Neuronal loss without amyloid‐β deposits in the thalamus and hippocampus in the late period after middle cerebral artery occlusion in cynomolgus monkeys

Abstract

Conflicting evidence exists regarding whether focal cerebral infarction contributes to cerebral amyloid‐β (Aβ) deposition, as observed in Alzheimer's disease. In this study, we aimed to evaluate the presence of Aβ deposits in the ipsilateral thalamus and hippocampus 12 months post‐stroke in non‐human primates, whose brains are structurally and functionally similar to that of humans. Four young male cynomolgus monkeys were subjected to unilateral permanent middle cerebral artery occlusion (MCAO), and another four sham‐operated monkeys served as controls. All monkeys underwent magnetic resonance imaging examination on post‐operative day 7 to assess the location and size of the infarction. The numbers of neurons, astrocytes, microglia and the Aβ load in the non‐affected thalamus and hippocampus ipsilaterally remote from infarct foci were examined immunohistochemically at sacrifice 12 months after operation. Thioflavin S and Congo Red stainings were used to identify amyloid deposits. Multiple Aβ antibodies recognizing both the N‐terminal and C‐terminal epitopes of Aβ peptides were used to avoid antibody cross‐reactivity. Aβ levels in cerebrospinal fluid (CSF) and plasma were examined using enzyme‐linked immunosorbent assay. The initial infarct was restricted to the left temporal, parietal, insular cortex and the subcortical white matter, while the thalamus and hippocampus remained intact. Of note, there were fewer neurons and more glia in the ipsilateral thalamus and hippocampus in the MCAO group at 12 months post‐stroke compared to the control group (all P < 0.05). However, there was no sign of extracellular Aβ plaques in the thalamus or hippocampus. No statistically significant difference was found in CSF or plasma levels of Aβ₄₀, Aβ₄₂ or the Aβ₄₀/Aβ₄₂ ratio between the two groups (P > 0.05). These results suggest that significant secondary neuronal loss and reactive gliosis occur in the non‐affected thalamus and hippocampus without Aβ deposits in the late period after MCAO in non‐human primates.

Introduction

Apart from physical disability, more than half of stroke survivors experience significant cognitive impairment, which poses a major impediment to their return to premorbid life 37. Cerebral amyloid‐β (Aβ) deposition, a major component of senile plaques and a pathological hallmark of Alzheimer's disease (AD), has been proposed as a pathologic substrate for post‐stroke cognitive decline 19, 33, 44. In rodent studies, robust evidence exists concerning altered amyloid precursor protein (APP) processing, the presence of Aβ deposits and loss of neurons in the ipsilateral thalamus following focal cerebral infarction, even at 9 months after the initial ischemic insult, whereas the only short‐term non‐human primate study using lissencephalic marmosets identified no Aβ pathology in the thalamus during a 45‐day follow‐up after cerebral infarction 14, 42, 47, 52, 53, 58, 59. In contrast, post‐mortem studies among long‐term stroke survivors or AD patients have generally reported a weak association or no association between the burden of cerebrovascular lesions and cerebral amyloid pathology compared to age‐matched controls, suggesting that cerebrovascular diseases may be just one of many contributors to Aβ deposition 1, 16, 31. Recently, a few more clinical studies used ¹¹C‐PiB positron emission tomography (PET) to assess the burden and the distribution of Aβ deposits after stroke in vivo 26, 28, 41, 51. It was reported that acute ischemic stroke does not lead to an increase in global neocortical ¹¹C‐PiB accumulation during an 18‐month follow‐up, while another 3‐year longitudinal study argued that stroke survivors with greater baseline Aβ levels suffer from a more rapid and severe cognitive decline 26, 28, 41, 51. Of note, these results from clinical researches were affected by many important confounding factors, including normal aging, pre‐existing AD pathology, diabetes mellitus and apolipoprotein E (APOE) gene variation 16. Therefore, whether acute focal ischemic stroke triggers cerebral Aβ deposition in the late post‐stroke period is still unknown, especially in high‐order non‐human and human species.

In this study, we performed permanent occlusion of the middle cerebral artery (MCAO) in young male cynomolgus monkeys (Macaca fascicularis), whose brains bear more genetic and anatomical similarities to that of humans than that of lissencephalic non‐human primates and rodents 8, 38. Using this model, we aimed to evaluate the presence of Aβ deposits and neuronal loss at 12 months post‐stroke in the thalamus and hippocampus, two remote cerebral areas connected to the primary ischemic insult but without ischemic lesions.

