulation, and subsequently catalyzing the formation of phosphatidylinositol (3,4,5)-trisphosphate, thereby initiating downstream pathways, leading to glucose uptake and glycogen synthesis [17], [18]. PI3K Class-IA p110α catalytic subunit is the primary insulin-responsive PI3K isoform [19], [20] and is important in growth factor and metabolic signaling [20], [21]. Mice with hepatic knockout of Pik3ca showed reduced insulin sensitivity, impaired glucose tolerance, and increased gluconeogenesis [20]. However, the relationships of PIK3CA mutations with IR in humans and whether any PIK3CA-nutrient interaction modulates IR are still unclear. The KCNMB3 gene, encoding potassium large conductance calcium-activated channel (BK channel) M beta member 3, is 5 kbp downstream of PIK3CA and is in a large linkage disequilibrium (LD) block with this gene. BK channels are a type of “large conductance” channel that transports potassium ions (K+) through cell membranes. As regulators of glucose homeostasis, these channels are important for β-cell function and regulate insulin secretion [22], [23]. To our knowledge, no studies have reported the association of KCNBM3 variants with IR and the potential interactions of these variants with dietary fatty acids.

The aim of the present study was to examine the associations between variants across the PIK3CA-KCNMB3 region and IR and other glucose-related traits with potential modifications by dietary fat in a non-Hispanic white population in the Genetics of Lipid Lowering Drugs and Diet Network (GOLDN) study [24]. Replication of the observed relationships was evaluated in an independent population of Puerto Rican adults living in Massachusetts and participating in the Boston Puerto Rican Health Study (BPRHS) [25].

Methods and Materials

Ethics Statement

For GOLDN, the protocol was approved by the Institutional Review Boards at the University of Alabama, the University of Minnesota, the University of Utah and Tufts University. For the BPRHS, the study protocol was approved by the Institutional Review Boards at Tufts University and Northeastern University. All the participants gave informed consent.

Study Populations

The GOLDN study (n = 820) consisted of 406 men and 414 women. All participants were of European origin, recruited from two National Heart, Lung, and Blood Institute Family Heart Study field centers (Minneapolis, MN, and Salt Lake City, UT) [26]. The study design and methodology were described previously [24]. Subjects were asked to take the lipid lowering drug (micronized fenofibrate, 160 mg) in a three-week unblinded clinical trial. The registration identifier is NCT00083369 (http://www.clinicaltrials.gov/ct2/show/N​CT00083369). Clinical and biochemical parameters were collected before and after the intervention, and all the fasting blood biochemical parameters used in the present study were from the second visit before the intervention.

The replication study included 844 participants (239 men, 605 women) from the BPRHS, with almost complete genotype and dietary assessment data. The study details were described in detail previously [25]. Briefly, the BPRHS is a longitudinal cohort study for stress, nutrition, health, and aging, for which participants who self-identified as Puerto Rican and lived in the Boston metropolitan area were recruited. Ancestry composition of the BPRHS is 57.2% European, 27.4% African, and 15.4% Native American [27].

Data Collection

In GOLDN, dietary information was collected by a Diet History Questionnaire (DHQ) developed by the National Cancer Institute, based on national dietary data (USDA Continuing Survey of Food Intakes by Individuals). The DHQ has been validated in two studies [28], [29]. In the BPRHS, dietary intake was assessed by a validated Food Frequency Questionnaire (FFQ) designed for and tested in this population [30], and nutrient analyses were linked to the Minnesota Nutrient Data system (1999, version 25). Importantly, PUFA intakes have been validated for both populations [27], [28], [29] and measures of erythrocyte membrane PUFA content in GOLDN is significantly correlated with dietary PUFA (r = 0.11, P = 0.002). Lifestyle and demographic information, medication use and medical history were obtained in each population by questionnaire. Anthropometric measurements were taken at the baseline visit. Height, body weight and waist circumference were all measured by standard techniques. Body mass index was calculated as weight (kg)/height (m)2. T2D was defined based on a fasting plasma glucose concentration ≥126 mg/dL or use of insulin or other diabetes medication.

Laboratory Methods

Blood samples were drawn after an overnight fast. In GOLDN, fasting insulin was measured by a commercial kit using a radioimmunoassay (Linco Research, St. Charles, MO), fasting glucose was determined by a hexokinase-mediated reaction on the Hitachi commercial kit (Linco Research, St. Charles, MO). In the BPRHS, measurements of fasting insulin, glucose and glycosylated hemoglobin (HbA1c) have been described in details previously [25]. Homeostasis model assessment of IR (HOMA, fasting glucose×fasting insulin/22.5) was used to assess IR in both populations.

