Neuron loss associated with age but not Alzheimer's disease pathology in the chimpanzee brain

Abstract

In the absence of disease, ageing in the human brain is accompanied by mild cognitive dysfunction, gradual volumetric atrophy, a lack of significant cell loss, moderate neuroinflammation, and an increase in the amyloid beta (Aβ) and tau proteins. Conversely, pathologic age-related conditions, particularly Alzheimer's disease (AD), result in extensive neocortical and hippocampal atrophy, neuron death, substantial Aβ plaque and tau-associated neurofibrillary tangle pathologies, glial activation and severe cognitive decline. Humans are considered uniquely susceptible to neurodegenerative disorders, although recent studies have revealed Aβ and tau pathology in non-human primate brains. Here, we investigate the effect of age and AD-like pathology on cell density in a large sample of postmortem chimpanzee brains (n = 28, ages 12–62 years). Using a stereologic, unbiased design, we quantified neuron density, glia density and glia:neuron ratio in the dorsolateral prefrontal cortex, middle temporal gyrus, and CA1 and CA3 hippocampal subfields. Ageing was associated with decreased CA1 and CA3 neuron densities, while AD pathologies were not correlated with changes in neuron or glia densities. Differing from cerebral ageing and AD in humans, these data indicate that chimpanzees exhibit regional neuron loss with ageing but appear protected from the severe cell death found in AD.

This article is part of the theme issue ‘Evolution of the primate ageing process’.

1. Introduction

As human longevity increases, distinguishing the neurological basis for age-related cognitive decline is imperative. Common cognitive deficits in elderly people include difficulties with complex tasks, word recall, episodic memory and processing speed [1]. Based on magnetic resonance imaging (MRI) studies, these cognitive alterations coincide with decreased brain weight, increased white matter hyperintensity, enlarged lateral ventricles and mild regional volumetric atrophy [2]. Accompanying these gross changes are modest modifications in neurons, dendritic spines, synapses and neurotransmitters along with increased glial activation, reduced cerebral blood flow and a weakening of the blood-brain barrier [3–5]. Ageing also contributes to an individual's risk for developing Alzheimer's disease (AD), the most prevalent form of dementia. AD is a progressive, irreversible brain disorder that results in extensive neocortical and hippocampal neuronal loss and atrophy, amyloid beta (Aβ) protein-containing plaques and vascular deposition, tau-associated neurofibrillary tangles (NFT), neuroinflammation and severe cognitive impairment [6,7].

Distinguishing the earliest stages of AD from healthy ageing remains an area of great interest and a difficult diagnostic problem, although certain metrics, such as regional neuron loss and glial activation, have been established [1]. Age-related decline in neuron numbers is modest in the dentate gyrus and subiculum, while the CA1-CA3, entorhinal cortex, as well as the neocortex are preserved [8–12]. In striking contrast, AD brains exhibit profound neuronal death in the prefrontal and temporal cortex, entorhinal cortex, CA1 of the hippocampus, dentate gyrus and subiculum [8,9,13–17]. Another discriminating factor between ageing and AD is the severity of neuroinflammation, which is assessed, in part, by changes in glial density, activation and morphology [18]. Total glial density does not change in the human neocortex during the normal course of ageing or AD [19–21]. However, glial subtypes, such as microglia and astrocytes, are altered in both conditions [22,23]. As the brain's primary immune cell, microglia activate and proliferate with age in the neocortex, including the hippocampus and entorhinal cortex, of healthy humans [23,24]. AD brains display greater microglial activation and proliferation concomitant with Aβ plaques, particularly in the hippocampus [25–27]. Astrocytes provide metabolic and structural support to neurons, regulate neurogenesis, and modulate synaptic activity and neurotransmitter homeostasis [28]. Astrogliosis, as indicated by the upregulation of glial fibrillary acidic protein (GFAP), hypertrophy of the soma and cellular processes, and loss of domain organization, has been associated with normal ageing, and to a larger extent in AD [28–32].

