Association of Plasma Aβ 40 Peptides,But Not Aβ 42,with Coronary Artery Disease and Diabetes Mellitus

INTRODUCTION

Alzheimer’s disease (AD) and coronary artery disease (CAD) have become major challenges for public health systems in an aging general population. In recent years, a growing body of evidence has argued for a substantial involvement of genetic and classical cardiovascular risk factors (CVRFs) in the pathophysiology and progression of AD. Beyond this notion, vascular inflammation has been lately proposed as a candidate for a common molecular pathway in these respective age-related diseases and amyloid-β (Aβ) peptides seem to be crucially involved in this process [1, 2].

The basis for formation of Aβ peptides, the hallmark components of amyloid plaques in AD, is enzymatic cleavage of the amyloid-β protein precursor (AβPP) by β- and γ-secretases [3] with generation of peptides of different lengths, in particular Aβ_1 - 40 and Aβ_1 - 42 as the most relevant ones. The tendency of Aβ peptides to misfold leads to aggregation and amyloid deposition in both brain parenchyma [4, 5] and vasculature [6, 7] and to overproduction or decreased degradation of Aβ which are proposed to be possible causes of cerebral amyloid plaques and cerebral amyloid angiopathy (CAA) [8].

Interestingly, Aβ_1 - 42 peptides seem to be more relevantly associated with parenchymal amyloid deposition [9, 10] and Aβ_1 - 40 peptides rather with vascular deposition [11, 12]. Furthermore, Aβ_1 - 40 peptide has been shown to be entangled in chronic vascular inflammation in numerous ways including endothelial activation [13], monocyte adhesion [14], macrophage activation [13], promotion of foam cells [15], and lipid oxidation [16] and can even be found as a component in atherosclerotic plaques [15]. Recently, Aβ_1 - 40 has been shown to additionally facilitate prediction of cardiovascular mortality and major adverse cardiovascular events and to be significantly associated with arterial stiffness progression, incident subclinical atherosclerosis, and incident congestive heart disease [2].

Altered levels of Aβ_1 - 40 and Aβ_1 - 42 peptides in cerebrospinal fluid (CSF) [17] in AD support the relevance of Aβ-pathology in AD. A number of studies indicate that specific alterations of these peptides also occur in blood in AD and pre-AD stages (mild cognitive impairment (MCI)) and therefore might serve as biomarkers. Baseline plasma Aβ_1 - 40 and Aβ_1 - 42 have been reported to be elevated in individuals later converting to AD [18], whereas elevated plasma Aβ_1 - 40 and low plasma Aβ_1 - 42 levels have been associated with an increased risk for dementia [19]. Decreased plasma ratio of Aβ_1 - 42 to Aβ_1 - 40 was shown to be associated with greater risk to develop MCI or AD [20] and with faster cognitive decline in AD [9, 10]. While a recent population-based study suggested that plasma Aβ may not be useful for clinical diagnosis of AD and evaluated plasma Aβ_1 - 40 merely as a moderate risk marker [12], another study revealed strongly decreased CSF levels of Aβ_1 - 40 and Aβ_1 - 42 in CAA and was thereby able to discriminate individuals with CAA from AD patients and controls [21].

Taken together, most studies are pointing toward increased plasma Aβ_1 - 40 accompanied by lowered plasma Aβ_1 - 42 levels and consequently a decrease in Aβ_1 - 42/Aβ_1 - 40 ratio as a predictive constellation for the development of MCI or AD.

Apolipoprotein E (APOE) ɛ4 allele is known as a genetic risk factor for both AD and cardiovascular disease (CVD) [22–24]. The mechanisms of APOE’s association with AD are not entirely illuminated, but there is profound evidence that APOE genotype, possibly through increased levels of APOE protein [24–26] and interaction of this protein with Aβ peptides might initiate or potentiate neurodegenerative mechanisms, i.e., breakdown of blood-brain barrier (BBB) and/or impaired Aβ-clearance from the brain parenchyma [27].

However, so far no study has investigated the relationships between plasma Aβ levels, APOE genotype, and CVD in a cohort with explicit and state-of-the-art neuropsychological exclusion of MCI, AD, and other neurodegenerative disorders. Thus, the focus of our cross-sectional study was to investigate the associations of plasma levels of Aβ_1 - 40 and Aβ_1 - 42 as well as the Aβ_1 - 42/Aβ_1 - 40 ratio with incident CAD, CVRFs, and APOE genotype in a well-characterized cohort of neurodegenerative healthy elderly individuals.