Materials and Methods

Animal groups and MCAO model

All experimental procedures were performed in accordance with the ARRIVE guidelines for the care and use of laboratory animals and were approved by the Institutional Animal Care and Use Committee (IACUC, Approval No: LD20130728). Eight young male cynomolgus monkeys (4 to 5 years old, 4.5 to 5.5 kg, provided by Guangdong Landao Biotechnology Co., Ltd, Guangzhou, China) were used in the present study. Before operation, each monkey was randomly assigned to receive permanent middle cerebral artery occlusion (MCAO) (n = 4) or sham procedures (n = 4). Monkeys were housed individually indoors at a controlled temperature of 24–28°C on a 12‐h light/dark cycle (07:00 to 19:00) with free access to water and a mixed diet. The details of the MCAO surgical procedure have been described previously 6. This permanent MCAO model of cynomolgus monkey has been proved to be suitable for the investigation of the secondary damage in remote regions after focal cerebral infarction, including the ipsilateral thalamus and hippocampus 6. Anesthesia was induced with ketamine (10 mg/kg, intramuscular injection) and maintained with isoflurane (1.5–3%, inhalation). The monkeys were fixed in a stereotaxic frame and a left frontotemporal pterional craniotomy was performed. Then, the left middle cerebral artery (MCA) was exposed and occluded in the distal M1 branch of the MCA by bipolar electrocoagulation. The occluded MCA was cut to avoid recanalization. Finally, the bone window was waxed, and the incision was sutured closed. Monkeys that were subjected to all the procedures described above, except coagulating the MCA, served as the sham group. After the surgery, the monkeys were placed in the recovery room on a heating pad and housed individually in a single cage. Penicillin was administered within 3 days after surgery (0.4 million IU, intramuscular injection, twice a day). Intensive care was given until the monkeys were able to self‐care, eat and drink.

Neurological assessment

Neurological assessments were performed before surgery, 1 week after surgery and prior to sacrifice by two investigators according to a standardized neurological deficit score 23. In brief, this scale includes the following four categories: consciousness, sensory system, motor system and skeletal muscle coordination. The points for each category are 28, 22, 32 and 18, respectively. Out of a total of 100 points, 0 corresponded to normal behavior and 100 indicates severe bilateral neurological function impairment.

Magnetic resonance imaging scanning and estimation of infarct volumes

Magnetic resonance imaging (MRI) scanning was performed at baseline, 7 days after operation and prior to sacrifice on a Siemens' 3.0‐T Trio system (Siemens, Germany). T1‐ and T2‐weighted and fluid attenuation inversion recovery (FLAIR) images were acquired in the axial plane according to our previous study 6. We also acquired T1‐weighted three dimensional magnetization prepared rapid acquisition gradient echo (T1WI‐3D‐MPRAGE) images using the following scan parameters: repetition time/echo time/flip angle = 2000 ms/2.98 ms/12°, 128 × 128 matrix dimensions, 45 slices and 1 × 1 × 1 mm³ voxel size. The ischemic lesion volume was measured from T1WI‐3D‐MPRAGE images using MRIcron and Matlab software by two observers, as described previously 57. Infarct volume was expressed as a percentage of the ipsilateral hemisphere using the following formula: (volume of infarct of the ipsilateral hemisphere/total volume of the ipsilateral hemisphere) × 100% 50.

Tissue preparation

Monkeys were sacrificed at 12 months after surgery. Monkeys were deeply anesthetized with pentobarbital (50 mg/kg, intramuscular injection) and transcardially perfused with 2 L of 0.9% of saline at 4°C, followed by 3 L of 4% of paraformaldehyde in phosphate buffer (0.1 mol/L, pH 7.4). The brains were removed, cut into 4‐mm‐thick coronal sections using a cynomolgus brain matrix (Neuroscience, Osaka, Japan) and post‐fixed in the above fixative for 24 h. Coronal sections cut from the anterior commissure (ac) −6 to −10 mm were used. Then, the ipsilateral and contralateral thalamus and the hippocampus were separated from these sections according to the cynomolgus brain atlas and sliced into 5‐μm‐thick paraffin‐embedded sections for further analysis 4, 45.

