Classical population genetics shows that varying permutations of genes and risk factors permit or disallow the effects of causative agents, depending on circumstance. For example, genes and environment determine whether a fox kills black or white rabbits on snow or black ash covered islands. Risk promoting effects are different on each island, but obscured by meta-analysis or GWAS data from both islands, unless partitioned by different contributory factors. In Alzheimer's disease, the foxes appear to be herpes, borrelia or chlamydial infection, hypercholesterolemia, hyperhomocysteinaemia, diabetes, cerebral hypoperfusion, oestrogen depletion, or vitamin A deficiency, all of which promote beta-amyloid deposition in animal models—without the aid of gene variants. All relate to risk factors and subsets of susceptibility genes, which condition their effects. All are less prevalent in convents, where nuns appear less susceptible to the ravages of ageing. Antagonism of the antimicrobial properties of beta-amyloid by Abeta autoantibodies in the ageing population, likely generated by antibodies raised to beta-amyloid/pathogen protein homologues, may play a role in this scenario. These agents are treatable by diet and drugs, vitamin supplementation, pathogen detection and elimination, and autoantibody removal, although again, the beneficial effects of individual treatments may be tempered by genes and environment.

1. Introduction

If there is one factor common to complex polygenic diseases it is the heterogeneity in both gene and risk factor association studies.

Although these have discovered key genes and risk factors, the results for most are invariably confounded by conflicting data [1]. In the genetic arena, the clear familial component of many diseases has driven the search for major genes using genome-wide association studies (GWAS) with large numbers of patients pooled from different regions [2]. Such studies have been able to discover rare variants that play a major role in a small percentage of patients, for example VIPR2 in schizophrenia [3]. However, in complex diseases, these have failed to find major genes relevant to all patients [4], instead unearthing yet more genes of small effect, whose risk promoting effects are yet again contested, as is the case with CR1 and PICALM, which have not been confirmed as risk factors for Alzheimer’s disease in Chinese patients [5] despite extensive evidence in Caucasian studies [6]. GWAS studies have, however, been more successful in uncovering larger numbers of genes of greater effect for simpler traits such as lipid levels [7].

Viruses and other pathogens have been implicated as risk factors in many diseases, although again, conflicting evidence leads to scepticism in many areas. For example, the involvement of the Epstein-Barr virus in multiple sclerosis is hotly contested [810].

Gene-gene and gene-environment interactions may play an important role in such inconsistency. For example, the risk promoting effects of genes can be better explained when using pathway analysis or combining the effects of genes with common function, rather than by studying single genes in isolation [11, 12]. Genes and risk factors can also act together, and in certain cases genes can be linked to environmental variables. For example, many of the genes implicated in schizophrenia or Alzheimer’s disease are involved in the life cycles of the pathogens involved in the diseases [13, 14]. Environment-environment interactions are also apparent. For example, the effects of vitamin E on lifespan, or on resistance to various infections can be null, deleterious, or protective, depending on confounding factors such as age, exercise, smoking, and vitamin C consumption [1517].

Complex diseases are also composed of many endophenotypes or underlying pathologies, and different genes or risk factors may contribute to any of these. Many different processes contribute to cell death in Alzheimer’s disease, for example, beta amyloid, glutamate, calcium, or free radical mediated toxicity [18, 19]. The efficiency of each of these subprocesses is controlled by genes, many of which have been implicated in association studies (see Table 1).

In genetic association studies, the drive has been to increase statistical power by increasing the numbers of subjects enrolled. This has resulted in the discovery of important genes and rare genetic variants, but has not delivered genes that confer a high degree of risk in the majority of patients. However, as illustrated below, more could perhaps be gained by a reanalysis of existing data in relation to other genetic and risk factor variables that could result in elucidation of the causes rather than the risks.

2. Methods

Alzheimer’s disease susceptibility genes and risk factors are stocked in an online database at http://www.polygenicpathways.co.uk/alzenvrisk.htm. KEGG pathway analysis of over 400 associated genes was performed [20], and the results of the exercise were posted at http://www.polygenicpathways.co.uk/alzkegg.htm. In these figures, yellow genes have been implicated in Alzheimer’s disease and red genes are also implicated in the herpes simplex life cycle. Other gene-risk factor relationships were identified by literature survey. B cell and T cell epitopes within the beta-amyloid peptide were identified using the immune epitope database server http://tools.immuneepitope.org/main/index.html which predicts the antigenicity of peptide sequences, based on their charge and hydrophobicity properties [21]. Sequence comparisons of the beta-amyloid peptide versus selected bacterial, fungal, or viral proteomes was performed using the NCBI blast server [22, 23].

