Putative Survival Advantages in Young Apolipoprotein ɛ 4 Carriers are Associated with Increased Neural Stress

INTRODUCTION

Alzheimer’s disease (AD) is a pathological condition adversely affecting the brain. AD has a large genetic component. The neuropathology is characterized by the presence of neurofibrillary tangles (NFTs) and neuritic plaques [1, 2]. The modern conceptualization of AD pathology dates from 1968 when Blessed et al. showed in elderly individuals that the NFTs correlated with the severity of the dementia, though the neuritic plaques, which contain cerebral amyloid-β (Aβ), did not [2]. A major advance in understanding AD pathology was the demonstration that AD pathology predominantly affects the posterior-temporal, inferior parietal, posterior cingulate, and medial temporal region [3], with a characteristic pattern of progression beginning in the entorhinal cortex involving neurofibrillary (NF) and microtubule associated protein-tau pathology rather than senile plaque and cerebral Aβ pathology [4], consistent with the earlier findings of Blessedet al. [2].

Dementia is a condition involving impairment of cognitive function, which has deteriorated from a prior higher level, and causes social impairment. There are a few types of dementia, mostly frontotemporal (rare) and temporal-parietal (common). Patients with temporal-parietal dementia subdivide into approximately one third with pure AD, one third not AD, and one third being mixed, with the prevalence of mixed cases increasing with age. This progression of AD (NF/NFT/tau pathology) is clearly reflected by both altered cerebral blood flow [5] and cerebral metabolism [6, 7], as well as by PET tracers targeting tau [8, 9], with characteristic regional and stage-specific variations [9]. These changes in the brain are closely related to the impairments of memory function that are so typical of the dementia associated with AD and its progression [10, 11]. This pattern has strongly suggested that AD pathology selectively attacks those neuroplastic brain systems which perform the functions of episodic memory [12, 13]. The neuritic plaques and Aβ are consistent components of AD pathology [14], which is the predominant cause of dementia. However, the distribution pattern of Aβ pathology, which is found at least as early and diffusely in the neocortex as the tau pathology, is found first in most regions of the neocortex [15] but is generally not or much less related to cognitive changes than the tau pathology [16 –18]. With a slowly progressing condition, mild cognitive impairment is a transition from normal cognition that precedes dementia, and is very poorly described [19].

Previously, we hypothesized that apolipoprotein E (APOE) ɛ4 allele-associated AD risk is consistent with increased lifetime exposure to a neurotoxic process [20]. Specifically, if the hippocampal neurons of two individuals possess the same susceptibility to an endogenous or exogenous stress factor, the neurons with the highest turnover of proteins, lipids, and other macromolecules might experience a larger integrated dose of detriment. Small differences in pharmacokinetic effects might be amplified by the extremely long pre-symptomatic phase of AD, i.e., average age of presentation for a homozygous ɛ4 is about 68 years of age [20]. Studies conducted across the age spectrum from infancy through senescence have suggested that APOE ɛ4-positive status is associated with increased brain activity and macromolecule turnover in young healthy individuals, with the reverse extant in elderly subjects. In the current study, we extend our analysis from the limited number of studies examined in Smith and Ashford [20] and attempt herein to comprehensively examine the literature on reported associations between clinical conditions, cognitive performance, and presence or absence of the ɛ4 allele in otherwise healthy young subjects.

The literature on associations between ɛ4 status in otherwise healthy young subjects falls into several categories including: cardiovascular responses (Table 1); reproduction and development (Table 2); co-morbidities (Table 3); resistance to infectious disease (Table 4); responses to head injury (Table 5); biochemical differences possibly related to neural stress (Table 6); brain structure-function (Table 7); and mental performance (Table 8). Possession of the ɛ4 allele is a strong risk factor for development of AD in the elderly, and all individuals possessing the ɛ4 allele are increasingly likely to develop AD the older they live, relative to those without this allele. An understanding of the clinical conditions and cognitive performance characteristics idiosyncratic to healthy young persons who possess the ɛ4 allele might assist in the complex task of disentangling the relative contributions of genetics and lifestyle that appear to play a role in how and when a given individual develops AD [21].

