Plasma Lipids as Biomarkers for Alzheimer's Disease: A Systematic Review

Alzheimer's disease (AD) is caused by several risk factors leading to dementia. It’s diagnosis usually depends on clinical presentation and certain biomarkers in the cerebrospinal fluid (CSF). The brain has a high content of cholesterol and the metabolism of cholesterol in the brain can be associated with beta-amyloid plaques formation, which is seen in Alzheimer’s disease. Given these implications, we studied if plasma lipid levels can vary in Alzheimer's disease and if these can be used as biomarkers to diagnose and predict the progression of Alzheimer's disease. Certain mutations in the brain cholesterol transport receptors and proteins and their association with Alzheimer's were also studied. This systematic review abides by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. We searched multiple databases, such as Pubmed, Google Scholar, Pubmed central, ScienceDirect, Web of Science, and Medline with the help of keywords like Alzheimer's disease, cognitive impairment, plasma lipid biomarkers, cholesterol, brain cholesterol metabolism separately and in combination with each other. We collected 49 quality appraised articles on the association between plasma lipids and Alzheimer's disease and the genetic mutations in alleles related to cholesterol metabolism and Alzheimer's disease by applying the inclusion and exclusion criteria. Based on the finding of the studies reviewed, we found an association between plasma lipids, polymorphisms in genes associated with cholesterol transport, and Alzheimer's disease. Increased serum low-density lipoprotein (LDL-C), triglycerides (TG), total cholesterol (TC), sphingolipids, 24S hydroxycholesterol (24S-HC), 27O hydroxycholesterol (27O-HC) was associated with Alzheimer's. Decreased high-density lipoprotein (HDL-C) and phospholipids were noticed. Genetic mutations in apolipoprotein E (ApoE), apolipoprotein B (ApoB), apolipoprotein A (ApoA), ATP binding cassette transporter 1 (ABCA1), ATP binding cassette transporter 7 (ABCA7), amyloid precursor protein (APP), cytochrome P450 family 46 subfamilies A member 1 (CYP46A1), presenilin 1 (PSEN1), presenilin 2 (PSEN2) are also associated with increased risk of Alzheimer's disease. This study found an association between plasma lipids and Alzheimer's, proving that plasma lipids can be used as biomarkers for early diagnosis of Alzheimer's disease. It may also help predict the prognosis and stage the disease severity. Further studies are needed to find out the exact mechanism behind these changes.

website of Neuropathology was included. A total of 49 articles were studied . This study includes 37 studies that proved the relationship between increased plasma cholesterol, triglycerides, 24S hydroxycholesterol, 27O hydroxycholesterol, sphingolipids and phospholipids, and Alzheimer's disease. Twelve studies prove that genetic mutations in ApoE, ApoB, ApoA, ABCA1, ABCA7, APP, and PSEN 1 and PSEN2 alleles are associated with AD. Seven studies explain the metabolism of cholesterol in the brain and the pathology associated with AD ( Figure 1).

Discussion
We studied 49 previously published articles, some stressing the association between Alzheimer's disease and plasma lipids and some about the genetic variants of various cholesterol transporters causing AD. In this study, we found that a dysregulation of brain lipid homeostasis can lead to cognitive disorders such as Alzheimer's disease. The exact pathway by which certain mutations happen is still not clear and more research is needed to confirm that.

