Plasma cholesterol in Alzheimer’s disease and frontotemporal dementia

2.1 Subjects

Ninety-three patients with dementia (63 AD and 30 bvFTD) were recruited from the cognitive impairment clinic of Tianjin Huanhu Hospital. All assessments, including neurologic examination, routine laboratory tests (especially vitamin B12 level, thyroid hormone level, syphilis, and HIV serology), and cognitive function tests, were done by trained doctors. To ensure the accuracy of the diagnosis, more than half of the dementia patients (including the patients with bvFTD, definite diagnosis, and atypical AD) underwent ^C11-PIB-PET and/or FDG-PET. All patients met the core clinical criteria of the new diagnostic criteria for probable bvFTD or AD [10,11,12], and clinical diagnosis was made by consensus between the neurologist and neuroradiologist. The criteria for exclusion from the study included significant neurological or psychiatric illness that could influence cognitive functions as well as significant unstable systemic illness or organ failure. Each subject was right-handed and underwent a battery of neuropsychological tests: the Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment (MoCA), Neuropsychiatric Inventory (NPI), and Activities of Daily Living (ADL) scale. In addition, 33 age- and sex-matched healthy NCs without neurological or psychiatric disorders were included for comparison. None of the subjects used either statins or other lipid-modifying agents. All subjects were scanned using a single 3-T MRI scanner. To avoid the disturbance of cerebrovascular pathology in this study, the subjects with a high burden of vessel disease [13], including the subjects with Fazekas scale 2 and 3, multiple lacunar infarction, PVS > 11, and so on, based on STandards for Reporting Vascular changes on nEuroimaging [14] were excluded through magnetic resonance imaging.

1. Ethical approval: The research related to human use has been complied with all the relevant national regulations and institutional policies, is in accordance with the tenets of the Declaration of Helsinki, and has been approved by the Medical Ethics Committee of Tianjin Huanhu Hospital.

2. Informed consent: Informed consent has been obtained from all individuals included in this study or their legal guardians (spouses or children).

2.2 APOE genotyping

The blood samples were collected in tubes containing ethylenediaminetetraacetic acid as anticoagulant, and DNA was isolated from blood cell nuclei. APOE genotypes were determined by restriction enzyme digestion according to the method of Wenham et al. [15].

2.3 Measurement of cholesterol levels

Collection of blood samples was done within 3 days after enrollment. Blood samples were collected after an overnight fast of >12 h. The concentrations of triglycerides (TGs), total cholesterol (TC), and high-density lipoprotein (HDL) were measured using an automatic chemical analyzer (Beckman AU5800, USA) [16,17]. Low-density lipoprotein (LDL) levels were calculated using the Friedewald formula: LDL = [Cholesterol − (Triglycerides/2.2)] − HDL [18,19].

2.4 Statistical analysis

Data were analyzed using SPSS statistics (version 19.0) and are presented as mean ± standard deviation or proportions. The Kolmogorov–Smirnov test was run to determine the suitability of variables for parametric analysis. Analysis of variance, followed by Tukey’s post hoc test, was used to explore the main effects of the variables age, TGs, TC, and LDL (P < 0.05 was regarded as significant) on the groups (NCs, bvFTD, and AD). MMSE, MoCA, ADL, and NPI scores and HDL levels were analyzed using the Kruskal–Wallis test followed by the post hoc Mann–Whitney test corrected for multiple comparisons (P < 0.01 was regarded as significant).

Allele frequencies were assessed by counting alleles and calculating proportions. AD, bvFTD, and NC groups were stratified according to the presence or absence of the APOE-ε4 allele; the number of APOE-ε2, ε3, and ε4 alleles; and APOE genotype (APOE-ε2/ε2, APOE-ε2/ε3, APOE-ε2/ε4, APOE-ε3/ε3, APOE-ε3/ε4, APOE-ε4/ε4). The χ² test and the post hoc Fisher exact test (P < 0.05 was regarded as significant) were applied to compare allele frequencies and sex among them. To investigate the relationship between lipid levels and psychological and behavioral abnormalities, we also explored Pearson’s correlations between lipid levels and the subscale scores (appetite and eating change) of the NPI. To clarify the potential effects of the APOE genotype, groups were additionally divided into two subgroups according to the presence or absence of at least one APOE-ε4 allele.