3. Results and Discussion

3.1. Kegg Pathway Analysis of Alzheimer’s Disease Susceptibility Genes

The overall results of this analysis are shown in Table 1. The pathways include many that are relevant to the known pathologies and risk factors of Alzheimer’s disease, including the Alzheimer’s disease pathway itself, primarily related to beta-amyloid and tau processing, but also to glutamate-related pathways (long-term potentiation and depression), apoptosis, insulin and diabetes pathways, neurotrophin signalling, oxidative stress (glutathione/oxidative phosphorylation), cerebral hypometabolism (oxidative phosphorylation, glycolysis and the Krebs cycle), arginine and proline metabolism (including nitric oxide), and folate, methionine and homocysteine metabolism, and steroid hormone synthesis (together with androgen and oestrogen receptors AR, ESR1, and ESR2). PPAR signalling regulates many lipoprotein-related genes and cholesterol/lipid pathways are dispersed in terpenoid backbone biosynthesis (FDPS, HMGCR, HMGCS2), steroid hormone biosynthesis (HSD11B1), steroid biosynthesis (DHCR24, LIPA, SOAT1), glycerolipid metabolism (ALDH2, LIPC, LPL), and bile acid biosynthesis (CH25H, CYP46A1) pathways. Immune, complement, and cytokine-related pathways figure prominently, as do several pathogen defence pathways including the DNA sensing retinoic acid inducible gene (RIG-1) pathways that react to viral DNA/RNA by increasing the expression of interferons and other antiviral genes, and the Toll receptor and NOD pattern recognition pathway that control immune and cytokine networks [2426]. Glutathione pathways were also present. Glutathione has potent viricidal and bactericidal properties and is often depleted by infections [2730]. A number of pathogen entry pathways are also concerned, and although C. Pneumoniae or C. Neoformans pathways are not specifically represented, many of these pathways can be considered as generic pathways relevant to many bacteria and other pathogens. The H. Pylori pathway contains only two genes, IL8 and CSK, but others can be added to this list including all members of the PI 3-kinase/AKT signalling network which is activated by the H. pylori protein CagA [31] or Toll receptor pathways that are activated by H. Pylori heat shock protein, HSP60 [32]. Similarly, there is no specific HSV-1 viral entry pathway in the KEGG database, but the virus uses actin pathways, endocytosis, protein processing, and DNA repair pathways during its life cycle, which are heavily represented [33].

3.2. Vitamin-A-Related Genes

These were identified by literature survey and the most directly relevant are shown in Table 2. Of particular interest is a close relationship between cholesterol/lipoprotein-related genes and vitamin A. Both retinols and cholesterol are transported by lipoproteins, and the clusterin receptor, LRP2/megalin, is a key retinol entry point. APOE4 is the isoform least able to bind to retinyl palmitate. ABCA1 is also involved in cholesterol and retinol transport. Several genes (ALDH2, CYP46A1, GSTM1, GSTP1, LIPA, LPL, and LRAT) are involved in Vitamin A metabolism, and 24-s hydroxycholesterol, the product of CYP46A1, is a ligand for retinoic acid receptors (RARA and RARG). Retinoid coreceptors and binding partners include RXRA, ESR1, KLF5 NPAS2, NR1H2, PARP1, PIN1, POU2F1, PPARA, PPARG, THRA, UBQLN1, and VDR. Retinoids modulate APP processing, via regulation of beta and gamma secretases while the RIG-1 pathway is crucial in viral defence. A large number of genes are also regulated by retinoids or retinoid receptors.

3.3. The Fox, the Rabbits, Gene, and Environmental Variables

This is an adaptation of Lees’s classical population genetics example of the peppered moth, whose dark or pale colouring confers advantage or disadvantage depending upon the degree of industrial pollution that covers trees with soot. It has served to illustrate the concept of natural selection where, over time, dark genes become more common in polluted areas, an effect that could eventually lead to speciation [34].

