AJCN. First published ahead of print October 14, 2009 as doi: 10.3945/ajcn.2009.28101.
Complement component 3 polymorphisms interact with polyunsaturated fatty acids to modulate risk of metabolic syndrome1–3 Catherine M Phillips, Louisa Goumidi, Sandrine Bertrais, Jane F Ferguson, Martyn R Field, Enda D Kelly, Gina M Peloso, L Adrienne Cupples, Jian Shen, Jose M Ordovas, Ross McManus, Serge Hercberg, Henri Portugal, Denis Lairon, Richard Planells, and Helen M Roche ABSTRACT Background: Complement component 3 (C3) is a novel determinant of the metabolic syndrome (MetS). Gene-nutrient interactions with dietary fat may affect MetS risk. Objectives: The objectives were to determine the relation between C3 polymorphisms and MetS and whether interaction with plasma polyunsaturated fatty acids (PUFAs), a biomarker of dietary PUFA, modulate this relation. Design: C3 polymorphisms (rs11569562, rs2250656, rs1047286, rs2230199, rs8107911, rs344548, rs344550, rs2241393, rs7257062, rs163913, and rs2230204), biochemical measurements, and plasma fatty acids were measured in the LIPGENE–SUpplementation en VItamines et Mine´raux AntioXydants (SU.VI.MAX) study in MetS cases and matched controls (n = 1754). Results: Two single nucleotide polymorphisms were associated with MetS. rs11569562 GG homozygotes had decreased MetS risk compared with the minor A allele carriers [odds ratio (OR): 0.53; 95% CI: 0.35, 0.82; P = 0.009], which was augmented by high plasma PUFA status (OR: 0.32; 95% CI: 0.11, 0.93; P = 0.04). GG homozygotes had lower C3 concentrations than those in AA homozygotes (P = 0.03) and decreased risk of hypertriglyceridemia compared with A allele carriers (OR: 0.54; 95% CI: 0.34, 0.92; P = 0.02), which was further ameliorated by an increase in long-chain n–3 (omega-3) PUFAs (OR: 0.46; 95% CI: 0.22, 0.97; P = 0.04) or a decrease in n–6 PUFAs (OR: 0.32; CI: 0.16, 0.62; P = 0.002). rs2250656 AA homozygotes had increased MetS risk relative to the minor G allele carriers (OR: 1.78; CI: 1.19, 2.70; P = 0.02), which was exacerbated by low n–6 PUFA status (OR: 2.20; CI: 1.09, 4.55; P = 0.03). Conclusion: Plasma PUFAs may modulate the susceptibility to MetS that is conferred by C3 polymorphisms, which suggests novel genenutrient interactions. This trial was registered at clinicaltrials.gov as NCT00272428. Am J Clin Nutr doi: 10.3945/ajcn.2009.28101.
INTRODUCTION
The metabolic syndrome (MetS) is a common, multicomponent condition characterized by insulin resistance, dyslipidemia, abdominal obesity, and hypertension that is associated with an increased risk of type 2 diabetes mellitus, cardiovascular disease, and atherosclerosis (1). The current global epidemic in the incidence of MetS and type 2 diabetes mellitus is an important illustration of the interaction between environmental and genetic factors and diet-related polygenic disorders. Chronic low-grade
inflammation plays a role in the pathogenesis of insulin resistance (2, 3). Elevated concentrations of complement component 3 (C3), a protein with a central role in the innate immune system, have been associated with insulin resistance, diabetes, and MetS. On activation, C3 is converted to its components, including acylation-stimulating protein (ASP), a key player in nonesterified fatty acid (NEFA) transport into adipocytes (4). A recent prospective population-based study identified C3 as a risk factor for the development of type 2 diabetes mellitus (5). C3 has also been positively associated with insulin resistance, obesity, fasting and postprandial triacylglycerol concentrations, hypertension, and cardiovascular disease (6, 7). More recently, a significant dose relation between C3 concentrations and the number of MetS components has been reported (8). Furthermore, after postprandial triacylglycerol, fasting C3 concentrations were the second most important determinant of MetS (8). Environmental factors, in particular dietary and plasma fatty acid composition, may alter the risk of MetS (9–12). In vitro and animal studies report insulin-sensitizing effects of long-chain n–3 polyunsaturated fatty acids (LC n–3 PUFAs) (13, 14). Epidemiologic studies report antiinflammatory effects of dietary fish, fish oil, and/or LC n–3 PUFA intake (15, 16). However, the findings of intervention trials to confirm the functional effects of dietary PUFAs are mixed (17–21), a phenomenon that might be 1 From the Nutrigenomics Research Group, UCD School of Public Health and Population Science, UCD Conway Institute, University College Dublin, Dublin, Ireland (CMP, JFF, and HMR); INSERM 476, Lipid Nutrients and Prevention of Metabolic Diseases, INRA 1260, Faculte´ de Me´decine, Universite´ de la Me´diterrane´e, Marseille, France (LG, HP, RP, and DL); INSERM U557, INRA:CNAM, Universite´ Paris 13, Bobigny, France (SB and SH); the Hitachi Dublin Laboratory, Dublin, Ireland (MRF and EDK); Boston University School of Public Health, Boston, MA (GMP and LAC); the Nutrition and Genomics Laboratory, Jean Mayer US Department of Agriculture Human Nutrition Research Center on Aging, Tufts University, Boston, MA (JS and JMO); and the Institute of Molecular Medicine, Trinity College Dublin, Dublin, Ireland (RM). 2 Supported by the European Commission, Framework Programme 6 (LIPGENE) (contract no. FOOD-CT-2003-505944). 3 Address correspondence to HM Roche, Nutrigenomics Research Group, UCD School of Public Health and Population Science, UCD Conway Institute, University College Dublin, Belfield, Dublin 4, Ireland. E-mail: helen.
[email protected]. Received May 20, 2009. Accepted for publication September 23, 2009. doi: 10.3945/ajcn.2009.28101.