4.1 The frequency of the ε4 allele in AD and bvFTD

The prevalence of the APOE-ε4 allele was higher in AD patients than in bvFTD patients and NCs, but the difference between the bvFTD patients and NCs did not reach statistical significance in our study. A higher number of APOE-ε4 alleles are now considered to be the most important risk factors for AD. However, the published data on APOE allele frequencies in FTD are conflicting. Some studies reported a higher ε4 allele frequency and found a significantly increased risk for the ε4 allele in patients with FTD than in controls [20]. However, several studies and our findings have not replicated this result. An autopsy study of various dementias found that the APOE-ε4 allele frequency was significantly elevated in AD but was comparable to the control frequency in FTD [21]. Another meta-analysis also found no significant result with the ε4 allele in FTD [22]. These differences in FTD were attributed to different ethnicities and selection effects, but we solely focused on bvFTD, a subtype of FTD, in Chinese patients. Another reason might be the relatively small sample size, and the larger case–control study only included 94 FTD patients [22]. Finally, all the dementia diagnoses were made clinically without neuropathological confirmation, although with the help of ¹¹C-PIB-PET and/or FDG-PET. The misdiagnoses might have reduced the power of the test.

4.2 Lipid levels in AD and bvFTD

It is well-known that the main component of senile plaque is Aβ, the main pathological basis of AD. Aβ is produced from the dual cleavage of amyloid precursor protein (APP) by BACE1 and γ-secretase. APP, β-secretase, and γ-secretase are located in lipid rafts, a kind of membrane-bound cholesterol, where APP metabolism occurs. Experiments in cell cultures indicated that cholesterol influences APP-secretase activity and that its accumulation in neurons might accelerate the decomposition of APPs into amyloidogenic components [23]. A growing number of studies have found that high cholesterol levels in the brain play an important role in the process of Aβ-induced AD [24]. In addition to high cholesterol in the brain, animal and human experiments have reported a link between serum cholesterol and Aβ deposition. For example, transgenic APP mice fed with a high cholesterol diet and hypercholesterolemic rabbits had accumulation of intracellular immunolabeled beta-amyloid protein in the brain [25,26]. At the same time, Reed and his colleagues found a direct relationship between cholesterol fractions in blood and amyloid deposition in the brain [27]. Another study also found significant positive associations of TC and LDL with the density of neuritic plaques, an AD hallmark [28]. In our study, higher levels of serum TC, HDL, and LDL were found in AD patients. However, it is important to recognize that CNS cholesterol is locally synthesized instead of coming from peripheral blood. The blood–brain barrier (BBB) separates cholesterol into the cerebral and extracerebral pools. In contrast to cholesterol itself, the side-chain oxidized metabolite 27-hydroxycholesterol (27OH) is able to pass the BBB and communicate between plasma cholesterol and the brain [29,30,31]. The level of 27OH in the circulation is proportional to the level of cholesterol, and there is a concentration-driven flux of 27OH from the circulation into the brain [29,32]. The accumulation of 27OH was found to be the most significant change in the sterol profile of the brains of patients with AD [33], and part of this accumulation is likely to have been affected by the level of cholesterol in the circulation. The high level of 27OH increases the formation of β-amyloid by antagonizing the suppressive effect of 24S-hydroxycholesterol [31,34] and suppresses the formation of activity-regulated cytoskeleton-associated protein by affecting the N-methyl-d-aspartate receptor [35]. Björkhem and his colleagues found that the flux of 27OH into the brain might be the link between hypercholesterolemia and AD [30,31]. In conclusion, higher cholesterol levels are associated with the further development of AD and represent an independent risk factor for AD [36].

