Aerobic fitness modulates the association between APOE genotypes and serum lipemia in adolescents

The purpose of this study was to analyze the association between APOE alleles and serum lipemia in adolescents with low and adequate aerobic fitness. The sample was comprised of 105 boys and 151 girls (49% and 46% from rural area) of European ancestry, aged 11 to 17 years, and classified according to: 1) APOE genotype: group ε2 (ε2/3+ε2/2), ε3 (ε3/3), and ε4 (ε3/4+ε4/4); 2) aerobic fitness: adequate or low; 3) serum lipemia: elevated total cholesterol (TC), low-density lipoprotein (LDL) and triglycerides, and low high-density lipoprotein (HDL). The results showed that aerobic fitness modulates the association between APOE alleles and serum lipemia in adolescents, suggesting that adequate aerobic fitness levels exert a greater effect of reducing TC and LDL in ε2 carriers, as well as of increasing HDL and reducing triglycerides in ε3 and ε4 carriers.


Introduction
Low levels of high-density lipoprotein (HDL) and elevated levels of triglycerides, total cholesterol (TC) and low-density lipoprotein (LDL) are risk factors for heart disease in adulthood.In children and adolescents, lipemia is directly related to atherosclerotic processes (Lamotte, Iliescu, Libersa, & Gottrand, 2011) and is inversely associated with physical activity level (Taimela, Lehtimäki, Porkka, Räsänen, & Viikari, 1996).Low serum lipemia is observed in children and adolescents with a high level of aerobic fitness (Kwon, Burns, & Janz, 2010).
The apolipoprotein E (APOE) gene exists in at least three different forms (alleles).The major alleles are called epsilon two (ε2), three (ε3) and four (ε4) and are important modulators of serum lipemia.In general, individuals carrying the ε2 allele present lower serum concentrations of TC and LDL (Alvim et al., 2010;Bernstein et al., 2002;Corella et al., 2001;França, Alves, & Hutz, 2004;Nghiem et al., 2004) whereas the opposite is observed for those carrying ε4 allele (Bernstein et al., 2002;Corella et al., 2001;Medina-Urrutia, Cardoso-Saldaña, Zamora-González, Liria, & Posadas-Romero, 2004;Nascimento et al., 2009).The association between APOE alleles and lipemia seems to be influenced by physical activity levels.Higher concentrations of HDL have been observed in physically active men with ε4 allele when compared to those with ε2 allele.In contrast, in inactive individuals higher concentrations of this lipoprotein were found among those carrying ε2 allele (Corella et al., 2001).In addition, inactive individuals, but not physically active individuals, carrying ε4 allele were found to have higher LDL concentrations than carriers of the other alleles (Pisciotta et al., 2003).
The metabolic responses to physical effort, as well as the capacity to exercise for prolonged periods of time, differ between children, adolescents and adults.Therefore, the observed relationship between APOE gene and physical activity on lipemia in adults cannot be extrapolated to other age groups.
Although aerobic fitness is one of the most important components of health-related physical fitness and is more strongly associated with cardiovascular risk factors than physical activity levels, its interaction on the relationship between the APOE gene and serum lipemia has not been investigated in adolescents.Therefore, the objective of the present study was to analyze the association between APOE alleles and serum lipemia in adolescents with low and adequate aerobic fitness.The hypothesis was that aerobic fitness modulates the association between APOE alleles and serum lipemia in adolescents.