Am J Clin Nutr doi: 10.3945/ajcn.2009.28101. Printed in USA. Ó 2009 American Society for Nutrition
Copyright (C) 2009 by the American Society for Nutrition
1 of 9
2 of 9
PHILLIPS ET AL
due to an interaction between an individual’s genetic background and dietary fat exposure, which affects his or her risk of MetS (22, 23). Fasting and postprandial triacylglycerol metabolism is disturbed in MetS (24). The postprandial C3 increment is related to triacylglycerol and NEFA metabolism (6), and chylomicrons are the strongest stimulators of adipocyte C3 production (25). It is not known whether plasma fatty acid composition directly influences C3 activation. However, PUFAs are ligands of the farnesoid X receptor (26), a nuclear receptor that regulates C3 expression (27). Although C3 clearly plays an important role relevant to both inflammation and lipid metabolism in MetS, no studies to date have investigated the association between genetic variants of C3 and those traits in this condition. Therefore, this study investigated the potential relation between common genetic polymorphisms of C3 and the MetS phenotype and whether this is modulated by plasma PUFAs, biomarkers of habitual dietary PUFA intake.
SUBJECTS AND METHODS
Subjects, MetS classification, and study design This study is part of a prospective case-control candidate gene study of LIPGENE, a European Union Sixth Framework Program Integrated Project, entitled “Diet, genomics and the metabolic syndrome: an integrated nutrition, agro-food, social and economic analysis.” Subjects were selected from an existing French SUpplementation en VItamines et Mine´raux AntioXydants (SU.VI.MAX) cohort, which included 13,000 subjects studied over 7.5 y, from 1994 to 2002 (28). The LIPGENE-SU.VI.MAX study is a nested case-control study of MetS, which consisted of women aged 35–60 y and men aged 45–60 y recruited from SU.VI.MAX. Additional ethical approval from the ethical committee of the Paris-Cochin Hospital included an additional clause (no. Am 2840–12-706) to perform the biochemical analysis and genetic analysis required for the LIPGENE study. LIPGENE subjects were informed of the study objectives and signed a consent form. Baseline and 7.5-y follow-up data, including plasma lipid profiles and full clinical examination records, were made available to LIPGENE. These data were used to identify cases, individuals who developed 3 characteristics of MetS over the 7.5-y followup period and control subjects. MetS cases were selected according to the National Cholesterol Education Program’s Adult Treatment Panel III criteria for MetS. MetS cases were required to fulfill 3 of the following 5 criteria: increased waist circumference (.94 cm for men or .80 cm for women), increased fasting blood glucose (5.5 mmol/L or treatment of diabetes), increased triacylglycerol (1.5 mmol/L or treatment of dyslipidemia), decreased HDL cholesterol (,1.04 mmol/L for men or ,1.29 mmol/L for women), and increased systolic/diastolic blood pressure (130/85 mm Hg or antihypertensive treatment). Cases were defined as men and women with 3 abnormalities, and controls were defined as men and women with no abnormalities or men with 1 abnormality. Cases and controls (n = 1754) were matched according to age (65 y) and sex. Biochemical analysis Fasting glucose, triacylglycerol, HDL cholesterol, and total cholesterol were measured as described previously (28). Insulin
and C-peptide were measured with the use of electrochemiluminescence immunoassays (Roche Diagnostics, Meylan, France). NEFAs and LDL cholesterol were measured with the use of enzymatic colorimetric methods (Randox Laboratories, United Kingdom, and Roche Diagnostics). Total plasma C3 and C-reactive protein (CRP) were measured on a Dade Behring BN II nephelometer (Dade Behring Diagnostics, Marburg, Germany). Homeostasis model assessment (HOMA), a measure of insulin resistance, was calculated as [(fasting plasma glucose · fasting serum insulin)/22.5] (29). The Quantitative Insulin-sensitivity ChecK Index (QUICKI), a measure of insulin sensitivity, was calculated as [1/(log fasting insulin + log fasting glucose + log fasting NEFA)] (30). Plasma fatty acid composition Plasma fatty acid composition (14:0 to 22:6n–3) was determined as a biomarker of habitual dietary fat intake. Fatty acids were extracted from plasma and transmethylated with boron trifluoride in methanol. Fatty acid methyl esters were analyzed by gas chromatography on a Perkin-Elmer Autosystem XL (PerkinElmer, Paris, France) and a Shimadzu GC2010 (Shimadzu, Kyoto, Japan) with random distribution of samples between machines. Heptadecanoic acid was used as an internal standard. Hydrogen and helium were used as carrier gases with the PerkinElmer Autosystem and Shimadzu GC2010, respectively. Carrier gas pressure was 0.8 psig. Injector and flame ionization detector temperatures were 250°C. The following temperature program was used: 215°C (43 min) and 10°C/min to 260°C (2.5 min). Fatty acids were identified by the comparison of the relative retention times of plasma fatty acid methyl esters with fatty acid methyl esters standards (SUPELCO, Saint Quentin Fallavier, France). Fatty acid mass was measured as a relative percentage of the total quantified fatty acids. PUFAs were measured as follows: PUFAs (n–3 + n–6 PUFAs), n–3 PUFAs (18:3n–3, 18:4, 20:4n–3, 20:5, 22:5, and 22:6), n–6 PUFAs (18:2, 18:3n–6, 20:3, 20:4n–6, and 22:4), and LC n–3 PUFAs (20:5 and 22:6). Saturated fatty acids (14:0, 16:0, and 18:0) and monounsaturated fatty acids (16:1n–7, 18:1n–9, and 20:1n–9) were also measured to complete the fatty acid profile. DNA extraction and genotyping DNA extraction from buffy coats was performed with the use of the Puregene protocol for DNA extraction, and samples were processed with the AutoPure LS automated system (Gentra Systems Inc, Minneapolis, MN). Low-yielding samples (,10 ng) were subjected to whole-genome amplification with the use of the REPLI-g kit (Qiagen Ltd, West Sussex, United Kingdom). C3 genotype data from HapMap, version 1.1 (www.hapmap. org), and Perlegen (www.perlegen.com) were uploaded into HITAGENE, a web-based combined database and genetic analysis software suite developed by Hitachi Dublin Laboratory (Dublin, Ireland). Haplotype frequencies were estimated by implementation of the expectation maximization algorithm. With the use of a 5% cutoff for individual haplotype frequency and .70% for the sum of all haplotype frequencies, haplotypetagged single nucleotide polymorphisms (SNPs) were identified with the use of SNP tagger (www.broad.mit.edu/mpg/tagger/ server.html). Together with SNPs from the literature (rs1047286
C3 GENOTYPE AND PUFAs MODULATE METABOLIC SYNDROME RISK
and rs2230199), 11 C3 SNPs (rs11569562, rs2250656, rs1047286, rs2230199, rs8107911, rs344548, rs344550, rs2241393, rs7257062, rs163913, and rs2230204) were genotyped as part of the entire genotyping component of the LIPGENE study by Illumina Inc (San Diego, CA), with the use of the Golden Gate Assay on a BeadStation 500G genotyping system (Illumina Inc). We achieved an average genotyping success rate of 99% and a call rate of 99%. Departure of genotype distributions from Hardy-Weinberg equilibrium was assessed with the use of chi-square tests in HITAGENE. All SNPs were in Hardy-Weinberg equilibrium (P . 0.01). Linkage disequilibrium between SNPs was assessed with the use of Golden Helix software (Golden Helix Inc, Bozeman, MT; available at www.goldenhelix.com). One pair of SNPs, rs1047286 and rs2230199, were in significant linkage disequilibrium (r2 = 0.85, D# = 0.96), as shown in Figure 1.