The mechanism through which inheritance of an allele of a protein associated with cholesterol metabolism, i.e., ɛ4, exerts an increased risk for the development of AD later in life is not understood. One current hypothesis emphasizes that vulnerability to AD is based on the very high rate of formation and removal of synapses in the brain. In healthy individuals, the number of synapses formed and actively removed is estimated to be one trillion per day, presumably associated with constant learning and establishment of new memories. The amyloid-beta protein precursor (AβPP) plays a central role in this aspect of neuroplasticity [13]. A considerable amount of evidence has related AβPP to AD [22], including a relationship between its Aβ product and APOE genotype [23]. AβPP can be cleaved either by the α-secretase, which leads to increased local synapse production and prevents formation of the synapto-toxic Aβ molecule or by the β-secretase, followed by γ-secretase, which leads to the removal of the affected synapse [22]. The control of the delicate balance between synapse creation and destruction is critical for learning and involves numerous mechanisms, with a positive neuronal action culminating in activation of the α-secretase, while activation failure leads to the default action, β-secretase cleavage, then γ-secretase cleavage of the AβPP, with the end-products being fatal to that synapse. Note that the γ-secretase appears to act only on the β-secretase product and not on the α-secretase product, producing both the Aβ and an AβPP intracellular domain (AICD) [24]. The AICD protein, acting intracellularly, leads to hyperphosphorylation of the critical structural protein of neuronal processes, microtubule associated protein tau [25]. The hyperphosphorylated tau appears to disrupt the normal flow in neuronal processes, leading to amputation [26]. The initial memory impairment of AD is likely due to the failure of the process for establishing new memory, while the dementia appears to develop when synaptic slaughter destroys large numbers of synapses [27], and loss of synapses is the factor most closely associated with the late progression of dementia [28]. Cholesterol metabolism, transport and processing might play a role in this process by affecting the composition of the neuronal synaptic membrane thereby altering its thickness. The extracellular region of AβPP is much longer than the intracellular region. Small differences in the availability of enzymatic cleavage sites influenced by membrane thickness might have profound effects on the ultimate production of AβPP sub-fragments [29, 30].

IMPORTANCE OF THE APOE GENOTYPE

The APOE genotype accounts for the vast majority of AD risk [31, 32]. There are three common alleles of the APOE gene, i.e., APOE ɛ2, APOE ɛ3, and APOE ɛ4 [33]. In the general US population, the ɛ4 allele prevalence is approximately 13% [34], constituted by about 2% being ɛ4 homozygous (2% of the US population) and 11% being heterozygous (22% of the US population). Possession of one ɛ4 allele increases the risk of developing AD by 3 to 4-fold, and possession of two ɛ4 alleles increases risk by 15-fold, as compared with the ɛ3/ɛ3 genotype, with a large part of the variation being related to substantially early age of onset, and over 60% of patients with non-familial AD carry the ɛ4 allele [35, 36].

This profound difference in AD risk results from only minor changes in the structure of the APOE molecule. The three isoforms of APOE differ in amino acid sequence at only chain positions 112 and 158: the APOE ɛ2 allele has cysteine at both positions; the APOE ɛ4 allele has arginine at both positions; and the APOE ɛ3 allele has cysteine at position 112 and arginine at 158 [37]. These small changes in amino acid sequence alter the biological activity of the APOE proteins in multiple ways, one of which is increased liver catabolism of the APOE ɛ4 lipoprotein as compared with the APOE ɛ3 lipoprotein [38, 39].

The strong positive association between possession of the ɛ4 allele and the development of late-onset AD has stimulated extensive investigation on young, healthy subjects differing in APOE allele status. Numerous ɛ4 allele-related associations have been reported in a series of epidemiological and clinical investigations on a variety of conditions, with the potential relevance of these associations to the pathogenesis of AD poorly understood at this time. Over the last several decades, numerous and elaborate interactions have been demonstrated between the nervous, immune, and endocrine systems [40]. This sophisticated system of sometimes centrally mediated biochemical cross-talk opens the possibility that an association between possession of the ɛ4 allele and a particular clinical characteristic, condition, susceptibility or outcome might provide mechanistic information as to why allelic variation in a cholesterol metabolism gene is so strongly associated with the loss of synapses in the brain.

In primarily young subjects, we have attempted to review a representative body of literature on ɛ4 allele-associations related to the following: cardiovascular responses; impacts on reproduction and fetal development; co-morbidities; resistance to infectious disease; responses to head injury; biochemical differences possibly related to neural stress; and brain structure-function differences. In addition, the literature on the association between the ɛ4 allele and cognitive performance has been reviewed comprehensively. In each summary table, the ethnic composition, age, clinical diagnosis, and number of ɛ4 carriers of the cohort studied are included.

LITERATURE SEARCH STRATEGY

Using Google and PubMed of the National Library of Medicine, the search terms apolipoprotein E or ɛ4 allele were cross-matched against the terms: infant, youth, child, children, adolescent, or young. This initial search gathered a large number of potentially relevant citations that were categorized by clinical condition including cardiovascular responses; reproduction and development; co-morbidities; resistance to infectious disease; responses to head injury; biochemical differences possibly related to neural stress; brain structure-function; and mental performance. Each citation was physically collected, and the reference sections were manually checked for additional relevant citations until no further citations were found. No papers on the relevant topics were excluded. A few studies on middle age subjects not yet presenting with mild cognitive impairment were included with the aberrant age ranges clearly marked in the tables. The studies in the tables were not amenable to meta-analysis as the protocols were too heterogeneous. There were four referenced articles ([15 , 130]) that included a meta-analysis. Some of the individual papers in the meta-analyses were also separately included in the review because the details in the individual papers were important.

METHOD FOR GRADING QUALITY OF EVIDENCE

As noted previously, the literature on associations between ɛ4 status in otherwise healthy young subjects can be subdivided into several categories including: cardiovascular responses (Table 1); reproduction and development (Table 2); co-morbidities (Table 3); resistance to infectious disease (Table 4); responses to head injury (Table 5); biochemical differences possibly related to neural stress (Table 6); brain structure-function (Table 7); and mental performance (Table 8). Each table contains seven headings including sequentially: Subject/Age; Sample Size; Sex; Study Scope; Major Finding; Citation; and Quality of Evidence. The first six headings summarize the study characteristics, its results, and citation. The seventh heading provides the reader an estimate of the quality of evidence in the study based on a recognized evidence-based rating system.

