Statement of significance

This review critically analyses an alternative pathway of AD known as the ‘vascular hypothesis’ and, for the first time, will identify SFA and APOE genotype as risk factors for AD through their putative roles in increase plasma L-sAβ.

Introduction

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder that accounts for approximately 70% of dementia and presently affecting in excess of 35 million [1]. Global prevalence of AD is increasing and is strongly associated with ageing [2]. There is also accumulating evidence that lifestyle choices including exercise, sleep and particularly diet are associated with heightened AD risk [3,4,5,6,7,8,9].

An equivocal diagnosis of AD is based on later-in-disease evidence of cerebral toxic lipofuscin aggregates within the central nervous system (senile plaque) that are enriched in the protein amyloid-beta (Aβ) neutral lipids and metals [10, 11]. However, microvascular disturbances and cognitive impairment are ordinarily indicated many years preceding frank amyloidosis and neutral lipid aggregation [12, 13], suggesting that a ‘soluble’-amyloid-microvascular-lipid axis triggers the onset and progression of sporadic AD.

An increasing number of studies report that blood measures of soluble-Aβ in cognitively healthy subjects can predict risk for AD at later age [14,15,16,17,18,19]. Specifically, the relative abundance of the Aβ1-42 relative to the Aβ1-40 isoform is reported by several laboratories to be a sensitive surrogate marker of AD risk in later life. Fandos et al. reported that in cognitively normal individuals, Aβ1-42/Aβ1-40 was associated with higher levels of cerebral Aβ plaques [20]. Plasma Aβ biomarkers were also described by Rembach et al. in predicting AD association, with baseline Aβ1-42/Aβ1-40 ratios associated with future cognitive decline [19].

Differences in the concentration of Aβ isoforms in blood may reflect changes in brain efflux of Aβ from cerebrospinal fluid to blood and therefore be indicative of central changes in amyloid metabolism. However, there is an accumulating body of evidence that plasma Aβ principally reflects changes in peripheral synthesis, secretion and metabolism of Aβ as an apoprotein of lipoproteins that are secreted from peripheral lipogenic organs [21,22,23,24,25,26,27,28].

This focussed review article considers contemporary evidence supporting the hypothesis that AD is associated specifically with aberrant peripheral metabolism of lipoprotein-Aβ. Herein, we extend the hypothesis to consider putative interactive effects of APOE genotype with specific dietary fats for onset and progression of AD.

Peripheral metabolism of amyloid-beta

Greater than 90% of blood Aβ1-40 and 97% of blood Aβ1-42 are associated with plasma lipoproteins [29]. Liver hepatocytes and small intestinal enterocytes (the site of dietary fat absorption) secrete Aβ associated with nascent very-low-density lipoproteins (VLDL) and chylomicrons, respectively [21, 30] (Fig. 1). Aβ is suggested to serve as a regulating apoprotein of triglyceride-rich lipoproteins (TRL), although few studies have investigated this directly. Consistent with the latter, in animal models, dietary-fat-induced lipogenesis was found to stimulate Aβ synthesis and secretion into blood and in other studies; obesity was reported to be positively associated with circulating Aβ and increased expression of amyloid precursor protein in adipocytes [31,32,33].

Fig. 1
figure 1

Proposed plasma lipoprotein-amyloid effects on the neurovascular unit. Dietary fats are absorbed as nonesterified fatty acids on the apical membrane of duodenal enterocytes, re-sterified and transiently stored as cytoplasmic lipid droplets. Chylomicron assembly is continuous but can be stimulated by accumulation of enterocytic lipids. Nascent chylomicrons (CM) are secreted into lymphatics with apoproteins, including amyloid-beta (Aβ), which regulate CM metabolism. In circulation, triglyceride-rich CM are progressively hydrolyzed by endothelial lipoprotein lipase, abundant on the plasma membrane of capillary endothelia. Triglyceride-depleted CM remnants (RM) are cleared from blood via an ApoE-dependent high-affinity processes, principally via the LDL receptor which is expressed in abundance on liver hepatocytes. The residual delivery of CM lipids stimulates genesis of nascent very-low-density lipoproteins (VLDL), which like CM are rich in triglyceride and share the same metabolic pathway of lipase-mediated hydrolysis and remnant clearance. VLDL-Aβ secretion may also be exaggerated because of genetic or endocrine-based comorbidities. Exaggerated abundance of lipoprotein-Aβ (CM-Aβ & VLDL-Aβ) is associated with capillary dysfunction, including attenuation of tight junction proteins and extravazation to brain parenchyme of the lipoprotein-amyloid. ApoE anchors the Aβ containing lipoprotein remnants to extracellular matrices and thereafter for receptor and phagocytic uptake of remnant lipoproteins by glia and monocyte-derived macrophages. Lysosomal degradation within the macrophage results in lipoprotein breakdown and protein hydrolylsis concomitant with the release of inflammatory cytokines and prooxidants. Focal inflammation and oxidative stress compromises neuronal cell integrity. Amyloid liberation within the cell and/or following macrophage cell death heightens propensity for Aβ aggregation. Diets enriched in saturated fatty acids may drive lipoprotein-Aβ synthesis and secretion, compromise the capillary endothelia and increase endoplasmic reticulum and mitochondrial stress. ApoE E4 relative to ApoE E3 may slow hydrolysis of triglyceride lipolysis and have lower affinity for hepatic receptor-mediated clearance pathways. Increased ApoE E4-mediated capillary exposure to lipoprotein-Aβ exacerbates capillary stress. ApoE E4 may regulate binding to matrices or regulate phagocytic uptake by inflammatory cells

