Frequency of Cardiovascular Genetic Risk Factors in a Calabrian Population and Their Effects on Dementia

INTRODUCTION

A growing body of evidence supports the assessment and treatment of cardiovascular risk factors in midlife as a preventable cause of cognitive decline, morbidity, and mortality in old age. With regards to early diagnosis and application of primary or secondary prevention principles, it is necessary to study all potential risk factors, including non-modifiable risk factors such as genetics and environmental modifiable factors (i.e., smoking, alcohol, overweight, hypertension) that have been reported to play an important role in the development and progression of neurodegenerative diseases [1]. Cerebrovascular diseases (CVD) and Alzheimer’s disease (AD) seem to share common risk factors including hypertension, diabetes mellitus, smoking, hypercholesterolemia, and hyperhomocysteinemia [2, 3]. Moreover, a number of single nucleotide polymorphisms (SNPs) in several genes playing a key role in cholesterol levels (Apolipoprotein E, APOE; Cholesteryl ester transfer protein, CETP), blood pressure (Angiotensin-converting enzyme, ACE; Plasma angiotensinogen, AGT), blood coagulation (Factor V, FV; Factor V Leiden, FV Leiden, Factor II, FII), and vascular dysfunction (Methylenetetrahydrofolate reductase, MTHFR) are known to increase the risk of CVD [4, 5] and AD [4, 7]. In particular, MTHFR has also been implicated as risk factor in the Italian population [8].

However, the association between these genetic variants and brain disease is controversial, and the many meta-analysis studies carried out on large numbers of samples in different populations have not provided clear results to date [6, 10]. One of the causes of these inconsistencies may be the result of trans-ethnic shift of risk genotypes described in different populations [11, 12]: hence, it is possible that the results of the meta-analyses, which merge individuals from different populations in order to obtain a larger number of samples, cannot be considered fully conclusive. To date, the only established genetic risk factor for AD and CVD across different ethnic groups is the ɛ4 allele of the APOE genotype [13].

The objectives of the present study were: 1) to determine the genotypes and allele frequencies of 13 polymorphisms most frequently reported in other populations to be correlated with cardio-cerebrovascular disease and AD in a Calabrian population of Southern Italy, to confirm previous findings, and 2) to investigate the possible association between these genetic risk factors and dementia in this population.

PATIENTS AND METHODS

RESULTS

Table 1 shows the characteristics of the 221 patients and 218 control subjects included in the study. As expected, cases and controls showed a comparable mean age and sex distribution. Compared to demented subjects, controls had significantly higher levels of education and MMSE scores (p < 0.0001 and p < 0.0001, respectively).

Table 2 shows the genotype and allele frequencies distributions for all analyzed SNPs in patients and control subjects. All the polymorphisms studied conformed to the HWE in the sample, with the exception of SNP FII G20210A (Table 2), which was expected. With exception of a weak LD detected between the variants located in the MTHFR gene (r2 = 0.332), the others analyzed SNPs were virtually unlinked from each other. After adjustments for education level, binary logistic regression analysis revealed no association between the polymorphisms FVL (R506Q), FV (H1299R, HR2), FII (G20210A), MTHFR (C677T), MTHFR (A1298C), CBS (844ins68, I/D), PAI 1 (4G/5G), GPIIIa (T1565C), ACE (I/D), AGT (M235T), or CYP7A1 (A278C) and the risk of dementia, considering the different diseases either together or separately. Conversely, APOE ɛ4 status and CETP (I405V) were significantly associated with the risk of dementia (Table 3).

For the APOE gene, carriers of at least one APOE ɛ4 allele resulted to be at higher risk for dementia [odds ratio (OR) (95% confidence interval [CI]) = 2.347(1.435–3.839)] (Table 3).

For CETP (I405V), subjects with the IV genotype were more likely to have dementia [OR(95% CI) = 1.783(1.138–2.795)] (Table 3) compared to those with the II-VV genotype.

Interaction analysis allowed us to verify that the genetic variation of the CETP gene affects the risk of dementia independently of APOE ɛ4 status. In this model, the interaction term was not significantly different from 0. In other words, the combined effect of these two genetic variables can be expressed as exp(0.853 + 0.578) = 4.184 (Table 3). In order to further confirm the absence of a statistical interaction between APOE and CETP genetic variants, we compared models with and without the interaction term by means of the likelihood ratio test. Also in this case, we found that the model including the interaction term did not significantly improve with respect the model without it (χ² = 2.611, p = 0.106).

As mentioned above, the 221 study participants affected by dementia were stratified according to dementia diagnosis. Following stratification, APOE ɛ4 status was found to be significantly associated with the risk of AD and VaD (Table 3). Indeed, carriers of at least one APOE ɛ4 allele resulted to be at higher risk of AD [OR(95% CI) = 2.721 (1.477–5.011)] and VaD [OR(95% CI) = 6.205 (2.356–16.342)] compared to non-carriers (Table 3). Individuals with the CETP (I405V) polymorphism having the IV genotype were more likely to have AD [OR(95% CI) = 2.427 (1.364–4.319)] and VaD [OR(95% CI) = 3.649 (1.455–9.152)] (Table 3) compared to subjects with the II-VV genotypes. Also in this case, we found that the model including the interaction between APOE and CETP genetic variants did not significantly improve, in comparison to the model without it, neither in AD patients group (χ² = 1.861, p = 0.173) nor in VaD patients group (χ² = 0.539, p = 0.463) (Table 3). No association was found between risk of FTD and any of the polymorphisms evaluated (Table 3). When the study samples were stratified by sex or age of onset, no significant differences were obtained.