A system developed by the Grading of Recommendations Assessment, Development and Evaluation Working Group (GRADE) for ranking the quality of evidence and the strength of recommendations of scientific literature and clinical practice guideline was applied to rate the studies found in Tables 1–8 [41]. The method described by Adkins et al. [41] was used to estimate the degree of confidence that can be placed on the evidence from research and clinical studies. The studies in the tables of this review represent a wide range of data (high to low) based on quality of evidence regarding main outcomes. The quality of evidence for each study was evaluated based on the following criteria.

Study design (e.g., observational versus randomized trials);

The following definitions from the GRADE Working Group were used in grading the quality of the evidence based on the criteria above [41]:

High = Further research is very unlikely to change our confidence in the estimate of effect.

It should be recognized that the ranking system is based on the quality of the evidence and not on the quality of the research, as all of the cited literature in this review was peer-reviewed.

CARDIOVASCULAR RESPONSES IN YOUNG APOLIPOPROTEIN ɛ4 CARRIERS (Table 1)

The relationship between APOE allele type and cardiovascular risk factors and responses has been extensively studied. In a large study on Finnish boys and girls (Table 1), ages 3, 6, 9, 12, 15, and 18, including 28 ɛ2/ɛ4 carriers, 483 ɛ3/ɛ4 carriers, and 50 ɛ4/ɛ4 carriers, the concentrations of serum total cholesterol, low density lipoprotein cholesterol, and apolipoprotein B increased with APOE phenotype in the order of ɛ2/ɛ2, ɛ3/ɛ2, ɛ4/ɛ2, ɛ3/ɛ3, ɛ4/ɛ3, and ɛ4/ɛ4 [42]. In a follow-up study on Finnish male and female newborns and 3-year-old children including five ɛ2/ɛ4, 147 ɛ3/ɛ4, and eight ɛ4/ɛ4 carriers, the concentrations of serum total cholesterol and LDL cholesterol increased with APOE phenotype in the order of ɛ3/ɛ2, ɛ3/ɛ3, ɛ4/ɛ3, and ɛ4/ɛ4, in both males and females (p < 0.001) in the 3-year-olds [43]. Another large study in 7- and 13-month-old Finnish infants included 36 ɛ4/ɛ4 and 209 ɛ3/ɛ4 carriers and reported that triglyceride concentrations were higher in infants with APOE ɛ4/ɛ4 or ɛ3/ɛ4 than in those with APOE ɛ3/ɛ3 (p-value for difference 0.01 and 0.009, respectively) [44]. A smaller follow-up study was conducted on healthy 13-month-old Finnish children and reported that 16 APOE ɛ4 children had 30% to 50% higher cholesterol-adjusted campesterol and sitosterol concentrations in serum than 20 APOE ɛ3/ɛ3 children (p = 0.002 and p = 0.02, respectively) [45]. In 44 ɛ4 carrier Finnish infants, fed high-fat, high-cholesterol human milk, the total and LDL-cholesterol concentrations and the LDL APOB concentration of those with the APOE phenotype ɛ4/ɛ4 or ɛ3/ɛ4 rose faster and to higher levels than in other infants [46]. In a large study on 16-year-old Finnish boys and girls including 28 ɛ2/ɛ4 carriers, 470 ɛ3/ɛ4 carriers, and 49 ɛ4/ɛ4 carriers, physical exercise did not affect LDL cholesterol, total cholesterol, or HDL/total cholesterol in APOE ɛ4 carriers [47]. In 1999, a study on Caucasian males in Perth, Australia, including eight pairs of ɛ3/ɛ3 and ɛ4/ɛ4 subjects matched for age and serum lipid levels, average age for the ɛ3/ɛ3 = 45 and ɛ4/ɛ4 = 42, suggested that inheritance of APOE ɛ4 is associated with an increased affinity of VLDL particles for LDL receptors on hepatocytes [48].

Cardiovascular endpoints in addition to blood lipid levels have also been examined. ɛ3/ɛ2 or ɛ3/ɛ3 carriers showed marginally significantly greater heart rate reactivity and significantly greater task levels of heart rate and heart rate variability during mental stress than ɛ4/ɛ2, ɛ4/ɛ3, or ɛ4/ɛ4 carriers among 28 healthy 16-year-old Finnish boys. The number of ɛ4 carriers in this study was limited at only 11 [49]. Ellis et al. [50] studied 8-year-old children from the Tasmanian Infant Health Survey including 75 ɛ4 carriers and found that the ɛ4 carriers had a lower body mass index and the effect was more evident among the less fit. In addition, Yu et al. [51] studied young adult Chinese men and women with an approximate mean age = 32; ɛ3/ɛ4 (N = 60), ɛ4/ɛ4 (N = 13). These authors found that VO_2max after exercise training increased significantly higher in carriers of ɛ3/ɛ4 in males (OR = 0.60, 95% CI = 0.09–1.11; p = 0.02) and females (OR = 0.62, 95% CI = 0.09–1.15; p = 0.02). In summary, as compared with ɛ3 carriers, young subjects possessing one or more ɛ4 alleles show significant differences in total serum cholesterol, LDL, triglycerides, heart rate variability during mental stress, and VO_2max. The directionality of the reported ɛ4-associated lipid profiles is consistent with an increased risk of atherosclerosis later in life.