On two islands one covered in snow and the other in black volcanic ash live an equal number of black and white rabbits and a family of foxes, who will find it easier to trap the black rabbits on the snowy island and the white rabbits on the island covered in black ash. Gene association studies would correctly identify the black and white genes as being protective or risk promoting depending upon the environment. The snow, the ash, or the fox, being equally present on each island, regardless of the toll of dead rabbits, could not be considered as risk factors. Genetic meta-analysis or pooled GWAS data would also rule out any genetic involvement, leaving no susceptibility genes, no risk factors, and no cause. However, a GWAS study, apportioning the genetic data in relation to ash, snow, and fox would be able to correctly surmise that the white gene is a risk factor on the ash-covered island, and the black gene a risk factor on the snowy island, as would have D. R. Lees. Again the fox is undetectable, being present in all compartments.

On other similar islands, live further populations of black and white rabbits with no fox, an equal number of deaths due to old age, and no reason to investigate either genes or risk factors. However, it is only by including this island, again partitioning GWAS data in relation to all variables, that the genes, the risk factors, and the cause can be correctly allotted their respective roles.

In this example, the genes and environmental variable are risk or protective factors for the cause as well as for the deaths, depending on circumstance. The genes or risk factors are not killing the rabbits, but are allowing the cause to do so. Nonstratified association studies would thus seem to be ill-suited to find important genes, risk factors, or causes, and the pursuit of greater statistical power may well be futile, although such strategies can find rare variants that may cause disease in a minority of patients, which is evidently useful. However, for the majority of cases, much could perhaps be gained from a reappraisal of existing data and by partitioning GWAS data in relation to the many known risk factors in each disease.

The situation is evidently more complex in polygenic diseases, where hundreds of interacting genes, many risk factors, and probably many causes are present. This is already appreciated, and several groups have analysed the statistical problems involved due to the mass of genes and risk factors [3537]. However, it is likely that an appropriate selection of genes and risk factors could markedly affect the degree of risk. For example, the odds ratio for APOE4 was shown to be 1.67 in Alzheimer’s disease patients without cerebral HSV-1 DNA and 16.8 in patients where viral DNA was detected [38]. The population genetics example, and the discussion below, suggests that certain susceptibility genes are restricted to risk factor subsets.

From the above, it would appear that a cause can be present in equal proportion in control and disease populations but should be able to produce the pathological features of the disease, and the disease incidence should be reduced where the causes are few. A cause can kill regardless of the genes (black or white) or risk factors (snow or volcanic ash) but its effects are tempered by a combination of the two (fox + snow + black gene or fox + ash + white gene = death and fox + snow + white gene or fox + ash + black gene, or no fox and any combination = life). The genes and risk factors are, however, both able to influence the cause.

In relation to Alzheimer’s disease, beta-amyloid deposition can be produced by herpes simplex [189191] or chlamydia pneumoniae infection [192, 193], hypercholesterolemia (which also causes cholinergic neuronal loss and memory deficits in rats [194197]), by or hyperhomocysteinaemia, an effect reversed by folate and vitamin-B12 [198], by NGF deprivation [199], by reduced cerebral perfusion (hypoxia, cerebral ischaemia, or carotid artery occlusion [200203]), as well as by experimental diabetes and streptozotocin [204, 205], oestrogen depletion [206], or vitamin A deficiency, which also reduces choline acetyltransferase activity in the forebrain [207]—all without the aid of any variant genes, in animal models. While none of the animal models faithfully reproduce the entire symptomatology of Alzheimer’s disease, in the clinical setting these risk factors coexist to varying degrees, rendering the situation rather more complex.

The risk factors in Alzheimer’s disease include herpes simplex infection [208] acting in combination with APOE4 [209], C. Pneumoniae [210, 211] or Helicobacter pylori infection [212, 213], mild hypercholesterolaemia [214], but declining cholesterol levels from midlife to late life [215], atherosclerosis of the carotid arteries, leptomeningeal arteries, or the circle of Willis, and stroke, leading to cerebral hypoperfusion [216219], hyperhomocysteinaemia [220], type 2 diabetes and modified insulin metabolism [221], age-related loss of sex steroid hormones in both women and men [222], but high total oestradiol levels [223, 224], and vitamin A deficiency [225]. These factors may be confounded by interrelationships, and in some cases by the fact that death due to other causes—for example, atherosclerosis-related myocardial infarction or stroke may lead to an apparent paucity of comorbid risk factors in Alzheimer’s disease groups at later ages.