Study population
In an observational, cross-sectional study, data were collected in 2008 in the municipality of Saudades, State of Santa Catarina, South of Brazil.The municipality has 9,016 inhabitants (Brazilian Institute of Geography and Statistics, 2010) and the human development index is 0.831 (United Nations, 2000).Its population mainly consists of German descendants and a minority of Russian and Italian descendants (Brazilian Institute of Geography and Statistics, 2010).Such a homogeneous sample in terms of ancestry helps preventing biases related to interethnic disparities among populations with different genetic backgrounds.
The adolescents were recruited from two public schools in the municipality, one located in the urban area (elementary and high school) and the other in the rural area (elementary school).All adolescents were invited, and those who agreed to participate in the study answered a recognition questionnaire and a blood sample was collected.The anthropometric measurements and aerobic fitness test were performed during regular school physical education classes.The criteria of exclusion were a) age < 10 and > 17 years; b) failure to complete one or more of the data collections; c) motor limitations that would prevent participation in the physical test; d) use of hypolipemic agents; e) pregnancy.The final sample consisted of 101 boys (49% from the rural area) and 151 girls (46% from the rural area).The study was approved by the local Ethics Committee.All adolescents were volunteers and their legal guardians signed the free informed consent form.

Anthropometric measurements
Body weight was measured with a portable digital scale (Soenhle®) to the nearest 0.1 kg and height was measured with a metric tape fixed to the wall to the nearest 0.1 cm.Triceps and medial calf skinfold thickness was measured with a Lange® caliper.The sum of these skinfolds was used to indicate total body fat.The individuals were classified as having adequate (girls: 16-36 mm; boys: 12-25 mm) and elevated fat (girls: > 36 mm; boys: > 25 mm) (AAHPERD, 1988).The three individuals with body fat below the adequate level were included in the adequate category.Abdominal perimeter was measured with a metric tape 2.5 cm above the umbilical scar.This measure was used as an indicator of abdominal fat, which was classified as normal or elevated (Katzmarzyk et al., 2004).

Aerobic fitness
Aerobic fitness was estimated by the one-mile walking test, with the individuals covering the distance in the shortest time possible.Aerobic fitness was classified as adequate or low according to gender and age (AAHPERD, 1988).

Pubertal stage
Pubertal stage was obtained by self-assessment of pubic hair development in each gender (Tanner, 1962).The volunteers were classified into two groups: pubertal (stages II, III, and IV) and postpubertal (stage V).One adolescent rated as stage I was included in the pubertal group.

Serum lipemia
Venous blood samples were collected in the morning after a 12-14 h fast into vacuum tubes containing clotting accelerator.Serum was separated from the samples and stored under refrigeration until the time for analysis.Total cholesterol, very low-density lipoprotein (VLDL), HDL and triglycerides were quantified by automated spectrophotometry (Cobas Mira Plus Roche ® ) using biochemical reagents from Labtest ® .LDL concentration was calculated as described elsewhere (Friedewald, Levy, & Fredrickson, 1972).LDL levels were measured directly in individuals who presented triglyceride concentrations > 400 mg/dL.Individuals with LDL and triglycerides < 100 mg/dL, TC < 150 mg/dL and HDL ≥ 45 mg/dL were classified as normal (I Guideline for Preventing Atherosclerosis in Childhood and Adolescence, 2005).Adolescents who presented concentrations outside these ranges were assigned to the elevated group.

Genetic analysis
Genomic DNA was extracted by the salting-out method (Miller, Dykes, & Polesky, 1988) from leukocytes of venous blood samples collected into vacuum tubes containing EDTA .The APOE gene polymorphism was genotyped by multiplex polymerase chain reaction (PCR) in a thermocycler (MJ96+ Biocycler ® ) (Donohoe, Salomäki, Lehtimäki, Pulkki & Kairisto, 1999).The PCR products were separated by agarose gel electrophoresis and the genotypes were directly identified according to the DNA fragments visualized under ultraviolet light.The adolescents were divided into three genotype groups for subsequent analysis: group ε2 (genotypes ε2/3 and ε2/2), group ε3 (genotype ε3/3), and group ε4 (genotypes ε3/4 and ε4/4).Adolescents carrying genotype ε2/4 (n = 8) were only considered in the analysis of allele and genotype frequencies, but ruled out the inferential analyses due to opposing effects that these two alleles exert on serum lipemia.