Humans are considered uniquely susceptible to neurodegenerative disease, such as AD, but several recent studies have revealed AD-like pathology in the brains of non-human primates [33–37]. Aged lemurs, prosimians, monkeys and great apes exhibit diffuse and neuritic amyloid plaques as well as vascular amyloid, although cognitive changes based on plaque burden were not observed in aged macaques [34,35,38–46]. Furthermore, the presence of hyperphosphorylated tau has been reported in lemurs, squirrel and rhesus monkeys, baboons, and gorillas adjacent to Aβ deposition, and African green monkeys and aged chimpanzees exhibited NFT [34,35,46–49]. Closest in phylogeny to humans, non-human primates also exhibit senescence-related changes [33,50,51]. Mouse lemurs, marmosets, rhesus monkeys and apes show mild cognitive variations with age in spatial memory and executive function [52–57]. Some MRI studies in ageing mouse lemurs and rhesus macaques show progressive volumetric atrophy in the prefrontal cortex (PFC) and decreased brain weight in chimpanzees [55,58,59]. However, a more recent, comprehensive MRI analysis found no overt atrophy of volume in the neocortex and white matter in chimpanzees [51,59]. Postmortem analyses in primates have detected mild age-related dendritic atrophy, synapse loss, white matter damage, gliosis (i.e. activation of astrocytes and microglia), neurotransmitter alterations, and increases of Aβ in brain parenchyma and in cerebral vasculature [34,60–63].

Previous studies of ageing in non-human primates reported a lack of neuron loss in the neocortex, including CA1, entorhinal cortex and subiculum, with age [64–69]. Additionally, microglial activation has been observed in the frontal cortex, cingulum bundle and primary visual cortex, although numbers of activated microglia did not increase significantly with age in the visual cortex, substantia nigra and ventral tegmental area of aged rhesus monkeys [50,70,71]. Heightened expression of major histocompatibility complex class II (MHC-II) antigen, a marker of activated microglia, also was identified in the cerebral cortex and white matter of pig-tailed macaques [72]. Astrocyte activation in the form of higher GFAP expression was detected in the aged rhesus monkey hippocampus and PFC, although astrocyte density did not vary [73].

Formerly, we analysed the brains of 20 aged chimpanzees for evidence of Aβ and tau lesions as well as microglia and astrocyte activation [34,74,75]. Aβ was observed in plaques and, most predominantly, in blood vessels, which correlated with increased tau. Tau lesions were found in the form of AT8-immunoreactive (ir) pretangles, NFT and neuritic clusters (NC) of aggregated dystrophic neurites, and NFT were observed in apes that exhibited plaques and moderate or severe cerebral amyloid angiopathy (CAA), a condition in which amyloid accumulates in the brain's vasculature. Age correlated with greater volumes of Aβ plaques and vessels, but not tau or activated microglia and astrocyte densities. Like AD, Aβ42 deposition was positively associated with higher hippocampal microglial activation in chimpanzees, while astrogliosis occurred in both the hippocampus and PFC layer I in conjunction with Aβ and tau proteins. Contrary to AD, activated microglia density was not significantly correlated with tau lesions and astrogliosis was not identified in other cortical layers in chimpanzees. Despite certain comparable age-related and AD-like pathologies identified in non-human primate brains, only humans exhibit major neuronal loss and severe cognitive decline as observed in clinical AD. Thus, building on our prior investigations, we quantified regional neuron density, glia density, and glia:neuron ratios in the dorsolateral PFC, middle temporal gyrus (MTG), and CA1 and CA3 hippocampal subfields of chimpanzees to determine if ageing or AD pathology correlates with regional neuron loss or glial activation.