Single-nucleotide Polymorphism Selection, Genotyping and Haplotype Analysis

Nine tag single-nucleotide polymorphisms (SNP) were selected using Tagger [31] based on HapMap Caucasian European Utah data (www.hapmap.org) in the Haploview program (version 4.2) with a minor allele frequency ≥0.05 and r2>0.80. The selected SNPs covered more than 80% of all SNPs in the region (160 kbp), ranging from 50 kbp upstream of the PIK3CA gene to 25 kbp downstream of the gene and covering KCNMB3. LD of the selected SNPs were estimated as correlation coefficients (r2) using Haploview (Figure S1). Haplotype of these SNPs were also estimated by Haploview.

DNA was isolated from blood samples using Gentra Puregene Blood Kits (Gentra Systems) in GOLDN and QIAamp DNA Blood Mini Kits (Qiagen) in the BPRHS. In GOLDN, genome-wide genotyping was conducted by using Affymetrix Genome-Wide Human SNP Array 6.0. MACH (Markov Chain based haplotyper V1.0.16, http://www.sph.umich.edu/csg/abecasis/Ma​CH/) was used to impute all autosomal SNPs using the phased haplotypes for CEU from HapMap (release 22, build 36) as a reference panel. In the BPRHS, genome-wide genotyping was conducted by using Illumina HumanOmni5-Quad GWAS arrays.

Selection of SNPs for Replication in the BRPHS and Population Admixture

Two SNPs: rs7645550 and rs1183319 were selected for replication in BPRHS based on evidence of significant relationships in GOLDN. One SNP, rs7645550, is a missense mutation in the KCNMB3 gene. The second SNP, rs1183319 is located in the 5′UTR of KCNMB3 gene. In the BPRHS population, the population admixture was estimated previously based on 100 SNPs that were selected as ancestry informative [27]. All the genotype-associated analyses were adjusted for the population admixture in the present study.

Statistical Analyses

Statistical analyses were performed using SAS (version 9.2) and R software (version 2.15.0). HOMA-IR, insulin, glucose and HbA1c were Box-Cox transformed [32] to achieve normal distribution prior to statistical analysis. Chi-square tests were conducted to examine whether the selected SNPs were in Hardy-Weinberg equilibrium. Due to the limited sample size, a dominant model was applied in the final analysis to increase statistical power for all SNPs, except for rs7645550, which followed a clear recessive model for HOMA-IR (HOMA-IR = 3.18, 3.67, and 3.71 for TT, CT and CC carriers, respectively). The minor allele was always treated as the reference allele in the models. The correlations of dietary fatty acids with glucose-related traits were assessed using a Pearson correlation coefficient after controlling for potential confounders. Dietary total fat, saturated fatty acid (SFA), monounsaturated fatty acids (MUFA), total PUFA, n-3 PUFA, n-6 PUFA and n-3: n-6 PUFA ratio were dichotomized based on the median intake of each population for the gene-nutrient interaction analysis. In GOLDN, the family relationships within the population were adjusted using the SAS GENMOD procedure, assuming an exchangeable correlation structure within pedigree. A multivariate interaction model was used to test the interactions of fatty acid intake with genetic variants after adjustment for age, sex, waist circumference, alcohol drinking, smoking status, physical activity, T2D and study center in the GOLDN population. In the BPRHS, using linear regression model, similar genotype associations and interactions were tested for selected SNPs while adjusting for age, sex, waist circumference, alcohol drinking, smoking status, physical activity, T2D and population admixture. Variance contribution of the genetic effect and the gene-nutrient interaction for HOMA-IR were estimated using a GWAF package [33] in R.

Meta-analysis was conducted with the Meta-Analysis Helper (METAL) (http://www.sph.umich.edu/csg/abecasis/me​tal/) under fixed-effects models, and combining the z-statistics across the two populations, weighted by sample size. Meta-analysis was only conducted to summarize the results from main genetic association, but not for gene-nutrient interaction, as the dietary intake range and FFQ used are substantially different between the two populations. To correct for multiple testing, Bonferroni adjustment was used. As our primary interest is insulin resistance, we only applied Bonferroni correction for HOMA-IR. The pertinent number of correction due to multiple testing is 63 (7×9), and the corrected P-value for a significant interaction is 0.0008 (0.05/63).