REPRODUCTION AND DEVELOPMENT IN APOLIPOPROTEIN ɛ4 CARRIERS (Table 2)

A limited number of studies have examined potential associations between APOE ɛ4 carrier status and fertility. In 2017, van Exel et al. [52] studied 100 women with a history of at least two first-trimester recurrent miscarriages in Tabriz, Iran, and 100 healthy women with at least two successful pregnancies and no miscarriages. This study incorporated 31 ɛ4 carriers and showed that APOE ɛ4 was associated with higher fertility in women exposed to high pathogen levels. In 2004, Corbo et al. [53] reported similar results in a population of Afro-Ecuadorian and Cayapa Indian women, average age 39; ɛ4 carriers (ɛ3/ɛ4 = 16; ɛ4/ɛ4 = 6; ɛ4/ɛ2 = 5). These authors reported that the ɛ3/ɛ4 genotype frequency (0.50) in the African-Ecuadorian women with 9–17 children was about three times that of the women with 0–8 children (0.14) (p = 0.02). In contrast with these reportedly positive effects on fertility in populations without access to birth control methods, Gerdeset al. [54] studied a cohort of 40-year-old married men residing in Aarhus Denmark; ɛ4 carriers (ɛ3/ɛ4 = 93; ɛ4/ɛ4 = 12; ɛ4/ɛ2 = 9) and found that on average, men with the ɛ3/ɛ3 genotype (n = 212) had 1.93 children, men with the ɛ3/ɛ4 or ɛ2/ɛ2 genotypes (n = 53) had 1.66 children. The potential confounding effects of the widespread availability of birth control methods in Denmark and the low average number of children per couple questions the clinical relevancy of these findings as per the association of the ɛ4 allele and fertility.

Several studies have investigated the possible relationship between APOE ɛ4 carrier status and occurrence of fetal miscarriages. A single study has reported a protective effect of the ɛ4 allele [55]. Some studies do not report an association between ɛ4 status and fetal loss [56 –58]. At least five studies and one meta-analysis report a positive association between possession of the ɛ4 allele and fetal loss [59 –65]. Several additional studies have examined the association between ɛ4 carrier status and a variety of pregnancy-related conditions. Gaynor et al. [66] studied White, Black, and Hispanic neonates and infants; ɛ4 carriers (N = 87) and reported that the APOE ɛ2 allele was associated with a lower Psychomotor Development Index (p = 0.038). Pregnant women with the ɛ4 allele displayed an increased risk to develop Pregnancy Induced Hypertension (OR 4.14, p = 0.013) and severe preeclampsia (OR 4.43, p = 0.019) in a group of pregnant Romanian women, average age 28; ɛ4 carriers (N = 46) [67]. Cerebrospinal fluid (CSF) was collected from 107 healthy Japanese subjects (70 males, 37 females) aged 1–86 by Hirayama et al. [68] who showed that the APOE phenotype does not affect the composition or concentration of CSF high density lipoprotein in children. In 2003, Gaynor et al. [69] studied Asian, Black, and White, male and female infants under 6 months undergoing cardiac surgery in Philadelphia; ɛ4 carriers (N = 52) and reported that patients with the ɛ2 allele had approximately a 7-point decrease in the Psychomotor Developmental Index [Bayley Scales of Infant Development] (p = 0.036).

In summary, the most notable findings related to fertility and the ɛ4 allele suggest a protective effect of ɛ4⁺ status and reproductive potential in populations exposed to high pathogen levels.

CO-MORBIDITIES IN YOUNG APOLIPOPROTEIN ɛ4 CARRIERS (Table 3)

A number of epidemiology studies have reported co-morbidities associated with possession of different APOE alleles (Table 1). Co-morbidities represent an important area of study as co-linearity in prevalence between clinical conditions of known etiology with clinical conditions whose current etiology is unknown (e.g., AD) might provide clues related to causation. A meta-analysis of 28 studies on schizophrenia reported a significant protective effect of ɛ3 in an Asian population [70]. Another meta-analysis found a significant decrease in risk associated with each copy of ɛ4 in all age-related macular degeneration sub-phenotypes [71]. A dermatology clinic in Spain reported that APOE ɛ4 carriers were significantly more frequent in patients with severe psoriasis compared to controls (p = 0.003) and to non-severe psoriasis (p = 0.017). Infants with hypoxic-ischemic encephalopathy did not show an association with APOE allele type [72]. In contrast, the percentage of children with at least one ɛ4 allele was significantly lower in non- Sudden infant death syndrome (SIDS) compared to SIDS (p = 0.016) in a study of infants from Scotland. Several studies have examined the potential relationship between APOE allele type and cerebral palsy. In children seen in Brasilia, Brazil, of whom 139 were ɛ4 carriers, the presence of the ɛ2 allele raised the probability of having cerebral palsy (OR 3.2; 95% CI 1.27–8.27) while the APOE ɛ4 allele was not significantly different among groups [73]. An earlier smaller study on Brazilian children that included only 13 ɛ4 carriers, reported a positive association between the ɛ4 allele and cerebral palsy [74], as did a study on children from Chicago that included 25 ɛ4 carriers [75].