2. Material and methods

(a). Specimens and sample processing

Postmortem brain samples from 12 male (ages 17–62 years) and 16 female (ages 12–58 years) chimpanzees (Pan troglodytes, electronic supplementary material, table S1) were acquired from Association of Zoos and Aquariums-accredited zoos and American Association for Accreditation of Laboratory Animal Care-accredited research institutions through the National Chimpanzee Brain Resource (NIH grant: NS092988). The chimpanzees included in this study did not participate in formal behavioural or cognitive testing and were maintained in accordance with each institution's animal care and use guidelines. Available health information for each animal has been included in electronic supplementary material, table S2. Depending on availability, samples were taken from the right or left hemispheres. Brains were collected postmortem after death from natural causes (approximately less than or equal to 20 h PMI), immersion-fixed in 10% buffered formalin for a minimum of 10 days, and transferred to a 0.1 M buffered saline solution containing 0.1% sodium azide at 4°C for storage. Samples were cryoprotected in a graded series of sucrose solutions (10, 20 and 30%) and cut frozen into 40 µm-thick sections in a coronal plane using a Leica SM2000R freezing sliding microtome (Buffalo Grove, IL). Sections then were placed into individual centrifuge tubes containing a cryo-protection solution (30% dH₂O, 30% ethylene glycol, 30% glycerol, 10% 0.244 M phosphate-buffered saline (PBS)), numbered sequentially, and stored at −20°C until histological or immunohistochemical processing. Every 10th section in each region was stained for Nissl substance with a 0.5% cresyl violet solution to reveal cell somata, define cytoarchitectural boundaries, and quantify neuron density (Nv), glia density (Gv) and glia to neuron (G : N) ratio (figure 1). Previously, immunohistochemistry was performed for markers of hyperphosphorylated tau (AT8), Aβ42, microglia (Iba1, ionized calcium-binding adaptor molecule 1), and astrocytes (GFAP, glial fibrillary acidic protein) in the same regions using the avidin-biotin-peroxidase method and 3,3′-diaminobenzidine (DAB) with nickel enhancement or NovaRed (figure 2) [34].

Figure 1.

Photomicrographs of Nissl staining in the chimpanzee brain: (a) classification of neurons (grey arrowhead) and glia (black arrowhead), (b) layer III in the MTG, (c) stratum pyramidale layer of the CA1 and (d) a photo montage delineating the CA1 and CA3 hippocampal subfields. Abbreviations: so, stratum oriens; sp, stratum pyramidale; sr, stratum radiatum. Scale bar = 25 µm (a), 250 µm (b–d). (Online version in colour.)

Figure 2.

Photomicrographs of Iba1-ir microglia (a), GFAP-ir astrocytes (b), Aβ42-ir plaque (c), leptomeningeal vessel (d), AT8-ir pretangle (e), NFT (f) and tau neuritic cluster (g) in the chimpanzee PFC (a–d,f) and hippocampus (e,g). Scale bars = 25 µm (a–c,e–g), 250 µm (d). (Online version in colour.)

(b). Regions of interest

We analysed layer III in Brodmann's areas 9 and 10 of the dorsolateral PFC and Brodmann's area 21 of the MTG, as well as the stratum pyramidale in the hippocampal subfields CA1 and CA3 (figure 1). Sampled areas were selected based on prior reports that demonstrated involvement of these regions in both ageing and AD pathology [76–78]. Humans with AD display extensive neuron and synapse loss in layers III and V of the neocortex and stratum pyramidale in the CA1 field, and neuritic Aβ plaques and NFT are most prevalent in these cortical layers [5,16,17,79,80]. Chimpanzees also display Aβ and tau pathologies in the neocortex and hippocampus [34,40].