Results

Characteristics of the Populations and Genetic Variants at PIK3CA-KCNMB3

Men had significantly higher total energy and MUFA intake than women in both GOLDN and the BPRHS populations (Table 1). In GOLDN, none of the nine SNPs deviated from Hardy-Weinberg equilibrium expectations (P>0.01), with the minor allele frequencies ranging from 0.07 to 0.49 (Table S1). In the BPRHS, the minor allele frequency of rs7645550 (T) and rs1183319 (G) was 0.36 and 0.45, respectively, and did not deviate from Hardy-Weinberg equilibrium expectations (P>0.01). SNPs rs7645550 and rs1183319 were in moderate LD in both GOLDN (r2 = 0.54) and BPRHS (r2 = 0.44). Three haplotypes were observed for this pair of SNPs: CG, TA, and CA, and the corresponding frequency was 0.45, 0.40 and 0.15 in GOLDN, and 0.45, 0.36 and 0.20 in BPRHS.

thumbnail

Table 1. Characteristics of participants in GOLDN and the BPRHS1.

doi:10.1371/journal.pone.0067394.t001

Associations between PIK3CA-KCNMB3 Variants and Glucose-related Traits

Meta-analysis indicated that SNP rs7645550 minor allele T homozygotes had significantly lower HOMA-IR (z-score = 2.86, pooled P-value = 0.004), insulin (z-score = 2.33, pooled P-value = 0.02) and glucose concentration (z-score = 2.83, pooled P-value = 0.005) compared with major allele C carriers (Table 2). For SNP rs1183319, minor allele G carriers had significantly higher HOMA-IR (z-score = 2.19, pooled P-value = 0.028) and glucose concentration (z-score = 3.14, pooled P-value = 0.002) compared with major allele A homozygotes.

thumbnail

Table 2. Associations between PIK3CA-KCNMB3 variants and glucose-related traits in the GOLDN and BPRHS participants1.

doi:10.1371/journal.pone.0067394.t002

Associations between Dietary Fatty Acids and Glucose-related Traits

In GOLDN, the ratio of dietary n-3: n-6 PUFA was positively correlated with the fasting glucose concentration (r = 0.081, P = 0.023) after controlling for age, sex, waist circumference, alcohol drinking, smoking status, physical activity, T2D and study center. In the BPRHS, this ratio was inversely correlated with HOMA-IR (r = −0.069, P = 0.05) and insulin (r = −0.073, P = 0.039), and dietary n-3 PUFA was inversely correlated with insulin (r = −0.072, P = 0.042) after controlling for age, sex, waist circumference, alcohol drinking, smoking status, physical activity, T2D and population admixture. Other dietary fatty acid intakes, such as total fat, SFA, MUFA, PUFA, n-3 PUFA, and n-6 PUFA, were not correlated with glucose-related traits in the GOLDN or in the BPRHS population.

Interactions of PIK3CA-KCNMB3 Variants with Dietary Fatty Acids on Fasting Glucose and HbA1c

Genotype by diet interactions between PIK3CA-KCNMB3 variants and dietary total fat, SFA, MUFA, PUFA, n-3 PUFA, n-6 PUFA and n-3: n-6 PUFA ratio for glucose-related traits were examined. Dietary fatty acids were expressed as percentage of total energy intake and dichotomized into categorical variables based on the median intake of the population. In GOLDN, SNP rs1183319 had significant and marginally significant interactions with dietary MUFA (P = 0.014) and PUFA (P = 0.051) for fasting glucose (data not shown). Other genetic variants did not interact with fatty acid intakes for glucose concentration.

In BPRHS, dietary n-3: n-6 PUFA ratio significantly modulated the association between rs1183319 and HbA1c (P = 0.034) (Table S2). The G allele carriers had higher HbA1c compared with A homozygotes when the n-3: n-6 ratio was high (>0.09, P = 0.011), but not when the ratio was low (≤0.09, P = 0.635).

Interactions of PIK3CA-KCNMB3 Variants with Dietary Fatty Acids for HOMA-IR and Insulin

In GOLDN, significant interactions between dietary PUFA and PIK3CA-KCNMB3 variants for HOMA-IR were observed (Table 3, Figure 1). SNPs rs7645550 and rs1183319 significantly interacted with the n-3: n-6 ratio to modulate HOMA-IR (P = 0.033 for rs7645550, P = 0.001 for rs1183319) and insulin (P = 0.018 for rs7645550, P = 0.004 for rs1183319). SNP rs7645550 minor allele T homozygotes had lower HOMA-IR than C allele carriers (P = 0.001) only when the n-3: n-6 ratio was low (≤0.11) (Figure 1). Minor allele G carriers of rs1183319 had significantly higher HOMA-IR than non-carriers when the dietary n-3: n-6 PUFA ratio was low (≤0.11, P<0.001), and not when the ratio was high (>0.11, P = 0.368) (Table 3).

thumbnail

Figure 1. Interactions between rs7645550 and dietary n-3: n-6 PUFA ratio on HOMA-IR and insulin in GOLDN.

