Association of PPAR Alpha Intron 7 G/C, PPAR Gamma 2 Pro12Ala, and C161T Polymorphisms with Serum Fetuin-A Concentrations

Background Both peroxisome activator proteins (PPARs) and fetuin-A play a role in lipid and glucose metabolism. Aims We investigated whether PPARα intron 7 G2468/C and PPARγ2 Pro12Ala and PPARγ exon 6 C161T polymorphisms are associated with serum fetuin-A concentrations. Patients and Methods The PPARα intron 7 G/C polymorphism was studied in cohort 1 (79 reference individuals, 165 postinfarction patients). The two PPARγ polymorphisms were investigated in cohort 2 (162 reference individuals, 165 postinfarction patients). Fetuin-A levels and PPAR polymorphisms were determined by radial immunodiffusion and polymerase chain reaction-restriction fragment length polymorphism techniques. Results The C allele variant of PPARα intron 7 G2467C was associated with higher fetuin-A levels (p = 0.018). Postinfarction status (p = 0.001), PPARα intron 7 GG/GC/CC genotypes (p = 0.032), and the C allele (p = 0.021) were the strongest determinants of fetuin-A concentration in a multiple regression model. Higher fetuin-A levels were associated with the Pro variant of PPARγ2 (p = 0.047). Postinfarction status (p = 0.041) and BMI (p < 0.001) but not PPARγ2 Pro were the strongest determinants of fetuin-A concentrations. PPARγ exon 6 C161T genotypes were not associated with fetuin-A levels. Conclusions Fetuin-A was determined mainly by the PPARα intron 7C allele and postinfarction status in cohort 1 and the BMI and postinfarction in cohort 2. The PPARα intron 7C and PPARγ2 Pro variants are associated with fetuin-A levels.

Peroxisome proliferator-activated receptors (PPARs) are members of the nuclear hormone receptor superfamily of ligand-activated transcription factors. The PPAR subgroups PPAR , PPAR / , and PPAR ( 1 and 2) play an important role in the pathogenesis of these processes, which has been extensively reviewed [10][11][12].
There are several observations suggesting a relationship between serum fetuin-A levels and activities of different PPARs. For example, the direct inhibitory effect of pioglitazone on hepatic fetuin-A expression has been observed in rats [13] and in humans [14]; the former was reversed by GW9602, direct PPAR inhibitor.
Polymorphisms of PPAR and PPAR have also been described and found to be associated with disorders of hyperlipidemia, glucose homeostasis, and diabetes. Thus the C allele of the PPAR intron 7 polymorphism was found to be more frequent in patients with myocardial infarction and dyslipidemia [15]. The T allele of the PPAR exon 6 C161T polymorphism was supposed to have protective role against coronary artery disease in Chinese population [16], whereas others found that this allele was associated with an increased 2 PPAR Research risk for coronary heart disease [12]. The association between polymorphic variants of PPAR and serum fetuin-A levels, however, has not been investigated yet.
In our study, we aimed to investigate whether polymorphisms of PPAR (PPAR 2 Pro12Ala and exon 6 C161T) and PPAR (intron 7 G2467C) are associated with or may affect serum fetuin-A levels in two cohorts.

Patients and Methods
Three-hundred and forty-two patients were originally involved in this study. Exclusion criteria were as follows: clinical or laboratory signs of acute vascular disease (myocardial infarction, stroke), acute infection, malignant tumor, hepatic disease, renal failure, immune suppression, severe medical or surgical conditions, and trauma. Finally, we had 327 patients (cohort 2) who had all comparable data (including successful genotyping for both PPAR polymorphisms). We were able to perform successful PPAR genotyping in a smaller number of patients (cohort 1, = 244). The genotyping success rate was greater than 99% in both cohorts. The genotypes were in the Hardy-Weinberg equilibrium.
Postinfarction patients had a history of STEMI myocardial infarction (6-24 months prior to the start of the study). Diabetes was diagnosed based on fasting plasma glucose > 7.0 mmol/l or the 2-hr OGTT > 11.1 mmol/l. Patients with diabetes were treated with diet, metformin, and bedtime insulin.
All persons gave their informed consent prior to their inclusion in the study. The study was approved by the local Ethics Committee of the Károlyi Sándor Municipality Hospital.