(c). Stereologic data acquisition

Quantitative analyses were performed using computer-assisted stereology with an Olympus BX-51 photomicroscope equipped with a digital camera and StereoInvestigator software v. 11 (MBF Bioscience, Williston, VT) by two observers blinded to age, sex and pathology. Initial subsampling techniques were performed for each probe to determine appropriate sampling parameters [81]. Regional Nv and Gv were obtained using the optical fractionator probe at 40× (N.A. 0.75) under Köhler illumination. Counting frames were set at 100 × 100 µm with a grid size of 250 × 250 µm, a disector height of 7 µm, and a guard zone of 2%. Beginning at a random starting point, three equidistant sections (every 10th section) per region of interest and animal were selected for analysis. A marker was placed on the nucleus of neurons and on glia when encountered within the optical disector frame, and mounted section thickness was measured at every fifth sampling site. Neurons were identified by the presence of a nucleus, nucleolus and axonal hillock, while glia were distinguished by their lack of nuclei and granules of heterochromatin, giving a speckled appearance. For each region, Nv and Gv (per mm³) were calculated as the population estimate divided by sum volume of the examined disectors, and the G : N ratio was calculated as Gv/Nv [82]. The percentage (%) of gain or loss between old and young animals was determined with the following equation: (aged density or ratio/young density or ratio × 100) − 100. To correct for tissue shrinkage in the z-axis, the height of the disector was multiplied by the ratio of section thickness to the actual weighted mean thickness after mounting and dehydration. No correction was necessary for the x and y dimensions, because shrinkage in section surface area is minimal [83]. The mean number of sampling sites per individual was 32 ± 6 (mean ± s.d.) for PFC and MTG and 65 ± 7 for the CA1 and CA3. The mean number of markers per individual for neurons was 375 ± 97 and for glia was 610 ± 227 in the PFC and MTG. In the CA1 and CA3, the mean number of markers per individual for neurons was 397 ± 33 and for glia was 960 ± 72. The average CE for all regional neuron densities was 0.06 and for glia densities was 0.05.

(d). Statistical analyses

Data were previously collected for Aβ42 plaque and vessel volume (%), AT8-immunoreactive (ir) pretangle, NFT and tau NC densities, Iba1-ir microglia densities, and GFAP-ir astrocyte densities from aged apes (i.e. greater than or equal to 37 years old; figure 2) [34,74,75]. Young chimpanzees (i.e. greater than or equal to 35 years) were assessed for AD pathology, which was absent with the exception of a few AT8-ir pretangles in the PFC of one 12-year-old female that died from peritonitis, and Iba1-ir microglia densities and GFAP-ir astrocyte densities [34,74,75]. All pathology densities and volumes were checked for linearity, and because of skewness of means close to zero, densities and volumes were transformed using the formula: arcsin (sqrt (density/1000)). To evaluate neuropathologic changes for each individual, a brain age value from 0 to 60 was computed using a pathology scoring system adapted from staging guidelines for Aβ and NFT deposition in AD and CAA [34]. Principal component analysis (PCA) was performed to reduce the number of pathological variables to the most relevant factors for brain age in chimpanzees. Four factors—AT8-ir NC densities, NFT density, and AB plaque and vessel volumes—explained 57% of the variance, and all variables had primary loadings between 0.67 and 0.87. Regression factors (PCA-generated pathology factor) created from this prior analysis were employed for further regression analyses with Nv and Gv in this study. Data were tested for a normal distribution and for outliers using the ROUT method (Q = 1), and outliers were excluded. Linear regression analyses were used to determine relationships between the dependent variables of regional Nv, Gv and G : N, and independent variables of chronological age, PCA-generated pathology factor, Aβ42 plaque and vessel volumes (%), and pretangle, NFT, and tau NC densities (per mm³). Sex and brain region variations were examined using two-way (sex) or mixed model (region) ANOVAs with Bonferroni post hoc tests. Statistical analyses were conducted using GraphPad Prism 8.3.0 (San Diego, CA), and the level of significance (α) was set at 0.05.

3. Results

Regional Nv, Gv and G : N were quantified for layer III of the PFC and MTG as well as the pyramidal layer of the CA1 and CA3 hippocampal subfields (figure 1). Average densities and ratios for young (n = 8, 12–35 years), aged (n = 20, 37–62 years), and all chimpanzees in addition to the per cent gain or loss between old and young animals for each region and variable are shown in electronic supplementary material, table S3. Using previously collected microglia and astrocyte densities from tissue in the same animals, along with the total glia density from the current study, the breakdown of glial subtypes was calculated for each region examined with an average across all areas of 81% oligodendrocytes, 8% astrocytes and 11% microglia (electronic supplementary material, table S4) [74,75].