When dietary n-3: n-6 PUFA ratio was low (≤0.11), there were significant differences by genotypes for HOMA-IR (P = 0.001) and insulin (P = 0.003), while no significant differences for HOMA-IR (P = 0.86) and insulin (P = 0.56) were observed among the rs7645550 genotypes when dietary n-3: n-6 PUFA ratio was high (>0.11). The number inside the bar indicates the number of participants with a given genotype.

doi:10.1371/journal.pone.0067394.g001
thumbnail

Table 3. Interactions between dietary fatty acids and PIK3CA-KCNMB3 variants on HOMA-IR in the GOLDN participants1.

doi:10.1371/journal.pone.0067394.t003

In addition to rs7645550 and rs1183319, SNP rs7642066 was found to interact significantly with dietary PUFA, MUFA, n-6 PUFA, and the n-3: n-6 ratio on HOMA-IR (Table 3). Furthermore, rs2677764 showed a significant interaction with dietary n-3 PUFA to modulate HOMA-IR (P = 0.020, data not shown) in GOLDN. No significant interactions between rs7645550 and rs118339 and dietary fat for HOMA-IR in the BPRHS were observed.

Variance Contributions by GxE Interaction for HOMA-IR in GOLDN

Genotype by environment (GxE) interactions may contribute substantially to phenotypic variation, as the main genetic effect usually inadequately explains phenotypic variation. In order to examine the portion of phenotypic variation explained by the PIK3CA-KCNMB3 variants and their interaction with diet, the GWAF package in R was used to examine the variance contribution in GOLDN, after controlling for potential confounders. We estimated the variance contribution by the GxE interaction of the n-3: n-6 PUFA ratio with the PIK3CA-KCNMB3 variants to the total HOMA-IR variance, as this ratio was the most important dietary factor affecting the main genetic effect for HOMA-IR in the present study. The total portion of HOMA-IR variation explained by PIK3CA-KCNMB3 variants (six independent SNPs, with r2<0.20 were selected: rs3975506, rs2677760, rs2677764, rs1170672, rs1183319, and rs7642066) was 1.91%, and that explained by the interaction of PIK3CA-KCNMB3 variants with dietary n-3: n-6 PUFA ratio was 1.78% (Table S3).

Discussion

In the present study, we found that genetic variants at the PIK3CA-KCNMB3 region associated with glucose-related traits, and had significant interactions with dietary PUFA to modulate these traits, especially HOMA-IR. SNP rs7645550 minor allele T homozygous carriers had lower HOMA-IR compared with major allele C carriers in both European-American (GOLDN) and Puerto Rican (BPRHS) populations. For rs1183319, minor allele G carriers had higher HOMA-IR compared with non-carriers in both populations. In addition, both of these two SNPs interacted with dietary n-3: n-6 PUFA ratio on HOMA-IR in the GOLDN population.

Previous studies reporting associations between IR, T2D and n-3 PUFA intake are inconclusive. Some studies found that n-3 PUFA was inversely associated with IR and risk of T2D [6], [8], [9], while others found n-3 PUFA was positively associated with IR and risk of T2D [5], [7] or had no effect on IR or T2D [4], [13]. Recent meta-analyses of randomized controlled trials found that n-3 PUFA supplementation had no significant effects on IR or glycemic control [11], [12]. Meta-analyses based on prospective studies consistently concluded that overall fish and n-3 PUFA intake were not associated with risk of T2D [10], [34], [35]. The number of genetic variants demonstrated to contribute to IR and T2D continues to expand [36], [37] and a growing number of variants have been shown to be influenced by interactions with dietary factors [36], [38][42]. Genetic variants in genes including ACSL1, ADIPOQ, ADIPOR1, CLOCK, PLIN1, TCF7L2, and LEPR [42] were found to interact with dietary fatty acids to modulate IR. Hence, inconsistent associations between n-3 PUFA and IR or T2D may be partly attributed to the influence of gene-nutrient interaction [14][16]. In the present study, the associations between dietary fatty acids and IR were inconsistent across the GOLDN and BPRHS populations. Dietary n-3: n-6 PUFA ratio was inversely correlated with IR in BPRHS, but not in GOLDN; while n-3: n-6 PUFA ratio was positively correlated with glucose in GOLDN. Interactions between PIK3CA-KCNMB3 variants and n-3: n-6 PUFA ratio may contribute to such inconsistencies. The different interaction patterns may be due to differences in fatty acid intake between the two populations. For example, the mean dietary intake of n-3: n-6 PUFA ratio (0.106±0.020) in GOLDN was higher than the mean intake of n-3: n-6 PUFA ratio (0.092±0.027) in the BPRHS. On the other hand, >40% of the participants in BPRHS had T2D and most of the T2D patients took diabetes medication, which may be an important source of the different observations between the two populations. In addition, distinction in both genetic background and culture, especially as related to food choices and preparation methods, between the BPRHS and GOLDN populations could affect the observed interactions [27].