Cholesterol Metabolism in the Brain and Pathology Related to AD
The brain has a rich lipid content. The majority of cholesterol in the central nervous system (CNS) is present in two places; one is the oligodendrocytes of the myelin sheath, and the other is the plasma membrane of astrocytes and neurons. Myelin comprises 70% lipids containing cholesterol, sphingolipids, and phospholipids [19]. Cholesterol, sphingolipids are essential components of the neuronal plasma membrane present in the form of lipid rafts. It participates in signal transduction, neurotransmitter release, synaptogenesis, and membrane trafficking [11,20]. It is believed that the developing brain produces a high level of cholesterol, but this gradually reduced in the adult brain. The adult brain synthesizes cholesterol by the astrocytes, then transported to the neurons to carry out its functions.
Cholesterol metabolism in the brain depends on brain cells' synthesis, transport across the cells, BBB, and catabolism [5]. Neurons depend on glial cells (astrocytes) for cholesterol. Astrocytes synthesize cholesterol from acetyl CoA by the HMG-CoA reductase enzyme. The transport of cholesterol to neural cells takes place with the help of ApoE which is also synthesized by astrocytes. The ApoE-cholesterol complex is transported with ABCA-1 and is taken up by the neurons via endocytosis through the LDL receptor-related protein (LRP-1) [5]. The endocytosed cholesterol in the neuron is further hydrolyzed to form free cholesterol. Free cholesterol undergoes esterification with acyl-coenzyme A cholesterol acyltransferase to form cholesterol esters stored in the neuron's cytoplasm. Some of the free cholesterol also controls the expression of cholesterol synthesizing enzymes and lipoprotein receptors such as liver X receptors (LXRs). These LXRs lead to increased expression of the ABCA1, thus mediating the transport of cholesterol from cells to apolipoproteins [5,6]. There is no degradation mechanism for any excess cholesterol in the brain. The excess cholesterol is converted to 24S-HC by the cytochrome P450 family 46 subfamilies A member 1 (CYP46A1) enzyme and is transported out of the brain to the blood-brain barrier [21].
The pathognomonic feature of AD is the building up of beta-amyloid plaques in the brain. In a healthy brain, the beta-secretase/beta-site amyloid precursor protein cleavage enzyme -1 (BACE-1) and gamma-secretase are present in the lipid rafts of the neuronal plasma membrane and cause cleavage of APP forming betaamyloid. However, in AD, increased cholesterol lipid rafts induce raft clustering and enhance BACE-1 and APP interaction leading to increased beta-amyloid production [5,20,22].
The BBB maintains CNS homeostasis by regulating the transport of solutes between blood and brain. The BBB allows diffusion of oxygen and carbon dioxide freely, although lipophilic molecules such as cholesterol enter through receptors or channels. The brain lipid nature's homeostatic balance helps in controlled betaamyloid production by APP cleavage, maintains the receptor channels, vesicle formation, secretion, signaling, inflammation, oxidation, membrane biosynthesis, and remodeling. Dysregulation in the brain lipid environment attributes to disturbed BBB, abnormal APP processing, abnormal cytosis, signaling, increased inflammation, oxidation. Long term, these can result in neuronal death, leading to AD (Figure 2) [22]. FIGURE 2: Demonstrating the process of synthesis, metabolism, and transport of brain cholesterol

