A Nutrigenetic Approach to Investigate ApoB EcoR1 Polymorphism-Dietary Acid Load Interactions on Lipid and Anthropometric-Related Outcomes in Adults with Dyslipidemic Type 2 Diabetes Open Access

Type 2 diabetes mellitus (T2DM) is among the most prevalent metabolic disorders worldwide and is one of the main causes of morbidity and mortality [1‒3]. The International Diabetes Federation has estimated a rapid growth in the number of individuals with T2DM from 463 to 700 million within 2019–2045 [4]. Diabetic dyslipidemia is the most common disorder in individuals with T2DM, which manifests as low levels of high-density lipoprotein (HDL) together with elevated low-density lipoprotein (LDL), serum triglyceride (TG), and total cholesterol (TC) [5, 6]. By considering the conclusions from multiple studies, a link is suggested between T2DM and increased inflammatory and oxidative stress markers, which contribute to the pathogenesis of diabetes mellitus and cardiovascular disease (CVD) [7‒9]. CVD is responsible for up to 80% of deaths in T2DM individuals [10, 11]. T2DM is considered a multifactorial disorder that is influenced by interactions between genetic and lifestyle factors [12].

Apolipoprotein B (ApoB) is a structural component of lipoproteins that facilitates the LDL receptor binding and contributes to the cellular uptake of cholesterol from cholesterol-rich lipoproteins [13‒15]. Various studies reported a role for ApoB polymorphisms, e.g., EcoR1 (rs1042031), in hypercholesterolemia, hypertension, atherosclerosis, coronary heart disease, and ischemic stroke [16‒18]. The replacement of lysine for glutamic acid in the EcoR1 single nucleotide polymorphism (SNP) of ApoB alters the constitution and tendency of the recognition site of LDL receptors and might lead to hypercholesterolemia [19]. Previous studies also revealed that EcoR1 is significantly related to an increased risk of ischemic stroke in both Asian and Caucasian populations [14]. Although the association between TG and EcoR1 polymorphism was insignificant through meta-analysis by Gu et al. [16] A allele carriers had a higher and lower level of LDL and HDL, respectively, compared to the GG genotype. Nevertheless, it was also reported that the AA genotype may reduce the risk of ischemic stroke by downregulating LDL levels [20].

Dietary pattern is also an important factor in the prevention of T2DM. Recent studies are assessing dietary patterns quality rather than specific nutrients for investigating their roles in health status. The dietary acid load (DAL) is a simple and useful tool for determining dietary pattern acidity which is predicated on the 2 methods of potential renal acid load (PRAL) and net-endogenous acid production (NEAP) [21, 22]. Consuming high amounts of acid-forming foods such as animal products along with low consumption of alkalizing foods such as fruits and vegetables are the main characteristics of the Western dietary pattern, and have been linked with DAL and may therefore contribute to the acid-base imbalance reported in previous studies [23]. Recently, the acid-base imbalance was proposed as a risk factor for metabolic disorders. Positive associations were observed between DAL, TC, LDL, body mass index (BMI), and waist circumference (WC) by Murakami et al. [24]. It was also suggested that a high DAL score might lead to an inflammatory state, hyperglycemia, and hypertriglyceridemia [25, 26]. Moreover, higher DAL tends to increase the likelihood of CVD as well as T2DM [27, 28]. While different studies examined the roles of DAL or the EcoR1 genotype on T2DM independently, it is unclear if the interaction between EcoR1 and DAL affects metabolic markers. Therefore, this study was designed to assess the interaction of EcoR1 and DAL on BMI, WC, TC, TG, HDL, and LDL in adults with type 2 diabetes.

This cross-sectional study was conducted on 492 randomly selected diabetic individuals (180 men and 312 women) from the Iranian Diabetes Society, Gabric Diabetes Association, and other Health Centers in Tehran during 2011–2012 [29]. Participants with the following characteristics were eligible for the present study: between the ages of 35 and 65 years together with a fasting blood glucose =126 mg/dL, without the use of anti-inflammatory and lipid-lowering drugs, multivitamin and mineral supplements, or insulin therapy, without following a special diet or changes in dietary habits during the last year, and participants without a history of chronic diseases such as thyroid, stroke, renal, hepatic, coagulation disorders, cancer, and inflammatory diseases. Demographic information, socioeconomic, and medical status were gathered with a standardized questionnaire. Participants were classified into 2 groups according to the definition of TC and TG lower than 200 and 150 mg/dL, respectively, for normolipidemia, and TC and TG higher than 200 and 150 mg/dL, respectively, for dyslipidemia [30].