Hematoxylin and eosin staining

For hematoxylin and eosin (HE; Richard‐Allan Scientific^®, Thermo Scientific) staining, air‐dried sections on slides were stained using standard histological procedures. The histological examination was performed using a light microscope (Leica DM6B).

Thioflavin S staining and Congo Red staining

For Thioflavin S (ThS) staining, brain sections were incubated with 1% of thioflavin S (T1892, Sigma Aldrich) dissolved in 70% of ethanol for 10 minutes and followed by two washes of 70% of ethanol for 10 minutes each 5. For Congo Red (CR) staining, brain sections were stained using Amyloid Stain, Congo Red, kit (HT60, Sigma Aldrich) according to the manufacturer's operating instructions 5. The brain sections from transgenic AD mice 5XFAD (9 months old) were used as a positive staining control. The results were examined using an Olympus BX63 fluorescence microscope, and an Olympus BX51 microscope, respectively.

Immunochemistry

Aβ staining was examined using multiple antibodies recognizing the amino‐terminal (Aβ_1‐16), central (Aβ_17‐24) and carboxy‐terminal epitope (Aβ₄₀, Aβ₄₂) within the Aβ peptide to avoid cross‐reaction with APP 2, 12, 18, 24. Briefly, citrate buffer pre‐treated sections (10 minutes, 85°C) were treated with 3% of hydrogen peroxide for 15 minutes for quenching endogenous peroxidase activity, and blocked with 5% normal goat serum for 1 h at room temperature. Then, the sections were incubated with the following primary antibodies at 4°C overnight: mouse anti‐Aβ_1‐16 (1:400, 6E10, Covance), mouse anti‐Aβ_17‐24 (1:400, 4G8, Covance), mouse anti‐Aβ₄₀ (1:200, Covance), mouse anti‐Aβ₄₂ (1:200, Covance), mouse anti‐APP (1:400, 22C11, Abcam), mouse anti‐phosphorylated tau (1:100, AT8, Thermo Scientific) and rabbit anti‐TAR DNA‐binding protein 43 (1:200, TDP‐43, Abcam) 43, 55. Brain tissue from AD patients and AD transgenic mice were used as positive control. Negative control sections were incubated with 0.01 M of phosphate buffered saline (PBS) instead of a primary antibody. After being rinsed with PBS for 10 minutes three times, sections were incubated for 1 h with a peroxidase‐marked rabbit/mouse secondary antibody (ready‐to‐use, #K5007, Dako). The signal was visualized using 3, 3‐diaminobenzidine (DAB) solution. The Aβ load was photographed in the thalamus and the hippocampus using an Olympus BX51 microscope.

Double‐labeled immunofluorescence

For double‐labeled immunofluorescence staining, sections were incubated with mixtures of rabbit and mouse primary antibodies overnight at 4°C: mouse anti‐Aβ_1‐16 (6E10), mouse anti‐Aβ_17‐24 (4G8), mouse anti‐Aβ₄₀, mouse anti‐Aβ₄₂, rabbit anti‐neuronal nuclei (NeuN; neuronal marker, 1:500, Millipore), rabbit anti‐ionized calcium‐binding adapter molecule 1 (Iba1; microglial marker, 1:500, Wako) or rabbit anti‐glial fibrillary acidic protein (GFAP; astrocytic marker, 1:500, Abcam). Then, sections were incubated for 1 h with secondary antibodies: Alexa Fluor 555‐conjugated goat anti‐mouse IgG or Alexa Fluor 488‐conjugated goat anti‐rabbit IgG (1:500, Cell Signaling Technology). Fluorescence signals were detected with a fluorescence microscope Olympus BX63.

Image analysis and quantification

The number of neurons, astrocytes and microglia immunopositive for NeuN, GFAP and Iba‐1, respectively, were counted within five non‐overlapping fields of the bilateral thalamus (the ventroposterior lateral nuclei, VPL) and the hippocampus (the CA4 sub‐region) at 200× magnification on each section, using Image J software (National Institutes of Health, Bethesda, USA). For quantitative analysis of Aβ deposition, the load of Aβ plaques was analyzed using ImageJ software in five non‐overlapping fields (using light microscopy at 200× magnification) on each section. The average number of Aβ plaques was then determined in the thalamus and hippocampus. The average number of Aβ plaques was categorized by applying the following ranked scale: minimal = fewer than 5 (+); moderate = 6 to 15 (++); or extensive = more than 15 (+++) 7.