RESISTANCE TO INFECTIOUS DISEASE IN APOLIPOPROTEIN ɛ4 CARRIERS (Table 4)

The potential relationship between resistance to infectious diseases and APOE allele status represents an important area of investigation as this environmental stressor can exert significant evolutionary pressures. The relationship between hepatitis C infection and APOE carrier status has been examined in several studies. Mueller et al. [76] enrolled 205 ɛ4 carriers in two combined cohorts selected from patients from Charite Berlin and the University of Leipzig diagnosed with either chronic or self-limited hepatitis C virus infection. The average age of the patients was 48.7. These authors found that APOE ɛ4 alleles were underrepresented in chronically hepatitis C-infected patients (10.2%) compared to 13% in healthy controls (p = 0.001).

Male and female Caucasian Italian patients with chronic hepatitis C, median age = 41; ɛ2/ɛ4 carriers (N = 3), ɛ3/ɛ4 carriers (N = 21), ɛ4/ɛ4 carriers (N = 1) were studied by Fabris et al. [77]. They reported that patients not carrying an ɛ3 allele, as well as carriers of a single ɛ3 allele with serum cholesterol concentration >190 mg/dL were more likely to have a favorable outcome regarding fibrosis progression with chronic hepatitis C. In an earlier study, Fabris et al. [78] investigated Italian male and female patients who underwent a cadaveric orthotopic liver transplantation, median age 55; ɛ4 carriers (N = 34). Their results showed that possession of an APOE ɛ4 allele to be associated with low fibrosis progression in recurrent hepatitis C infection, and with an idiosyncratic APOE-associated lipid profile. Price et al. [79] found that the APOE ɛ2 and APOE ɛ4 alleles were both associated with a reduced likelihood of chronic infection (hepatitis C virus). For ɛ2, OR = 0.39[95% CI = 0.211–0.728] (p = 0.003), and for ɛ4, OR = 0.6[95% CI = 0.38–0.96] (p = 0.032), in a study population of British and Irish Caucasian hepatitis C patients; ɛ2 carriers (N = 48), ɛ4 carriers (N = 84), 11 carriers were both, i.e., ɛ2/ɛ4. Twenty-four male and female patients, median age = 53.5, with recurrent hepatitis C following cadaveric orthotopic liver transplantation; ɛ4 carriers (N = 12) of 48 donors, ɛ4 carriers (N = 17) of 48 recipients were studied by Toniutto et al. [80]. These authors found that recipient (but not donor) carriage of at least one ɛ4 allele was associated with improvement in the staging score due exclusively to the contribution given by male recipients. Similarly, Wozniak et al. [81] observed British men and women with hepatitis C viral infection, mean age = 41; ɛ4 carriers (N = 42) and concluded that in chronically hepatitis C virus-infected subjects grouped according to extent of fibrosis, necroinflammation, and total Knodell score, an overrepresentation of the APOE ɛ4 allele was found in those whose livers were mildly affected. These authors suggested that carriage of an APOE4 ɛ4 allele protects against severe liver damage induced by hepatitis C virus.

Although not as extensively studied as hepatitis C infection, the relationship between resistance to malaria and APOE allele status has also been examined in several studies. Fujioka et al. [82] showed that human plasma samples from APOE ɛ4/ɛ4 but not APOE ɛ3/ɛ3 donors inhibited growth and disrupted morphology of P. falciparum (malaria) in seven APOE ɛ4/ɛ4 and six APOE ɛ3/ɛ3 donors from Cleveland, OH, USA. Aucan et al. [83] observed Gambian children ages 1–10; ɛ4 carriers (N = 244) and concluded that the APOE ɛ3/ɛ4 genotype was found to be more common in children with both cerebral malaria and severe malarial anemia (42.9%) than in controls (24.8%) and mild malaria cases (27.2%). When corrected for the number of clinical groups (4) compared with controls in this study, this finding was not statistically significant. An earlier study on African infants was reported by Wozniak et al. [84]. This group studied infants from Prampram, 50 km east of Accra on the south coast of Ghana; ɛ2/ɛ2 carriers (N = 4), ɛ4 carriers (N = 47). Based on small numbers, APOE ɛ2 homozygotes became infected with malaria at an earlier age than those carrying other genotypes.

At least one study on a potential relationship between APOE allele status and risk for development of herpes simplex encephalitis has been conducted in the United Kingdom [85]. These authors examined specimens from the brain or spleen of 14 patients with herpes simplex encephalitis and from the CSF of seven patients with HSV1 in their CSF detected by PCR. Lin et al. [85] concluded that APOE ɛ3 and ɛ4 allele frequencies did not differ significantly between the two groups, and that APOE ɛ2 is a risk factor for herpes simplex encephalitis. In summary, possession of the ɛ4 allele might provide some measure of protection against hepatitis C and malaria.

