| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Lipids, Lipoproteins, and Cardiovascular Risk Factors |
1 Unit of Internal Medicine, Department of Clinical and Experimental Medicine, 2 Institute of Clinical Chemistry, and 3 Section of Biology and Genetics, Department of Mother and Child and Biology-Genetics, University of Verona, Verona, Italy.
aAddress correspondence to this author at: Dipartmento Medicina Clinica e Sperimentale, Cattedra di Medicina Interna, Università di Verona, Policlinico Borgo Roma, 37134 Verona, Italy. Fax 39-45-580111; e-mail oliviero.olivieri{at}univr.it.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Methods: We investigated the possibility of an interactive effect between the APOC3 gene variant and erythrocyte n-3 PUFAs, suitable markers of dietary intake of fatty acids, on apo C-III concentrations in a population of 848 heart disease patients who had coronary angiography.
Results: In the population as a whole, apo C-III concentrations were significantly inversely correlated with total erythrocyte PUFAs, but the correlation was not significant when only –455CC homozygous individuals were taken into account. In the total population and in subgroups with the –455TT and –455CT genotypes, the relative proportions of individuals presenting with increased apo C-III (i.e., above the 75th percentile value calculated on the entire population after exclusion of individuals taking lipids-lowering medications) decreased progressively as the n-3 PUFA and docosahexaenoic acid concentrations increased. The opposite situation was observed in the homozygous –455CC subgroup, in whom increasing erythrocyte n-3 PUFA and docosahexaenoic acid concentrations were associated with higher proportions of individuals with high apo C-III. A formal interactive effect between genotype and n-3 PUFAs was confirmed even after adjustment for possible confounding variables [age, sex, body mass index, smoking, coronary artery disease (CAD)/CAD-free status, or use of lipid-lowering medications] by logistic models.
Conclusion: Patients homozygous for the –455C APOC3 variant are poorly responsive to the apo C-III-lowering effects of n-3 PUFAs.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Different genetic and acquired factors influence serum apo C-III concentrations (10). Several polymorphic variants have been described in the APOC3 gene promoter, affecting protein transcription and synthesis (11); among them, promoter variants at positions –455 and –482 have been studied more extensively because of their altered affinity for the nuclear transcription factors that mediate the insulin response (12). APOC3 is transcriptionally down-regulated by insulin concentrations (13), but the presence of mutant sequences seems to reduce the inhibitory modulation of the hormone ("insulin resistance" at the gene level) (12).
Dietary factors, such as the consumption of long-chain n-3 polyunsaturated fatty acids (PUFAs) contained in fish and fish oil, have been also described to affect serum apo C-III concentrations through a mechanism similar to that exerted by fibrate lipid-lowering medications, which involves the activation of specific nuclear receptors, i.e., the so-called "peroxisome proliferator-activated receptor-
" (PPAR) (14). APOC3 is one of the target genes transcriptionally down-regulated by PPAR activation, thus contributing to the lipid- and lipoprotein-lowering properties of fish or fish oil intake (14). However, not all individuals within a population seem to gain the beneficial effects of a fish-rich diet. Genetic factors may render individuals differently susceptible as either "dietary responsive" or "dietary nonresponsive" (15).
Recently we reported that the T–455C variant on the APOC3 gene promoter is associated with increased TG and apo C-III concentrations (16)(17) and represents an independent susceptibility factor for CAD (16), particularly in the presence of metabolic syndrome (17). However, in patients with high dietary intake, n-3 PUFAs could act as mediating factors able to substantially reduce the "over time impact" of APOC3 gene variants.
The fatty acid (FA) content in the erythrocyte membrane reflects previous intake over a relatively long time period (months), and an analysis by gas chromatography can provide information on multiple FAs, an approach superior to traditional dietary assessment methods (18). For this reason, erythrocyte FAs are considered as suitable biological markers for dietary intake, particularly for nutritional epidemiology purposes.