Full size image

There is a paucity of studies that have investigated plasma lipoprotein-Aβ homeostasis in the context of AD risk per se. In a relatively small study, it was reported that patients with AD had greater net abundance of lipoprotein-Aβ in blood compared to aged-matched controls, particularly indicated within the triglyceride-rich fraction of plasma lipoproteins [34]. Moreover, the authors reported that in response to a dietary fat challenge, the AD subjects had a fourfold exaggerated post-prandial chylomicron response, compared to aged-matched controls. Additionally, several studies have reported dietary affects on plasma Aβ levels [35,36,37]; therefore, repeated cycles of potentially exaggerated postprandial hyperamyloidemia generated through diet could notionally accelerate lipoprotein-Aβ-induced breakdown and inflammation of the neurovascular unit (Fig. 1).

Nutrition and Alzheimer’s disease

There is a substantial body of evidence through population; clinical and preclinical studies that demonstrate nutritional status influence risk for and progression of Alzheimer’s disease. Broad mechanisms may include modulation of neurovascular inflammation [7], through epigenetic factors and neurotransmitter modulation [38]. Nutritional homeostasis is critical for cell metabolism and bioenergetics, and indirect modulation of diet and AD risk includes maintaining a healthy body weight and reducing vascular risk factors [39]. Contemporary studies also suggest a gut/brain axis via dietary modulation of the gut microbiota [40].

Saturated fatty acids and Alzheimer’s disease risk

Diets enriched in fat have been associated with heightened AD risk [41, 42]. Previous longitudinal studies have identified that individuals with a higher intake of saturated fatty acids (SFA) relative to unsaturated fatty acids had an increased risk of developing mild-cognitive impairment (MCI) and AD later in life [43,44,45,46]. The cardiovascular risk factors, ageing and dementia study aimed to identify a link between dietary fat consumption at midlife and subsequent effects on cognitive function. Eskelinen et al. discovered after a mean follow-up time of 21 years that chronic consumption of SFA-rich foods was associated with a reduction in global cognition, prospective memory and an increased risk of MCI [43]. Moreover, a recent meta-analysis of prospective cohort studies identified a link between SFA and AD, with individuals chronically consuming a diet richer in SFA having a greater risk of developing AD [47].

To investigate potential mechanisms underlying a putative link between dietary SFA, the peripheral metabolism of lipoprotein-Aβ and risk for AD, in preclinical studies, wild-type mice were randomized to chronically consume diets enriched in either saturated, polyunsaturated or unsaturated triglyceride. It was found that mice ingesting the SFA-rich diet had significantly increased Aβ abundance colocalized with apolipoprotein (Apo) B within the perinuclear region of enterocytes, the site of chylomicron synthesis [30]. Apo B is an obligatory protein necessary for secretion of nascent TRL from hepatocytes and enterocytes. Exaggerated chylomicron-Aβ secretion in SFA fed mice was also associated with breakdown of the blood-brain barrier (BBB), resulting in brain parenchymal extravasation of plasma proteins including plasma-derived lipoprotein-amyloid, neurovascular inflammation, neuronal degeneration and cognitive impairment, indicated by activation of astrocytes and microglia [48] (Fig. 1). In contrast, mice randomized to a diet enriched in unsaturated, or polyunsaturated fatty acids, had no evidence of increased lipoprotein-Aβ genesis and with preservation of cerebral capillary integrity and cognitive function. The differential effects of nonesterified fatty acids in the wild-type mice provide insight as to the potential mechansism underlying the association of SFA with AD reported in population studies.