The Relation Between Plasma Lipids and AD
Many researchers have found a link between cholesterol homeostasis and AD; however, the exact pathogenesis remains unclear. We know that brain is rich in cholesterol content. The cholesterol in the brain and periphery are two separate units as they are well separated by a blood barrier, which restricts the entry of peripheral cholesterol. The cholesterol present in the lipid rafts of plasma neuronal membrane causes cleavage of APP. In the case of increased cholesterol in the lipid rafts, cholesterol enhances the activity of BACE-1 and Gamma-secretase and causes cleavage of APP leading to increased beta-amyloid production. These beta-amyloid plaques are characteristic hallmarks of AD. Increased plasma LDL-C, TC, triglycerides (TG), and decreased HDL-C was associated with increased beta-amyloid plaques causing AD. A cohort study by Pappolla et al. emphasized the association between high plasma cholesterol and patients with AD [23]. They found that increased cholesterol in the plasma was leading to increased beta-amyloid production. Many other studies found the same association between cholesterol and beta-amyloid plaques in the brain [4,24]. Some studied proved that beta-amyloid isoforms 1-40, 1-42 are associated explicitly with plasma cholesterol [7,13]. The study by Iqbal et al. was a systematic review showing that increased LDL-C, TC, TG were increasing beta-amyloid production. In contrast, decreased HDL-C was associated with AD [24].
Apart from amyloid plaques, AD is also associated with atrophy of the left/right hippocampal and entorhinal cortex. These changes are seen in the early stage of AD even before symptoms could arise. A case-control study by Proitsi et al., including 300 patients, also found that increased plasma lipids led to amyloid plaques in the brain and hippocampal and entorhinal cortex atrophy patients with AD [25]. Another study by Wolf et al. also studied the hippocampal atrophy of the brain and found its link with cholesterol; however, it was only associated with HDL-C [26]. The role of HDL-C in the disease process of AD is controversial. HDL is found to reduce the build-up of beta-amyloid plaques, thus reducing inflammation. Many researchers studied this protective role of HDL-C. Formiga et al. included 321 patients in a cohort study and proved the association between decreased HDL and AD [27]. Few other studies also proved the same objective [26,28]. Physical activity can decrease HDL levels and has proven to improve symptoms and progression of AD by a randomized control trial of 170 patients [28].
Cholesterol and triglycerides are also known to be a predictive marker for cardiovascular diseases. These lipids build up and obstruct the arteries supplying the heart. A similar mechanism was assumed for AD. Scientists believed that AD happens due to the brain's poor oxygenation due to the lipids' clogged vessels. Some studies determined the relation between these serum lipids and AD and if it is similar to the cardiac risk profile. A cohort study of nearly 4000 subjects by Helzner et al. shows that plasma lipids are associated with AD and other vascular diseases [29]. However, the mini-mental state exam (MMSE) was not affected, and hence it is still unclear how the cardiac risk profile affects cognition [15,30]. The relation between cholesterol and low MMSE was seen in a study by Hall et al. [31].
Unlike the studies already discussed, some studies say that only TC has an association with AD. The theory remains unclear, though. A study by He et al. proved that LDL, HDL, and triglycerides levels remain normal and only increased TC was seen in AD [32]. He included 130 patients in his case-control study. Similar studies by other authors also proved the same concept [33,34]. These studies by Solomon et al. and Anstey et al. suggest that total cholesterol has a bidirectional association with AD. It was seen to rise to midlife, suggesting cognitive impairment, and later declined with age in patients with AD. Solomon et al. proved this by conducting a large-scale case-control study, including 1321 patients with AD and 1203 controls [34]. Some other studies also found that only a single lipoprotein-low-density lipoprotein was elevated in AD, serum levels of rest were normal.
LDL-C is seen to cause vascular and neurotoxic effects in the brain [35,36]. Another study by Zhon, a systematic review including nearly 6500 patients, found the same association; however, LDL-C was high in patients with AD mostly around 60-70 years of age, gradually reduced with ageing [37]. Some studies found an association between AD and only LDL-C, TC. Many hypotheses are present. Some say LDL-C, TC is associated with increased tau concentration, according to some LDL-C, TC cause increased amyloid buildup, and some found that LDL and TC disrupt the cell cycle [38]. A study by Liu et al., including around 2333 AD patients and 3615 healthy controls, also suggested the association between LDL-C, TC, and AD [39]. Two studies also found that cholesterol remains normal, and only the serum triglyceride level is increased in the case of AD [22,40].
It is known that de novo synthesis of cholesterol occurs in the brain, and any disruption in this mechanism can lead to AD. 24S Hydroxycholesterol is the elimination product of neuronal cholesterol that leaves the brain and enters the periphery by crossing the BBB. 24S-HC and 27O hydroxycholesterol in the plasma indicates the degree of beta-amyloid production, loss of active grey matter, phosphorylated tau accumulation, and brain atrophy, thus indicating AD [11]. A case-control study by Popp et al., including 200 patients, found an association between increased plasma 24S hydroxycholesterol, 27O hydroxycholesterol, and AD [41]. Another study also showed a similar association [30].
Another group of lipids associated with AD is the sphingolipids and phospholipids. Sphingolipids such as sphingomyelin, ceramide, sulfatide, and sphingosine are major constituents of the neuron's plasma membrane. They are present in the lipid rafts, hence have a role in enhancing the activity of BACE-1 and gamma-secretase that causes cleavage of APP, forming beta-amyloid [20]. Phospholipids such as phosphatidylcholine, plasmogens, phosphatidylinositol, too, are a part of the membrane-forming lipid [20]. An association between serum and CSF levels of sphingolipids and phospholipids and AD has been studied. A study by Wong et al. showed that CSF sphingomyelin levels increase in the prodromal stage in AD, CSF ceramide increases in AD, CSF sulfatide levels decrease in AD, CSF phospholipids levels also increase. Whereas in the blood, ceramide level increases, sphingomyelin levels decrease, and phospholipids decrease in AD [20]. Another study by Kosicek et al. also found similar results [17]. A case-control study by Costa et al. studied ten membrane phospholipids and their association with AD. These were found to decrease in AD ( Table 1) [42].

