Alzheimer's associated amyloid and tau deposition co-localizes with a homeostatic myelin repair pathway in two mouse models of post-stroke mixed dementia

Abstract

The goal of this study was to determine the chronic impact of stroke on the manifestation of Alzheimer's disease (AD) related pathology and behavioral impairments in mice. To accomplish this goal, we used two distinct models. First, we experimentally induced ischemic stroke in aged wildtype (wt) C57BL/6 mice to determine if stroke leads to the manifestation of AD-associated pathological β-amyloid (Aβ) and tau in aged versus young adult wt mice. Second, we utilized a transgenic (Tg) mouse model of AD (hAPP-SL) to determine if stroke leads to the worsening of pre-existing AD pathology, as well as the development of pathology in brain regions not typically expressed in AD Tg mice. In the wt mice, there was delayed motor recovery and an accelerated development of cognitive deficits in aged mice compared to young adult mice following stroke. This corresponded with increased brain atrophy, increased cholinergic degeneration, and a focal increase of Aβ in areas of axonal degeneration in the ipsilateral hemisphere of the aged animals. By contrast, in the hAPP-SL mice, we found that ischemia induced aggravated behavioral deficits in conjunction with a global increase in Aβ_, tau, and cholinergic pathology compared to hAPP-SL mice that underwent a sham stroke procedure. With regard to a potential mechanism, in both models, we found that the stroke-induced Aβ and tau deposits co-localized with increased levels of β-secretase 1 (BACE1), along with its substrate, neuregulin 1 (NGR1) type III, both of which are proteins integral for myelin repair. Based on these findings, we propose that the chronic sequelae of stroke may be ratcheting-up a myelin repair pathway, and that the consequent increase in BACE1 could be causing an inadvertent cleavage of its alternative substrate, AβPP, resulting in greater Aβ seeding and pathogenesis.

Keywords: Alzheimer’s disease; Amyloid; BACE1; Mixed dementia; Myelin repair; Neuregulin; Stroke; Tau.

Fig. 1

Motor recovery is impaired and there is accelerated onset of cognitive impairment in aged versus young adult C57BL/6 mice following stroke. a Study design: 3 and 18 month-old (mo) C57BL/6 mice were assessed on the ladder rung and object relocation tests prior to a distal hypoxic (DH) stroke or sham surgery at the indicated timepoints. Following behavioral testing, mice were euthanized and brains were harvested for histology and immunostaining at either 8 or 12 weeks post-surgery. b Motor ability on the ladder rung test was assessed at 1 week pre-surgery, and at 2 days, 2 weeks, 4 weeks, and 6 weeks post-surgery. At 1 week prior to a stroke, there was no difference in motor performance between 3 and 18 mo mice, as both age groups displayed a similar number of correct (nearly 100%) foot placements on the rungs. At 2 days and 2 weeks post-surgery, 3 and 18 mo stroked mice exhibited a motor deficit, as both age groups displayed a significantly fewer number of correct foot placement relative to age-matched sham mice. However, at 4 and 6 weeks post-surgery, 3 mo stroked mice exhibited motor recovery, as they displayed an indistinguishable number of correct foot placements as age-matched sham mice, which was once again nearly 100% correct. 18 mo stroked mice, however, continued to exhibit a motor deficit relative to age-matched sham mice at 4 and 6 weeks post-surgery. c Cognitive function using the object relocation test was assessed at 1 week pre-surgery, and at 1 week, 4 weeks, and 7 weeks post-surgery. At 1 week prior to a stroke, there was no difference in cognitive function between 3 and 18 mo mice, as both age groups displayed significantly more interactions (sniffs and rears) with moved (novel location) verses unmoved objects relative to age-matched sham mice. However, at 4 weeks post-surgery, the 18 mo stroked mice exhibited cognitive dysfunction, as they were now interacting with the two sets of objects equally, indicating that they could no longer distinguish between the moved and unmoved objects. Not until 7 weeks post-surgery did the 3 mo stroked mice exhibit cognitive dysfunction. Data represent mean ± SEM. *p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001