Four PIK3CA-KCNMB3 variants were found to interact with n-3: n-6 PUFA ratio or n-3 PUFA in the GOLDN population. The threshold for the Bonferroni correction was 0.0008, and the interaction of rs1183319 with the n-3: n-6 PUFA ratio on HOMA-IR (P = 0.001) marginally passed the Bonferroni correction. Furthermore, these results of gene-nutrient interactions were confirmed with variance contribution analysis, and variance contribution analysis indicated that the interaction of PIK3CA-KCNMB3 variants with n-3: n-6 PUFA ratio contributed a similar proportion of the HOMA-IR variance as did the main genetic effects. The mechanisms through which dietary n-3 PUFA modulated the association of these variants with IR and other glucose-related traits are still unclear. It is well known that n-3 PUFA and their metabolites are natural ligands for peroxisome proliferator receptor activator (PPAR) γ, and PPAR-γ is reported to regulate insulin sensitivity and glucose homeostasis through different mechanisms [43], [44]. It is predicted by the MAPPER transcription factor binding site prediction tool (http://mapper.chip.org) that the KCNMB3 promoter has a PPAR-γ -retinoid X receptor complex binding site, through which n-3 PUFA may influence KCNMB3 gene expression and interact with KCNMB3 variants to influence IR. In addition, n-3 PUFA (docosahexaenoic acid) was reported to activate vascular BK channels dependent on the cytochrome P450 epoxygenase activity [45], and this mechanism may support the interactions of n-3 PUFA with genetic variants at KCNMB3 to influence activity of BK channels. However, more research is needed to elucidate a precise mechanism.

Despite the different genetic backgrounds of the two populations, PIK3CA-KCNMB3 variants (rs7645550 and rs1183319) were found to be associated with glucose-related traits in both populations in the current study. Both SNPs are located in the KCNMB3 gene: rs7645550 is a missense mutation (Ala to Thr) and rs1183319 maps to the 5′-UTR. Importantly, the missense mutation of rs7645550, which changes an alanine conserved from mammals to amphibians to a threonine is predicted by PolyPhen-2 [46] to be “probably damaging”. Previous GWAS [47] indicated a tendency of association between rs1183319 (P = 0.08), rs7645550 (P = 0.198) and HOMA-IR (http://www.magicinvestigators.org/downlo​ads/). These non-significant associations may be due to the influence of dietary factors, as suggested by this study. It is reasonable to postulate that KCNMB3 variants may affect IR and glucose metabolism through BK channels. KCNMB3, encodes BK channel subunit 3, and variants at KCNMB3 were shown to affect BK channel activity [48]. An animal study [22] suggested that BK channel knockout (BK-KO) mice displayed markedly impaired glucose tolerance and insulin secretion through effects on β-cell function. BK channels were also reported to regulate insulin secretion in mouse pancreatic beta-cells [23]. However, no study so far has reported associations between KCNMB3 variants and IR or other glucose-related traits. Studies in additional populations are needed to further replicate these associations. Apart from KCNMB3, PIK3CA variant (rs2677760) was also reported to be associated with IR in the GOLDN population. PIK3CA is critical in the insulin transduction pathway and dysfunction of PIK3CA may lead to IR and impaired glucose tolerance, as suggested by evidence from the hepatic Pik3ca knockout mice [20]. However, rs2677760 is in the intron of the PIK3CA gene, and its function is still unclear.

Although there were significant associations between PIK3CA-KCNMB3 variants and glucose-related traits in both GOLDN and BPRHS populations, and PIK3CA-KCNMB3 variants were found to interact with PUFA, our results should be interpreted cautiously. First, the moderate sample sizes in both populations limit statistical power. Second, the associations between PIK3CA-KCNMB3 variants and glucose-related traits and the modulation by dietary fatty acids were reported for the first time in the current study. Further replication in other populations is warranted.

In conclusion, we observed that PIK3CA-KCNMB3 variants were associated with the glucose-related traits in populations of European-American and Puerto Rican ancestry. Dietary PUFA, especially the n-3: n-6 PUFA ratio, modulated the effects of PIK3CA-KCNMB3 variants on the glucose-related traits.

Supporting Information

Figure S1.

Linkage disequilibrium (LD) plot across PIK3CA-KCNMB3 region in CEU population. The horizontal white bar depicts the 160-kb DNA segment analyzed in the sample. An LD plot is presented at the bottom of the figure: each diamond represents the magnitude of LD for the single pair of markers. The numbers under the variant name and above the plot indicates the order of the corresponding variants among all the variants in the region. The numbers inside the diamonds indicate the r2 value.

doi:10.1371/journal.pone.0067394.s001

(DOCX)

Table S1.

Description of selected variants in PIK3CA-KCNMB3 region in GOLDN.

doi:10.1371/journal.pone.0067394.s002

(DOCX)

Table S2.