Determination of PPAR Polymorphic Variants.
The determination of the PPAR and PPAR variants was performed by PCR-RFLP technique.
The PPAR G2467C intron 7 polymorphism (rs 4253778) was studied by PCR-RFLP technique, using a 5 forward primer of ACA ATC ACT CCT TAA ATA TGG TGG and a 3 reverse primer of AAG TAG GGA CAG ACA GGA CCA GTA. The PCR product was digested with Taq1 (New England Biolabs, Boston, MA, USA) resulting in one fragment of 266 bp of the carriers of the wild-type allele and two fragments of 216 and 50 bp in the carriers of the mutant allele (thermocycles 94 ∘ C 15 min, 30 × 94 ∘ C 30 sec, 50 ∘ C 20 sec, and 72 ∘ C 30 sec) [17].
For PPAR Pro12Ala (rs 1801282) polymorphism, we used a 5 forward primer of GCC AAT TCA AGC CCA GTC and a mutagenic 3 reverse primer of GAT ATG TTT GCA GAC AGT GTA TCA GTG AAG GAA TCG CTT TCC G. The PCR product was digested with Bst U1 enzyme (New England Biolabs, Boston, MA, USA) resulting in one fragment of 270 bp in the carriers of wild-type and two fragments of 227 and 43 bp in carriers of mutant allele (thermocycles 95 ∘ C 15 min, 35 × 94 ∘ C 30 sec, 65 ∘ C 45 sec, and 72 ∘ C 1 min) [18].
The exon 6 polymorphism C161T of PPAR (rs 3856806) was investigated by PCR-RFLP technique using a 5 forward primer of CAA GAC AAC CTG CTA CAA GC and a 3 reverse primer of TCC TTG TAG ATC TCC TGC AG. The PCR product was digested with Pml1 enzyme (New England Biolabs, Boston, MA, USA) resulting in two fragments of 120 and 80 bp in carriers of the wild-type allele and only one fragment of 200 bp in the carriers of the mutant allele (thermocycles 94 ∘ C 15 min, 30 × 94 ∘ C 30 sec, 56 ∘ C 30 sec, and 72 ∘ C 30 sec) [19].

Determination of Serum Fetuin-A Concentration.
Serum fetuin-A concentrations were determined by radial immunodiffusion using the commercially available product (antifetuin-A, IgG fraction, Incstar, cat. number 81931, 13.7 mg/ml, in a final concentration of 84 l/11.5 ml gel), as previously described [20].

Determination of Insulin Resistance Parameters.
Plasma glucose and insulin were determined by the routine HK-G6P-DH and ELCIA methods, respectively. The Homeostasis Model Assessment-Insulin Resistance (HOMA-IR) model was calculated according to Matthews et al. [21].

Statistical Analysis.
Statistical analysis was carried out using the SPSS v.21 statistical software (SPSS Inc., Chicago, IL, USA). Nonparametric methods, including the Bonferroni (Dunn) post hoc test, were used.
values < 0.05 were considered as significant.

Subject Characteristics.
The characteristics of the study participants are shown in Table 1.
Sixty-five per cent of the postinfarction patients received statins and 70% of them aspirin. Serum fetuin-A concentrations did not differ statistically between patients treated and not treated with these two medications (687 ± 122 versus 636 ± 81 mg/l, = 0.204 for statins and 665 ± 0.120 versus 672 ± 124 mg/l, = 0.795 for aspirin, resp.).

PPAR Intron 7 G/C, PPAR 2 Pro12Ala, and PPAR Exon 6 C161T Allele Distribution in Postinfarction Patients and
Reference Individuals. The distribution of PPAR and PPAR alleles is shown in Table 2. PPAR Pro12Ala and PPAR intron 7 G/C alleles did not differ significantly between postinfarction patients and reference subjects. Postinfarction patients, however, had a significantly higher T allele frequency of the PPAR C161T compared to reference subjects.