(a). Region and sex

Mixed model ANOVA with Bonferroni multiple comparison tests revealed that PFC and MTG Nv are significantly higher than CA1 (PFC: t₂₇ = 12.01, MTG: t₂₆ = 13.54, ps ≤ 0.01) and CA3 (PFC: t₂₇ = 4.97, MTG: t₂₆ = 5.97, ps ≤ 0.01) but do not differ from each other (t₂₆ = 0.17, p = 0.99; figure 3a). CA3 Nv was higher than CA1 (t₂₇ = 12.44, p ≤ 0.01). Glia density was significantly higher in the PFC and MTG than CA1 (PFC: t₂₅ = 5.77, MTG: t₂₄ = 8.69, ps ≤ 0.01), but not CA3 (PFC: t₂₅ = 1.43, MTG: t₂₄ = 0.11, ps ≥ 0.99; figure 3b). PFC Gv did not vary from MTG (t₂₆ = 2.14, p = 0.25). As with Nv, CA3 Gv was greater than CA1 (t₂₅ = 16.26, p ≤ 0.01). Both PFC and MTG G : N ratios were significantly lower than CA1 (PFC: t₂₅ = 5.14, MTG: t₂₂ = 5.94, ps ≤ 0.01) and CA3 (PFC₂₄: t = 8.18, MTG: t₂₁ = 9.48, ps ≤ 0.01), while CA3 was significantly greater than CA1 (t₂₄ = 5.76, p ≤ 0.01; figure 3c). G : N in the PFC did not differ from MTG (t₂₄ = 1.01, p = 0.99). Two-way ANOVA revealed no sex differences in Nv (F_1,103 = 1.61, p = 0.21), Gv (F_1,99 = 0.24, p = 0.63) or G : N (F_1,97 = 0.61, p = 0.44; figure 3d–f) for any region examined.

Figure 3.

Scatter plots of regional neuron density (a), glia density (b) and glia:neuron ratios (c) in the chimpanzee brain. Although variations were found by region (a–c), sex differences were not observed (d–f; solid circle, females; open circle, males). *p ≤ 0.05.

(b). Age

Age was associated with significantly decreased Nv in both CA1 and CA3 (figure 4, ps ≤ 0.02; electronic supplementary material, table S5). Age-related changes were not observed in PFC or MTG Nv (electronic supplementary material, figure S1A-B, ps ≥ 0.29). There were no age-related changes for Gv (electronic supplementary material, figure S1C-F, ps ≥ 0.07), or G : N in any of the regions examined (electronic supplementary material, figure S1G-J, ps ≥ 0.10; electronic supplementary material, table S5).

Figure 4.

A decrease in CA1 and CA3 neuron densities (Nv, mm³; a,b) was correlated with older age in the chimpanzee brain (ps ≤ 0.02). Photomicrographs of Nissl staining in hippocampal subfields CA1 (c,d) and CA3 (e,f) in a young chimpanzee (c,e) and an old (d,f) chimpanzee brain. Scale bar = 250 µm. (Online version in colour.)