Interactions between dietary PUFA and rs1183319 on fasting glucose and HbA1c in the BPRHS participants.

doi:10.1371/journal.pone.0067394.s003

(DOCX)

Table S3.

Variance contributions of PIK3CA-KCNMB3 variants and their interaction with dietary n-3: n-6 PUFA ratio for HOMA-IR in GOLDN.

doi:10.1371/journal.pone.0067394.s004

(DOCX)

Acknowledgments

We thank the participants and research staff of the Genetics of Lipid Lowering Drugs and Diet Network (GOLDN) study and the Boston Puerto Rican Health Study (BPRHS).

Author Contributions

Conceived and designed the experiments: CQL JMO DKA IBB KLT. Analyzed the data: JSZ CQL. Wrote the paper: JSZ CQL DL. Manuscript editing: YCL YM CES LDP KR.

References

  1. 1. Shaw JE, Sicree RA, Zimmet PZ (2010) Global estimates of the prevalence of diabetes for 2010 and 2030. Diabetes Res Clin Pract 87: 4–14. doi: 10.1016/j.diabres.2009.10.007
  2. 2. Kelly GS (2000) Insulin resistance: lifestyle and nutritional interventions. Altern Med Rev 5: 109–132.
  3. 3. Riserus U (2008) Fatty acids and insulin sensitivity. Curr Opin Clin Nutr Metab Care 11: 100–105. doi: 10.1097/mco.0b013e3282f52708
  4. 4. Woodman RJ, Mori TA, Burke V, Puddey IB, Watts GF, et al. (2002) Effects of purified eicosapentaenoic and docosahexaenoic acids on glycemic control, blood pressure, and serum lipids in type 2 diabetic patients with treated hypertension. Am J Clin Nutr 76: 1007–1015.
  5. 5. Mostad IL, Bjerve KS, Bjorgaas MR, Lydersen S, Grill V (2006) Effects of n-3 fatty acids in subjects with type 2 diabetes: reduction of insulin sensitivity and time-dependent alteration from carbohydrate to fat oxidation. Am J Clin Nutr 84: 540–550.
  6. 6. Abete I, Parra D, Crujeiras AB, Goyenechea E, Martinez JA (2008) Specific insulin sensitivity and leptin responses to a nutritional treatment of obesity via a combination of energy restriction and fatty fish intake. J Hum Nutr Diet 21: 591–600. doi: 10.1111/j.1365-277x.2008.00902.x
  7. 7. Djousse L, Gaziano JM, Buring JE, Lee IM (2011) Dietary omega-3 fatty acids and fish consumption and risk of type 2 diabetes. Am J Clin Nutr 93: 143–150. doi: 10.3945/ajcn.110.005603
  8. 8. Nanri A, Mizoue T, Noda M, Takahashi Y, Matsushita Y, et al. (2011) Fish intake and type 2 diabetes in Japanese men and women: the Japan Public Health Center-based Prospective Study. Am J Clin Nutr 94: 884–891. doi: 10.3945/ajcn.111.012252
  9. 9. Villegas R, Xiang YB, Elasy T, Li HL, Yang G, et al. (2011) Fish, shellfish, and long-chain n-3 fatty acid consumption and risk of incident type 2 diabetes in middle-aged Chinese men and women. Am J Clin Nutr 94: 543–551. doi: 10.3945/ajcn.111.013193
  10. 10. Zheng JS, Huang T, Yang J, Fu YQ, Li D (2012) Marine n-3 polyunsaturated fatty acids are inversely associated with risk of type 2 diabetes in Asians: a systematic review and meta-analysis. PLoS ONE 2012 7: e44525 doi:10.1371/journal.pone.0044525.
  11. 11. Hartweg J, Perera R, Montori V, Dinneen S, Neil HA, et al. (2008) Omega-3 polyunsaturated fatty acids (PUFA) for type 2 diabetes mellitus. Cochrane Database Syst Rev: CD003205.
  12. 12. Akinkuolie AO, Ngwa JS, Meigs JB, Djousse L (2011) Omega-3 polyunsaturated fatty acid and insulin sensitivity: A meta-analysis of randomized controlled trials. Clin Nutr 30: 702–707. doi: 10.1016/j.clnu.2011.08.013
  13. 13. Kroger J, Zietemann V, Enzenbach C, Weikert C, Jansen EH, et al. (2011) Erythrocyte membrane phospholipid fatty acids, desaturase activity, and dietary fatty acids in relation to risk of type 2 diabetes in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam Study. Am J Clin Nutr 93: 127–142. doi: 10.3945/ajcn.110.005447
  14. 14. Lee YC, Lai CQ, Ordovas JM, Parnell LD (2011) A Database of Gene-Environment Interactions Pertaining to Blood Lipid Traits, Cardiovascular Disease and Type 2 Diabetes. J Data Mining Genomics Proteomics 2: 106. doi: 10.4172/2153-0602.1000106