Fig. 2

Brain atrophy and cholinergic degeneration is more pronounced in aged versus young adult C57BL/6 mice following stroke. a Representative 4× images of Nissl-stained whole brain sections from 3 and 18 month-old (mo), sham- and stroke-operated C57BL/6 mice at 8 weeks post-surgery. b Representative 2× images of Nissl-stained cortical layers I-VI of the ipsilateral primary somatosensory cortex in 3 and 18 mo, sham- and stroke-operated mice at 8 weeks post-surgery. Scale bar, 250 μm. c Quantification of the lateral ventricle (left graph) in the ipsilateral hemisphere revealed significant ventricle enlargement in the 18 mo stroked mice relative to the age-matched sham-operated mice. Furthermore, the 18 mo stroked mice had a significantly larger ventricle area compared to the 3 mo stroked mice. Quantification of cortical thickness (right graph) in the ipsilateral hemisphere revealed significant tissue loss or shrinkage in both the 3 and 18 mo stroked mice relative to age-matched sham-operated mice. Furthermore, the 18 mo stroked mice had significantly more tissue loss compared to the 3 mo stroke mice. d Representative 5× images of choline acetyltransferase (ChAT)-immunolabeled neuronal cell bodies, their neurites, and innervating projection fibers in the medial septum of the basal forebrain in 3 and 18 mo, sham- and stroke-operated C57BL/6 mice at 8 weeks post-surgery. Many neurites (arrows) in the stroked mice displayed qualitative degenerative changes, including decreased length relative to sham mice. Scale bar, 100 μm. e Quantification of cholinergic somas, neurites, and fibers revealed a significant reduction of ChAT+ staining in both the 3 and 18 mo stroked mice relative to the age-matched sham mice. Furthermore, the 18 mo stroked mice exhibited significantly more loss of ChAT+ staining compared to the 3 mo stroked mice. Data represent mean ± SEM. *p<0.05, **p<0.01, and ****p<0.0001

Fig. 3

Stroke causes β-amyloid (Aβ) and phosphorylated (p) tau deposition in the white matter tracts of aged wildtype (wt) mice compared to young adult mice. Representative 10× images of Aβ₄₂-immunolabeled deposits (arrows) in the white matter tracts of the (a) internal capsule, b thalamus, and (c) corpus callosum of the 3 and 18 mo, sham- and stroke-operated C57BL/6 mice at 8 weeks post-surgery. Scale bar, 100 μm (internal capsule and thalamus), 50 μm (corpus callosum). Nissl-stained sections to the left of each image delineate where representative images were taken. d Quantification of the ipsilateral hemisphere revealed a significant deposition of Aβ₄₂ in the internal capsule (top graph), thalamus (middle graph), and corpus callosum (bottom graph) of the 18 mo stroked mice relative to the age-matched sham-operated mice. Furthermore, the 18 mo stroked mice had significantly more Aβ₄₂ accumulation in three of the brain regions analyzed compared to the 3 mo stroked mice. e-h Representative 10× images of (e) Aβ₄₂- and (g) p-tau-immunolabeled deposits (arrows) in white matter tracts (thalamus-internal capsule) of the 18 mo mice at 12 weeks after sham or stroke surgery (Equivalent = area imaged in wt-sham mice that is equivalent to the ipsilateral hemisphere imaged in wt-stroke mice; Contralateral = area imaged in the contralateral hemisphere of wt-stroke mice that is equivalent to the ipsilateral hemisphere of wt-stroke mice). Scale bar, 125 μm (Aβ₄₂ and p-tau). Quantification of the ipsilateral and contralateral hemispheres revealed significantly more deposits of (f) Aβ₄₂ and (h) p-tau in the white matter tracts of the 18 mo stroked mice compared to the age-matched sham-operated mice. Furthermore, there was also significantly more Aβ₄₂ and p-tau accumulation in the white matter tracts of the ipsilateral versus the contralateral hemisphere. No Aβ₄₂ signal was detected in (i) astrocytes (GFAP, green; n=3 mice/experimental group) or (j) microglia (Iba1, green; n=3 mice/experimental group). Scale bar, 125 μm. Data represent mean ± SEM. *p<0.05, **p<0.01, and ***p<0.001