RESPONSES TO HEAD INJURY IN YOUNG APOLIPOPROTEIN ɛ4 CARRIERS (Table 5)

Several studies have been conducted on the association between APOE allele frequency and either susceptibility to head injury, or recovery from head injury. In a 2018 study by Terrell et al. [86] on a cohort of male football and male and female soccer players (average age 19.85) from four US southern universities that enrolled 35 ɛ4 carriers, IL-6R CC was associated with a three times greater concussion risk and APOE ɛ4 with a 40% lower risk. In a similar cohort with 62 ɛ4 carriers, Tierney et al. [87] found no significant association between carrying the ɛ4 allele and history of concussion. However, Tierneyet al. [87] stated that they assessed the medical history via researcher-assisted paper and pencil assessment, attempting to indicate the athlete’s concussion history. Collegiate student athletes from the University of Toronto, average age 20.5, with 79 ɛ4 carriers were studied by Kristman et al. [88] with a small statistically significant positive association reported as stated by the authors of an unadjusted hazard ratio for concussion in the APOE ɛ4 carriers of 1.18 (95% CI: 0.52, 2.69) compared to non-carriers. Adjusting for sex, weight, height, and team type resulted in an only slightly lower hazard ratio of 1.06 (95% CI: 0.41, 2.72), indicating little effect from confounding factors. A somewhat larger study that relied upon self-reported concussion history over the previous eight years was conducted by Terrell et al. [89] on Black and White, male and female, college athletes from 23 schools, average age 19.7; ɛ3/ɛ4 (N = 225), ɛ4/ɛ4 (N = 20). Terrell et al. [89] summarized their results as showing no substantial evidence of an association between a history of concussion and APOE genotypes and haplotypes. However, the authors also noted that the cell sizes for some of the APOE genotypes were so small that meaningful analysis was not possible. Also, compared to those with the APOE ɛ3/ɛ3 genotype, those with the ɛ2/ɛ3 genotype were at a 60% higher risk for concussion, but the results were not statistically significant (OR, 1.6; 95% CI, 0.5 to 4.8). In summary, the association between possession of the APOE ɛ4 allele and risk of concussion is unclear at this time due to methodological limitations in determination of concussion history and limited sample sizes.

Han et al. [90] conducted a study on active duty military personnel with a recent history of mild to moderate traumatic brain injury, average age 22.6; ɛ4 carriers (N = 16). Their analysis showed comparable performances on most neuropsychological measures and better performances by ɛ4 carriers on select measures of attention, executive functioning and episodic memory encoding. Merritt et al. [91] reported data on 53 veterans with a history of mild traumatic brain injury and 46 military controls and found that traumatic brain injury ɛ4 carriers had relative impairments in memory function and speed of processing but not on executive function relative to traumatic brain injury-affected veterans without an ɛ4 allele. However, there was no ɛ4-related difference among the non-traumatic brain injury military controls.

In a relatively large study (ɛ4 carriers N = 324) enrolling consecutive head injury admissions (men and women) to a regional neurosurgical unit in West Scotland, average age 35, Teasdale et al. [92] found no overall association between APOE genotype and outcome. Thirty-six percent of APOE ɛ4 carriers had an unfavorable outcome compared with 33% of non-carriers of APOE ɛ4. This relatively large study was a follow-up to a smaller study conducted by Teasdale et al. in 1997 [93] on only 30 ɛ4 carriers. Another small study on 16 ɛ3/ɛ4 subjects recruited from a Canadian traumatic brain injury clinic who experienced mild to moderate traumatic brain injury, mean age 33, was conducted by Chamelian et al. [94]. These authors reported no association between the presence of the APOE ɛ4 allele and poor outcome across all measures. In 2005, Blackman et al. [95] reviewed the childhood literature on APOE and brain injury up through 2005 and concluded that results from the limited studies in children were contrary to the adult experience with ɛ4 seeming to confer protection for the brain whereas ɛ2 posed a risk. Further larger studies are needed to definitively determine the role of APOE status on recovery from brain injury in young subjects.

BIOCHEMICAL DIFFERENCES POSSIBLY RELATED TO NEURAL STRESS IN APOLIPOPROTEIN ɛ4 CARRIERS (Table 6)

The strongest direct evidence that some of the macromolecular components comprising the synapses of ɛ4 carriers might be turned over at a higher rate than comparable macromolecular synaptic components in non-ɛ4 carriers has been provided by Yassine et al. [96]. This group studied 22 middle-aged healthy adults (mean age 35 years, range 19–65 years) and found that k*, the mean global gray matter DHA incorporation coefficient, was significantly higher (16%) among ɛ4 carriers (n = 9) than among non-carriers (n = 13, p = 0.046). Also in 2017, the same group [13] reviewed original articles, systematic reviews, and meta-analyses of omega-3 studies in AD that were published before August 20, 2016 and concluded that while randomized clinical trials of omega-3 in symptomatic AD reported negative findings, several observational and clinical trials of omega-3 in the pre-dementia stage of AD suggest that omega-3 supplementation might slow early memory decline in ɛ4 carriers. Tambini et al. [97] conducted an in vitro experiment whose results are consistent with the increased lipid metabolism observed by Yassine et al. [96]. Using an astrocyte-conditioned media model, this group measured the synthesis of phospholipids and cholesteryl esters and reported a significant increase in cells treated with APOE ɛ4-containing astrocyte-conditioned media as compared to those treated with APOE ɛ3-containing-astrocyte-conditioned media.