Enzyme‐linked immunosorbent assay

Prior to sacrifice, the moneys' fasting venous blood and cerebrospinal fluid (CSF) were sampled for the measurement of Aβ concentrations 39. Serum was separated within 30 minutes after sampling, and CSF was obtained by lumbar puncture. Serum and CSF were stored at −80°C until further analysis. The total Aβ₄₀ and Aβ₄₂ levels in serum and CSF were quantified using commercial sandwich enzyme‐linked immunosorbent assay (ELISA) kits according to the manufacturer's instructions (Wako, Japan). The Aβ₄₀ and Aβ₄₂ concentrations were determined by comparison with the standard curve. The standard curve was linear over the range 0 to 100 pg/mL with R ² ≥ 0.98. Samples and standards were measured in duplicate and read with a spectrophotometer at 450 nm. The means of the duplicates were used for statistical analysis. Investigators were blinded to information about the samples.

Statistical analysis

All statistical analyses were performed in SPSS 23.0 for windows (SPSS Inc., Chicago, IL, USA). Numeric data were presented as mean ± standard error of the mean (SEM). Statistical analyses were conducted using two‐tailed unpaired Students' t tests for two‐group comparisons. Differences with a P‐value of < 0.05 was considered as statistically significant.

Results

Neurological deficits

All monkeys in the MCAO group exhibited similar neurological deficits after operation. They displayed severe motor difficulty in the contralateral (right) hand. The mean neurological score 1 week and 12 months after MCAO was 39.5 ± 3.0 and 21.0 ± 0.6, respectively. All monkeys in the sham‐operated group manifested no neurological deficits.

Infarct volume

The T2‐weighted images demonstrated a subcortical and cortical infarction in the left MCA distribution 1 week after MCAO. No apparent hyper‐intense signal was found in the thalamus or hippocampus in any of the images (Figure 1). The corresponding average infarct volume was 21.97 ± 1.28% of the ipsilateral hemisphere (Table 1). Obvious infarction with intact thalamus and hippocampus could be directly observed by HE staining in the MCAO group on gross brain examination (Figure 2). No infarct was detected on any T2‐weighted images or sections with HE staining in the control group.

Figure 1.

The ischemic stroke model of cynomolgus monkeys indicated by MRI and gross brain images. A–D, MRA images showed the occlusion of the distal M1 segment of the left MCA (indicated by white arrowhead) from 4 monkeys 1 week after MCAO. E–H, T2‐weighted images showed the cerebral infarction (indicated by the hyper‐intense areas) from 4 monkeys 1 week after MCAO. I–J, Representative gross brain images from a representative monkey in the MCAO group and the sham‐operative group after sacrifice. There is a gross view of the cerebral infarct (indicated by white arrow). Scale bar = 2 cm. Abbreviations: MCA, middle cerebral artery; MCAO, middle cerebral artery occlusion; MRA, magnetic resonance angiography; MRI, magnetic resonance imaging.

Table 1.

Infarct volume calculated based on TIWI‐3D‐MPRAGE images in the MCAO group.

Number	Volume of infarct (mm3)	Volume of ipsiltaral hemisphere (mm3)	Volume of infarct/Volume of ipsilateral hemisphere (%)
1	7245	28 394	25.52
2	7224	36 879	19.59
3	5642	27 107	20.81
4	6991	31 850	21.95

Abbreviations: TIWI‐3D‐MPRAGE = T1‐weighted three dimensional magnetization prepared rapid acquisition gradient echo; MCAO = middle cerebral artery occlusion.

Figure 2.

The representative imaging shows the location of the infarct, the intact ipsilateral thalamus and hippocampus. A, The macaque brain map shows areas of the thalamus and hippocampus delineated in red curve 4. B, The infarct on the left cerebral hemisphere was demarcated from the surrounding healthy tissue by the loss of cortical tissue in HE‐stained sections, indicated by the asterisk mark (**). The ipsilateral thalamus and hippocampus remained intact, delineated by the dark curve. Scar bar = 1 cm. Abbreviations: ac, anterior commissure; HE, hematoxylin and eosin; Hi, hippocampus; MCAO, middle cerebral artery occlusion; Th: thalamus.