(c). Alzheimer’s disease pathology

To determine an overall Aβ and tau score for each chimpanzee, a PCA-generated pathology factor was calculated as previously described [34]. Linear regression analyses were used to investigate the association of regional Nv, Gv or GN with PCA-generated pathology factors and regional AD pathology measurements previously collected in the 20 oldest apes (i.e. greater than or equal to 37 years [34]). Pathologic markers included Aβ42-ir plaque and vessel volumes (%) and AT8-ir pretangle, NFT and NC densities (per mm³) collected in the PFC, MTG, CA1 and CA3. Briefly, plaques were defined as extracellular accumulations of insoluble Aβ42, while vascular amyloid was quantified when Aβ42 deposition was present in the vessel walls. Pretangles were characterized as having an intact cell soma, the presence of diffuse punctate hyperphosphorylated tau (AT8) immunostaining in the cytoplasm, well-preserved dendrites and a nucleus. NFT were identified based on intraneuronal aggregates of hyperphosphorylated tau, and distorted, shortened or absent dendrites and axons. Tau NC contained clusters of dystrophic neurites, consisting of AT8-ir swollen axons and dendrites, or diffuse, punctate staining.

Regional Nv, Gv and G : N were not correlated with PCA-based pathology factors or AD pathologies with two exceptions (electronic supplementary material, table S6, ps ≥ 0.07). In the MTG only, Aβ42 vessel volume was associated with decreased Nv (R² = 0.23, p = 0.05; electronic supplementary material, figure S2A), even after excluding two outliers with significantly high levels of amyloid deposition (R² = 0.28, p = 0.04; electronic supplementary material, figure S2B,E-F). However, to correct for multiple comparison testing error, we performed a Bonferroni correction (p = 0.002), and the correlation of MTG Aβ42 vessel volume and Nv was no longer significant. Additionally, pretangle density was negatively correlated with Nv in the CA1 (R² = 0.19, p = 0.05; electronic supplementary material, figure S2C), but after removal of two outliers with high pretangle numbers, the association was non-significant (R² = 0.00, p = 0.88; electronic supplementary material, figure S2D,G-H).

4. Discussion

Several studies identified senescence-related changes and AD-like pathology in the brains of non-human primates [33–35,37,38]. Yet, none established whether Aβ or tau pathologies were associated with the profound neuron loss and neuroinflammation observed in the human AD brain. To address this knowledge gap, we quantified Nv, Gv and G : N in postmortem tissue obtained from a large cohort of chimpanzee brains with markers of AD pathology.

Neuron density decreased moderately with age in the CA1 (12%) and CA3 (19%) hippocampal subfields in this sample of chimpanzees. These data diverge from prior investigations in older non-human primates and humans [5,12,65,84]. Previously, no changes in neuron number with age were found in CA1 and CA3 for chimpanzees and rhesus macaques [65,84]. However, both studies were limited in the number of animals (chimpanzees = 6; macaques = 8) compared to the current study of 28 apes. Neuron numbers were reported rather than densities, and both the polymorphic and pyramidal layers were measured in macaques. In elderly humans, the majority of studies of CA1 and CA3 neurons demonstrated preservation during physiological ageing [5,9,10,12,85]. Although the human and chimpanzee hippocampus appears to be differentially affected by ageing, this region is vulnerable to neuronal death in both species with age and neurodegeneration.

A non-significant trend of neuron loss (20-24%) was associated with Aβ42 vascular deposition in the temporal cortex of chimpanzees. Previously, we found that Aβ42 was threefold higher in neocortical vessels compared to the hippocampus [34]. In addition, Aβ plaque and vessel volumes were significantly greater in older apes, suggesting that age-related increases contribute to neuronal toxicity in MTG of aged chimpanzees. This concept is further supported by the lack of neuron loss with age in this region. Notably, pretangle densities were highest in the MTG compared to the hippocampus, although pretangles did not correlate with neuron loss [34,40]. AD pathologic markers in great apes demonstrate similar regional staging progression as humans with Aβ deposits occurring first in the neocortex, whereas NFT initiate in the medial temporal lobe and brainstem [86,87]. Thus, MTG may represent an area in which Aβ and tau pathologies converge in chimpanzees. Moreover, MTG exhibits the largest, albeit nonsignificant, decrease in Gv (18%) and G : N (34%) with age and the lowest density of microglia in these apes (electronic supplementary material, table S3). In AD, microglia play an important role in the removal of Aβ peptides through phagocytosis, and physiological senescence in microglia is accompanied by the release of inflammatory cytokines, which have detrimental effects on neurons [88]. Consequently, in chimpanzees, the low number of microglia in MTG may result in decreased Aβ42 phagocytosis, exacerbated by an age-associated increase in Aβ42-ir vasculature and high numbers of pretangles, leading to neuronal death.