Weight, height, and WC were measured in this study. The measurement of weight (kg) and height [31] was based on the standard protocols with an accuracy of 100 g and 0.5 cm, respectively. These values were used to calculate BMI. WC was determined by measuring mid-way between the lowest rib and the iliac crest with an accuracy of 0.5 cm when a participant stood firmly.

Assessment of dietary intake for the last year was based on a previously validated semi-quantitative food frequency questionnaire (FFQ) comprising 147 items. In a validation study, a random sample of 200 cohort members reported their consumption frequency of a given serving of each food item during the last year, on a daily, weekly, or monthly basis. Thereafter, each food item was converted to a daily intake. Portion sizes of consumed foods were converted to grams. Dietary data was also obtained utilizing 24 h dietary recalls, repeated monthly for 1 year by the same trained dietitians using a standardized protocol that lasted 20 min on average. Next, the means and standard deviations for energy and all nutrient intakes were calculated for both FFQ and the twelve 24 h dietary recalls. Suitable statistical analyses were done to confirm the validity and reliability of the FFQ [32]. DAL was calculated based on the PRAL and NEAP. The PRAL was determined according to the approach suggested by Remer et al. [22] as follows:

PRAL (mEq/day) = 0.4888 × protein intake (g/day) + 0.0366 × phosphorus (mg/day) – 0.0205 × potassium (mg/day) – 0.0125 × calcium (mg/day) – 0.0263 × magnesium (mg/day)

Moreover, the following formula formed NEAP given by Frassetto et al. [21, 33].

Blood samples were collected following an overnight fast (8–14 h). Enzymatic measurements for TC, TG, LDL, and HDL were determined using commercially available kits (Pars Azmoon, Iran). The ELISA technique (Shanghai Crystal Day Biotech Co., Ltd.) was used to measure the serum concentration of 8-isoprostane F2a (PGF2a). Total antioxidant capacity [34] and serum superoxide dismutase activity were quantified by spectrophotometry and colorimetric methods (Cayman Chemical Co., USA), presented by Rice Evans & Miller, respectively [35].

Peripheral blood samples were collected for genomic DNA extraction as previously described [36]. PCR reactions were carried out in 20 µL mixtures containing 50 ng of genomic DNA, 0.2 mM of each primer (Forward: 5'-CTGAGAGAAGTGTCTTCGAAG-3'; Reverse: 5'-CTCGAAAGGAAGTGTAATCAC-3'), and 2x PCR master mix containing 1.5 mm MgCl2 (Amplicon A/S, Odense, Denmark) in 35 cycles of denaturation at 95°C for 30 s, annealing at 56°C for 30 s, and extension at 72°C for 40 s, using a peqSTAR thermocycler (Peqlab, Erlangen, Germany). PCR products were analyzed by 1% agarose gel electrophoresis. The restriction fragment length polymorphism method was used to detect the rs1042031 (GAA>AAA) polymorphism by EcoR1 restriction enzyme (ThermoScientific, Rochester, USA) according to the manufacturer’s protocol, followed by running the products on 2% agarose gel electrophoresis. EcoR1 cuts the 480bp PCR product containing the A allele into 253bp and 227bp fragments. Therefore, GG, GA, and AA genotype cut into one fragment (480bp), three fragments (480bp, 253bp, and 227bp), and two fragments (253bp and 227bp), respectively.

The sample size was determined based on the TG level in GA, GG, and AA genotypes with at least 50 mg/L differences between groups to support the TG level equivalence hypothesis given a type I error of a = 0.05 and type II error of ß = 80%:

S_p² = ((n1 – 1) × SD + (n2 – 1) × SD) / ((n1 – 1) + (n2 – 1) – 2);

S_p: pooled variance

d = (µ1 – µ2) / (v2 × S_p); d: precision

N = (Z_1–a/2 + Z_1–ß)² / d

Given that the GG genotype was less frequent in different populations and the frequency of the genotype has not been previously reported in the Iranian population, we considered 1% as the frequency for the GG genotype in this population. Thus, the sample size was estimated at 500 participants (5/0.01 = 500) to increase statistical power.