Lipoprotein-amyloid beta metabolism and Alzheimer’s disease

Strong evidence of a lipoprotein-Aβ cascade hypothesis for cerebral capillary corruption and cognitive decline was very recently demonstrated in C57BL/6J mice that were genetically engineered to secrete exclusively from the liver, human-Aß (hAß) as an apoprotein of nascent VLDL [21]. The liver-specific amyloid transgenic mice had VLDL-hAβ at concentrations in blood that were physiologically relevant but nonetheless showed marked neurovascular inflammation and astrogliosis (Fig. 2) and cerebral accumulation of amyloid compared to control mice. Liver-specific amyloid transgenic mice also had accelerated evolution of otherwise naturally occurring, but potentially cytotoxic, age-associated lipid-inclusion bodies (Fig. 2), particularly within the hippocampal formation, a region critical to episodic memory. The liver-specific amyloid transgenic mice had chronically exaggerated rates of neurodegeneration across their lifespan, resulting in premature and significant cognitive decline compared to aged-matched controls.

Fig. 2
figure 2

Three-dimensional confocal immunomicrographs of Apo B, GFAP, Iba1 and lipids. Apo B (yellow), activated astrocytes (GFAP; green) and activated microglia (Iba1; magenta) were all measured in the hippocampus of wild-type (WT) control mice and hepatic-specific-human amyloid (HSHA) mice (cell nuclei indicated in blue), with lipid inclusion bodies (LIB’s) measured in the cerebral cortex. Apolipoprotein (Apo) B, an exclusive marker of hepatic and intestinally derived lipoproteins, was measured at 8 months whilst glial fibrillar acidic protein (GFAP) GFAP, ionized calcium-binding adaptor molecule 1(Iba1) and lipids were all measured at 6 months. Apo B, GFAP and Iba1 all use same scale; white scale bar = 50 μm. LIBs of the cortex; black scale bar = 20 μm). WT, wild type; HSHA, hepatocyte-specific human amyloid; Apo B, apolipoprotein B; GFAP, glial fibrillary acidic protein; Iba1, ionized calcium-binding adaptor molecule 1; LIBs, lipid-inclusion bodies

Full size image

Clinical evidence of a peripheral lipoprotein-Aβ hypothesis as a risk factor for AD is supported in post-mortem analysis of AD patient brains. Immunohistochemical staining of Apo B, which is synthesized exlusively by hepatocytes and enterocytes, was clearly indicated within senile Aβ-rich plaques, consistent with causality [49].

Apolipoprotein E4 and synergistic effects with SFA may influence the metabolism of circulating L-Aβ

In humans, ApoE exists in 3 isoforms, E2, E3 and E4, with a frequency of 7%, 78% and 15%, respectively [50]. ApoE E4 is an established risk factor for AD, increasing risk for early onset by 2.8 and 8-fold for hetero- and homozygosity respectively [51, 52]. The mechanisms underlying the APOE ε4 association are not yet established. However, evidence of interactive effects of APOE genotype with diet is suggested by findings that younger APOE ε4 carriers in preclinical stages may benefit mostly from lifestyle interventions, whereas older APOE ε4 noncarriers with dementia may show the most pronounced effects [53].

ApoE is predominantly synthesized in the liver and in blood and is lipidated. ApoE serves as the primary ligand for receptor-mediated clearance for triglyceride-depleted post-hydrolysed remnants of VLDL and chylomicrons [54] (Fig. 1). However, the APOE ε4 allele is associated with a distributional shift of ApoE E4 to lipoproteins richer in triglyceride, interfering in lipolysis and delaying catabolism and clearance from blood of the lipoprotein moiety [55,56,57,58]. We contend that it is a reasonable proposition to suggest that an ApoE E4-mediated delay in metabolism and clearance of TRL remnant-Aβ may exacerbate age-associated microvascular sequale that lead to capillary breakdown and neurovascular inflammation (Fig. 2). Several recent studies support this hypothesis. Montagne et al. reported that individuals bearing APOE ε4 (with the ε3/ε4 or ε4/ε4 alleles) displayed breakdown of the BBB in the hippocampus and medial temporal lobe [59]. The finding was indicated in cognitively unimpaired APOE ε4 carriers and preceded classical AD pathology of frank cerebral amyloid deposition but was more severe in those with cognitive impairment. Indirect evidence comes from Liu et al., who showed that mice genetically engineered for human ApoE E4 restricted exclusively to liver compromised synaptic plasticity and cognition by compromising the cerebrovasculature [60]. Moreover, the APOE ε4 allele exacerbated amyloid brain pathology when cross-bred with amyloid transgenic mice. However, it remains to be determined whether heightened risk for AD associated with APOE ε4 reflects specifically a lipoprotein-Aβ-induced corruption of the cerebral microvasculature.