Humans with AD differ from chimpanzees in that significant neuronal loss occurs in the PFC and CA1 as well as the temporal cortex [5,14,16,89,90]. The absence of cell death in the chimpanzee hippocampus could be due to the increase in severity of NFT pathology in AD relative to milder NFT densities in aged chimpanzees. In AD, NFT correlates strongly with neuronal loss and cognitive decline [91,92]. Therefore, we compared CA1 Nv in apes with pretangles (13 581 mm⁻³) or NFT (115 959 mm⁻³) to those without (15 025 mm⁻³), and observed a trend of mild decline with tau pathologies. Moreover, a 57-year-old male chimpanzee with the highest CA1 pretangle and NFT densities exhibited the lowest Nv (6902 mm⁻³) of all the apes and was an outlier in our original analysis, which showed a negative correlation of CA1 pretangle density and Nv (electronic supplementary material, figure S2C). Once this animal was removed, the association no longer remained (electronic supplementary material, figure S2D). Furthermore, like humans, Aβ plaque and vascular levels were higher in the neocortical regions compared to the hippocampal subfields in these chimpanzees.

Neocortical and hippocampal glia densities were not associated with age or AD pathology in chimpanzees. Our previous findings support the lack of age and pathology-related variation in microglia and astrocyte densities in the same apes and brain regions [74,75]. Human studies also found that astrocyte numbers did not change with age or AD pathology, and microglia density did not differ in the temporal cortex of control and AD brains [19,20,93]. Instead, expression of GFAP and MHC-II, markers of astrocyte and microglia activation, increased in AD patients, implicating a phenotypic change instead of a marked proliferation [93]. Greater microglial activation was also detected in the hippocampus in elderly non-demented subjects [23]. Non-human primates demonstrate a similar pattern. Rhesus monkeys exhibited both senescence-associated expression of MHC-II and activated microglia concomitant with fibrillar Aβ plaques with clusters of phosphorylated tau-containing swollen neurites [72,94,95]. Likewise, Aβ oligomers trigger astrocyte and microglial activation in long-tailed macaques, and microglial activation was observed in the brains of aged common marmosets in conjunction with Aβ and hyperphosphorylated tau deposition [46,96]. In chimpanzees, greater volumes of Aβ42 plaque and vessel deposition were correlated with higher hippocampal microglial activation, while tau deposition was significantly increased in activated, intermediate microglia [34]. This evidence suggests that overall glia densities are not impacted by age or AD pathology in great apes, although glial activation increases in proximity to Aβ and tau deposition.

Unexpectedly, the breakdown of glial subtypes in chimpanzees (81% oligodendrocytes, 8% astrocytes and 11% microglia) diverges from estimates in humans (males: 75% oligodendrocytes, 20% astrocytes, 5% microglia), who have a higher percentage of astrocytes and lower percentage of microglia [20]. Rhesus monkey brains have a distribution of 35% oligodendrocytes, 57% astrocytes and 7% microglial cells in the cortex [97]. Although potentially a result of different quantification methods, the species-related variation in glial makeup may be an important difference in humans and could play a role in the reduced neuroinflammation observed in ape brains during ageing, although further analyses are required.

Glia : neuron ratios were not associated with age or pathology in chimpanzees. Average G : N were higher in the hippocampus (2.36) compared to the neocortex (1.57). These results are in accordance with previous studies in other non-human primates and humans. A neocortical G : N of 1.70 was reported in rhesus monkeys and 1.98 in chimpanzees, while human cortical grey matter ranges from 0.6 to 4.0 with an average of 1.79 [72,98,99]. In addition, a G : N of 1.37 was found in the neocortex of AD patients and controls, suggesting the ratio does not change with disease [21].