Analysis of Association between PPAR Intron 7 G2467C Variants and Serum Fetuin-A Concentrations. Serum fetuin-
A levels of individuals with the CC genotype were higher than those of GG genotype (Table 3). In the dominant model (C versus non-C nucleotide), individuals with the minor variant C allele had significantly higher serum fetuin-A concentrations than those with the non-C. In a recessive model (G versus non-G nucleotide), there was no difference between the two variants (651 ± 107 mg/l, = 238 versus 662 ± 171 mg/l, = 6, = 0.702). Except for age, serum fetuin-A concentrations showed significant correlations with parameters listed in Table 4. Serum fetuin-A levels associated weakly but significantly with PPAR intron 7 GG/GC/CC genotypes and the C allele but not with the G allele. During partial correlation analysis, however, the correlation between fetuin-A concentrations and PPAR intron 7 GG/GC/CC genotype lost significance The results of the univariate linear regression analysis between independent variables (predictors) and serum fetuin-A concentration (dependent variable) are shown in Table 5. Serum fetuin-A levels showed weak but statistically significant data with all investigated potential predictors, including PPAR intron 7 G/C genotypes and the C allele but not with age. Thus age was excluded from further analysis.
Next we investigated whether PPAR intron 7 G/C genotype and the C allele may determine serum fetuin-A concentration in a multiple regression model (Table 6). In the model containing all independent parameters, we investigated the PPAR intron 7 GG/GC/CC genotype and   the C allele were the only statistically significant predictors of serum fetuin-A levels. In a backward stepwise regression model, the postinfarction status, the PPAR intron 7 GG/GC/CC genotypes, and the C allele proved to be the strongest determinants of fetuin-A concentration. Serum fetuin-A concentration was significantly associated with BMI, HOMA-IR, gender, and the Pro allele but not with the diabetes and postinfarction status, PPAR Pro/Pro, Pro/Ala, and Ala/Ala genotypes or the Ala allele (Table 7). Thus these two latter parameters were left out from further analysis. The correlation between fetuin-A concentration and the Pro allele was lost following correction for BMI and gender but not with HOMA-IR (Table 8).

Analysis of Association between PPAR 2 Pro12Ala Variants and Serum Fetuin-A Concentrations. Patients with
Univariate regression analysis showed that serum fetuin-A (dependent variable) weakly but significantly correlated with BMI and the PPAR Pro allele (independent variable, Table 9). This latter independent variable lost its predictor role when BMI was included in the regression model.

Analysis of Association between PPAR Exon 6 C161T Variants and Serum Fetuin-A Concentrations.
We found no significant differences among serum fetuin-A concentrations of individuals with different PPAR exon 6 C161T genotypes, nor between the C and non-C or T and non-T groups (Table 10). Fetuin-A levels did not correlate with the PPAR C161T genotypes, C, and T alleles, either (data not shown).