Taking into account all of these considerations, we analyzed apo C-III concentrations, erythrocyte FA concentrations, and APOC3 genotypes with the aim of evaluating possible interactions among these factors in determining circulating apo C-III concentrations in a large cohort of heart disease patients examined as part of the Verona Heart Project.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The study was approved by our Institutional Review Boards. Either written or oral informed consent was obtained from all patients.
biochemical analyses
Samples of venous blood were drawn from each patient in the free-living state, after an overnight fast. Serum lipids and the other common biochemical indices were measured as described previously (16). Apo A1, apo B, and apo E were measured by commercially available nephelometric immunoassays; antisera, calibrators, and the BNII nephelometer were from Dade Behring. Apo C-III was measured by a fully automated turbidimetric immunoassay. The reagent were obtained from Wako Pure Chemical Industries, and the procedure recommended by the manufacturer was implemented on a RxL Dimension Analyzer (Dade International Inc.). Imprecision was assessed on three pools of control sera with low, medium, and high concentrations of apo C-III. For the low, medium, and high concentrations, the intraassay CV was 1.8%, 2.0%, and 2.0%, respectively, and the interassay CV was 4.4%, 3.4%, and 2.3%.
mutation analysis
The APOC3 T–455C polymorphism was analyzed as described previously (16).
statistical analysis
All computations were performed with use of the STATA 8.0 statistical package (Stata Corp.). Distributions of continuous variables are reported as the mean (SD). Influences on apo C-III concentrations were first analyzed with log-transformed apo C-III concentrations as the variable of interest (outcome). Given the asymmetric and bimodal distribution of this variable (Fig. 1
), gene-diet interaction effects were assessed by quantile regression models (19). Subsequently, interaction effects were analyzed by logistic models using as outcome the binary variable high apo C-III concentration (h-apo C-III), with the value 1 (and 0 otherwise) assigned to patients with an apo C-III concentration above than the 75th percentile of the apo C-III concentrations of the entire sample, after exclusion of the individuals taking lipid-lowering medications such as statins and/or fibrates (122 mg/L; see also Fig. 1
, dashed line).
|
In both quantile and logistic models, the explanatory variables were the APOC3 genotype (in the recessive model, a binary variable takes the value 1 for homozygous –455CC individuals and 0 otherwise), the erythrocyte FA concentration (continuous variable), and the product of the two preceding variables (interaction term). The set of independent variables contained the following potential confounders: gender, age, CAD/CAD-free status, use of lipid-lowering medications (statins or other medications, including fibrates), smoking, and body mass index (BMI) (10)(20). In the logistic models, the statistical significance of the interaction term was tested by the likelihood ratio test.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
.
|
As expected, several features associated with cardiovascular risk were differently distributed between individuals with CAD or without CAD, including the T–455C genotypes (16). Because the aim of the study was to detect possible gene-diet interaction effects in the whole population, these differences were not the object of specific analyses [with this in mind, compare Refs. (16) and (17)], but CAD/CAD-free status was taken into account as a confounder in the final logistic models (see below).
Apo C-III concentrations in the total patient population showed a bimodal and asymmetric distribution (median value, 105 mg/L; skewness value, 1.83) with a long tail for the highest values, as shown in Fig. 1
. For this reason, the variance-stabilizing logarithmic transformation of apo C-III was used in all subsequent analyses.
Total erythrocyte PUFAs were significantly correlated with log-transformed apo C-III concentrations [correlation coefficient (
) = –0.12; 95% confidence interval (CI), –0.18 to –0.05]. When the analysis was performed on the subgroups based on genotype, correlations remained significant for the patients with the –455TT ( = –0.12; 95% CI, –0.23 to –0.01) and –455TC ( = –0.15; 95% CI, –0.24 to –0.05) genotypes and in the combined TT + TC subgroup ( = –0.14; 95% CI, –0.21 to –0.06) but not in the –455CC homozygous individuals ( = –0.02; 95% CI, –0.19 to 0.15). Following this observation, we based further analyses on the assumption of a recessive model of interaction (patients carrying or not carrying the gene variant in homozygosity).