Comparing differences in the brains of humans and great apes can enhance our understanding of the selective vulnerability of humans to neurodegenerative diseases, such as AD (table 1). In the current study, we found that chimpanzees experience limited decreases in Nv in association with physiologic and pathologic ageing, although not to the same severity as AD. A caveat of this investigation, though, is the lack of antemortem cognitive testing, which is necessary to determine whether neuronal loss is associated with cognitive decline in aged chimpanzees. Moreover, the contribution of life-history factors, such as social environment, diet, body metabolism and physical activity, that may affect neurological variation were not examined in these animals. These data highlight and further support the role for non-human primates as models of ageing and neurodegenerative diseases due to their long lifespans, genomic similarities, and complex physiology and cognition.

Table 1.

Summary of neuropathologic changes in ageing and Alzheimer's disease in chimpanzees and humans.

pathology	Alzheimer's disease		ageing
	chimpanzee	human	chimpanzee	human
Aβ	Aβ is primarily in blood vessels and occurs prior to plaque formation	Aβ is primarily in plaques, although CAA occurs in 80% of AD patients	chimpanzees demonstrate increased intravascular Aβ deposition with age	humans exhibit increased Aβ deposition in plaques and vessels with age
	Aβ pathology is associated with increased tau pathology
tau	neuritic clusters contain dystrophic tau neurites but lack an Aβ core	neuritic plaques contain an Aβ core	pretangles in the neocortex increase with age	NFT increase with age in the hippocampus
neuro-inflammation	Aβ42 is correlated with increased microglial activation (i.e. plaques in humans, vessels in chimpanzees)		age is not associated with increased microglial or astrocyte density or activation in chimpanzees	humans display both increased microglial activation and density with age
	microglial activation is correlated with Aβ but not NFT lesions	microglial activation is associated with Aβ and NFT pathology
neuron loss	neuron loss occurs in the temporal cortex in association with Aβ deposition in the brain's blood vessels	selective neuronal loss occurs in the prefrontal cortex and hippocampus	neuron loss was observed in the hippocampus of chimpanzees	selective neuronal loss occurs in the subiculum and dentate gyrus but not CA1-CA3 subfields
cognitive impairment	antemortem cognitive testing in chimpanzees with AD pathology has not been performed yet	severe memory, cognitive and behavioural deficits are observed	cognitive testing is rare in aged apes; mild cognitive deficits in spatial memory, attention, executive function, and cognitive flexibility have been noted	processing speed, attention, memory and cognitive flexibility gradually decline during ageing

Ethics

All samples were acquired postmortem from American Zoo and Aquarium-accredited zoos and American Association for Accreditation of Laboratory Animal Care-accredited research institutions through the National Chimpanzee Brain Resource.

Data accessibility

Data will be available on the National Chimpanzee Brain Resource website: https://www.chimpanzeebrain.org.

Authors' contributions

All authors have approved this manuscript. M.K.E., P.R.H. and M.A.R. designed research; C.C.S., W.D.H., J.J.E., J.M.E. and P.R.H. provided samples; M.K.E. and E.L.M. collected data; M.K.E., R.S.M. and M.A.R. analysed data; and M.K.E., C.C.S., R.S.M., W.D.H., J.J.E., J.M.E., E.J.M., P.R.H. and M.A.R. wrote the manuscript.

Competing interests

The authors declare no competing interests.

Funding

This work was supported and funded by the NSF (BCS-1316829, M.A.R.), NIH (NS042867, NS073134, NS092988, W.D.H., C.C.S.; AG017802, J.J.E.; AG014308, J.M.E.; AG005138, P.R.H.; AG014449, AG043375, E.J.M.), James S. McDonnell Foundation (220020293, C.C.S.), Sigma Xi (M.K.E.), KSU Research Council (M.A.R.) and KSU Graduate Student Senate (M.K.E.).