Fig. 4

Stroke induces β-secretase (BACE) 1 and neuregulin (NRG) 1 type III expression in the white matter tracts of aged wildtype (wt) mice compared to young adult mice. a-c Western blotting with a BACE1 (~70 kDa) or NRG type III (~50 kDa) antibody detected a single band at the appropriate molecular weight in mouse brain lysates for each protein. Representative 20× images (n=3 mice/experimental group) of (a) BACE1+ and (b) NRG1 type III+ immunostaining (arrows) in the white matter tracts (thalamus-internal capsule) of 18 mo mice at 12 weeks after sham or stroke surgery (Equivalent = area imaged in wt-sham mice that is equivalent to the ipsilateral hemisphere imaged in wt-stroke mice; Contralateral = area imaged in the contralateral hemisphere of wt-stroke mice that is equivalent to the ipsilateral hemisphere of wt-stroke mice). Scale bar, 50 μm (BACE1 and NRG1 type III). Images revealed BACE1 and NRG1 type III expression in the ipsilateral white matter tracts of stroked, but not in sham-operated mice. c Representative 10× immunofluorescence images (n=3 mice/experimental group) of total tau+ immunostaining in the ipsilateral and contralateral white matter tracts of wt mice after stroke surgery. Scale bar, 125 μm. d Image: Representative 40× Fluoro-Jade staining in the ipsilateral white matter tracts of wt mice after stroke confirms that this is an area of axonal degeneration following DH stroke. Scale bar, 100 μm. Graph: Relative to naïve mice, there is significant Fluoro-Jade staining starting at 1 week (wk) post-stroke, continuing into at least 8 wk post-stroke. Data represent mean ± SEM from n=5 mice/experimental group. *p<0.05

Fig. 5

Stroke exacerbates behavioral deficits in aged hAPP-SL mice on tests of motor, cognition, and anxiety. a Study design: 18 month-old (mo) hAPP-SL mice were assessed on the ladder rung, Y-maze spontaneous alternation behavior (SAB), novel object recognition, and light dark transition tests prior to a distal hypoxic (DH) stroke or sham surgery at the indicated timepoints. Mice were also weighed at each indicated timepoint. Mice were euthanized and brains were harvested for histology and immunostaining at 12 weeks post-surgery, along with spleen for organ weight. b Motor ability on the ladder rung test was assessed at 1-week pre-surgery, and at 1 week, 6 weeks, and 11 weeks post-surgery. At 1 week prior to a stroke, there was no difference in motor performance in the naïve hAPP-SL mice, as mice displayed a similar number of acceptable (<12% error) baseline foot placements on the rungs. These naïve mice would later be assigned into the sham or stroke experimental groups. At 1 week post-surgery, stroked hAPP-SL mice exhibited an enhanced motor deficit, as they displayed a significantly fewer number of correct foot placements relative to sham-operated hAPP-SL mice. Significant motor deficits continued to manifest in stroked hAPP-SL mice at 6 and 11 weeks post-surgery compared to sham-operated hAPP-SL mice at those timepoints. c Cognitive function using the Y-maze SAB test was assessed at 1 week pre-stroke, and at 1 week, 6 weeks, and 11 weeks post-surgery. At 1 week prior to surgery, there was no difference in cognitive function in naïve hAPP-SL mice, as mice displayed similar levels of spontaneous alternations. At 1 and 6 weeks post-surgery, there was no difference in the cognitive status of sham- and stroke-operated hAPP-SL mice. However, at 11 weeks post-surgery, stroked hAPP-SL mice exhibited aggrevated short-term spatial memory impairment, as they displayed significantly less spontaneous alternations compared to sham-operated hAPP-SL mice. d Cognitive function using the novel object recognition test was assessed at 1 week pre-stroke, and at 1 week, 6 weeks, and 11 weeks post-surgery. At 1 week prior to surgery, there was no difference in cognitive function in naïve hAPP-SL mice, as mice displayed similar recognition indexes, which corresponds to similar exploration time for an unfamiliar (novel) and a familiar object. At 1 and 6 weeks post-surgery, there was no difference in the cognitive status of the sham- and stroke-operated hAPP-SL mice. However, at 11 weeks post-surgery, the stroked hAPP-SL mice exhibited worsened intermediate recognition memory impairment, as they displayed significantly lower recognition indexes calculated from less time spent distinguishing and exploring an unfamiliar/novel object compared to the sham-operated hAPP-SL mice. e Using the light dark transition test, we assessed mice on the anxiety-impulsivity spectrum of behavior at 1 week pre-surgery, and at 1 week, 6 weeks, and 11 weeks post-surgery. At 1 week prior to surgery, there was no difference in the amount of time spent in the light, open (intimidating space) versus the dark, enclosed (safe space) arenas of the chamber in the naïve hAPP-SL mice. At 1 week prior to surgery, the amount of time spent in each arena remained similar between sham- and stroke-operated hAPP-SL mice, and this pattern was seen at 6 weeks post-surgery. However, at 11 weeks post-surgery, stroked hAPP-SL mice spent significantly more time in the light arena than sham-operated hAPP-SL mice, suggesting that stroke initiated behavioral impulsivity or a lack of inhibition in the hAPP-SL mice by reducing their anxiety of open spaces. f No significant weight changes were seen in the sham- and stroke-operated experimental hAPP-SL mouse groups at pre- and post-surgery timepoints. g There was no spleen weight difference between experimental groups at 12 weeks post-surgery. h There was no significant difference in any of the selected frailty outcomes depicted, with the exception of kyphosis, in the 18 mo sham- versus stroke-operated hAPP-SL mice at pre- and post-surgery timepoints. Data represent mean ± SEM. *p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001