The effects of APOE ε4 genotype on modulating the metabolism of hepatically derived VLDL is mirrored in the catabolism of intestinally derived postprandial chylomicrons, because they share the same catabolic cascade (Fig. 1). Limited studies suggest postprandial amyloidemia may occur in APOE ε4 versus APOE ε3 subjects. Six hours after ingestion of a lipid-rich meal, APOE ε4 had threefold greater abundance in blood of intestinally derived chylomicrons (indicated as Apo B48) and 1.5-fold greater abundance of hepatically derived Apo B100, consistent with previous reports of reduced clearance rates of post-hydrolysed remnants [61]. Interestingly, a similar postprandial triglyceride excursion was indicated for individuals with APOE ε4 and APOE ε3, consistent with the proposition that Aβ may have been elevated as a consequence of decreased catabolic processes. Less clear is potential differential effects of the apo E-dependent interaction with brain parenchymal cellular matrices and uptake by glia.

Conclusion

There is an accumulating body of evidence that microvascular disruption is the first pathological feature realized in AD. Loss of barrier function is associated with neurovascular inflammation and neurodegeneration. Less clear is what accelerates age-associated cerebral capillary breakdown. Several studies suggest that peripheral blood homeostasis of Aβ, particularly associated with lipoproteins rich in triglyceride, regulate capillary integrity. Exaggerated vascular exposure to lipoproteins-Aβ is associated with loss of tight junction proteins, brain parenchymal extravasation of plasma proteins including lipoprotein-Aβ, brain atrophy and cognitive impairment.

In preclinical studies, it has been shown that dietary SFA increase the plasma abundance of lipoprotein-Aβ concomitant with capillary dysfunction. Clinical studies showing a delayed catabolism of nascent triglyceride-rich lipoproteins, which principally chaperone soluble Aβ in blood, suggest potential synergistic amplification of amyloidemia in blood and, by extension, microvascular disruption. The collective body of evidence suggests that attenuating the flux of systemic lipoprotein amyloid, particulary the remnants of triglyceride-rich lipoproteins, may confer microvascular protection and reduce risk for AD. Given the substantial knowledge of modulating peripheral metabolism of lipoproteins by lifestyle changes, lipid-lowering and apoprotein-targeted pharmacotherapies, new opportunities to reduce risk for AD and slow progression may potentially be realized.

Contemporary dietary recommendations to reduce Alzheimer’s disease risk in the context of peripheral metabolism of lipoprotein-Aβ

Authoratative reviews consistently recommend healthy diets such as Mediterranean style, featuring greater ingestion of polyunsaturated, monounsaturated and omega-3 fatty acids being associated with decreased inflammation, increased insulin sensitivity and brain-derived neurotrophic factor [62,63,64,65,66,67]. Preclinical studies would predict that a Mediterranean diet would result in decreased synthesis and secretion of lipoprotein-Aβ and better preservation of the neurovascular junction [31, 48, 68, 69]. Population studies investigating dementia and AD risk confirm the health benefits of a Mediterranean diet and conversely demonstrate that Western style diets richer in SFA promote neurovascular inflammation and suppress production of brain-derived neurotrophic factor (BDNF), an important molecule involved in learning and memory [70,71,72,73,74]. However, to date, population studies have not reported if chronic dietary behaviour influences peripheral lipoprotein-Aβ homeostasis and neurovascular integrity per se.

Excessive intake of carbohydrate, particularly high glycaemic index food commodities, is associated with dyslipidemia, as a consequence of increased lipogenesis and secretion of nascent TRL [75, 76]. Notionally, excessive carbohydrate intake may also be associated with exaggerated secretion into blood of lipoprotein-Aβ; however, this remains to be proven.

Adequate intake of micronutrients, polyphenols and antioxidants is associated with healthy ageing and is relative to dementia and AD risk [77,78,79]. The Mediterranean and other tailored diets such as DASH and MIND are associated with better cognitive functioning and slower cognitive decline [80,81,82]. However, there are a paucity of studies that can shed insight into impact of specific micronutrients on lipoprotein-Aβ metabolism and the cerebral microvasculature.

Collectively, there is an accumulating and strong body of evidence that adherence to brain-healthy diets can reduce risk for AD. However, whether the mechanisms include positive modulation of the lipoprotein-Aβ/capillary axis remains to be reported.