3 of 9
peated with the use of the median concentration of control subjects to dichotomize fatty acids and examine associations below and above the fatty acid median. Generalized estimating equation linear regression (31) was used to investigate interactions between continuous MetS phenotypes and PUFAs. Potential confounding factors used in the adjusted multivariate analysis included age, sex, smoking status, physical activity, alcohol and energy intake, and use of medications including lipid-lowering, hypertension, and diabetes treatments. A P value ,0.05 was considered significant. To account for multiple testing for associations and interactions, false discovery rates (FDRs) (32) were also computed, and we report FDR-adjusted P values that were calculated separately for each test. FDRs 0.05 were considered significant.
RESULTS
Statistical analysis Statistical analysis was performed with the use of SAS for Windows, version 9.0 (SAS Institute Inc, Cary, NC). Data are expressed as means 6 SEMs. After skewness and kurtosis were checked for, glucose, insulin, NEFA, triacylglycerol, QUICKI, and HOMA were normalized by logarithmic transformation. Genotype frequencies were compared between cases and controls in HITAGENE with the use of Fisher’s exact test. Conditional logistic regression determined associations between genotypes and MetS and its risk phenotypes (high triacylglycerol, low HDL cholesterol, high C3 concentrations, high HOMA, low QUICKI, fasting hyperglycemia, abdominal obesity, and high blood pressure). Cutoffs for these MetS risk phenotypes were determined by MetS criteria. HOMA, QUICKI, and C3 values were dichotomized on the basis of control-subject medians. Three genotype groups were considered first, to check different inherent models (additive, dominant, and recessive). Where a dominant or recessive effect existed, analysis was repeated with a comparison between carriers and noncarriers of that particular allele. To determine modulation by plasma fatty acids, logistic analyses were re-
The C3 polymorphisms studied are detailed in Table 1. Genotype frequencies were different between cases and controls for rs11569562 (P = 0.02), rs2250656 (P = 0.01), rs1047286 (P = 0.006), and rs2230199 (P = 0.007). Further investigation revealed that these differences were even more significant in the recessive models (P = 0.02, P = 0.005, P = 0.002, and P = 0.005, respectively), and all remained significant after multiple testing was corrected for (FDRs: 0.046, 0.019, 0.019, and 0.019, respectively). However, in the adjusted multivariate model, only 2 of these SNPs (rs11569562 and rs2250656) were still shown to be associated with MetS (FDRs , 0.05). The clinical characteristics, inflammatory markers, and plasma fatty acid profiles of the subjects according to the rs11569562 genotype are presented in Table 2. In terms of their phenotype, the A allele carriers had lower HDL-cholesterol and higher triacylglycerol, insulin, and C-peptide concentrations (P , 0.05) than did the GG homozygotes. AA homozygotes also had raised inflammatory status, with elevated C3 and CRP concentrations, compared with the GG homozygotes (P , 0.05). Age, sex distribution, alcohol intake, and medication use were not different between groups. Plasma fatty acid composition showed some minor differences. AA homozygotes had slightly higher
FIGURE 1. Pairwise linkage disequilibrium between the C3 single nucleotide polymorphisms (SNPs) estimated as the correlation coefficient (r2), the strength of which is indicated by the color scale: black, absolute linkage disequilibrium; dark gray, intermediate linkage disequilibrium; and pale gray, lack of any linkage. One pair of SNPs, rs1047286 and rs2230199, were in significant linkage disequilibrium (r2 = 0.85, D´ = 0.96) (Golden Helix software; Golden Helix Inc, Bozeman, MT; www.goldenhelix.com).
4 of 9
PHILLIPS ET AL
TABLE 1 Investigation of C3 polymorphisms and a comparison of genotype frequencies between metabolic syndrome (MetS) cases and controls (n = 1754)1 P values MetS cases
Controls
Locus
Location
Genotype
Occurrence
Frequency
Occurrence
Frequency
Fisher’s
Dominant model
Recessive model
rs11569562
Intron 38
0.079
0.005
0.006
0.031
0.004
rs2230199
Exon 3 (R102G)
0.007
0.144
0.002
rs8107911
Intron 17
0.058
0.020
0.296
rs344548
Intron 29
0.086
0.033
1.000
rs344550
Intron 33
0.664
0.809
0.374
rs2241393
Intron 29
0.251
0.880
0.137
rs7257062
Intron 29
0.686
0.520
0.797
rs163913
5# UTR
0.095
0.0509
0.220
rs2230204
Exon 14 (V564V)
0.2372 0.4744 0.2884 0.4953 0.4052 0.0995 0.0285 0.2800 0.6916 0.0344 0.2969 0.6686 0.5917 0.3560 0.0524 0.6992 0.2759 0.0250 0.1012 0.4548 0.4440 0.3903 0.4840 0.1257 0.1812 0.4696 0.3492 0.6326 0.3282 0.0392 0.4939 0.4203 0.0858
0.011
Exon 9 (P314L)
no. 199 399 243 417 341 84 24 235 582 29 250 562 498 299 44 588 232 21 85 382 373 328 407 106 152 395 294 532 276 33 415 353 72
0.016
rs1047286
0.2267 0.5370 0.2363 0.5382 0.3998 0.0621 0.0489 0.3258 0.6253 0.0490 0.3548 0.5962 0.6468 0.3126 0.0406 0.7464 0.2297 0.0239 0.1055 0.4724 0.4221 0.3945 0.4544 0.1511 0.1685 0.4886 0.3429 0.6790 0.2936 0.0274 0.5048 0.4207 0.0745
0.643
Intron 2
no. 191 452 199 453 336 52 41 274 526 41 298 501 544 263 34 628 193 20 89 397 355 332 382 127 142 411 288 571 247 23 425 354 63
0.020
rs2250656
A/A A/G G/G A/A A/G G/G A/A A/G G/G C/C C/G G/G A/A A/G G/G C/C C/G G/G C/C C/G G/G C/C C/G G/G A/A A/G G/G A/A A/G G/G C/C C/T T/T
0.685
0.657
0.415
1 UTR, untranslated region. Genotype frequencies were compared between cases and controls by using HITAGENE (Hitachi Dublin Laboratory, Dublin, Ireland) with Fisher’s exact test.