All of the statistical analyses were done with IBM SPSS version 21. By considering the median amount of PRAL and NEAP, participants were divided into 2 groups. “Low” represents a more negative score and a healthier dietary pattern, while “high” represents a more positive score and an unhealthier dietary pattern. The normality of variables was checked by the Kolmogorov-Smirnov test and variables were squared or log-transformed if they were not normally distributed. The independent Student’s t tests or analysis of variance (ANOVA) were used to examine the mean difference of continuous variables based on EcoR1 genotype along with DAL. Finally, an analysis of covariance (ANCOVA) and general linear model was performed to examine the interaction between of the EcoR1 polymorphism and DAL on the abovementioned variables after adjustment for confounders including age, gender, smoking habits, physical activity, alcohol intake, energy intake, and familial history of diabetes. p = 0.05 was assigned for statistical significance.

In this study, 608 participants had baseline characteristics and lipid profiles available, whereas 492 participants had genetic data available. 344 people were taken into account for the final interaction analysis (Fig. 1). Table 1 shows the general characteristics and lipid markers based on dyslipidemic and normolipidemic groups. Individuals with dyslipidemia had higher anthropometric measurements, including BMI (p = 0.04) and WC (p = 0.002), compared to normolipidemic participants. Additionally, we acknowledge that the differences in lipid markers between normolipidemic and dyslipidemic groups were not as substantial as expected. In particular, serum levels of LDL, TC, TG, and TG/HDL in dyslipidemic individuals were significantly higher than in the normolipidemic group (p < 0.001). No significant difference was identified in physical activity, dietary intake, and other lipid profiles between the two groups (p > 0.05) (Table 1). This study was performed based on the dominant genetic model (risk allele carriers (GA, AA) versus homozygous non-risk allele) [37]. The frequency of EcoR1 polymorphism was not different between dyslipidemic and normolipidemic groups (p > 0.05).

A statistical analysis of the baseline information of study participants, according to DAL (NEAP and PRAL), is presented in Table 2 and Table 3. All participants were divided into two groups, based on their NEAP and PRAL scores. Participants with higher NEAP median intake were more likely to be male (p < 0.001). An individual with higher adherence to NEAP had a higher BMI (p = 0.02). There were no significant associations found for other basic characteristics and biochemical parameters between the NEAP and PRAL groups.

The interaction between the EcoR1 polymorphism and NEAP on anthropometric indices (BMI and WC) and lipid profiles are presented in Table 4. In the dyslipidemic group, the interaction between EcoR1 polymorphism and NEAP (BMI, WC, LDL/HDL, TG, and TG/HDL) was significant. The results of the analysis showed that among dyslipidemic individuals with the GG genotype, the mean BMI (p = 0.03) and WC (p = 0.02) with higher NEAP intake was higher than those with lower median intake. However, in normolipidemic individuals, no significant difference was observed in the obesity indices between EcoR1 genotype groups. Moreover, TG concentration (p = 0.007), LDL/HDL (p = 0.03) and the TG/HDL ratio (p = 0.03) were significantly higher in A allele carriers with higher than the median intake of NEAP, in comparison with GG homozygotes. The statistical adjustment did not change the statistical significance of these gene-diet interactions (p < 0.05) (Table 4). Further, carriers of the GA/AA genotype who were in the higher median intake of PRAL had higher TG concentration (p = 0.006) and a TG/HDL ratio (p = 0.01) compared to the lower median intake in the dyslipidemic group. These significant interactions persisted even after adjusting potential confounders (age, gender, smoking habits, physical activity, alcohol intake, energy intake, and familial history of diabetes) (p = 0.005, p = 0.01), respectively (Table 5).

The present study indicated a significant interaction between ApoB EcoR1 polymorphism and NEAP in terms of BMI, WC, LDL/HDL, TG, and TG/HDL among individuals with dyslipidemia. We found high levels of TG, TG/HDL, and LDL/HDL in individuals with dyslipidemia with the GA/AA genotypes who had higher NEAP/PRAL intake. Remarkably, the result showed that the relationship between ApoB EcoR1 polymorphism and TG level can be influenced by DAL in diabetic individuals with dyslipidemia.

DAL (NEAP/PRAL) is determined by the balance of acid-inducing foods and is directly linked to lipid metabolism [38]. It has been suggested that the Western dietary pattern, which is rich in red meat, eggs, and refined grains, promotes the risk of hypertriglyceridemia and hypo-HDL-cholesterolemia by affecting acid-base balance [39‒41]. Although the association between higher TG levels and higher NEAP/PRAL scores is not clear, it has been attributed to the central role of insulin in lipid metabolism and hepatic clearance of lipoproteins [42, 43]. Increased levels of lipolysis, newly synthesized TG and very low-density lipoprotein (VLDL) secretion are the result of insulin resistance (IR), which leads to hypertriglyceridemia in diabetic individuals [44, 45]. In addition, higher NEAP/PRAL intake contributes to the occurrence of dyslipidemia by stimulating cortisol secretion and decreasing insulin sensitivity [23, 42, 45‒47].