Statistical analysis
Allele frequencies were obtained by gene counting.Fisher's exact test was used to evaluate the Hardy-Weinberg equilibrium and to compare allele and genotype frequencies between genders.The normality of data distribution was evaluated in each group by the Kolmogorov-Smirnov and Shapiro-Wilk tests.Variables that showed no normal distribution were log 10 transformed.Groups of ε2, ε3, and ε4 carriers were compared by one-way analysis of variance (ANOVA), with results expressed in the original scale.If the difference was statistically significant, the Bonferroni post-hoc test was applied.Variables showing no normal distribution even after logarithmic transformation were compared by the Kruskal-Wallis test.The frequencies of adequate aerobic fitness, adequate body fat and normal abdominal perimeter were compared by the chi-square test.In view of the high prevalence of the dependent variable (elevated serum lipemia), Poisson regression with robust variance was used.The association between elevated serum lipemia and APOE alleles was evaluated according to aerobic fitness level (adequate and low).For this purpose, the APOE alleles were used as the independent variable and age, gender, pubertal stage, total body fat, and abdominal fat as the controlled variables.Prevalence ratios (PR) and the respective confidence intervals (CI) were estimated.All analyses were carried out using the Statistical Package for the Social Sciences (SPSS), version 15.0.A level of significance of p < .05 was adopted for all tests.

Characteristics of the sample
Boys carrying ε2 allele presented lower TC and LDL concentrations than those carrying the ε3 and ε4 alleles (p < .05).No difference was observed for the other variables (p > .05).Girls carrying ε2 allele presented lower TC concentrations than those carrying the ε3 and ε4 alleles (p < .05).Lower LDL levels were observed in girls carrying ε2 allele when compared those carrying ε4 allele.In addition, girls carrying ε3 allele were taller than those carrying ε4 allele (p < .05).No difference in the other variables was observed among alleles (p > .05).Most boys and girls presented a waist circumference (68.3% and 51.7%, respectively) and aerobic fitness (55.4% and 86.8%) within the recommended range.At the same time, only 43.6% of boys and 39.7% of girls had adequate total body fat (Table 1).