Secondary degeneration in the ipsilateral thalamus and hippocampus

Secondary neuronal damage within the ipsilateral thalamus and hippocampus, as characterized by reduced NeuN+ neurons, was evident in the MCAO group compared to the control group (all P < 0.05). In contrast, the numbers of GFAP+ and Iba‐1+ cells within the ipsilateral thalamus and hippocampus were higher in the MCAO group than in the control group (all P < 0.05). (Figures 3 and 4).

Figure 3.

Secondary damage of the ipsilateral thalamus 12 months after MCAO. Representative immunostained images indicated the loss of neurons (A) and the proliferation of astrocytes (B) and microglia (C) in the ipsilateral thalamus (the ventroposterior lateral nuclei, VPL) after in MCAO group compared with control group. Data are expressed as mean ± SEM. n = 4. *P < 0.05 compared with the sham‐operated controls, #P < 0.05 compared with the contralateral thalamus. Scale bar = 50 μm. Abbreviation: MCAO, middle cerebral artery occlusion, SEM, standard error of the mean.

Figure 4.

Secondary damage of ipsilateral hippocampus 12 months after MCAO. Representative immunostained images indicated the loss of neurons (A), the proliferation of astrocytes (B) and microglia (C) in the ipsilateral hippocampus (the CA4 sub‐region) in the MCAO group compared with sham‐operated control group. Data are expressed as mean ± SEM. n = 4. *P < 0.05 compared with the sham‐operated controls. #P < 0.05 compared with the contralateral hippocampus. Scale bar = 50 μm. Abbreviation: MCAO, middle cerebral artery occlusion, SEM, standard error of the mean.

Aβ load in the thalamus and hippocampus

Chemical staining of Aβ identified no sign of ThS‐positive or CR‐positive plaques were observed in the ipsilateral thalamus and hippocampus at 12 months after MCAO (Figure 5). In the cortical tissue of post‐mortem AD brains that served as a positive control, all Aβ antibodies, including 4G8, 6E10, Aβ₄₀ and Aβ₄₂, recognized the extensive extracellular Aβ plaques in the cerebral cortex. In the four monkeys of the control group, no positive extracellular Aβ plaques were detected in the ipsilateral thalamus or hippocampus. At 12 months post‐stroke, both the thalamus and hippocampus were devoid of any sign of extracellular diffuse or dense plaque‐like Aβ deposits recognized by all Aβ specific antibodies. However, there were some intracellular immunoreactive granules in the bilateral hippocampus and thalamus detected by 4G8 and 6E10 antibodies, but these deposits did not react with Aβ₄₀ or Aβ₄₂ antibodies (Figure 6). To further investigate their expression pattern, double‐labeling immunofluorescence was performed using antibodies against 4G8 as well as the cell type‐specific markers NeuN, GFAP and Iba‐1 (Figure 7). Notably, double‐staining indicated that the immunoreactive granules were co‐localized with NeuN, but not GFAP or Iba‐1. Moreover, the intracellular deposits were found to be widely distributed in neurons at different thalamic and hippocampal regions, similar to the normal distribution of the APP protein recognized by 22C11 antibody (Figure 8). No cytoplasmic accumulation of phosphorylated tau protein or TDP‐43 was detected, too (Figures 5 and 9).

Figure 5.

Thioflavin S staining, Congo Red staining and phosphorylated tau protein immunostaining of ipsilateral thalamus and hippocampus at 12 months after MCAO. (A and D) Aβ plaques in brain sections from AD transgenic mice were recognized by thioflavin S (ThS) staining and congo red (CR) staining (indicated by black arrow). No sign of ThS‐positive or CR‐positive plaques were observed in the ipsilateral thalamus (B and E) and hippocampus (C and F) at 12 months after MCAO. There are scattered AT8‐positive neurons detected in the cortex (G, indicated by black arrow) in the cortex section from AD transgenic mice. No sign of AT8‐positive cells were seen in the ipsilateral thalamus (H) and hippocampus (I). Scale bar = 50 μm. Abbreviations: AD, Alzheimer's disease; MCAO, middle cerebral artery occlusion.

Figure 6.