total PUFAs (P , 0.05) than did the GG homozygotes, which reflects their increased n–6 PUFAs (P , 0.05). Logistic regression analysis identified a genotype effect of rs11569562 on MetS risk (P = 0.015; FDR: 0.03), particularly when GG homozygotes were compared with minor A allele carriers [odds ratio (OR): 0.53; 95% CI: 0.35, 0.82; P = 0.009; FDR: 0.02]. The GG homozygotes also had approximately half the risk of having high plasma triacylglycerol concentrations compared with A allele carriers (OR: 0.54; 95% CI: 0.34, 0.92; P = 0.02; FDR: 0.04). When total PUFAs were stratified according to the control median among all subjects with the top 50th percentile of PUFA, GG homozygotes had an even lower MetS risk than that of A allele carriers (OR: 0.32; 95% CI: 0.11, 0.93; P = 0.04; FDR: 0.13). MetS risk was not altered when n–6, n–3, or LC n–3 PUFAs were analyzed separately. Interestingly, the risk of high triacylglycerol was modified by n–6 and LC n–3 PUFA status; GG homozygotes with the lowest n–6 PUFAs and the highest LC n–3 PUFA concentrations had further decreased risk compared with the A allele carriers (OR: 0.32; 95% CI: 0.16, 0.62;
P = 0.002; FDR: 0.006; and OR: 0.46; 95% CI: 0.22, 0.97; P = 0.04; FDR: 0.15, respectively). Interaction analysis confirmed the latter gene-nutrient interaction, whereby higher LC n–3 PUFA status was predictive of decreased plasma triacylglycerol concentrations in GG homozygotes (P = 0.02; FDR: 0.048), as shown in Figure 2. Clinical characteristics and inflammatory and PUFA profiles according to rs2250656 genotype are also shown in Table 2. Major A allele carriers displayed a classic MetS profile: they had higher triacylglycerol and lower HDL cholesterol concentrations (P , 0.05), higher body mass index (in kg/m2) (P , 0.05) and abdominal obesity (P = 0.006), and were less insulin sensitive (P = 0.035) than were the GG homozygotes. A allele carriers also had elevated inflammatory status with higher plasma C3 (P = 0.005) and CRP concentrations (P = 0.02) than those of the GG homozygotes. Age, sex distribution, alcohol intake, and medication use were not different between groups; nor were there any differences in PUFA composition across genotypes. Logistic regression analyses revealed a genotype effect for
5 of 9
C3 GENOTYPE AND PUFAs MODULATE METABOLIC SYNDROME RISK TABLE 2 Clinical characteristics, inflammatory markers, and plasma fatty acid profiles according to rs11569562 and rs2250656 genotypes1 rs11569562 AA
AG
rs2250656 GG
AA + AG
AA
AG
GG
AA + AG
n 390 851 441 1241 869 677 136 1546 Cases/controls (%/%) 49/512 53/473 45/55 52/482 52/484 50/502 38/62 51/494 M/F (%/%) 59/41 61/39 60/40 60/40 59/41 62/38 57/43 61/39 58.14 6 0.20 58.20 6 0.47 58.15 6 0.13 Age (y) 58.49 6 0.265 57.83 6 0.18 58.49 6 0.25 58.04 6 0.15 58.18 6 0.18 25.91 6 0.22 26.13 6 0.15 25.85 6 0.20 26.06 6 0.13 26.00 6 0.142 26.14 6 0.172 25.20 6 0.41 26.06 6 0.112 BMI (kg/m2) Waist (cm) 87.77 6 0.64 88.77 6 0.44 87.44 6 0.62 88.46 6 0.36 88.21 6 0.432 88.68 6 0.494 85.27 6 1.11 88.42 6 0.324 SBP (mm Hg) 130.80 6 0.76 131.39 6 0.55 130.61 6 0.75 131.2 6 0.44 131.64 6 0.53 130.60 6 0.61 129.21 6 1.31 131.19 6 0.40 DBP (mm Hg) 82.10 6 0.47 81.96 6 0.32 82.46 6 0.44 82.00 6 0.26 82.57 6 0.31 81.70 6 0.36 81.22 6 0.76 82.19 6 0.24 Total cholesterol 5.75 6 0.05 5.69 6 0.03 5.72 6 0.04 5.71 6 0.03 5.71 6 0.03 5.70 6 0.03 5.71 6 0.08 5.71 6 0.02 (mmol/L) 1.51 6 0.02 1.46 6 0.012 1.46 6 0.012 1.47 6 0.012 1.55 6 0.04 1.47 6 0.012 HDL-C (mmol/L) 1.46 6 0.02 1.46 6 0.012 LDL-C (mmol/L) 3.59 6 0.06 3.51 6 0.04 3.52 6 0.05 3.54 6 0.03 3.55 6 0.04 3.52 6 0.04 3.41 6 0.11 3.54 6 0.03 1.20 6 0.03 1.29 6 0.022 1.29 6 0.022 1.26 6 0.03 1.14 6 0.05 1.27 6 0.022 Triacylglycerol 1.27 6 0.04 1.30 6 0.022 (mmol/L) NEFA (mmol/L) 0.90 6 0.04 0.93 6 0.03 0.89 6 0.04 0.92 6 0.02 0.90 6 0.03 0.95 6 0.03 0.80 6 0.06 0.92 6 0.02 0.73 6 0.02 0.77 6 0.012 0.77 6 0.01 0.75 6 0.01 0.72 6 0.03 0.76 6 0.01 C-peptide (nmol/L) 0.77 6 0.02 0.77 6 0.012 Glucose (mmol/L) 5.27 6 0.05 5.26 6 0.03 5.24 6 0.05 5.26 6 0.03 5.26 6 0.03 5.26 6 0.04 5.17 6 0.0 5.26 6 0.03 6.88 6 0.26 7.59 6 0.172 7.60 6 0.20 7.22 6 0.21 6.84 6 0.55 7.43 6 0.15 Insulin (mU/L) 7.46 6 0.25 7.64 6 0.222 HOMA 1.82 6 0.08 1.90 6 0.07 1.69 6 0.07 1.87 6 0.06 1.87 6 0.06 1.79 6 0.07 1.69 6 0.18 