The genes implicated in Alzheimer’s disease are related to the herpes simplex life cycle [13, 226], bacterial and viral entry pathways, viral and pathogen defence (Table 1), and the immune network [227], cholesterol and lipoprotein pathways [11, 228], folate and homocysteine pathways [229], and insulin, or neurotrophin signalling pathways, steroid metabolism and receptors (Table 1 and 2), and vitamin A metabolism and function (Table 2).

The genes, risk factors and agents known to increase beta-amyloid deposition all concur, suggesting that Alzheimer’s disease is multifactorial with many foxes, each with their respective genes and risk factors, any of which can lead to beta-amyloid deposition in multifarious ways. Each risk factor can act independently of any gene or other risk factor variant, in animal models—as with the fox. This in turn might suggest that it is not the risk promoting polymorphisms in the Alzheimer’s disease patients that are crucial, as the risk factors can in any case promote beta-amyloid deposition, but the equivalent polymorphisms in the control group, that are providing protection; a subtle distinction that awaits characterisation of the functional effects of many different gene variants. The reasoning also suggests that beta-amyloid deposition is the consequence and not the cause of the many factors able to promote Alzheimer’s disease. Anoxia, ischaemia, hypoglycaemia, hypercholesterolaemia, and vitamin A deficiency are all able to kill neurones, in some cases including cholinergic neurones, without the aid of beta amyloid.

3.4. Low Incidence of Alzheimer’s Disease and Protective Factors

In relation to Alzheimer’s disease, there is an island where longevity is increased, related to the nun study [231234]. Nuns do not have children, (the number of pregnancies is a risk factor in Alzheimer’s disease [235]), do not consume high concentrations of saturated fats (low cholesterol), and are unlikely to have sexually transmitted diseases or viral and other common pathogen diseases vectored by childhood infections (Herpes and chlamydia, inter alia). Their vitamin A levels and general health are sustained by a healthy diet, regular fish on Fridays, and exercise.

There are few strategies that have been shown to reduce the severity of Alzheimer’s disease, once established. It has, however, been shown that Helicobacter pylori elimination increases the cognitive abilities and the lifespan of Alzheimer’s disease patients [236]. In addition, two separate case reports have shown complete reversal of dementia in two patients diagnosed with Alzheimer’s disease, by identification and eradication of the fungal pathogen Cryptococcus Neoformans [237, 238].The TNF antagonist, etanercept, has also been reported to produce a striking remedial effect on symptomatology, following perispinal application [239]. However, the use of TNF antagonists is also associated with an increased incidence of opportunistic bacterial, fungal, and viral infestations, including cytomegalovirus, and cryptococcal infections [240], perhaps a contraindication for their prolonged use.

A number of epidemiological studies have shown that the incidence of Alzheimer’s disease can be reduced, although, once the disease is established, there is little evidence for any curative effects of any treatment other than the above. These protective factors are in most cases the obverse of the risk factors and include diets rich in fish or polyunsaturated fatty acids [241, 242], the Mediterranean diet [243] and the use of statins [244], which are counter to the effects of high cholesterol. A diet rich in fruit and vegetables is associated with reduced dementia incidence [245] and is able to sustain Vitamin A levels and reduce homocysteine levels in the elderly population [246]. High folate intake, which reduces homocysteine levels [247], and the use of nonsteroidal anti-inflammatories have also been reported to reduce risk [248]. Again these are related to the risk factors and to the genes, which may condition their success (cf. Vitamin E).