We also found an elevated level of HDL among individuals with higher NEAP/PRAL intake. However, the difference was not significant. Our observations were corroborated by the results of studies in which a direct link between higher HDL levels and higher NEAP/PRAL scores was found [27, 48, 49]. We found no significant relationship between serum lipid levels and NEAP/PRAL intake in either normal and dyslipidemic groups. Consistent with this result, numerous studies have demonstrated the lack of correlation between DAL and TG [24, 26, 27, 49‒59]. In contrast, a recent meta-analysis reported a positive relationship between high DAL content and higher TG concentrations [26]. These conflicting outcomes reinforce the importance of considering gene-diet interactions. Previous studies have focused on finding a possible correlation between DAL and lipid profile [26, 50, 53, 58, 59] or between ApoB gene polymorphisms and lipoprotein levels among different populations [60‒65].

To our knowledge, no prior study has investigated the interactions between DAL and ApoB EcoR1 polymorphism. Interestingly, we found a significant interaction between serum lipid profile and higher NEAP intake among A-allele carriers with dyslipidemia. The serum TG level was significantly higher in dyslipidemic individuals with the AA/GA genotype who had a higher intake of NEAP/PRAL compared to the GG genotype. We found that EcoR1 polymorphism may notably increase the effect of DAL on lipid profile. Also, the result showed that the interaction between ApoB EcoR1 polymorphism and DAL is different among diabetic individuals with or without dyslipidemia. ApoB has a crucial role in the occurrence and development of hyperlipidemia by modulating human lipoprotein metabolism [43, 66]. While the underlying mechanism is not completely known, it is hypothesized that the EcoR1 polymorphism influences the binding of ApoB to LDL receptors, as well as affecting LDL catabolic rate [16, 60, 67]. Several studies have confirmed that multiple SNPs in ApoB are associated with hyperlipidemia and CVDs [37, 65, 68‒70]. Some of them have reported elevated levels of TC and TG in response to different ApoB polymorphisms [71, 72].

Our result suggest that the A allele could be considered a risk factor for obese diabetic individuals due to the development of dyslipidemia. Consistent with our results, a study in obese children with hyperlipidemia revealed that children with GA genotype for ApoB EcoR1 polymorphism showed higher TC and LDL levels compared to those with the GG genotype. The authors of this past study suggested that children with obesity carrying the A allele have a higher risk of hyperlipidemia in their adult life [60]. A meta-analysis revealed that the A risk allele for the ApoB EcoR1 polymorphism was found to be linked with elevated LDL and lower HDL compared with the GG genotype [16, 73, 74]. A study that investigated the relationship between the EcoR1 SNP and blood lipids in two different Chinese populations reported a higher level of TG and a lower level of HDL among A allele carriers compared to those carrying the GG genotype in one population and only elevated level of TG in the other population [75]. This past study suggests that differences in blood lipid parameters may be related to gene-diet interactions due to the lack of differences in genotype frequencies between the two populations. Numerous studies have confirmed the role of ApoB polymorphisms in response to diet [76‒83]. A case-control cohort study investigated the interaction between SNPs in APOB and dietary cholesterol on plasma cholesterol level and showed significantly higher TC level as a result of the interaction between EcoR1 polymorphism and a high intake of cholesterol. It has been reported that genetic and dietary variations modulate TC levels simultaneously via the absorption and transportation of dietary cholesterol [84]. Finally, a recent meta-analysis showed a higher TC level in individuals with the AA genotype compared to GA/GG genotypes [76]. The overall effect analysis of several dietary intervention studies showed increased levels of TC and LDL among individuals with the AA genotype when compared to those with the GG genotype who had a high fat or cholesterol diet [77, 78, 85]. Furthermore, a recent interaction study indicated that a higher dietary phytochemical index (DPI) score is associated with lower TC levels in the GG genotype [86].