Discussion
In the present study, the frequencies of the ε2, ε3 and ε4 alleles (5.0%, 78.9%, and 16.1%, respectively) were similar to those reported for samples of the Swedish (7.0%, 82.0%, and  The sum of skinfolds, abdominal perimeter and time in the onemile walking test did not differ among ε2, ε3 and ε4 carriers (Table 1).These findings, together with the fact that total and abdominal fat (Dai et al., 2009), as well as, aerobic fitness (Kwon et al., 2010) are associated with serum lipemia, suggest that differences in lipemia among alleles are not due the differences in these variables.
In boys and girls, TC concentrations were lower in ε2 carriers when compared to other alleles carriers (Table 1).The same has been reported for children (Nghiem et al., 2004) and other adolescents (França et al., 2004).In addition, the probability of elevated TC was 27% and 32% higher among ε3 and ε4 carriers than ε2 carriers (Table 2).The same has been reported for adults (Alvim et al., 2010), demonstrating that ε2 protects against hypercholesterolemia. Considering the level of aerobic fitness, only adolescents with adequate aerobic fitness benefited from the ε2 lipid-lowering property, resulting in ε3 and ε4 individuals with higher probability (39% to 47%) of elevated TC when compared to ε2 carriers (Table 3).In agreement with this finding, TC concentrations did not differ among inactive adults carrying the different APOE alleles (Hagberg et al., 1999;Thompson et al., 2004).The inverse association between physical activity and TC was found to be stronger among boys with ε2 allele (Taimela et al., 1996).These findings suggest that the protective effect of allele ε2 against elevated TC is context-sensitive, being enhanced by adequate aerobic fitness.
Lower LDL cholesterol concentrations were observed in boys with ε2 alleles when compared to the other groups, whereas girls with ε2 and ε4 alleles exhibited a similar phenotype (Table 1).In agreement with this finding, lower concentrations of this lipoprotein have been demonstrated in children (Nghiem et al., 2004) and other adolescents (França, Alves & Hutz, 2004) with ε2 allele.The probability of elevated LDL cholesterol ranged from 10% to 44% among adolescents carrying ε4 allele compared to the other alleles and was 30% higher in ε3 than ε2 carriers (Table 2).Similar findings have been reported for adults (Alvim et al., 2010), suggesting that allele ε2 protects against an atherogenic lipemic profile.Considering the level of aerobic fitness, the probability of elevated LDL cholesterol was only higher among adolescents with adequate aerobic fitness carrying ε4 allele (15% to 48%) when compared to the other groups, and among those carrying ε3 allele (30%) when compared to those carrying ε2 allele (Table 3).These findings seems to indicate that in adolescents the protective effect of ε2 in relation to ε4 depends on adequate aerobic fitness, in agreement with the study of Pisciotta et al. (2003).
HDL cholesterol concentrations did not differ between adolescent boys or girls with the different alleles (Table 1).This finding agrees with other studies involving adolescents (França, Alves & Hutz, 2004;Nascimento et al., 2009), but disagrees with the studies of Medina-Urrutia et al. (2004) and Nghiem et al. (2004).In addition, the APOE alleles were not associated with low HDL cholesterol (Table 2), as also observed in adults (Alvim et al., 2010).Only adolescents with adequate aerobic fitness carrying the ε3 and ε4 alleles presented a lower probability (20% to 19%) of low HDL cholesterol than ε2 carriers (Table 3).Lower HDL cholesterol concentrations have been observed in physically active men with ε2 allele when compared to those with ε4 allele (Corella et al., 2001).Similar concentrations have been detected 11.0%) (Bernstein et al., 2002) and French (8.0%, 80.0%, and 12.0%) (Marques-Vidal et al.,2003) population.Furthermore, no difference in allele or genotype frequency was observed between genders, in agreement with other studies (Correla et al., 2001;França et al., 2004;Nascimento et al., 2009).
The frequencies (Table 1) of adequate total body fat (boys: 43.6%, girls: 39.7%) and aerobic fitness (boys: 55.4%, girls: 86.8%) were similar to those obtained a decade ago for adolescents of the same age group from the same municipality (Glaner, 2002).In the present study, 68.3% of boys and 51.7% of girls had an adequate abdominal perimeter, which was similar to that seen in other developing countries (De Moraes et al., 2011).Considering that aerobic fitness has an inverse relationship with body fat (Kwon, Burns, & Janz, 2010), it was expected that prevalence of high levels of body fat and abdominal perimeter were lower among girls, which did not occur.These findings have been reported in other study (Byrd-Williams et al., 2008), suggesting that, among girls, others factors may be more important determinant to increase adiposity levels, like nutritional ones, because poor eating habits are more prevalent among girls than boys (Neutzling, Assunção, Malcon, Hallal, & Menezes, 2010).