3.5. Relevance of These Factors to the Genes Identified in Genome-Wide Association Studies

Four major genes have been discovered prior to and from GWAS studies, APOE4, clusterin, complement receptor 1, and PICALM [6, 249]. The close relationships between these genes and herpes simplex infection have been the subject of a previous article [226]. APOE4 also favours the binding of C. Pneumoniae elementary bodies to host cells [250]. It is also a risk factor for hypercholesterolaemia, per se [251], and for carotid artery atherosclerosis in men with diabetes [252]. APOE4 is also the isoform least able to promote lipid efflux from neuronal cells [253], a factor that may enhance the cholesterol dependent cleavage of beta-amyloid by beta and gamma secretase [254]. It is also the least able isoform binding the vitamin A precursor retinyl palmitate [255] (see below). Complement receptor 1 is a pathogen receptor for both herpes simplex [256], and C. Neoformans [257] and also for the atherogenic pathogen, P. Gingivalis [258], a key cause of periodontitis/gum disease, which has also been implicated as a risk factor in dementia [259]. Both Helicobacter pylori and C. Pneumoniae [260] use the mannose-6-phosphate IGF2 receptor (inter alia) for entry. This binds to clusterin and its endocytosis is controlled by PICALM [226]. PICALM (phosphatidylinositol binding clathrin assembly protein), as its name implies, is involved in clathrin-mediated endocytosis [261], a process used by C. Pneumoniae to gain cellular entry [262], for the internalisation of herpes simplex glycoprotein D [263] or the cytomegalovirus chemokine receptor [264] and for the uptake of outer membrane vesicles from certain strains of H. pylori, into gastric epithelial cells [265]. Clathrin also associates with HHV-6 virions [266]. Clusterin is an inhibitor of the membrane attack complex that is inserted into microbial membranes causing death by lysis [267]. It is also a ligand for the retinol/lipoprotein receptor LRP2 [268], and the gene contains a retinoid response element (Table 2). Thus it would seem that the key role of these genes may be related to their ability to target multiple aspects of diverse risk factor networks.

3.6. Relationships between Risk Factors and a Key Role for Herpes Simplex Activation (Figure 1)

Hypercholesterolaemia can evidently be related to other dietary risk factors such as saturated fat consumption, and to atherosclerosis. Docosahexaenoic acid increases total plasma cholesterol levels in hymans, but only in APOE4 carriers, an effect that may negate the cardioprotective effects of fish oil supplementation [269]. Helicobacter pylori infection also causes malabsorption of vitamin B12 and folate, leading to increased homocysteine levels, that can be restored by H. pylori eradication [270]. Homocysteine metabolism is also related to glutathione synthesis via the transsulfuration pathway (homocysteine→cysteine→glutathione). Increased levels of homocysteine and reduced levels of glutathione in Alzheimer’s disease suggest impairments in the transsulfuration pathway [271]. Glutathione is a potent antiviral and bactericidal agent with effects targeted at herpes simplex and C. Pneumoniae, inter alia both of which also diminish glutathione levels in infected cells [27, 272, 273]. H. pylori expresses an enzyme, gamma-glutamyltranspeptidase that enables it to metabolise the host’s extracellular glutamine and glutathione which are hydrolysed to glutamate, which is fed into the H. pylori Krebs cycle, resulting in diminished glutathione levels that can be restored by H. pylori elimination [27, 273]. Glutathione levels are reduced in Alzheimer’s disease and many others [274].

In both coronary artery disease and carotid artery atherosclerosis, high plasma levels of homocysteine are positively correlated with C. Pneumoniae seropositivity suggesting a role for the bacterium in promoting high homocysteine levels. [275, 276]. Indeed carotid artery atherosclerosis is correlated with antibodies to C. Pneumoniae, and to a lesser extent with antibodies to H. pylori, but in this case, only in individuals with low social status [277].

The growth of C. Neoformans is attenuated by diethylstilbestrol and oestradiol but not by progesterone or testosterone [278]. Helicobacter pylori adsorbs a number of steroids including pregnenolone and two androgens (dehydroepiandrosterone and epiandrosterone and 3-hydroxylated oestrogens (oestrone and oestradiol). These are glucosylated and the glucosyl-steroid hormone derivatives used as membrane lipid components [279]. Oestradiol, androstenedione, and progesterone are all able to inhibit the growth of H. pylori [280].

These complex interactions, of which there are likely many more, suggest that in addition to epistasis and gene-environment interactions, environment-environment interactions have to be factored in to an already complex equation (cf. vitamin E).

Factors known to reactivate herpes simplex include heat [281], 17-beta oestradiol [282] and the inflammatory cytokine interleukin 6 [283] where a role for corticosterone has been proposed [284]. NGF deprivation [285] also reactivates the virus and NGF promotes viral latency via the TrkA receptor [286] (cf. neurotrophin signalling). Vitamin A supplementation in rats increases the cerebral levels of both NGF and BDNF [129] while oestrogen deficiency lowers cerebral NGF levels, an effect reversed by 17-beta oestradiol [287]. Transient cerebral ischaemia lowers NGF levels [288]. Hypoxia is also able to increase the replication of herpes simplex [289].