1.83 6 0.05 0.34 6 0.01 0.36 6 0.02 0.33 6 0.022 QUICKI 0.33 6 0.01 0.33 6 0.01 0.35 6 0.01 0.33 6 0.01 0.33 6 0.0052 2 4 2 C3 (g/L) 1.58 6 0.03 1.52 6 0.02 1.48 6 0.03 1.54 6 0.02 1.54 6 0.02 1.48 6 0.02 1.37 6 0.05 1.52 6 0.024 CRP (mg/L) 2.56 6 0.272 2.17 6 0.13 1.92 6 0.17 2.30 6 0.12 1.87 6 0.063 1.87 6 0.062 1.39 6 0.20 2.26 6 0.112 Taking lipid-lowering 20.51 19.98 17.78 20.15 20.02 19.35 19.71 19.73 medication (%) Taking antidiabetic 2.56 3.17 3.17 2.98 2.76 3.25 3.68 2.98 medication (%) Taking hypertensive 20.77 22.56 19.05 22.00 20.94 21.71 19.85 21.28 medication (%) Alcohol intake 6.03 6 0.36 6.34 6 0.27 6.22 6 0.36 6.19 6 0.22 6.37 6 0.25 6.50 6 0.29 6.26 6 0.60 6.43 6 0.20 (% of energy) Fatty acids (%) 43.62 6 0.23 43.60 6 0.16 43.98 6 0.52 Total PUFA 44.11 6 0.322 43.55 6 0.20 43.27 6 0.30 43.73 6 0.17 43.58 6 0.21 39.19 6 0.21 39.21 6 0.15 39.36 6 0.48 n–6 PUFA 39.73 6 0.292 39.12 6 0.19 38.86 6 0.28 39.31 6 0.16 39.23 6 0.20 n–3 PUFA 4.38 6 0.10 4.43 6 0.06 4.38 6 0.09 4.42 6 0.05 4.35 6 0.06 4.412 6 0.07 4.38 6 0.05 4.63 6 0.15 LC n–3 PUFA 3.31 6 0.08 3.37 6 0.06 3.31 6 0.08 3.35 6 0.05 3.32 6 0.06 3.30 6 0.06 3.32 6 0.04 3.56 6 0.14 1 SBP, systolic blood pressure; DBP, diastolic blood pressure; HDL-C, HDL cholesterol; LDL-C, LDL cholesterol; TAG, triacylglycerol; NEFA, nonesterified fatty acid; HOMA, homeostasis model assessment; QUICKI, quantitative insulin-sensitivity check index; C3, complement component 3; CRP, C-reactive protein; PUFA, polyunsaturated fatty acid; LC, long-chain. 2–4 Significantly different from GG homozygotes: 2P , 0.05, 3P , 0.01, 4P , 0.005 (ANOVA). 5 Mean 6 SEM (all such values).
rs2250656, with major A allele carriers having .2-fold higher MetS risk than GG homozygotes (OR: 2.59; 95% CI: 1.12, 5.80; P = 0.032; FDR: 0.05). In particular, the AA subjects had a higher MetS risk compared with carriers of the minor G allele (OR: 1.78; 95% CI: 1.19, 2.70; P = 0.02; FDR: 0.042), and especially compared with GG subjects (OR: 2.70; CI: 1.20, 6.10; P = 0.002; FDR: 0.01), which resulted in a significant P value for the additive model (P = 0.005; FDR: 0.008). We investigated whether PUFA status modified MetS risk and showed that AA homozygotes with the lowest 50th percentile of n–6 and total PUFAs further increased their MetS risk by ’30% over that of the minor G allele carriers (OR: 2.20; 95% CI: 1.09, 4.55; P = 0.03; and OR: 2.27; 95% CI: 1.02, 4.76; P = 0.04, respectively). Adjustment for multiple testing yielded FDRs of 0.08. C3 concentrations were also influenced by PUFA composition. As plasma total PUFAs (P = 0.03), particularly plasma n–6 PUFA concentrations (P = 0.02), increased, plasma C3 concentrations
were predicted to decrease in G allele carriers, with the opposite effect observed in AA homozygotes (Figure 3). Adjustment for multiple testing yielded FDRs of ,0.05. n–3 and LC n–3 PUFA constituents did not appear to modulate this gene-nutrient interaction. Homogeneity of the genetic effects on MetS risk was assessed by stratification according to sex. The MetS associations were derived primarily from female subjects for both rs11569562 (OR: 0.37; 95% CI: 0.10, 0.88; P = 0.02, GG homozygotes relative to the A allele carriers) and rs2250656 (OR: 2.42; 95% CI: 1.13, 5.40; P = 0.03, A allele carriers compared with GG homozygotes). Adjustment for multiple testing yielded FDRs of ,0.05. Although the effects were in the same direction in the male subjects, they were not significant. Finally, plasma C3 and CRP concentrations were strongly correlated (r2 = 0.52, P , 0.0001) and exhibited a dose relation with the number of components of MetS (P for trend: ,0.0001) (Figure 4).
6 of 9
PHILLIPS ET AL
FIGURE 2. The gene-nutrient interaction between plasma long-chain n–3 polyunsaturated fatty acids (LC n–3 PUFA) and rs11569562 genotype on plasma triacylglycerol (TAG) concentrations in all subjects. Increasing LC n–3 PUFA status was predictive of decreased plasma TAG concentrations in GG homozygotes (closed circles; n = 441) but not in either of the A allele carriers (open circles; n = 1241) (P = 0.02; FDR: 0.048). A generalized estimating equation linear regression (31) model that contained LC n–3 PUFAs, genotypes, their interaction term, and potential confounders was used for the analysis.