Variables
Prevalence ratio a (95% confidence interval)   2005;Hagberg et al., 1999).As reported for adults with increased physical activity level (Bernstein et al., 2002), these findings suggest a greater increase of HDL cholesterol in adolescents carrying ε4 and ε3 alleles in response to improved aerobic fitness when compared to those with ε2 allele.However, no effect of the interaction between physical activity and the APOE gene on HDL cholesterol has been seen in young people (Taimela et al., 1996).Furthermore, the results of experimental studies on adults comparing HDL cholesterol responses between APOE alleles after an exercise program were not consistent (Hagberg et al., 1999;Leon et al., 2004;Thompson et al., 2004).
As observed for children (Nghiem et al., 2004) and other adolescents (França, Alves, Hutz, 2004;Medina-Urrutia, Cardoso-Saldaña, Zamora-González, Liria, Posadas-Romero, 2004;Nascimento et al., 2009), triglyceride levels did not differ between APOE alleles in either gender (Table 1).This lack of an association with any of the APOE alleles suggests that these alleles do not stimulate an increase of triglycerides, irrespective of other factors (Table 2).However, analysis according to the aerobic fitness levels showed a higher probability (45% to 56%) of hypertriglyceridemia only in adolescents with low aerobic fitness who carry the ε3 and ε4 alleles when compared to those carrying ε2 allele (Table 3).These findings suggest lower triglyceride levels in adolescents with ε4 and ε3 alleles as a result of better aerobic fitness, when compared to those with ε2 allele.However, no difference in triglyceride concentrations among APOE alleles was found in inactive men (Hagberg et al., 1999;Thompson et al., 2004).In children and adolescents, the interaction between physical activity and the APOE gene exerted no effect on triglyceride levels (Taimela et al., 1996).Furthermore, the results of investigations comparing triglyceride responses to physical activity between the different APOE alleles are not consistent, with some studies demonstrating a reduction of triglyceride levels in carriers of alleles ε2, ε3 (Leon et al., 2004) and ε4 (Bernstein et al., 2002), whereas no difference between alleles was reported in another study (Thompson et al., 2004).
The present study has some limitations: 1) the small number of volunteers per alleles, especially for allele ε2, which may have prevented the detection of some associations, and 2) the lack of investigation of eating habits that could have influenced the magnitude and direction of some associations since fat intake is known to influence triglyceride and LDL concentrations in carriers of alleles ε2 and ε4 (Andrade et al., 2010).In contrast, the strengths of this study were the use of a Brazilian sample with the same ethnic background and with similar life habits as a result of the strong cultural tradition of the population in this geographic area; the use of rigorous reference criteria (AAHPERD, 1988) for the classification of aerobic fitness, and the control of variables that could influence serum lipemia (age, pubertal stage, total body fat, and abdominal fat).Another strength of this study was that puberty stage (stage II or III or IV) was observed in most part of the sample (83.8% and 98.1% of the boys and girls, respectively), a fact minimizing the effects of biological differences.The findings suggest that adequate levels of aerobic fitness have a greater effect of reducing TC and LDL in ε2 carriers, as well as of increasing HDL and reducing triglycerides in ε3 and ε4 carriers.To our knowledge, this is the first report demonstrating this association and further studies are needed to elucidate the physiological mechanisms underlying the relationship between aerobic fitness, APOE gene and serum lipemia.

Conclusion
In conclusion, the aerobic fitness modulates the association between APOE alleles and serum lipemia in adolescents by reducing TC and LDL levels in ε2 carriers as well as by increasing HDL and reducing triglycerides in ε3 and ε4 carriers.Therefore, aerobic fitness level was found to be a confounding factor in the analysis of the association between APOE alleles and serum lipemia, and should be taken into account in investigations considering these variables, which may help to interpret some divergent/ambiguous findings.From the public health perspective, policies addressing the improvement and maintenance of adequate levels of aerobic fitness e adiposity should be implemented in childhood and adolescence in order to minimize or prevent the cardiovascular disease.

Recommendations
Study results showed that adolescents with adequate aerobic fitness who carried the ε2 allele have a protective effect against serum lipemia.Thus, this physical quality should be considered in investigation related to APOE gene and serum lipemia, which may help to interpret some divergent/ambiguous findings.

Table 1 .
Characteristics of boys and girls according to APOE allele.
TR+CA = sum of triceps and medial calf skinfolds; AP = abdominal perimeter; 1-mile time = time in the one-mile walking test; LDL = low-density lipoprotein; HDL = high-density lipoprotein.Means in the same row followed by different superscript letters differed significantly (p < .05).† Data compared after logarithmic transformation.‡ Data compared by the Kruskal-Wallis test.

Table 2 .
Association between APOE alleles and elevated serum lipemia.

Table 3 .
Association between APOE alleles and elevated serum lipemia according to aerobic fitness level.