Dose et al. [98] conducted a mini-review of APOE genotype and stress responses. From their analysis of the evidence on APOE isoform-dependent oxidative stress and mitochondrial function these authors concluded that APOE4 is associated with an increased stress response. Ramassamy et al. [99] conducted a brain tissue study whose results are consistent with the observations of Dose et al. [98]. They obtained human brain tissue from the Douglas Hospital Research Centre Brain Bank, Canada, average age 75–79; ɛ4 carriers (N = 18). Among ɛ4 carriers with AD, the levels of thiobarbituric acid-reactive substances were found to be higher among ɛ4 carriers while the APOE protein concentrations were lower.

At least one small study has examined the potential association between vitamin D and APOE allele status [100]. In a sub-group of 93 subjects from a general population sample of 699 subjects (age of subjects unknown), multivariate adjusted modeling showed a positive association (p = 0.072) of the APOE ɛ4 allele with 25(OH)D [vitamin D] levels. Another small study was conducted in 2012 by Ringman et al. [101] wherein thirty-three subjects were studied including six ɛ2/ɛ3, six ɛ3/ɛ4, and 21 ɛ3/ɛ4 allele combinations. Plasma levels of APOE and superoxide dismutase 1 were lowest in ɛ4 carriers, intermediate in ɛ3 carriers, and highest in the ɛ2 carriers. In contrast, multiple plasma interleukins were highest in ɛ4 carriers and demonstrated significant negative correlations with age. Larger studies similar to those conducted by Yassine et al. would be helpful in clarifying the association of neural stress/macromolecular turnover rate with allele subtype.

BRAIN STRUCTURE-FUNCTION DIFFERENCES IN YOUNG APOLIPOPROTEIN ɛ4 CARRIERS (Table 7)

MENTAL PERFORMANCE IN YOUNG APOLIPOPROTEIN ɛ4 CARRIERS (Table 8)

At least 32 studies or meta-analyses have examined mental performance in young APOE ɛ4 carriers. No effect of the APOE allele was reported in six studies [112 –117]. All six of the study populations reporting no effect were Caucasian and lived in the United States, United Kingdom, or Western Europe. A single study [118] on only 33 ɛ4 carriers reported inferior performance associated with possession of ɛ4. Significant differences were found on the Rey-Osterrieth Complex Figure Test, with ɛ2-positive children scoring 29.2% relative to ɛ3/3 at 8.9% and ɛ4-positive children at 6.1% (p = 0.12).

Three studies reported improved performance in spatial tasks in ɛ4 carriers [102 , 120]. A large British study (ɛ3/4 = 542, ɛ4/4 = 43) showed faster reaction times in association with the ɛ4 allele [121]. In a series of studies, Oria et al. [122 –124] have shown that possession of the ɛ4 allele protects against long-term cognitive deficits associated with severe diarrhea in Brazilian shanty-town children. In young Finnish men and women (mean age = 28), ɛ4 carriers (N = 20), ɛ4⁺ status was associated with good performance in mental arithmetic and the association was dependent on LDL cholesterol level [125].

Two studies have reported a higher IQ in ɛ4 carriers. In a large British study [126], there was a consistent pattern that ɛ2/ɛ2 and ɛ4/ɛ4 girls had higher IQ scores (from 3 to 7 points) compared with ɛ3/ɛ3 girls. A small study (31 ɛ4 carriers) on Han Chinese female nursing students showed a modest increase in performance IQ and N100 amplitude for ɛ4 carriers (p = 0.038 and 0.068, respectively) [127]. Using the same cohort of nursing students reported in Yu et al. [127], ɛ4⁺ status did not affect tridimensional personality questionnaire scores [128]. In contrast with the lack of correlation with personality reported in the small cohort of Tsai et al. [128], a large study on Finnish children, adolescents, and young adults [ɛ4/ɛ3 (N = 483) and ɛ4/ɛ4 (N = 50)] reported that motor activity and even hyperactivity in childhood, and mental vitality in adolescence and young adulthood increased significantly in the order of ɛ2/ɛ2, ɛ3/ɛ2, ɛ4/ɛ2, ɛ3/ɛ3, ɛ4/ɛ3, and ɛ4/ɛ4 [129]. In addition, ɛ4 carriers displayed a wider field of attention in two tasks [130].

Several studies have reported better performance on memory tasks in young subjects possessing the ɛ4 allele [131 –136]. Comparatively enhanced verbal fluency has also been observed in ɛ4 allele carriers [137, 104]. In a study on 53 infants in Mexico City who were ɛ4 carriers, ɛ4⁺ status had a 4.4 point higher 24-month Mental Development Index on the Bayley Scale [138]. A small study from the Czech Republic on 23 ɛ4⁺ subjects suggested that early life cognitive advantages might persist through young adulthood as 87% of ɛ4⁺ carriers reached higher education [139].