We first analyzed the possible gene-diet interactions able to influence apo C-III concentrations by quantile regression models using 50th, 65th, 75th, and 85th percentiles of the log-transformed apo C-III concentrations. By this approach, statistically significant interactions emerged when we modeled the 75th (or greater) percentile. The proportion of h-apo C-III individuals was therefore considered as a binary variable of interest (see the section on statistical analysis), and it was analyzed in relation to the different (low, intermediate, high) FA concentrations categorized according the tertile distributions in the population as a whole (below the 33th, 33th–66th, and above the 66th percentile, respectively, after exclusion of patients taking lipid-lowering medications). Of note, h-apo C-III individuals were characterized by an unfavorable lipid profile (increased TG, total cholesterol, and LDL-cholesterol concentrations) despite being similar in age, BMI, gender, and smoking status to the remaining population (data not shown). The proportions of h-apo C-III patients, plotted vs erythrocyte total PUFA concentrations, in the total population are shown in Fig. 2A
; the corresponding distributions in the subgroups of individuals either carrying or not carrying the –455CC genotype are shown in Fig. 2
, B and C. Both in the total population and in patients with the –455TT or –455CT genotype, an increasing erythrocyte PUFA concentration was associated with a progressively minor proportion of h-apo C-III patients (Fig. 2
, A and B). This was not the case for –455CC homozygous individuals, in whom no substantial differences in the proportions of h-apo C-III patients were observed in connection with increasing total PUFA concentrations (Fig. 2C
). In addition, we observed a surprisingly opposite association when the erythrocyte concentrations of n-3 PUFAs or docosahexaenoic acid (C22:6) were considered rather than total PUFAs; the proportion of h-apo C-III patients was highest in the subgroup of –455CC individuals with increased intake rather than in the subgroups with low or medium intake of n-3 PUFAs or C22:6 (Fig. 3
).
|
|
We analyzed the effects of n-3 PUFA concentrations and APOC3 genotype and their interaction on the risk of having or not having increased apo C-III concentrations (h-apo C-III) by appropriate logistic models after adjustment for possible confounding variables (gender, age, CAD/CAD-free status, use of lipid-lowering medications, smoking status, and BMI). Shown in Tables 2
and 3
are the models for erythrocyte total PUFA, n-3 PUFA, and C22:6 concentrations and APOC3 genotypes estimated on the whole population and on the subsample of patients not taking lipid-lowering medications, respectively. We observed a significant interaction between –455CC homozygosity and erythrocyte C22:6 concentrations in both samples; in contrast, the interaction between genotype and n-3 PUFA concentration was statistically significant in patients not taking lipid-lowering medications but not in the total population. When similarly tested, we found no statistically significant results for other FAs in the n-3 family (e.g., C20:5 and C18:3).
|
|
To look for possible interactive effects on TG concentrations, we also applied the same statistical approach using either log-transformed TG concentrations or the proportion of patients with higher TG values (at or above the 75th percentile of the entire population after exclusion of patients taking lipid-lowering medications) as the dependent variable. However, we observed no statistically significant interactions between erythrocyte FA concentrations and APOC3 genotype.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
85%), i.e., in individuals not carrying the –455CC genotype, increasing concentrations of these FAs in erythrocyte membranes were associated with a lower probability of having high concentrations of apo C-III; the finding was particularly clear in the subgroup of patients not receiving lipid-lowering treatment (Table 3
). Such an association not only disappeared but had an opposite trend in –455CC homozygotes (Fig. 3C
). Despite statistical significance and biological coherence of the interaction reported, our findings should be viewed with some caution, taking into account the limitation that our statistical interactions are based on a relatively small number of participants and that we looked at several aspects of apo C-III distribution before concluding that there was a threshold effect. To the best of our knowledge, this is the first report showing an APOC3 gene-diet interaction on apo C-III concentrations. Only a few studies have investigated the relationship between APOC3 gene polymorphisms and FA intake in determining the concentrations of plasma lipids, but the consequences on the protein product of that gene, i.e., the apo C-III values, have never been analyzed.
In 1996, Humphries et al. (21) reported that the APOC3 C–1100T polymorphism affects the consistency and magnitude of changes in plasma cholesterol in response to a diet high in polyunsaturated fats. In a study by Lopez-Miranda et al. (22), the SstI polymorphism, which arises from a cytosine-to-guanosine substitution in the 3'-untranslated region of the APOC3 gene, was shown to be associated with the changes in total and LDL-cholesterol induced by a diet rich in monounsaturated FAs. The same polymorphic variant was also reported to interact with smoking in determining plasma lipid responses to dietary changes (23). More recently, Brown et al.(24) demonstrated that a diet low in saturated fat, compared with a diet rich in saturated fat, was associated with a beneficial lipid profile (lower concentrations of apo B, total cholesterol, and LDL-cholesterol) only among individuals homozygous for of the APOC3 promoter 455T-625T polymorphism, whereas carriers of the APOC3 455C-625del allele were not responsive to a similar diet. Although different in study design, the report by Brown et al. (24), which is the only one reporting on the same APOC3 gene polymorphism as in our study, suggested that carriers of the APOC3 455C-625del allele receive no evident benefit from a diet poor in saturated FAs (and, as a necessary consequence, rich in unsaturated FAs). Unfortunately, data on apo C-III concentrations were not available in the report by Brown et al. (24) so that comparison with our findings can only be indirect.