METHODS

RESULTS

Arterial hypertension and hyperlipidemia were significantly more frequent in CAD (p <  0.001; Table 1). The mean number of pack years was higher in patients with CAD (p <  0.001). Heart rate as well as systolic and diastolic blood pressure did not significantly differ between individuals with CAD and controls.

Individuals with CAD had significantly higher plasma Aβ_1 - 40 levels than those without (p = 0.043; Table 1; p = 0.010 in age group > 63 years; Table 2). Plasma Aβ_1 - 42 levels and Aβ_1 - 42/Aβ_1 - 40 plasma ratio did not significantly differ between both groups. Effect sizes of the association of CAD and plasma Aβ levels (partial η² = 0.008) did not relevantly change after inclusion of study-specific risk factors for neurodegeneration (hyposmia, REM sleep behavior disorder, depression).

Plasma Aβ_1 - 42 levels were significantly associated with APOE genotype (adjusted for age and gender, p = 0.004), while plasma Aβ_1 - 40 was only associated with age, gender, DM, and incident CAD but did not show any relevant association with the APOE genotype (Table 3).

To additionally investigate interaction effects of Aβ_1 - 40 and APOE genotype on the presence of CAD, these factors as well as age and gender were entered in a logistic regression model but revealed no significant interaction effects.

We further examined the association of number CVRFs (hypertonus, hyperlipidemia, current smoking status, DM, and physical inactivity) on Aβ_1 - 40 levels and found only a non significant trend toward higher Aβ_1 - 40 levels in the group with more than one CVRF after adjustment for age and gender (0 CVRF, n = 205, mean 193.10 pg/ml, standard deviation [SD] 38.18; ≥1 CVRF, n = 321, mean 200.52 pg/ml, SD 40.83; p = 0.063).

In the univariate analysis of variance of single possible confounding CVRFs on plasma Aβ levels, DM type 2 was found to be highly significantly associated with Aβ_1 - 40 after adjustment for age and gender (p = 0.001; Table 3), however, the other CVRFs were not associated (p >  0.1). When introducing incident CAD and APOE genotype to the model, the significant interaction was still present (p = 0.004). Also, after including the neurodegenerative risk factors to the model, effect sizes did not relevantly change (partial η² = 0.022).

Correspondingly, the Aβ_1 - 42/Aβ_1 - 40 ratio was significantly associated with age, gender, and presence of DM (Table 3).

DISCUSSION

By investigating the associations of Aβ peptides with CAD, CVRF, and APOE genotype, we confirm recent evidence of higher plasma Aβ_1 - 40 levels in individuals with CAD [2].

Extending these findings, our results demonstrate the association of elevated plasma Aβ_1 - 40 levels with CAD in a cohort of neurodegenerative healthy elderly with no clinical/neuropsychological symptoms indicating clinically relevant Aβ pathology being associated with CAD in a cohort without any clinical evidence of co-existing subtle or overt AD pathology. In this context, our additional finding that plasma Aβ_1 - 42 peptides are not associated with CAD argues in favor of differential systemic roles of Aβ peptides.

Furthermore, plasma Aβ_1 - 42 levels were significantly higher in APOE ɛ4 allele carriers. Interestingly, we found no difference in plasma Aβ_1 - 40 levels between APOE ɛ4 allele carriers compared to non-carriers.

Elevation of plasma Aβ_1 - 40 levels most likely reflects increased efflux of this peptide from the brain via the BBB. Reasons for increased Aβ efflux can be: (i) elevated amyloid burden in the brain with extended clearance of Aβ_1 - 40 [32] and/or (ii) increase in Aβ-trafficking caused by cerebral dysfunction [33, 34]. A recent population-based study demonstrated that even cognitively normal, elderly individuals without a neurodegenerative AD signature in FDG-PET or reduced hippocampal volume on MRI, show amyloid deposition in amyloid PET with ¹¹C Pittsburgh compound B in 24% at 74 years of age [35]. Furthermore, the origin of Aβ peptides in peripheral blood include the liver as well as activated platelets, endothelial cells, and macrophages of which the last three are centrally involved in the pathogenesis of atherosclerosis [36].

Existence of atherosclerotic disease, i.e., indicated by clinical CAD, represents atherosclerotic changes in the entire vascular system including cerebral atherosclerotic pathology with subsequent cerebral dysfunction, which is characterized by decreased cerebral blood flow, increased production of radical oxygen species and dysfunctional BBB with extended Aβ trafficking [33, 34]. Indirect evidence for an involvement of Aβ_1 - 40 in vascular pathology is based on investigations in CAA indicating decreased CSF Aβ_1 - 40 levels [21].