Representative images of Aβ pathology in ipsilateral thalamus and hippocampus across different Aβ antibodies at 12 months after MCAO. Aβ was identified by antibodies 4G8, 6E10, Aβ₄₀ and Aβ₄₂, respectively. No extracellular Aβ plagues were found in ipsilateral thalamus (the ventroposterior lateral nuclei, VPL) and the hippocampus (the CA4 sub‐region) at 12 months after MCAO. Autopsy cortical brain cortical tissues from AD patients were used as positive control. n = 4. Scale bar = 50 μm. Abbreviations: Aβ, β‐amyloid; AD, Alzheimer's disease; MCAO, middle cerebral artery occlusion.

Figure 7.

Double‐immunofluorescence labelling for Aβ with NeuN, GFAP and Iba1 in the ipsilateral thalamus (A) and hippocampus (B) at 12 months after MCAO. Arrows indicate co‐localization of neurons with 4G8. n = 4. Scale bar = 50 μm. Abbreviations: Aβ, β‐amyloid; MCAO, middle cerebral artery occlusion.

Figure 8.

Expression of APP in ipsilateral thalamus and hippocampus 12 months after MCAO. APP immunostaining of ipsilateral thalamus and hippocampus in sham‐operated group (A and C) and MCAO group (B and D) showed positive expression detected by 22C11 antibody in the cytoplasm of cells in ipsilateral thalamus and hippocampus. Scale bar = 50 μm. Abbreviations: APP, amyloid precursor protein; MCAO, middle cerebral artery occlusion.

Figure 9.

Expression of TDP‐43 in ipsilateral thalamus and hippocampus 12 months after MCAO. TDP‐43 immunostaining of ipsilateral thalamus and hippocampus showed that TDP‐43 remained in the nuclei of cells (indicated by black arrow). No cytoplasmic inclusion of TDP‐43 was detected. Scale bar = 100, 50 μm. Abbreviations: MCAO, middle cerebral artery occlusion; TDP‐43, TAR DNA‐binding protein 43.

Aβ levels in CSF and plasma

First, we compared the CSF Aβ levels (pmol/L) in the MCAO and control groups. There was no significant change in CSF levels of Aβ at 12 months post‐stroke relative to that of the control group (Aβ₄₀: 1371.0 ± 243.4 vs. 1191 ± 109.4, P = 0.523; Aβ₄₂: 143.7 ± 23.4 vs. 131.1 ± 11.7, P = 0.649). No significant difference was found in the CSF levels of Aβ₄₀/Aβ₄₂ at 12 months (8.79 ± 0.98 vs. 9.07 ± 0.11, P = 0.787) in the stroke group compared with the control group. We next investigated the difference in plasma Aβ levels between the stroke and control groups. Similar to the CSF results, no significant difference was observed in the plasma levels of Aβ₄₀, Aβ₄₂ or the Aβ₄₀/Aβ₄₂ ratio at 12 months in the stroke group compared to the control group (Aβ₄₀:7.88 ± 1.39 vs. 8.41 ± 0.30, P = 0.720; Aβ₄₂: 1.14 ± 0.06 vs. 1.35 ± 0.08, P = 0.086; Aβ₄₀/Aβ₄₂: 6.88 ± 1.11 vs. 6.36 ± 0.53, P = 0.683). (Figure 10)

Figure 10.

Intergroup analyses of Aβ₄₀, Aβ₄₂ and the ratio Aβ₄₀/Aβ₄₂ levels in CSF and plasma at 12 months after MCAO. No significant differences were seen between MCAO group and control group in CSF Aβ₄₀, Aβ₄₂ and the ratio Aβ₄₀/Aβ₄₂ levels (A, P > 0.05). No significant differences were seen for plasma Aβ₄₀, Aβ₄₂ and the ratio Aβ₄₀/Aβ₄₂ levels as well (B, P > 0.05). n = 4. The error bars indicate the mean ± SEM. Abbreviations: Aβ, β‐amyloid; CSF, cerebrospinal fluid; SEM, standard error of the mean.

Discussion

In the present study, we showed that young cynomolgus monkeys displayed secondary neurodegeneration characterized by neuron loss and gliosis, with no sign of extracellular amyloid deposits in the ipsilateral thalamus or hippocampus at 12 months following permanent MCAO. This was surprising considering that progressive Aβ plaque accumulation was originally anticipated. To our knowledge, this study provides the first evidence from high‐order gyrencephalic non‐human primates to indicate that Aβ might not be the pathological substrate of secondary neurodegeneration following MCAO.