  15. 15. Corella D, Ordovas JM (2012) Interactions between dietary n-3 fatty acids and genetic variants and risk of disease. Br J Nutr 107: S271–S283. doi: 10.1017/s0007114512001651
  16. 16. Cornelis MC, Hu FB (2012) Gene-Environment Interactions in the Development of Type 2 Diabetes: Recent Progress and Continuing Challenges. Annu Rev Nutr 32: 245–259. doi: 10.1146/annurev-nutr-071811-150648
  17. 17. Cantley LC (2002) The phosphoinositide 3-kinase pathway. Science 296: 1655–1657. doi: 10.1126/science.296.5573.1655
  18. 18. Boura-Halfon S, Zick Y (2009) Phosphorylation of IRS proteins, insulin action, and insulin resistance. Am J Physiol Endocrinol Metab 296: E581–591. doi: 10.1152/ajpendo.90437.2008
  19. 19. Knight ZA, Gonzalez B, Feldman ME, Zunder ER, Goldenberg DD, et al. (2006) A pharmacological map of the PI3-K family defines a role for p110alpha in insulin signaling. Cell 125: 733–747. doi: 10.1016/j.cell.2006.03.035
  20. 20. Sopasakis VR, Liu PX, Suzuki R, Kondo T, Winnay J, et al. (2010) Specific Roles of the p110 alpha Isoform of Phosphatidylinsositol 3-Kinase in Hepatic Insulin Signaling and Metabolic Regulation. Cell Metabol 11: 220–230. doi: 10.1016/j.cmet.2010.02.002
  21. 21. Foukas LC, Claret M, Pearce W, Okkenhaug K, Meek S, et al. (2006) Critical role for the p110alpha phosphoinositide-3-OH kinase in growth and metabolic regulation. Nature 441: 366–370. doi: 10.1038/nature04694
  22. 22. Dufer M, Neye Y, Horth K, Krippeit-Drews P, Hennige A, et al. (2011) BK channels affect glucose homeostasis and cell viability of murine pancreatic beta cells. Diabetologia 54: 423–432. doi: 10.1007/s00125-010-1936-0
  23. 23. Houamed KM, Sweet IR, Satin LS (2010) BK channels mediate a novel ionic mechanism that regulates glucose-dependent electrical activity and insulin secretion in mouse pancreatic beta-cells. J Physiol 588: 3511–3523. doi: 10.1113/jphysiol.2009.184341
  24. 24. Corella D, Arnett DK, Tsai MY, Kabagambe EK, Peacock JM, et al. (2007) The-256T>C polymorphism in the apolipoprotein A-II gene promoter is associated with body mass index and food intake in the genetics of lipid lowering drugs and diet network study. Clin Chem 53: 1144–1152. doi: 10.1373/clinchem.2006.084863
  25. 25. Tucker KL, Mattei J, Noel SE, Collado BM, Mendez J, et al. (2010) The Boston Puerto Rican Health Study, a longitudinal cohort study on health disparities in Puerto Rican adults: challenges and opportunities. BMC Public Health 10: 107. doi: 10.1186/1471-2458-10-107
  26. 26. Higgins M, Province M, Heiss G, Eckfeldt J, Ellison RC, et al. (1996) NHLBI Family Heart Study: objectives and design. Am J Epidemiol 143: 1219–1228. doi: 10.1093/oxfordjournals.aje.a008709
  27. 27. Lai CQ, Tucker KL, Choudhry S, Parnell LD, Mattei J, et al. (2009) Population admixture associated with disease prevalence in the Boston Puerto Rican health study. Hum Genet 125: 199–209. doi: 10.1007/s00439-008-0612-7
  28. 28. Subar AF, Thompson FE, Kipnis V, Midthune D, Hurwitz P, et al. (2001) Comparative validation of the Block, Willett, and National Cancer Institute food frequency questionnaires : the Eating at America's Table Study. Am J Epidemiol 154: 1089–1099. doi: 10.1093/aje/154.12.1089
  29. 29. Thompson FE, Subar AF, Brown CC, Smith AF, Sharbaugh CO, et al. (2002) Cognitive research enhances accuracy of food frequency questionnaire reports: results of an experimental validation study. J Am Diet Assoc 102: 212–225. doi: 10.1016/s0002-8223(02)90050-7
  30. 30. Tucker KL, Bianchi LA, Maras J, Bermudez OI (1998) Adaptation of a food frequency questionnaire to assess diets of Puerto Rican and non-Hispanic adults. Am J Epidemiol 148: 507–518. doi: 10.1093/oxfordjournals.aje.a009676