Another novel finding from the present study is the significant interaction between ApoB EcoR1 polymorphism and NEAP intake on BMI and WC in individuals in the dyslipidemic group. There was a greater mean BMI and WC in participants with higher NEAP intake compared to those with lower intake in individuals with the GG genotype. It has been suggested that the association between ApoB gene polymorphisms and body weight is highly related to affecting lipid metabolism [87] and possibly due to the substitution of glutamic acid for lysine which is expected to have an important effect on ApoB protein function [67]. A study that investigated the effect of ApoB EcoR1 polymorphism on abdominal obesity revealed that the amount of visceral adipose tissue, as measured by computed tomography (CT), was significantly higher in A allele carriers. Although the mechanism by which ApoB influences adiposity is undefined, it has been hypothesized that this relationship may be triggered by different environmental factors such as dietary intake [87]. An in vitro experiment using different tissues from the chicken revealed that the ApoB gene may contribute to fat deposition in the subcutaneous and abdominal fat due to a higher amount of ApoB gene expression in the liver [88]. However, there is a study in Asian children which reported no relationship between ApoB EcoR1 polymorphism and anthropometric measures or obesity between normolipidemic and dyslipidemic groups [60].

As previously mentioned, dietary intake can modify the adiposity effect of ApoB protein. We hypothesize that DAL plays a synergistic role in increasing fat accumulation and risk of obesity [89, 90]. Numerous studies have shown a higher BMI or WC in the higher category of DAL indices [31, 48, 58, 91‒95]. A current meta-analysis reported that high DAL content was associated with higher obesity prevalence [26]. A high intake of animal-based foods by following a Western dietary pattern with high DAL content is potentially considered an acid-producing diet. Acidosis status can lead to loss of muscle mass due to reducing protein synthesis and stimulating the catabolism of proteins in the muscles and amino acid oxidation [48]. This mechanism may be mediated by altering insulin-like growth factor 1 (IGF-1) signaling [96]. Consequently, it seems that a higher prevalence of obesity might occur through reducing lean body mass and stimulating fat synthesis as a result of impaired acid-base balance [97]. Therefore, measurement of lean body mass (LBM) or fat-free mass (FFM) instead of BMI might be more effective in elucidating the obesity-DAL associations [26].

Oxidative stress and insulin resistance also result from a high-caloric Western diet, and diabetes causes the accumulation of reactive oxygen species (ROS) and lipid intermediates such as diacylglycerols (DAG) that contribute to the mitochondrial and adipocyte dysfunction associated with impaired lipid metabolism and fat accumulation [98‒100]. Also, insulin resistance may increase ApoB gene expression and corresponding protein levels, which may lead to an increased rate of de novo synthesis of hepatic lipids and fat accumulation [101]. Therefore, the polymorphism effect might be mediated by environmental factors such as eating habits [102].

To date, no previous study has investigated the effect of DAL on the relationship between ApoB EcoR1 SNP and lipid-lipoprotein levels and anthropometric indices in individuals with varying in lipid status. However, further research is required to identify the possible mechanisms of the observed interactions in this study. There are some limitations in the current study. First, serum ApoB protein level and its activity and also other polymorphisms in this gene were not investigated due to time and financial limitations. Second, the recall bias by FFQ may affect estimations of DAL. As a consequence, the dyslipidemic participants with T2DM and GG genotype who consumed a higher intake of NAEP were more likely to have increased BMI and WC. Moreover, individuals with the GA/AA genotype who had a high intake of NEAP/PRAL have experienced an increased TG concentration in the dyslipidemic group. This may mean that diabetic individuals with dyslipidemia and GA/AA genotypes should consider limiting the consumption of a diet rich in acidogenic foods. In addition, the A allele may be a risk factor for obese diabetic individuals with dyslipidemia that might contribute to the development of lipid abnormalities.

The authors thank all participants in this study.

This study was conducted according to the guidelines laid down in the Declaration of Helsinki and the protocol number of the study which was approved by the Ethics Committee of Tehran University of Medical Sciences is IR.TUMS.VCR.REC.1395.15060. Written informed consent was obtained for participation in this study.

The authors declare that there is no conflict of interest regarding the publication of this paper.

This work was supported by the Tehran University of Medical Sciences (Grant number 15060, 2015).

Zeinab Naeini: contribute to writing original paper, updating reference lists; Faezeh Abaj: interpreting results, extracting and analyzing data; Zahra Esmaeily: contribute to writing original paper; Ehsan Alvandi: performing the experiments; Masoumeh Rafiee: responsible for review and editing, project administration; Fariba Koohdani: conducting the search, review, and editing.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.

Masoumeh Rafiee and Fariba Koohdani contributed equally to this work.