Our study also uncovered high cholesterol levels in bvFTD patients, and the LDL level was correlated with the scores of appetite and eating. The changes in eating behavior are core symptoms of the diagnostic criteria for bvFTD, and abnormal eating behaviors are present in >60% of the patients, even >80% over the course of the disease [37]. The abnormal eating behaviors mainly include overeating, stuffing food into the mouth, and changes in food preferences (craving for carbohydrates) [38]. These eating behaviors are likely to contribute to high cholesterol levels. Recent studies revealed that eating abnormalities were partly related to hypothalamic degeneration and to potential disintegration of the network connections between the hypothalamus and reward pathways [38,39]. Therefore, the study of lipid metabolism might provide invaluable data for understanding the pathogenesis of bvFTD. Consistent with our results, another sample of Chinese bvFTD patients presented higher TC and LDL levels compared with NCs, and LDL level was associated with the diagnosis of FTD [40]. However, research from other countries found high TGs and high body mass index (BMI) in bvFTD, but not cholesterol levels [5,18]. Among the reasons involved in this variation, the first was the significant difference in tastes, diets, and perceptions from ethnic differences. Another reason was that lipid levels might be influenced not only by APOE genotype but also by the timing of lipid measurements in relationship to the stage of dementia [41]. High blood pressure and BMI have been associated with an increased risk of dementia for >10 years of follow-up [42], but there is a null or opposite relationship for <10 years [43]. The range of findings might hint at the fact that several years before dementia, onset blood pressure and BMI began to decline, possibly as a result of the ongoing neurodegenerative pathology, suggesting that the same may be true for TGs. Furthermore, large-sample and longitudinal research is required to determine the lipid levels in different diseases and stages. Overall, we suggest that adjusting eating behavior and dietary habits are possible targeting treatments that can modify eating behavior, metabolic abnormalities, and disease progression [44].

Pathological aggregation of the microtubule-associated protein Tau is a common feature of many neurodegenerative diseases, such as AD and bvFTD. The role of cholesterol in Tau phosphorylation or Tau proteostasis is not clear. The disturbance in cholesterol metabolism might impair the degradation of Tau aggregates [45]. The injection of Aβ42 into rat cortices caused a significant increase in hyperphosphorylation of Tau protein [46]. Some researchers argued that Aβ can induce Tau phosphorylation, and excess cholesterol may indirectly promote the production of NFTs. However, a new study implicated cholesterol as a dual upstream regulator of pTau and Aβ and suggested that reducing cholesterol potently decreases the neuronal pTau level through proteasomal upregulation and degradation of pTau in an APP- and Aβ-independent manner [47].

4.3 Effects of APOE genotype on lipid levels

Many candidate genes were considered to be related to lipid metabolism, and among them, the APOE genotype was the major source of genetically determined variation in lipid metabolism. For the subjects who carried the APOE-ε4 allele, the levels of TC and/or LDL were higher than in those without the APOE-ε4 allele in the AD and NC groups. Our findings reflected previous research that APOE-ε4 carriers are associated with increased concentrations of serum TC and LDL [48]. APOE plays significant roles in binding the low-density lipoprotein receptor (LDLR) and the LDLR-related protein and in facilitating cellular uptake of lipoprotein particles [49]. The difference in the amino acid sequences of ε2, ε3, and ε4 alleles results in different receptor-binding affinities. It could be hypothesized that the increase in TAG-rich lipoproteins of APOE in ε4 carriers could increase the affinity to bind LDLR. Ultimately, this could reduce LDL uptake and increase circulating plasma cholesterol [50]. The presence of the ε4 allele has been correlated with the risk of coronary heart disease [51] and with the risk of AD. This is attributable partly to the impact of APOE polymorphism on cholesterol metabolism. Compared with AD, bvFTD is a highly heterogeneous disease, and lipid metabolism is affected by many different factors. Our results show that changes in eating behavior might mediate changes in lipid levels and result in the lack of correlation between APOE-ε4 and cholesterol level in bvFTD.