  31. 31. de Bakker PI, Yelensky R, Pe'er I, Gabriel SB, Daly MJ, et al. (2005) Efficiency and power in genetic association studies. Nat Genet 37: 1217–1223. doi: 10.1038/ng1669
  32. 32. Box GEP, Cox DR (1964) An Analysis of Transformations. J R Stat Soc B 26: 211–252.
  33. 33. Chen MH, Yang Q (2010) GWAF: an R package for genome-wide association analyses with family data. Bioinformatics 26: 580–581. doi: 10.1093/bioinformatics/btp710
  34. 34. Wallin A, Di Giuseppe D, Orsini N, Patel PS, Forouhi NG, et al. (2012) Fish consumption, dietary long-chain n-3 fatty acids, and risk of type 2 diabetes: systematic review and meta-analysis of prospective studies. Diabetes Care 35: 918–929. doi: 10.2337/dc11-1631
  35. 35. Xun P, He K (2012) Fish Consumption and Incidence of Diabetes: Meta-analysis of data from 438,000 individuals in 12 independent prospective cohorts with an average 11-year follow-up. Diabetes Care 35: 930–938. doi: 10.2337/dc11-1869
  36. 36. Dedoussis GV, Kaliora AC, Panagiotakos DB (2007) Genes, diet and type 2 diabetes mellitus: a review. Rev Diabet Stud 4: 13–24. doi: 10.1900/rds.2007.4.13
  37. 37. Watanabe RM (2010) The genetics of insulin resistance: Where's Waldo? Curr Diab Rep 10: 476–484. doi: 10.1007/s11892-010-0143-1
  38. 38. Marin C, Perez-Jimenez F, Gomez P, Delgado J, Paniagua JA, et al. (2005) The Ala54Thr polymorphism of the fatty acid-binding protein 2 gene is associated with a change in insulin sensitivity after a change in the type of dietary fat. Am J Clin Nutr 82: 196–200.
  39. 39. Corella D, Qi L, Tai ES, Deurenberg-Yap M, Tan CE, et al. (2006) Perilipin gene variation determines higher susceptibility to insulin resistance in Asian women when consuming a high-saturated fat, low-carbohydrate diet. Diabetes Care 29: 1313–1319. doi: 10.2337/dc06-0045
  40. 40. Smith CE, Arnett DK, Corella D, Tsai MY, Lai CQ, et al. (2010) Perilipin polymorphism interacts with saturated fat and carbohydrates to modulate insulin resistance. Nutr Metab Cardiovasc Dis 22: 449–455. doi: 10.1016/j.numecd.2010.09.003
  41. 41. Marin C, Perez-Martinez P, Delgado-Lista J, Gomez P, Rodriguez F, et al. (2011) The insulin sensitivity response is determined by the interaction between the G972R polymorphism of the insulin receptor substrate 1 gene and dietary fat. Mol Nutr Food Res 55: 328–335. doi: 10.1002/mnfr.201000235
  42. 42. Richardson K, Lai CQ, Parnell LD, Lee YC, Ordovas JM (2011) A Genome-wide survey for SNPs altering microRNA seed sites identifies functional candidates in GWAS. BMC Genomics 12: 504. doi: 10.1186/1471-2164-12-504
  43. 43. Liao W, Nguyen MT, Yoshizaki T, Favelyukis S, Patsouris D, et al. (2007) Suppression of PPAR-gamma attenuates insulin-stimulated glucose uptake by affecting both GLUT1 and GLUT4 in 3T3-L1 adipocytes. Am J Physiol Endocrinol Metab 293: E219–227. doi: 10.1152/ajpendo.00695.2006
  44. 44. Picard F, Auwerx J (2002) PPAR(gamma) and glucose homeostasis. Annu Rev Nutr 22: 167–197. doi: 10.1146/annurev.nutr.22.010402.102808
  45. 45. Wang RX, Chai Q, Lu T, Lee HC (2011) Activation of vascular BK channels by docosahexaenoic acid is dependent on cytochrome P450 epoxygenase activity. Cardiovasc Res 90: 344–352. doi: 10.1093/cvr/cvq411
  46. 46. Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, et al. (2010) A method and server for predicting damaging missense mutations. Nat Methods 7: 248–249. doi: 10.1038/nmeth0410-248
  47. 47. Dupuis J, Langenberg C, Prokopenko I, Saxena R, Soranzo N, et al. (2010) New genetic loci implicated in fasting glucose homeostasis and their impact on type 2 diabetes risk. Nat Genet 42: 105–116.
  48. 48. Hu S, Labuda MZ, Pandolfo M, Goss GG, McDermid HE, et al. (2003) Variants of the KCNMB3 regulatory subunit of maxi BK channels affect channel inactivation. Physiol Genomics 15: 191–198.