Extracellular Aβ plaques are one of the neuropathological hallmarks of AD 3, 46. Cerebral infarction has been reported to trigger accelerated amyloid deposition by increasing its production or interfering with its clearance, as shown in AD brains 14, 15, 29, 34, 40, 47, 49. The effect of cerebral infarction on Aβ deposition might help explain the association between ischemic stroke and AD observed in some epidemiological studies 17, 36. Moreover, post‐stroke Aβ pathology has been suggested as the possible underlying mechanism of post‐stroke cognitive impairment 19, 44. Nevertheless, the past two decades of preclinical and clinical studies have yielded conflicting results.

An important explanation for the current discrepancy is the species difference between rodents and high‐order primates. To avoid this problem, our present study used a model of MCAO in gyrencephalic non‐human primates, whose brain is structurally and functionally similar to that of humans 38. The structure of rodent APP is not similar to that of humans. As a result, the proteolytic processing of APP and the metabolism of Aβ peptides may be different. By contrast, the homology of APP in monkeys and humans has been well documented 38. Moreover, it has been shown that Aβ peptides do not aggregate in murid rodents and, therefore, are non‐plaque‐forming in nature. On the contrary, Aβ peptides in non‐human primates have the capacity to aggregate and form senile plaques, like in humans 10. The maximum life span of old‐world monkeys is about 40 years, and spontaneous and age‐related presence of extracellular Aβ plaques and other AD‐associated proteins in the brain have been shown in aged cynomolgus monkeys over 20 years old 13, 21, 22, 56. Thus, gyrencephalic non‐human primates, like cynomolgus monkeys, provide a more applicable animal model to study cerebral Aβ pathology and other fundamental molecular mechanisms of human neurodegenerative diseases 10, 38. Moreover, we used young monkeys in the present study to avoid the confounding effect of aging on cerebral Aβ deposition 21, 22, 56.

There is evidence of the concomitant presence of Aβ plaques in post‐mortem human brains with verified ischemic lesions, but the reported association is weak 1, 16, 19, 31, 44. To explore the relationship further, ¹¹C‐PiB PET scanning has been used recently to assess cerebral Aβ after stroke onset in vivo in several clinical studies 32. A study with a follow‐up of 18 months found no significant increase in ¹¹C‐PiB accumulation in the ipsilesional hemisphere after acute ischemic stroke 41. No difference in the prevalence of Aβ deposition was found between patients with post‐stroke cognitive impairment and cognitively healthy stroke survivors in a study of patients followed for 6 months 51. Another cross‐sectional study among older people concluded that cerebrovascular diseases and Aβ aggregation appear to be two independent processes within the spectrum of normal aging 30. On the contrary, a prospective PET study with a 3‐year follow‐up suggested that, if present, amyloid positivity may be associated with more rapid and severe cognitive decline after ischemic stroke and transient ischemic attack (TIA) 26. Together, the preliminary evidence from human studies tends to draw a negative conclusion on the relationship between stroke and Aβ pathology, though this is unconfirmed. Moreover, these clinical studies are limited by their relatively small sample sizes, and patient heterogeneity in terms of age, stroke type, location and etiopathogenesis. The relationship is also blurred by many complex confounding factors like diabetes mellitus, apolipoprotein E (APOE) gene variation, white matter lesions and pre‐existing AD pathology before the stroke occurs 16, 20, 54. Besides, ¹¹C‐PiB PET scanning can only detect insoluble Aβ deposits in the form of extracellular plaques 34. In our study, we did not perform in vivo ¹¹C‐PiB PET scanning in cynomolgus monkeys for technical reasons.