Fig. 6

Stroke increases cholinergic neurodegeneration and levels of tau phosphorylation (p) in aged hAPP-SL mice. a Representative 10× images of choline acetyltransferase (ChAT)-immunolabeled dystrophic neurites (arrows) in the primary somatosensory cortex of 18 mo sham- or stroke-operated hAPP-SL mice (Equivalent = area imaged in wt-sham mice that is equivalent to the ipsilateral hemisphere imaged in wt-stroke mice). Scale bar, 125 μm. b Quantification revealed that relative to sham-operated hAPP-SL mice, the area occupied by cholinergic dystrophic neurites in the cortex was significantly higher in the stroked hAPP-SL mice; no significant difference in the amount of cholinergic dystrophic neurites was found between the ipsilateral versus contralateral cortex. c Representative 10× images of AT8 (p-tau^{Ser202/Thr205})-immunolabeled dystrophic neurites (arrows) in the primary somatosensory cortex of 18 mo sham- or stroke-operated hAPP-SL mice (Equivalent = area imaged in wt-sham mice that is equivalent to the ipsilateral hemisphere imaged in wt-stroke mice). Scale bar, 125 μm. d Quantification revealed that relative to sham-operated hAPP-SL mice, the area occupied by p-tau+ dystrophic neurites was significantly higher in the cortex (top graph), thalamus (middle graph), and internal capsule (bottom graph) of the stroked hAPP-SL mice; no significant difference in the amount of p-tau+ dystrophic neurites was found between the ipsilateral versus contralateral cortex and thalamus, although the ipsilateral internal capsule showed significantly more p-tau+ dystrophic neurites than the contralateral region. **p<0.01 and ***p<0.001