Author
Year of

Genetic Mutations in Cholesterol-Related Genes and AD
Apolipoprotein E is produced by the astrocytes and is a carrier transport protein, shifting cholesterol from astrocytes to neurons. Apolipoprotein E also has an affinity for beta-amyloid in the presence of cholesterol. Some studies found that ApoE is required for the clearance of beta-amyloid. Thus, any mutation in this could lead to a decrease in the clearance and building up of plaques [6]. There are three isoforms of ApoE: ApoE2, ApoE3, ApoE4. ApoE3 is the commonest isoform present in the majority of the population; however, ApoE4 has the strongest established association with AD. Any individual homozygous for this allele and old can develop AD [6]. Since it has such a strong association, any mutations in ApoE4 can result in AD. According to studies, ApoE4 is an independent risk factor for AD [10]. A Cohort study by Kivipelto et al., including 1449 AD patients, suggests that ApoE 4 is an independent risk factor for AD [43]. Another cohort study by Toro et al. also found the same association [44]. Other isoforms of ApoE are not as closely related to causing AD as ApoE4. ApoE2 does not have any role in AD, and ApoE3 has protective effects against AD. A systematic review by Agarwal et al. suggests the above and found that all alleles of ApoE4 -ApoE 2/4,3/4,4/4 are associated with AD [45].
Some other types of apolipoproteins are also related to AD. ApoB is another lipoprotein associated with AD; the exact reason remains unclear. In the brain, most of the cholesterol is present in density similar to that of HDL-C and transported via ApoE or ApoA [5]. Any mutation in the ApoA gene can also cause AD. A study conducted by Raygani et al. found that apart from ApoE 4, increased ApoB and decreased ApoA1 are associated with Alzheimer's disease [46]. Early-onset AD has also been associated with mutations in the genes coded for enzymes and transporters involved in beta-amyloid metabolisms, such as APP, PSEN1, and PSEN2. APP cleaves to form beta-amyloid, and presenilin is a protein present in the gamma-secretase complex [8]. Studies have shown the association between AD and mutation in the gene of APP, PSEN1, PSEN2, and ApoB [47].
Some studies have found the association between another cholesterol transporter -ABCA1 and AD. ABCA1 is essential for the cholesterol efflux from the CSF to the serum. ABCA1 is also found to reduce beta-amyloid accumulation. Lack of gene/defect in the gene of ABCA1 is associated with building up of amyloid plaques, causing AD [16]. A study by Li et al. studied the association between ABCA1 and Cholesterol efflux causing AD [48]. The same study also found the association between another genotype, ABCA7, and AD. By the metabolism of cholesterol, we have studied that cholesterol does not cross BBB; hence it is metabolized to 24S hydroxycholesterol with the help of the enzyme CYP46A1 and further eliminated via the BBB. Any mutation in the gene coding for CYP46A1 can lead to defects in elimination, causing increased cholesterol and more amyloid plaques leading to AD. A study by Lukiw et al. found the association between 24S-HC, cholesterol synthesizing enzyme, and AD (Table 2, Figure 3) [49].

Limitations
Even though the study clearly shows the relation between AD and plasma cholesterol, there are many aspects of cholesterol metabolism in the brain that are still not clearly understood. Different studies included in this study have different theories about AD pathology due to high plasma cholesterol. The exact reasons for polymorphism in many genes and associated AD also remain unclear. We were also unable to find if these biomarkers are altered by gender and physical activity. Only studies published in English were included; we did not include articles that were in other languages. Case reports, case series, letter to the editor, editorials were not included. Similarly, studies that did not find any association between AD and cholesterol were also excluded.

Conclusions
We studied many articles on the association between AD and plasma lipids. Based on those, we found that levels of plasma cholesterol, triglycerides, sphingolipids, phospholipids are altered in AD, and genetic variants of the various cholesterol metabolism-related proteins also lead to AD. The brain is made up of high lipid content, and thus any alteration in the cholesterol metabolism in the brain can cause dysregulation of the brain lipid homeostasis. Cholesterol is associated with the beta-amyloid build-up in the brain, thus increased plasma cholesterol results in increased beta-amyloid plaques, which is pathognomonic of AD. We also found out that ApoE4 and mutations in many other transporters of cholesterol in the brain are linked with increasing the chances of AD. This article thus points out all the lipid-associated risk factors causing AD. This will help us improve our knowledge and scope of the disease. Plasma lipids biomarkers can also be studied by a simple blood test making diagnosis and prediction of AD much easier. However, more advanced studies and research must understand the exact pathology behind this, as brain and peripheral cholesterol are two different entities and separated by a strict BBB.

Conflicts of interest:
In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.