Animal studies allow for better control of confounding factors and direct examination of brain tissue. In earlier rodent studies, post‐stroke Aβ deposits were found to be common, while the only published non‐human primate study using marmosets identified no Aβ pathology in the thalamus in a 45‐day follow‐up post‐stroke 14, 25, 35, 43, 47, 52, 53, 58, 59. We have previously demonstrated consistent post‐stroke Aβ deposits in the thalamus of rats at 7 and 14 days after permanent MCAO 52, 53, 58, 59. Another study with a longer follow‐up period of 9 months also showed the persistent presence and aggregation of Aβ peptides to dense plaque‐like deposits in the thalamus of rats subjected to transient MCAO 47. In addition, increased APP expression has also been observed in the ischemic penumbra as well as the thalamus in experimental models of stroke 14. However, the post‐stroke amyloid pathology might be more dynamic than expected. One hypothesis proposes that Aβ is an acute‐phase protein produced in stressful situations and that it is sufficiently eliminated with time 7, 11, 14, 48. Some rodent studies have found that the Aβ deposits could be induced shortly after stroke but vanish within a few weeks, possibly caused by an increased clearance of pre‐existing Aβ deposits, suggesting that cerebrovascular lesions only induce transient but not permanent Aβ deposition 11, 14. A clinical study using ELISA to examine serum Aβ levels found that serum Aβ₄₀ and Aβ₄₂ levels were increased at day 1, reached peak levels at day 3 and decreased to pre‐stroke levels at day 7, indicating that the elevation of serum Aβ levels was also transient 27. In our study, no significant difference was found between monkeys with MCAO and controls in terms of Aβ₄₀ and, Aβ₄₂ levels, or the Aβ₄₀/Aβ₄₂ ratio in CSF and plasma at 12 months after MCAO. The dynamic nature of post‐stroke amyloid deposits and clearance may help explain the absence of Aβ plaques months to years after stroke in our models of young cynomolgus monkeys. A limitation of our study is that we did not collect brain tissues, CSF, or blood samples at earlier time points after stroke, which prevented us from investigating the dynamic change in Aβ. Moreover, we did not examine the protein levels of Aβ degrading enzymes, like insulin degrading enzyme and neprilysin 11. No specific experiments were carried out to test amyloid clearance along the interstitial fluid drainage pathways in our study either 11.

In addition, the cross‐reactivity of commonly used Aβ antibodies leads to significant uncertainty in the interpretation of results and may explain the inconsistent results to some extent 2, 12, 18, 24. Some previous rodent studies of stroke have only used a single Aβ antibody, like 6E10, which recognize the N‐terminal epitope of Aβ 52, 53, 58, 59. The brain Aβ immunoreactivity after stroke observed in these previous studies could be compromised by full‐length APP, as antibodies like 6E10 and 4G8 have been reported to cross‐react with APP, which is normally much more abundant than Aβ in the brain 18. To unambiguously confirm the presence of Aβ, especially intracellular Aβ, multiple antibodies, including antibodies to the N‐terminal and C‐terminal neoepitopes of the Aβ peptide, must be used 2, 12, 18, 24. We also used brain tissues from AD post‐mortem brains as positive controls to demonstrate the specificity of antibody staining and to increase confidence that the immunoreactive material is indeed Aβ 18. We detected intraneuronal immunoreactive granules in the bilateral hippocampus and thalamus using 4G8 and 6E10 antibodies, but these deposits did not react with Aβ₄₀ or Aβ₄₂ antibodies that are specific for C‐terminal epitopes; thus, the immunoreactivity revealed by 4G8 and 6E10 is not necessarily intraneuronal Aβ. Moreover, the intracellular deposits are found to widely distribute in neurons at different thalamic and hippocampal regions bilaterally, similar to the normal distribution of the APP protein 9. We suspect that this is likely because of the cross‐reactivity with full length APP or other APP proteolytic fragments.

In conclusion, in contrast to the results obtained in rodents, our present study found no sign of extracellular Aβ plaque deposits in the bilateral thalamus or hippocampus in a young cynomolgus model of MCAO during a long‐term follow‐up period. This study provides the first evidence from high‐order gyrencephalic non‐human primates and may help resolve the discrepancy between rodent and human studies. The absence of Aβ plaque deposition in long‐term survivors of ischemic stroke fails to provide an AD‐related pathological substrate for cognitive impairment following stroke. Other neurodegenerative mechanisms might be involved, and this requires further studies. An obvious limitation of the present study was the small sample size and the limited methodology. Due to the small sample size, we failed to collect brain homogenate of thalamus and hippocampus to do additional experiments using techniques for Aβ quantification, including western blot, ELISA and polymerase chain reaction 5. Moreover, we only investigated one time‐point in the present study and we cannot rule out the possibility that Aβ deposits would occur at a later period after stroke 25. Therefore, researches with larger sample size, more comprehensive methodology and longer follow‐up period are still needed to help add further evidence of post‐stroke Aβ pathology.

Disclosure

None.

DATA AVAILABILITY STATEMENT

Data available on request from the corresponding author upon reasonable request.