As dysfunction of nitric oxygen oxidase and increased release of endothelin can also be found in cerebrovascular dysfunction [37], Aβ-related disorders and CAD share common mechanisms known to contribute to vascular and cardiovascular pathology, in particular arterial hypertension, DM, and hypercholesterolemia [38]. A recent study demonstrated that Aβ peptides have the property to activate α₁-adrenergic cardiovascular receptors in neonatal cardiomyocytes with the effect of subsequent vasoconstriction [39]. Thus, it is conceivable that circulation of Aβ peptides that migrate through dysfunctional BBB or that are peripherally produced, might not solely affect cerebral vessels but also coronary arteries resulting in CAD. Described effects of Aβ peptides in the brain contributing to well-known mechanism of vascular pathology might even affect every vessel in the human body as indicated by several in vitro and mouse model studies [40–42].

Finally, our results revealed evidence that presence of DM type 2 is independently associated with plasma Aβ_1 - 40 levels. This is interesting in the light of the contribution of DM as a risk factor for CAD. Consequently, the question arises if presence of DM and elevation of plasma Aβ_1 - 40 levels have synergistic effects on vascular pathology and development of CAD, and if one affects or induces the other.

Aβ_1 - 40 as well as diabetic metabolism is reported to cause cerebrovascular dysfunction [33]. In case of DM, cerebral insulin resistance is often referred to as “type 3 diabetes” or “brain diabetes”, which has also been shown to be triggered by Aβ oligomers [43, 44]. Functional insulin signaling and functional insulin receptors are known to have protective effects against Aβ pathology [44], though impaired insulin signaling and decrease in number of insulin receptors are followed by production of reactive oxygen species and release of TNFα and are associated with progression in Aβ pathology [44, 45]. Moreover, a recent study indicates that Aβ oligomers might induce hepatic insulin resistance [46]. In summary, these studies suggest that Aβ-related AD pathology is aggravated by diabetic metabolism [47] but might also aggravate diabetic metabolism itself [45, 48].

Although not particularly investigated in this study, the receptor for advanced glycation endproducts (RAGE) might be a candidate for a common pathway in that respect, as it is a well-known target for both Aβ peptides and advanced glycation end products as ligands [49, 50]. Based on recent literature, Aβ-RAGE interaction promotes permeability of BBB with transcytosis of Aβ and is also involved in upregulation of pro-inflammatory genes inter alia NF-κB [50], subsequent vasoconstriction, and migration of monocytes and macrophages [14]. The prolonged maintenance of such chronic inflammatory state may lead to micro- and macrovascular damage [51] that could result in CVD.

The present study has some limitations. First, we do not have data of an independent validation cohort. However, we argue that the large sample investigated here can serve as a profound basis for future hypothesis-driven studies. Second, assessment of CVRF and assignment to the CAD group were made by dichotomous interview and/or existing medical reports. Nevertheless, we used an approach as standardized as possible for the assessment of CVRF and CAD. Third, DM was also not detected by laboratory test, but solely based on interview and/or medical reports and therefore some cases of DM could have remained undetected. Last, we only found small effect sizes. However we were able to confirm recent findings of the association of plasma Aβ on presence of CAD in a cohort stratified for multiple factors known to be associated with altered Aβ levels as well as CVRFs (e.g., MCI, PD, AD, hyposmia, rapid eye movement sleep behavior disorder, depression). Fourth, we had no data available on estimated glomerular filtration rate, which is known to have effects on plasma Aβ levels. Nonetheless, we argue that our study cohort is comprised of mainly healthy individuals without clinical diagnosis of chronic kidney disease based on the available medical records.

In summary, our findings argue for a differential association of Aβ peptides with CAD and DM type 2, bridging the gap between AD and cardiovascular disease burden. We provide additional evidence for the association of high plasma Aβ_1 - 40 levels but not plasma Aβ_1 - 42 and APOE genotype with CAD in a large cohort of elderly, healthy individuals without clinical or neuropsychological signs of relevant Aβ pathology. Moreover, we show for the first time that plasma Aβ_1 - 40 levels are independently associated with DM type 2. Therefore future investigations should also focus on DM as it is an important factor in neurodegenerative as well as in vascular processes, and possibly a target for therapeutic strategies in the respective disease.

Finally, plasma Aβ_1 - 40 peptides might also play a relevant role in the pathophysiology of CVD and atherosclerosis, probably independent of AD pathology. Additional studies with a distinct evaluation and categorization of cohorts are needed to elucidate this association and not least to clarify the prognostic value of blood-based Aβ peptides in dementia development in AD patients and elderly people.