Fig. 7

Stroke exacerbates amyloid plaque burden and soluble Aβ₄₂ levels in aged stroke hAPP-SL mice. a Representative 4× stitched images of Thioflavin S (ThioS)-stained coronal brain sections of 18 mo sham- or stroke-operated hAPP-SL mice. b Quantification revealed that relative to sham-operated hAPP-SL mice, the area occupied by ThioS-labeled amyloid plaques was significantly greater in the cortex (top graph), hippocampus (middle graph), and thalamus (bottom graph) of the stroked hAPP-SL mice; no significant difference in the amount of amyloid plaques was found between the ipsilateral versus contralateral hemisphere. c Representative 10× images of Aβ₄₂-immunolabeled deposits (arrows) in the primary somatosensory cortex of the 18 mo sham- or stroke-operated hAPP-SL mice (Equivalent = area imaged in wt-sham mice that is equivalent to the ipsilateral hemisphere imaged in wt-stroke mice). Scale bar, 125 μm. d Quantification revealed that relative to sham-operated hAPP-SL mice, the area occupied by Aβ₄₂+ deposits was significantly higher in the cortex (top graph), thalamus (bottom middle graph), and internal capsule (bottom graph) of the stroked hAPP-SL mice (p value for the hippocampus shown in the top middle graph was 0.0751); no significant difference in the amount of Aβ₄₂+ deposits was found between the ipsilateral versus contralateral hemisphere. e A single-plex immunoassay of tissue samples from the ipsilateral cortex or thalamus/internal capsule regions showed that in hAPP-SL mice, significantly higher levels of soluble Aβ₄₂ are found in stroked mice compared to sham-operated mice. Data represent mean ± SEM. *p<0.05 and **p<0.01

Fig. 8

Stroke increases β-secretase (BACE) 1 and neuregulin (NRG) 1 type III immunostaining in aged hAPP-SL mice. a Representative 10× images of BACE1+ immunostaining (arrows) in the primary somatosensory cortex of 18 mo sham- or stroke-operated hAPP-SL mice. Scale bar, 125 μm. b Quantification revealed that relative to sham-operated hAPP-SL mice, the area occupied by BACE1+ staining was significantly higher in the cortex (top graph), hippocampus (middle graph), and thalamus (bottom graph) of the stroked hAPP-SL mice; no significant difference in the amount of BACE1+ staining was found between the ipsilateral versus contralateral hemisphere. c Representative 20× images of NRG1 type III+ immunostaining (arrows) in the primary somatosensory cortex of 18 mo sham- or stroke-operated hAPP-SL mice. Scale bar, 125 μm. d Quantification revealed that relative to sham-operated hAPP-SL mice, the area occupied by NRG1 type III+ staining was significantly higher in the cortex (top graph), hippocampus (middle graph), and thalamus (bottom graph) of the stroked hAPP-SL mice; no significant difference in the amount of NRG1 type III+ staining was found between the ipsilateral versus contralateral hemisphere. Data represent mean ± SEM. *p<0.05, **p<0.01, and ****p<0.0001. e Representative 40× fluorescence images (n=3 mice/experimental group) from the 18 mo stroked hAPP-SL mouse sections showing NRG1 type III staining (green) within cells staining for the astrocytic marker, glial fibrillary acidic protein (GFAP, magenta), in the ipsilateral and contralateral white matter tracts (thalamus-internal capsule). Magnified outsets show GFAP+ astrocytes colocalized with NRG1 type III. Scale bar, 50 μm

Fig. 9

Schematic representations summarizing key findings of our study. a In this study, we provide evidence that there is overlap of pathology (amyloid plaques, and Aβ₄₂ and p-tau deposition) associated with AD following stroke with key components (BACE1 and NRG1 type III) of a myelin repair pathway in aged stroked wt and stroked hAPP-SL mice. b Working model proposing AD-associated pathology as a chronic sequela of ischemic stroke: chronic inflammation, chronic BBB dysfunction, axonal degeneration, impaired paravascular clearance, and transneuronal degeneration are all chronic sequalae of stroke that may result in stressed myelin homeostasis. BACE1’s role as a protease that cleaves NRG1 type III at its β-site is critical in the formation of myelin sheaths by oligodendrocytes, and is a key regulator of myelination of axons in the central nervous system [10]. BACE1’s cleavage of NRG1 type III initiates γ-secretase cleavage of NRG1 type III at its γ-site. However, in addition to NRG1 type III, BACE1 has high cleavage affinity for the Aβ precursor protein (AβPP). Therefore, the chronic sequelae of stroke may be causing myelin repair mechanisms involving NRG1 type III and BACE1 to be racheted up, resulting in the inadvertent cleavage of AβPP by BACE1 and γ-secretase, and the abnormal generation of amyloidogenic Aβ peptides