Discussion
In our study, we investigated whether PPAR intron 7 G/C, PPAR 2 Pro12Ala, and PPAR C161T variants are associated with serum fetuin-A concentration. Since subjects in our groups had several parameters that are known to affect fetuin-A levels such as age, gender, BMI, parameters of insulin resistance (diabetes status, HOMA-IR), and postinfarction status [22], we chose regression model to estimate the impact of these variables. PPAR has been termed as a lipid sensor and is involved in microsomal -oxidation and mitochondrial and peroxisomal -oxidation resulting in energy burning and reduced fat storage [11]. The minor variant of the PPAR intron 7C has been considered to have a decreased activity compared to G, the major variant. The C haplotype promotes the early development of type 2 diabetes [17] and is more frequent among postinfarction patients [12,15]. Doney et al. have found that the risk of myocardial infarction is higher in the presence of the C allele [23]. We also have found that the C allele is associated with higher fetuin-A levels and the multiple regression revealed that fetuin-A levels are strongly determined by the postinfarction status and remarkably by the PPAR intron 7 G/C polymorphism, as well. Although the C allele was not more frequent among our postinfarction patients, the higher fetuin-A concentration may also have deleterious effects in them. Elevated fetuin-A concentration is a marker of fatty liver, characterized by decreased -oxidation of fatty acids [4].
Fetuin-A is synthesized almost exclusively by the hepatocytes in adults [24] and the PPAR is mainly expressed in the liver, as well. Fetuin-A is known as an endogenous ligand 6 PPAR Research that binds to free fatty acids and functions as an endogenous ligand for the Toll-like receptor TLR-4 thereby linking metabolic diseases (hyperlipidemia, insulin resistance) and subclinical inflammation [25]. This "missing link" character of fetuin-A has been further supported by the clinical studies of Stefan and Häring [26]. These findings are in line with our observation that postinfarction status, an endpoint of subclinical inflammation, and PPAR intron 7C allele were the strongest determinants of fetuin-A in our subjects. Compared to Ala metabolically disadvantageous characteristics are attributed to the Pro variant of PPAR 2 Pro12Ala [27]. Indeed, we found only the Pro allele among individuals with BMI over 25 kg/m 2 and only lean (BMI ≤ 25 kg/m 2 ) subjects had the Ala/Ala homozygous variant. Nevertheless, even among individuals with the Ala variant, obesity was associated with higher fetuin-A levels compared to lean ones. Patients with diabetes had higher fetuin-A levels, the difference being significant in Pro/Pro major allele homozygotes (704 ± 124 mg/l, = 64 versus 673 ± 132 mg/l, = 183, = 0.020). Since fetuin-A is known to be the natural inhibitor of the insulin receptor tyrosine kinase, the Pro allele may convey increased insulin resistance. Indeed, we found minor variant Ala to be more frequent among nondiabetics (68/251 = 27.1%) compared to diabetics (12/76 = 15.8%, = 0.044). This finding is in accordance with that of Vergotine, who observed that the Pro allele increases insulin resistance, along with IRS1Gly972 [28]. Conversely, the minor allele Ala seemed to be protective in Iranian and Chinese populations [29,30]. In our model, however, the relationship of fetuin-A levels with BMI and postinfarction status was much stronger than the one with insulin resistance (diabetes status or HOMA). The Pro allele was associated with higher fetuin-A in the nondiabetic group, as well, which is reflected by the weak correlation with diabetes status and HOMA-IR. Given the property of fetuin-A to increase insulin resistance, its weak association with the Pro allele can contribute to the deleterious effects of this allele.
Analysis of the effect of the PPAR C161T variants on serum fetuin-A concentrations yielded the least consistent results. We had 3 minor homozygotes (TT) among the postinfarction patients but not among the reference subjects.
This finding is in accord with the observation of Qian et al., who found that coronary heart disease was not associated with the T-carrier state but these individuals had higher risk for acute coronary heart syndrome [12]. Wu et al. found mild protective effect of the T allele only in the Chinese but not in other populations [16]. We found no marked associations of C161T genotypes and alleles neither in postinfarction patients nor in the reference group, and not during the analysis of nonobese, nondiabetic individuals. This suggests that the C161T has the weakest association with fetuin-A levels out of the three PPAR polymorphisms we studied. Since PPAR is expressed mainly in the fat tissue, its association with the levels of the liver secretory protein fetuin-A cannot be as close as that of PPAR .
Although not yet entirely clarified, several observations suggest the molecular basis of the association between PPAR variants and fetuin-A synthesis. The PPAR agonist fibrates decrease fetuin-A expression in obese patients with or without type 2 diabetes mellitus [31]. The PPAR agonist pioglitazone strongly inhibits fetuin-A expression [14]. The upregulation of both PPAR and PPAR results in the downregulation of fetuin-A and NF B and upregulation of the AMPK kinase activities. Palmitate, the oxidation of which is highly induced by PPAR , has been shown to stimulate NF B binding to the fetuin-A promoter [9]. Thus a less functional variant of PPAR could finally result in enhanced fetuin-A expression.
Our study has its limitations. First, the sample size is not big enough to allow for analysis of a comparable number of minor variants. Second, our cohorts were not controlled for environmental factors and prescribed medication and dietary saturated and polyunsaturated fat as it has been suggested [32].

Conclusion
In summary, our results indicate a relatively close relationship between PPAR intron 7 G/C and PPAR 2 Pro12Ala variants and serum fetuin-A concentrations reflecting higher levels in the presence of the C allele of the former and the Pro allele of the latter one. It is very likely that these associations are obscured by obesity and/or diabetes. Larger scale studies are needed to further determine the biological and clinical significance of the PPAR polymorphisms on fetuin-A levels.