With regards to environmental/modifiable vascular risk factors, it was not possible to carry out the case-control study since complete data for control subjects were not available. The frequency of modifiable VRFs in patients is shown in Table 4. The prevalence of smoking, hypertension, cardiopathy, and diabetes was not higher when compared to the general population from the same ethnic and geographic background, as already reported in previous epidemiological studies (http://www.cuore.iss.it/fattori/CuoreDataInfo.asp). Conversely, in our sample the prevalence of alcohol consumption and hypertriglyceridemia is lower, whereas hypercholesterolemia is higher (Table 4).

In contrast, the evaluation of the influence of genetic VRFs on modifiable VRFs in patients affected by dementia showed that the presence of carrying at least one APOE ɛ4 was significantly associated with hypercholesterolemia [p = 0.009, OR(95% CI) = 2.118 (1.20–3.73)], which was also observed in AD patients APOE ɛ4+ [p = 0.02, OR(95% CI) = 2.435 (1.16–5.13)]. No correlation was found between genetic or modifiable VRFs or the age of onset among patients (data not shown).

DISCUSSION

To our knowledge, this is the first study investigating the allele frequency of 13 polymorphisms reported to be associated with an increased risk of cerebrovascular disease and their potential association with dementia in a Calabrian population. We found that the allele frequencies of each SNP are in line with worldwide population frequencies (Table 2).

Our results reflect those of previous reports showing individuals with at least one APOE ɛ4 allele have an increased risk of dementia [24], and in particular of AD [13] and VaD [25] (Table 3).

The association analysis of the CETP I405V variant showed cases with the heterozygous IV genotype had an increased risk of dementia, AD and VaD, compared to the II and VV genotypes. These data suggested an overdominant effect of this CETP I405V polymorphism, which was confirmed by the relevant statistical model (Table 3).

Our study showed that subjects carrying at least one APOE ɛ4 allele combined with the IV genotype had a significantly higher risk of dementia, AD and VaD. Moreover, our findings also suggest that CETP polymorphisms may modify the risk for dementia, including AD and VaD, with a synergistic effect by APOE, an observation also supported by other studies [26]. However, these data also appear to contrast with those of other studies showing that the CETP genotype does not influence the risk of either AD [26, 27] or VaD [28], and other reports indicating the CETP I or V allele as an at-risk allele [6, 29]. Conversely, the protective effect of the VV genotype has been reported in previous papers suggesting that I405V valine homozygosity is associated with slower age-associated memory decline and a lower risk of incident dementia and AD [30, 31]. It has been proposed that CETP may modify the risk of AD by altering high-density lipoprotein (HDL) concentrations in the brain, and that an interaction between APOE and CETP may influence HDL concentrations [27, 32]. Thus, it is reasonable to suggest that the effect of CETP on brain cholesterol metabolism and structure may be modulated by APOE activity within the brain, and that a dysfunction in cholesterol metabolism involving both proteins may play a central role in AD pathology [13]. With regards to the other SNPs analyzed that have previously been reported to confer increased risk of cardio-cerebrovascular disease and AD, in the present study there was no evidence of any association in our population. However, this may be a consequence of an insufficient number of samples analyzed from our population and is contrast with what is observed in other populations, as for, for example, the MTHFR C677T polymorphism [5, 33]. Nonetheless, the lack of association may be a result of a population-specific genetic background that is not in line with the general Italian and European populations, a difference we previously demonstrated for the Calabrian population [34, 35]. As well as the lower prevalence of alcohol consumption and hypertriglyceridemia observed in our patients (Table 4) compared to the population sample reported in the CUORE project (http://www.cuore.iss.it/fattori/CuoreDataInfo.asp), this may also be the consequence of an environmental/cultural/epigenetic effect (such as differences in the lifestyle, diet, and so on).

Similarly, we also failed to identify a relationship between the evaluated SNPs and the risk of developing FTD, as is the case for the APOE ɛ4 allele we previously investigated in the same population [36]. This may be a consequence of the fewer number of FTD patients evaluated in the present study or that these cardiovascular risk factors are likely not implicated in this type of dementia [37], although very few studies examining this association in FTD have been performed [37].

The APOE ɛ4 allele also resulted to be significantly associated with hypercholesterolemia in subjects affected by dementia and AD, confirming the relationship between the APOE gene and AD and VaD [24, 38]. This association could be to explain the higher prevalence of hypercholesterolemia in our patients (Table 4) compared to the population sample reported in the CUORE project (http://www.cuore.iss.it/fattori/CuoreDataInfo.asp), together with a different lifestyle. Moreover, this might be due also to a family history for hypercholesterolemia, not analyzed in this paper, depending on the particular genetic background of the Calabrian population.

Moreover, our data confirm the well documented key relationship between education and dementia (Table 1, p < 0.0001) [39], thus underlining the importance of a higher level of education in the prevention of dementia.

A limitation of this study may also derive from the difficulty in evaluating the genetic contribution of single polymorphisms to heterogeneous disorders such as dementia and AD, which are likely impacted by the interaction of multiple genes, proteins, and non-genetic factors. Although a sample size of 218 elderly controls and 221 patients affected by dementia may seem inadequate to identify a genetic risk factor for dementia susceptibility, our dataset is considered representative of the genetically isolated Calabrian population [34, 35] with complete genetic and clinical data available for such analyses. Indeed, we believe that genetic studies on isolated populations such as ours could be very significant, given the controversial results that have been reported by different meta-analysis studies performed to date. In fact, these inconsistent results are likely due to the ethnic heterogeneity among different populations [9] and to the ensuing trans-ethnic shift of the risk genotypes [11, 12]. Thus, we believe that the data on specific relationships between genetic variants and diseases are significant in isolated populations, and are specific to their genetic background and are thus not easily generalizable.