Fevers induced by diverse infections might thus be expected to reactivate herpes simplex, as well as cerebral hypoperfusion. IL6 plasma and CSF levels have been reported to be increased in Alzheimer’s disease and the secretion of IL6 from monocytes is increased [290292]. IL6 plasma levels are raised by infection with C. Pneumoniae [293] or Helicobacter pylori [294], and IL6 production in monocytes is stimulated by C. Neoformans [295]. Cortisol levels are also increased in the ageing population and in Alzheimer’s disease [296, 297]. High levels of total oestradiol have been reported as a risk factor for Alzheimer’s disease in both women and men [223, 224].

As so many other risk factors seem able to reactivate the virus, this may be a key precipitant for the final curtain. Herpes simplex viral DNA is present in Alzheimer’s disease plaques [298], and the plaques and tangles in Alzheimer’s disease contain a very high proportion of herpes simplex interacting proteins, as well as immune-related components. The presence of the complement membrane attack complex in neurones suggests that the neuronal destruction in Alzheimer’s disease might well be related to the consequences of battle between the virus and the immune network that has eliminated the virus at a terrible cost of collateral neuronal damage [33]. IgM+ antibodies, which preferentially index HSV-1 reactivation, have been shown to be able to predict the future risk of developing Alzheimer’s disease [208], and the ability of other risk factors, particularly other infections, to reactivate the virus suggests a complex interplay of genes and risk factors that funnel towards viral reactivation and plaque and tangle formation.

3.7. The Importance of Vitamin A

Low vitamin A levels are a problem in the ageing population, and even in successfully ageing persons can be observed in 50% of the population over the age of 80–85 [299] (cf. the fox). Low vitamin A levels are also a risk factor for Alzheimer’s disease [225]. Vitamin A plays an important role in maintaining the immune system [300], many genes of which are implicated in Alzheimer’s disease. APP is involved in the vitamin A arena, as a gamma57 gamma secretase cleavage product suppresses retinoid signalling [75].

The vitamin A derivative, retinoic acid, inhibits herpes simplex replication [301, 302] as well as chlamydial infection and growth [303]. Vitamin A also stunts the growth of Helicobacter pylori [304]. The effects of vitamin A on C. Neoformans do not appear to have been examined. However, glucuronoxylomannan, the polysaccharide component of the capsular material of cryptococcus neoformans, exhibits potent immunosuppressive properties. This compound downregulates TNF-alpha and IL-1beta, and upregulates the inhibitory cytokine IL-10, but also inhibits retinoic receptor (RORG) synthesis [305]. Retinoic acid is also able to lower plasma homocysteine levels via the induction of hepatic glycine N-methyltransferase. Homocysteine in contrast inhibits retinoic acid synthesis [306, 307].

Vitamin A levels are in fact higher in hypercholesterolemia patients [308]. This may perhaps be due to the fact that retinyl palmitate, the vitamin A precursor, like cholesterol, is also transported by lipoproteins in the blood, mainly in the VLDL fraction (which primarily consist of APOC2/APOE) and the LDL fraction (which primarily consists of APOB) [41]. Retinyl palmitate concentrations in the blood are affected by APOE polymorphisms and radiolabelled retinyl palmitate binding in total plasma and nonchylomicron fractions is least in APOE4+/+ carriers [255]. The aortic concentrations of triglycerides, total cholesterol, free and esterified cholesterol, and phospholipids are increased in vitamin A deficient rats, an effect reversed by vitamin A supplementation [144]

A large number of Alzheimer’s disease genes are related to vitamin A (Table 2), and an even larger number responsive to retinoic acid via the action of RAR or RXR transcriptional control.

3.8. Autoantibodies to the Antimicrobial Peptide Beta Amyloid: Likely Derivation from Antibodies to Pathogens

Beta-amyloid is a potent antimicrobial peptide. Although not tested against C. Neoformans, H. pylori, or C. Pneumoniae, it was found to have broad spectrum activity against a variety of yeasts and bacteria, effects that were attenuated by anti-Aβ antibodies. [309]. Lactotransferrin is also an antimicrobial peptide that colocalises to plaques and tangle in the Alzheimer’s disease brain [310] and other antimicrobial peptides include the susceptibility gene products cystatin C, defensin DEFB122, myeloperoxidase, and transferrin [88]. beta-amyloid, like acyclovir, also attenuates the stimulatory effects of HSV-1 on miRNA-146a levels in neuronal cells [311]. The antimicrobial and antiviral properties of beta amyloid are, however, likely to be abrogated by the presence of beta-amyloid antibodies in the sera of the ageing population [312] and in Alzheimer’s disease [313].