DISCUSSION
In this study we observed that common genetic variants at the C3 locus were associated with risk of MetS and its phenotypes, including dyslipidemia, abdominal obesity, and insulin sensitivity. Plasma PUFA composition and status appeared to modulate these genetic influences. Whereas plasma fatty acid composition reflects the combination of dietary fat consumption and endogenous de novo fatty acid biosynthesis and metabolism, such gene-nutrient interactions suggest that dietary fat may modify genetic susceptibility to MetS, which warrants further investigation. This is the first case-control study to report an association between C3 polymorphisms and MetS risk. We noted sex differences for some of the associations reported in this study; although the effects were in the same direction in the male subjects, they did not reach statistical significance, which may reflect lack of statistical power. Higher C3 concentrations have been reported in women (33), and the C3/ASP pathway is thought to be more active in subcutaneous adipose tissue (34), which is more abundant in women than in men. Thus, our findings may reflect sex-specific differences in intra- and extraperitoneal adipose tissue mass.
The protective rs11569562 GG genotype may be accounted for by its lower C3 concentrations and decreased risk of high triacylglycerol compared with that in A allele carriers. MetS risk was subject to a significant effect modification by PUFAs, with the greatest protection from MetS achieved by GG homozygotes with the highest PUFA status. Likewise, GG homozygotes with the highest LC n–3 PUFAs had the lowest risk of hypertriglyceridemia. One interpretation could be that individuals who are genetically resilient to MetS and high triacylglycerol (27% of this population) are most sensitive to PUFAs, such that either high LC n–3 or low n–6 PUFA concentrations further ameliorate their protection. The increased MetS risk associated with the rs2250656 A allele may be explained by its classic MetS profile and elevated inflammatory status. Interestingly, plasma PUFAs modified MetS risk, whereby the combination of carrying 2 A alleles and having the lowest n–6 or total PUFA status exacerbated MetS risk, which suggests that these individuals, who represent approximately half of the population and who are genetically predisposed to MetS, are also more sensitive to PUFAs. Gene-nutrient interactions also modulated C3 concentrations, whereby increasing plasma n–6 PUFAs was predicted to decrease C3 concentrations in GG homozygotes,
C3 GENOTYPE AND PUFAs MODULATE METABOLIC SYNDROME RISK
7 of 9
FIGURE 3. The predicted plasma complement component 3 (C3) values by C3 rs2250656 genotypes according to plasma concentrations of n–6 polyunsaturated fatty acids (PUFAs) in all subjects. As plasma n–6 PUFA concentrations increased, plasma C3 concentrations were predicted to decrease in the G allele carriers (closed circles; n = 813), with the opposite effect observed in the AA homozygotes (open circles; n = 869) (P = 0.02, FDR: 0.05) A generalized estimating equation linear regression (31) model that contained n–6 PUFAs, genotypes, their interaction term, and potential confounders was used for the analysis.
with the opposite effect observed in AA homozygotes. To summarize, these 2 C3 gene variants modulate MetS risk in association with plasma PUFA status. Because the mechanisms that underlie these somewhat contradictory interactions between n–3 and n–6 PUFAs and C3 genotype are currently unknown, these results should be taken with some caution and require further investigation. We attempted to replicate our findings in a separate, independent LIPGENE MetS-case-only cohort (n = 464) (35) and report that the “protective” rs11569562 GG genotype was as-
sociated with enhanced insulin sensitivity, compared with that in A allele carriers (QUICKI: 0.65 6 0.06 compared with 0.58 6 0.02; P = 0.035). GG homozygotes were also more responsive to LC n–3 PUFAs. After a 12-wk, low-fat (28% energy), highcomplex-carbohydrate diet intervention supplemented with 1.24 g LC n–3 PUFAs/d, GG homozygotes displayed beneficial changes to their lipid profile (10% decrease in NEFA, P = 0.027; 8% nonsignificant decrease in triacylglycerol, 5% decrease in total cholesterol, P = 0.02; and 17% decrease in LDLcholesterol concentrations, P = 0.002), compared with those in
FIGURE 4. The dose relation between the concentrations of (A) complement component 3 (C3) and (B) C-reactive protein (CRP) and the number of components of the metabolic syndrome (MetS) in the LIPGENE–SUpplementation en VItamines et Mine´raux AntioXydants (SU.VI.MAX) study (n = 1754) (unadjusted P for trend ,0.0001, repeated-measures ANOVA).
8 of 9
PHILLIPS ET AL
A allele carriers. No changes were observed between genotypes when subjects on the same diet received 1 g high-oleic-acidcontrol supplement/d. In addition, the “at risk” rs2250656 A allele carriers had decreased insulin sensitivity as assessed by the Insulin Sensitivity Index (2.65 6 0.11 compared with 3.18 6 0.15, P = 0.003) and increased body mass index (32.82 6 0.28 compared with 31.90 6 0.29, P = 0.037), relative to the GG homozygotes. Genetic influences were modified by LC n–3 PUFA supplementation; the A allele carriers achieved a 35% improvement (P = 0.02) in insulin sensitivity after intervention, whereas no changes were noted between genotypes after oleic acid supplementation. Notwithstanding the study design differences, these data help strengthen our findings; however, functional studies are needed to ascertain their biological significance. Epidemiologic research indicates that n–3 PUFA consumption inversely correlates with biomarkers of inflammation (15, 16). We observed significant but weak inverse correlations between plasma n–3 PUFAs and both C3 and CRP concentrations (data not shown). The fact that PUFAs are also ligands of farnesoid X receptor (26), a nuclear receptor that regulates C3 expression (27), raises the possibility that alteration of C3 expression via modulation of farnesoid X receptor is a potential mechanism by which gene-nutrient interaction of C3 genotype and dietary PUFAs could influence C3 concentrations and thus MetS risk. Although this is speculative, it may be worthy of further investigation