The results from two studies suggest that the cognitive response to particular pharmacologic agents might be affected by APOE ɛ4 status. Twenty-seven ɛ4 carriers who were healthy nonsmokers, aged 18–30 recruited from Sussex University demonstrated that the ɛ4 allele confers a cognitive advantage on tasks mediated by the frontal lobes. In addition, young carriers of the ɛ4 allele show larger cognitive benefit from procholinergic nicotinic stimulation [140]. In the second study, poor Brazilian children with below median height (mean age = 8.6), including 37 ɛ4 carriers reported that ɛ4⁺ children receiving glutamine supplementation showed short-term gains in HAZ (height for age Z score), WAZ (weight for age Z score) and WHZ (weight for height Z score) that were correlated with better performance in long-term cognitive testing [141].

In this review of cognition and APOE allele genotype, our emphasis has been on ɛ-related clinical and performance associations in young rather than middle-age subjects. Several recent studies on middle age cohorts might address the issue of transition from ɛ4-associated cognitive advantage to cognitive deficit. In an older cohort with an average age for ɛ4 carriers of 58.0 and non-carriers of 61.4, Caselli et al. [142] sought to determine the age at presentation of ɛ4-related declines in memory. They enrolled 815 subjects: 317 ɛ4 carriers, 79 of whom were ɛ4/ɛ4 and 238 ɛ3/ɛ4. The non-carrier group contained 498 subjects. Carriers were followed for a longer period (5.3 versus 4.7 years, p = 0.01), with an equivalent duration of formal education (15.4 years) and proportion of women (69%). Longitudinal decline in memory in ɛ4 carriers began before age 60 and showed greater acceleration than in non-carriers (p = 0.03). There was a possible ɛ4 dose effect (p = 0.008). In 2017, Lancaster et al. published a systematic and meta-analytic review of 36 studies on subjects ranging from 35–60 years old investigating APOE-related differences in cognition in mid-adulthood [143]. The average effect size of ɛ4 status was non-significant across cognitive domains. Sinclair et al. [144] studied 114 participants with the allelic combinations of ɛ3/ɛ3 (39 subjects), ɛ3/ɛ4 (27 subjects), ɛ4/ɛ4 (15 subjects), ɛ3/ɛ2 (26 subjects), and ɛ2/ɛ2 (7 subjects). The primary outcome was performance on the Rey Auditory Verbal Learning Test (RAVLT). ɛ2 carriers displayed slightly better episodic memory performance (p = 0.016), somewhat improved n-back accuracy and better executive functioning (p = 0.005).

CONCLUSIONS

The weight-of-the-evidence presented in Tables 1–8 supports the hypothesis that the possession of the ɛ4 allele in youth may have a positive differential impact on fitness during different life stages [145]. Young subjects having at least one copy of the ɛ4 allele reportedly possess a number of advantages that might facilitate survival in harsh environments including the following among others: more rapid improvement in VO_2max following exercise [51]; increased fertility at high pathogen burdens [52, 53]; more rapid infant development [105]; prophylaxis against cognitive deficits associated with severe diarrhea [122 –124]; resistance to certain infections, e.g., hepatitis [76] and malaria [83]; faster reaction times [121]; better spatial memory [111 , 120]; and slight superiority in direct or indirect measures of IQ [135 , 141]. Some of these advantages appear to come at the expense of differences in neural processing that might place higher metabolic demands per unit time on the brains of young ɛ4 carriers [143 , 96].

Until 300,000 years ago, ancestors of modern humans were ubiquitously ɛ4/ɛ4 and then the ɛ3 allele mutated from the ancestral ɛ4 allele [146]. The ɛ3 allele displayed a competitive survival advantage sufficiently robust to result in the current predominance of the ɛ3/ɛ3 genotype which is now found in over 60% of the US population, presumably because of its protection for memory loss and dementia in progressively older age ranges [23]. Similarly, the ɛ2 allele mutated from the ɛ3 allele about 200,000 years ago, but this protective allele has remained relatively rare with the homozygous ɛ2/ɛ2 variant less than 1%, and the ɛ3/ɛ2 heterozygote in about 11% of the population [144].

Given the ancestral primacy of the ɛ4 allele, and the evolutionary trade-off of superior performance in youth versus additional years beyond historical lifespans, the abnormality of the ɛ4 allele is somewhat a matter of perspective. If part of the APOE ɛ4-associated neurotoxic susceptibility is based on pharmacokinetic rather than toxicant receptor interactions on a per mole basis [20], future therapies that slow down synaptic pruning might carefully consider differential effects based on APOE allele subtype. Current knowledge of potential sources of AD patient heterogeneity is lacking. Reducing at least one important source of inter-subject heterogeneity, i.e., APOE ɛ4 allele carrier status, is advisable. Early attempts at shifting the balance away from synaptic pruning might consider enrolling early stage AD patients possessing at least one ɛ4 allele.

DISCLOSURE STATEMENT

Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/18-1089r3).