As shown in Table 3, a number of pathogens implicated in Alzheimer’s disease or its attendant risk factors, express proteins with a high degree of homology to beta-amyloid. These include proteins from HSV-1, HHV-6, the cytomegalovirus, C. Neoformans, H. pylori, C. Pneumoniae, B. Burgdorferi and, P. Gingivalis. Antibodies to B. Burgdorferi [314] and C. Pneumoniae [210] and Herpes simplex and HHV-6 viral DNA [298, 315] have all been reported in or around Alzheimer’s disease plaques, and antibodies to H. pylori recovered from Alzheimer’s disease serum and cerebrospinal fluid [316]. A recent study showed that P. Gingivalis antibodies, cross-reactive with human HSP60, were observed in 100% of a sample of 20 atherosclerosis patients [317]. The immune system is not designed to raise antibodies to a self-protein, and this high degree of homology suggests that the autoantibodies to beta-amyloid are created by antibodies to homologous pathogens’ proteins.

Several other autoantibodies have been reported in Alzheimer’s disease, including targets such as nerve growth factor [318], cholinergic neurones [319], the choroid plexus [320], and neurofibrillary tangles, inter alia [321]. The very extensive sharing of viral and bacterial protein sequences with the human proteome [230, 322324] suggests that these too might be derived from cross-reactive microbial antigens. Again the ability to create autoantibodies is conditioned by genes, particularly HLA-antigens [325, 326]. Autoantibodies are often regarded as an epiphenomenon, but their ability to traverse the blood-brain barrier [327], and the recent recognition of their ability to enter cells [328] casts them in a new light as pathological immunopharmacological agents able to block the function of their target proteins: this has indeed been shown for Alzheimer’s disease-derived ATP synthase autoantibodies which block ATP synthesis and cause apoptosis in neuroblastoma cells [329].

4. Summary

The classical population genetics example of the foxes and rabbits illustrates how genes and risk factors can differentially permit or disallow the effects of a causative agent, depending on a permutation of circumstance. Applying this model to Alzheimer’s disease also suggests that many of the environmental risk factors in Alzheimer’s disease are in fact causative agents, at least in terms of an ability to produce beta-amyloid deposition, per se, as shown in nontransgenic animal models. Their effects are also clearly related to other risk factors and genes. This would infer that the susceptibility genes in Alzheimer’s disease permit the actions of these agents but, perhaps more importantly, that polymorphisms in the control population do not. Functional characterisation of these control variants may provide important clues to overall methods of protection.

Many of these risk factors are avoidable or amenable to therapy. Diet is already known to be an important risk/protective factor with regard to the incidence of Alzheimer’s disease. For example there is clear evidence that the omega-3 fatty acid docosahexaenoic acid (DHA), a component of fish and the Mediterranean diet, also protective factors [243, 330], is associated with a reduced risk of dementia [331], although a recent study with DHA in mild to moderate Alzheimer’s disease, failed to show disease arrest or diminution [332]. However, in agreement with epidemiology, DHA significantly benefited two measures of cognition in mild to moderate non-ApoE4 carriers [333]. High vitamin A and low homocysteine levels are related to a high intake of fruit and vegetables in elderly patients [246]. Fruit and vegetable juice consumption is also associated with reduced Alzheimer’s disease incidence [334].

Given the fact that the potential causes of Alzheimer’s disease appear to be multifactorial, perhaps a multifactorial therapeutic effort is also needed. Such approaches might include a greater attention to diet, homocysteine and cholesterol levels, vitamin A supplementation where necessary, and the regular detection and elimination of herpes simplex, B. Burgdorferi, C. Pneumoniae, H. pylori, C. Neoformans, and other pathogens in the ageing population. The removal of beta-amyloid antibodies and of others prevalent in Alzheimer’s disease might also be of benefit. These are simple preventive measures, requiring public health attention, whose corporate instigation might markedly reduce the incidence, and perhaps halt or reverse the progression, of Alzheimer’s disease.