to help elucidate the molecular basis of such genenutrient interactions and their effect on markers of inflammation and insulin resistance. n–3 PUFAs are associated with insulin sensitivity (36, 37) and lower triacylglycerol concentrations (38). We observed that individuals with the lowest median of plasma LC n–3 and n–3 PUFAs were more insulin resistant and had higher triacylglycerol concentrations than those with the highest median of these fatty acids (data not shown). The C3/ ASP pathway stimulates triacylglycerol synthesis and glucose transport (4, 39), which may account for these associations. Conflicting reports have failed to establish a clear correlation between C3 and ASP concentrations, and it unknown whether ASP is influenced by C3 polymorphisms. We report elevated C3 concentrations in MetS with certain genotypes that affect C3 concentrations. Furthermore, we report a dose relation between C3 concentrations and the number of MetS abnormalities, which concurs with the literature (5, 7, 8). One is tempted to speculate that in conditions in which insulin sensitivity is impaired, further potential disruption to the C3/ASP pathway that arises from C3 polymorphisms may interfere with the positive feedback mechanism provided by ASP, thereby driving the metabolic phenotype. The potential of dietary fat to attenuate this remains an intriguing possibility that would be worthwhile to address. No functional data on these C3 polymorphisms exist; thus, we can only speculate about the mechanisms that underlie our findings. The intronic location of these SNPs has the potential to affect messenger RNA stability or modulate C3 gene transcriptional activity. It is also possible that these SNPs may be surrogate markers for other functional C3 SNPs in the region. Furthermore, FASTSNP, a functional analysis tool (40), identified rs11569562 as lying in an intronic enhancer sequence homologous to a binding site for the transcription factor AP-1, which plays an important role in the transcriptional regulation of the C3a receptor (C3aR), a prerequisite for C3a, a cleavage product of C3, to participate in inflammation (41). Interestingly,
C3aR knockout mice fed high-fat diets displayed resistance to diet-induced obesity and insulin resistance. Examination of their adipocytes revealed decreased macrophage infiltration and proinflammatory status (42). These data provide evidence that the C3aR is responsive to dietary fat and that the C3aR plays an important role in insulin resistance and obesity. In conclusion, this study provides new data on C3 status, C3 genotype, and plasma PUFAs in MetS. Furthermore, we replicated some of these results in an independent MetS cohort. Further investigation of these novel associations and genenutrient interactions may help improve the therapeutic efficacy of dietary recommendations with a “personalized nutrition” approach, wherein genetic profile may determine choice of dietary therapy to aid responsiveness to dietary fatty acid interventions. We acknowledge V Pirisi, B Gleize, and AM Lorec for handling of plasma biochemical analyses. We also acknowledge the LIPGENE Dietary Intervention Centres [Nutrigenomics Research Group, University College Dublin, Belfield, Dublin 4, Republic of Ireland; Lipid and Atherosclerosis Unit, Reina Sofia University Hospital, School of Medicine, University of Cordoba, CIBER Physiopathology of Obesity and Nutrition (CB06/03), Spain; INSERM 476, Lipid Nutrients and Prevention of Metabolic Diseases; INRA 1260; Universite´ de la Me´diterrane´e, Faculte´ de Me´decine, Marseille, France; Hugh Sinclair Unit of Human Nutrition, University of Reading, Reading, United Kingdom; Department of Nutrition, Institute of Basic Medical Sciences, and Department of Endocrinology, Aker University Hospital, University of Oslo, Oslo, Norway; Department of Human Biology, Nutrition and Toxicology Research Institute Maastricht (NUTRIM), Maastricht, Netherlands; Department of Clinical Biochemistry, Jagiellonian University Medical College, Kopernika, Krakow, Poland; and Department of Public Health and Caring Sciences/Clinical Nutrition and Metabolism, Uppsala University, Sweden] for permission to include the replication data generated from analyzing their independent MetS cohort. The authors’ responsibilities were as follows—CMP: plasma fatty acid composition, DNA extraction, data analysis, and draft of manuscript; LG: DNA extraction; SB: generation of LIPGENE cohort; JFF: plasma fatty acid analysis and DNA extraction; MRF, EDK, and JMO: development of Hitagene software and database for the LIPGENE study; GMP, LAC, JS, and JMO: statistical advice and assistance; SH: access to the SU.VI.MAX cohort for selection of LIPGENE subjects; HP: plasma fatty acid composition determination; DL, RP, RM, and HMR: design and supervision of the study; and all authors: manuscript revision. None of the authors had a financial or personal conflict of interest to disclose.
REFERENCES 1. Moller DE, Kaufman KD. Metabolic syndrome: a clinical and molecular perspective. Annu Rev Med 2005;56:45–62. 2. Spranger J, Kroke A, Mohlig M, et al. Inflammatory cytokines and the risk to develop type 2 diabetes: results of the prospective population-based European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam Study. Diabetes 2003;52:812–7. 3. Hu FB, Meigs JB, Li TY, Rifai N, Manson JE. Inflammatory markers and risk of developing type 2 diabetes in women. Diabetes 2004;53: 693–700. 4. Yasruel Z, Cianflone K, Sniderman AD, Rosenbloom M, Walsh M, Rodriguez MA. Effect of acylation stimulating protein on the triacylglycerol synthetic pathway of human adipose tissue. Lipids 1991; 26:495–9. 5. Engstrom G, Hedblad B, Eriksson KF, Janzon L, Lindgarde F. Complement C3 is a risk factor for the development of diabetes: a populationbased cohort study. Diabetes 2005;54:570–5. 6. Halkes CJ, van Dijk H, de Jaegere PP, et al. Postprandial increase of complement component 3 in normolipidemic patients with coronary artery disease: effects of expanded-dose simvastatin. Arterioscler Thromb Vasc Biol 2001;21:1526–30. 7. Muscari A, Massarelli G, Bastagli L, et al. Relationship of serum C3 to fasting insulin, risk factors and previous ischaemic events in middle-aged men. Eur Heart J 2000;21:1081–90.