Our results are consistent with the in vitro experimental evidence of an apo C-III-lowering effect generally exerted by n-3 FAs (14). This effect was attributable mainly to docosahexaenoic acid, the quantitatively most relevant FA of the n-3 family in our patients. The mechanism by which this FA acts on apo C-III production is not completely clear because the in vitro demonstration of activation of PPAR receptors as a necessary step to lower apolipoprotein synthesis has recently been refuted by the results of a study on PPAR-deficient animals (25). Our data concerning the insulin-resistant –455C APOC3 gene variant in humans support the hypothesis that n-3 FAs may interfere with the mechanism of APOC3 gene transcription and that it may be mediated to some extent by insulin or by nuclear factors operating on the APOC3 insulin-responsive element on the gene promoter. However, the matter is complex because individuals homozygous for –455C, –482T on the insulin-responsive element of the APOC3 gene should have lower insulin secretion after an oral glucose test (26). Although in our patients fasting insulin concentrations were not genotype related (17), a decreased incremental insulin response to physiologic stimuli could over time affect APOC3 gene transcription and apolipoprotein synthesis. Further studies are necessary to better clarify this point.
In combination with other positive biological effects, the hypolipidemic properties of n-3 PUFAs have been pharmacologically exploited to reduce CAD risk, and fish oil capsules are now recognized as useful medications in TG-associated dyslipidemia (27). For this reason, the conclusions derived from the present study may be of pharmacogenomic interest. The sample of heart disease patients recruited in the Verona Heart Project was not representative of the general population because many patients were males affected by CAD. The validity of our conclusions is therefore limited to this specific clinical setting, and it needs to be confirmed in a healthy population. In interpreting these results, however, it is important to emphasize that this is not a case-control study between CAD and CAD-free patients, but rather an analysis of interactions in a population in which CAD/CAD-free status is one of the confounders. Adjustment for possible confounding variables (including CAD/CAD-free status) did not modify the results of the interaction models, supporting the view that the relationship between apo C-III concentrations, erythrocyte n-3 PUFA concentrations, and genotype is independent of the concurrent modifiers. Moreover, because the only firmly established therapeutic recommendation for n-3 PUFA supplementation is for patients with documented CAD (27), such as those enrolled in the present study, the value of the present results could be of clinical relevance.
Although this is not an intervention study, there is no logical reason to presume that dietary intake of n-3 PUFAs or fish oil capsules would yield qualitatively different effects. On the contrary, quantitative differences may play a role because the amounts of PUFAs ingested in fish oil capsules are generally much higher than those consumed in the diet. For example, the difference between the lowest and highest tertiles of the erythrocyte n-3 PUFA distribution in our population was 17–18%, whereas in intervention studies the corresponding difference between controls and treatment groups was 300–400%, depending on the duration and dose of n-3 PUFA supplementation (28)(29). The inconsistency for an interactive diet-genotype effect on TG concentrations could be explained by this aspect, considering that a quantitatively more relevant change in n-3 PUFA intake is necessary to observe significant TG-lowering effects independently from any genetic influence (27).
In this respect, the beneficial n-3 PUFA effects could have been under- or overrated in the past, depending on the relative proportions of genetically nonresponsive individuals present in the different populations treated with fish oil capsules. Similarly, it is possible that in genetically selected patients, the advantages achieved by n-3 PUFA supplementation may in the future be shown to be even bigger than the effects demonstrated to date. Further genotype-tailored studies will be necessary to evaluate this possibility.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
, peroxisome proliferator-activated receptor-; FA, fatty acid; h-apo C-III, high apolipoprotein C-III; BMI, body mass index; and CI confidence interval. 
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
is not rate-limiting for the lipoprotein-lowering action of fish oil. J Biol Chem 2001;276:4634-4639.The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  N. Martinelli, D. Girelli, G. Malerba, P. Guarini, T. Illig, E. Trabetti, M. Sandri, S. Friso, F. Pizzolo, L. Schaeffer, et al. FADS genotypes and desaturase activity estimated by the ratio of arachidonic acid to linoleic acid are associated with inflammation and coronary artery disease Am. J. Clinical Nutrition, October 1, 2008; 88(4): 941 - 949. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