C3 GENOTYPE AND PUFAs MODULATE METABOLIC SYNDROME RISK 8. van Oostrom AJ, Alipour A, Plokker TW, Sniderman AD, Cabezas MC. The metabolic syndrome in relation to complement component 3 and postprandial lipemia in patients from an outpatient lipid clinic and healthy volunteers. Atherosclerosis 2007;190:167–73. 9. Kabagambe EK, Tsai MY, Hopkins PN, et al. Erythrocyte fatty acid composition and the metabolic syndrome: a National Heart, Lung, and Blood Institute GOLDN study. Clin Chem 2008;54:154–62. 10. Vessby B. Dietary fat, fatty acid composition in plasma and the metabolic syndrome. Curr Opin Lipidol 2003;14:15–9. 11. Warensjo E, Riserus U, Vessby B. Fatty acid composition of serum lipids predicts the development of the metabolic syndrome in men. Diabetologia 2005;48:1999–2005. 12. Warensjo E, Sundstrom J, Lind L, Vessby B. Factor analysis of fatty acids in serum lipids as a measure of dietary fat quality in relation to the metabolic syndrome in men. Am J Clin Nutr 2006;84:442–8. 13. Storlien LH, Baur LA, Kriketos AD, et al. Dietary fats and insulin action. Diabetologia 1996;39:621–31. 14. Storlien LH, Jenkins AB, Chisholm DJ, Pascoe WS, Khouri S, Kraegen EW. Influence of dietary fat composition on development of insulin resistance in rats. Relationship to muscle triglyceride and omega-3 fatty acids in muscle phospholipid. Diabetes 1991;40: 280–9. 15. Lopez-Garcia E, Schulze MB, Manson JE, et al. Consumption of (n–3) fatty acids is related to plasma biomarkers of inflammation and endothelial activation in women. J Nutr 2004;134:1806–11. 16. Madsen T, Skou HA, Hansen VE, et al. C-reactive protein, dietary n–3 fatty acids, and the extent of coronary artery disease. Am J Cardiol 2001;88:1139–42. 17. Blok WL, Deslypere JP, Demacker PN, et al. Pro- and anti-inflammatory cytokines in healthy volunteers fed various doses of fish oil for 1 year. Eur J Clin Invest 1997;27:1003–8. 18. Hartweg J, Perera R, Montori V, Dinneen S, Neil HA, Farmer A. Omega-3 polyunsaturated fatty acids (PUFA) for type 2 diabetes mellitus. Cochrane Database Syst Rev 2008;CD003205. 19. Lovejoy JC, Smith SR, Champagne CM, et al. Effects of diets enriched in saturated (palmitic), monounsaturated (oleic), or trans (elaidic) fatty acids on insulin sensitivity and substrate oxidation in healthy adults. Diabetes Care 2002;25:1283–8. 20. Luu NT, Madden J, Calder PC, et al. Dietary supplementation with fish oil modifies the ability of human monocytes to induce an inflammatory response. J Nutr 2007;137:2769–74. 21. Mayer-Davis EJ, Monaco JH, Hoen HM, et al. Dietary fat and insulin sensitivity in a triethnic population: the role of obesity. The Insulin Resistance Atherosclerosis Study (IRAS). Am J Clin Nutr 1997;65: 79–87. 22. Phillips C, Lopez-Miranda J, Perez-Jimenez F, McManus R, Roche HM. Genetic and nutrient determinants of the metabolic syndrome. Curr Opin Cardiol 2006;21:185–93. 23. Szabo de Edelenyi F, Goumidi L, Bertrais S, et al. Prediction of the metabolic syndrome status based on dietary and genetic parameters, using Random Forest. Genes Nutr 2008;3:173–6. 24. Lopez-Miranda J, Williams C, Lairon D. Dietary, physiological, genetic and pathological influences on postprandial lipid metabolism. Br J Nutr 2007;98:458–73.
9 of 9
25. Verseyden C, Meijssen S, van Dijk H, Jansen H, Castro Cabezas M. Effects of atorvastatin on fasting and postprandial complement component 3 response in familial combined hyperlipidemia. J Lipid Res 2003;44:2100–8. 26. Zhao A, Yu J, Lew JL, Huang L, Wright SD, Cui J. Polyunsaturated fatty acids are FXR ligands and differentially regulate expression of FXR targets. DNA Cell Biol 2004;23:519–26. 27. Li J, Pircher PC, Schulman IG, Westin SK. Regulation of complement C3 expression by the bile acid receptor FXR. J Biol Chem 2005;280:7427–34. 28. Hercberg S, Galan P, Preziosi P, et al. The SU.VI.MAX Study: a randomized, placebo-controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med 2004;164:2335–42. 29. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28:412–9. 30. Perseghin G, Caumo A, Caloni M, Testolin G, Luzi L. Incorporation of the fasting plasma FFA concentration into QUICKI improves its association with insulin sensitivity in nonobese individuals. J Clin Endocrinol Metab 2001;86:4776–81. 31. Liang KY, Zeger SL. Regression analysis for correlated data. Annu Rev Public Health 1993;14:43–68. 32. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Roy Statist Soc Ser B 1995;57:289–300. 33. Szeplaki G, Prohaszka Z, Duba J, et al. Association of high serum concentration of the third component of complement (C3) with preexisting severe coronary artery disease and new vascular events in women. Atherosclerosis 2004;177:383–9. 34. Maslowska M, Vu H, Phelis S, et al. Plasma acylation stimulating protein, adipsin and lipids in non-obese and obese populations. Eur J Clin Invest 1999;29:679–86. 35. Shaw DI, Tierney AC, McCarthy S, et al. LIPGENE food-exchange model for alteration of dietary fat quantity and quality in free-living participants from eight European countries. Br J Nutr 2009;101:750–9. 36. Feskens EJ, Loeber JG, Kromhout D. Diet and physical activity as determinants of hyperinsulinemia: the Zutphen Elderly Study. Am J Epidemiol 1994;140:350–60. 37. Feskens EJ, Bowles CH, Kromhout D. Inverse association between fish intake and risk of glucose intolerance in normoglycemic elderly men and women. Diabetes Care 1991;14:935–41. 38. Harris WS. n3 Fatty acids and serum lipoproteins: human studies. Am J Clin Nutr 1997;65:1645S–54S. 39. Maslowska M, Sniderman AD, Germinario R, Cianflone K. ASP stimulates glucose transport in cultured human adipocytes. Int J Obes Relat Metab Disord 1997;21:261–6. 40. Yuan HY, Chiou JJ, Tseng WH, et al. FASTSNP: an always up-to-date and extendable service for SNP function analysis and prioritization. Nucleic Acids Res 2006;34:W635-41. 41. Schaefer M, Konrad S, Thalmann J, et al. The transcription factors AP-1 and Ets are regulators of C3a receptor expression. J Biol Chem 2005; 280:42113–23. 42. Mamane Y, Chung Chan C, Lavallee G, et al. The C3a anaphylatoxin receptor is a key mediator of insulin resistance and functions by modulating adipose tissue macrophage infiltration and activation. Diabetes 2009;58:2008–17.
