Introduction
Amyloid beta (Aβ) plaques with extracellular fibrillar deposits of Aβ peptide and neurofibrillary tangles (NFTs) composed of phosphorylated tau represent the key histopathological hallmarks of Alzheimer’s disease [
22,
33,
58]. Aβ peptides are generated via sequential proteolytic cleavage of the amyloid precursor protein (APP) [
47]. Initial cleavage of APP by β-secretase generates a membrane-bound C-terminal fragment of APP (β-CTF) [
14,
66] which is sequentially processed by γ-secretase resulting in the formation of longer species, like Aβ40 and Aβ42/43, in addition to several other C-terminally truncated Aβ variants [
39,
56]. A full set of Aβ isoforms with both N- and C-terminally truncated Aβ species can be detected in the human brain and cerebrospinal fluid (CSF) [
28,
41,
42,
45,
62]. N-terminally truncated Aβ isoforms show an increased aggregation propensity and have previously been associated with the deposition of dense-core amyloid plaques in the brain [
18,
35,
40,
46,
49,
51]. In contrast, C-terminally truncated shorter Aβ isoforms show a higher solubility in vitro and are more commonly detected in the cerebral vasculature [
11,
18,
28,
43]. One of the C-terminally truncated isoforms, Aβ34, can be generated as an intermediate of multiple degradation pathways including Aβ degradation by the matrix metalloproteases (MMP), MMP2 and MMP9 [
25], as well as stepwise proteolytic cleavage of both Aβ40 and Aβ42 by γ-secretase [
39]. In addition to its role as the major β-secretase in APP processing, BACE1 (beta-site amyloid precursor protein cleaving enzyme 1) has been shown to cleave longer Aβ isoforms at position 34 and this pathway has been identified as the major source of Aβ34 [
19,
50]. This additional BACE1 mediated cleavage can only take place with Aβ peptides as substrates implying that BACE1 indeed acts as an Aβ40/42 degrading enzyme to generate Aβ34 [
19,
50]. In a recent study, we were able to show that Aβ34 levels are directly affected by over-expression or inhibition of BACE1 both in vitro and in vivo, and that cerebral BACE1 is the critical factor for Aβ34 generation [
30]. Supporting its proposed role as a biomarker for clearance in sporadic AD, CSF Aβ34 levels were found to be elevated in subjects with mild cognitive impairment (MCI) who later converted to AD dementia [
30]. Moreover, we also discovered a positive correlation between CSF Aβ34 levels and overall Aβ clearance rates in individuals with biomarker evidence of cerebral amyloid deposition [
30].
In the present study, we investigated the tissue distribution of the C-terminally truncated isoform Aβ34 in post-mortem brain samples from healthy elderly individuals and AD patients of various AD neuropathological stages according to Braak [
8,
9]. Aβ34 immunoreactivity was detected exclusively in the vasculature and was most prominent in brain capillaries of healthy elderly individuals in the early Braak stages. In AD patients and more advanced Braak stages, however, this capillary associated Aβ34 immunoreactivity was largely lost. Colocalization studies in postmortem brain tissue and isolated human microvessels revealed a unique association of Aβ34 with the platelet derived growth factor receptor beta (PDGFR-β) which serves as a marker for brain pericytes [
1,
31]. Further in vitro biochemical studies in cultured human primary pericytes confirmed that the C-terminally truncated isoform Aβ34 is released by these cells via BACE1 mediated enzymatic cleavage of Aβ40. Collectively, the results of our work suggest the existence of a novel pericyte mediated Aβ clearance pathway, which generates Aβ34 as a stable intermediate. Early impairment of this pathway may contribute to the overall reduced Aβ clearance associated with sporadic AD [
34].
Materials and methods
Post-mortem human samples
Paraffin-embedded hippocampus and middle-frontal cortex (gyrus frontalis medius) tissue blocks and frozen middle frontal cortex samples were provided by the Netherlands Brain Bank (NBB), the Netherlands Institute for Neuroscience, Amsterdam, the Netherlands. Post-mortem samples were collected from donors with a written informed consent for a brain autopsy and the use of the material for research purposes was obtained by the NBB. Subjects’ demographics, neuropathological and clinical evaluations are listed in Table
1. In brief, subjects are classified according to their Braak stages from stage I to VI [
8,
9]. Age, gender, apolipoprotein E (APOE) genotype, CERAD (Consortium to Establish a Registry for Alzheimer’s Disease) score for amyloid load and clinical diagnosis (non-demented control (NDCNTRL) or AD patient (AD)) are provided for each subject [
27,
36].
Table 1
Characteristics of post-mortem human brain samples
Brain Region: Hippocampus (paraffin-embedded) |
Braak Stage | Number of Subjects | Age at Deatha | Genderb | APOE4 Statusc | CERAD Scored | Diagnosise |
I | 6 | 79 ± 6.8 | M (1) F (5) | + (1), − (2) | O(2), A(2), B(2) | NDCNTRL |
II | 6 | 85.7 ± 1.8 | M (1) F (5) | + (2), − (4) | A(1), B(3), C(2) | NDCNTRL |
III | 6 | 82 ± 5.8 | M (5) F (1) | + (3), − (2) | A(1), B(2), C(3) | NDCNTRL |
IV | 7 | 92.3 ± 3.3 | M (1) F (6) | + (2), − (5) | C | NDCNTRL (2), AD(5) |
V | 4 | 82.3 ± 10.1 | M (0) F (4) | + (3), − (1) | C | AD |
VI | 6 | 86 ± 6.8 | M (0) F (6) | + (4), − (2) | C | AD |
Brain Region: Cortex (paraffin-embedded) |
Braak Stage | Number of Subjects | Age at Death | Gender | APOE4 Status | CERAD Score | Diagnosis |
I | 4 | 82.5 ± 9.3 | M (2) F (2) | + (0), − (4) | A(1), B(2), C(1) | NDCNTRL |
II | 3 | 85.3 ± 0.5 | M (1) F (2) | + (1), − (2) | A(1), B(3), C(2) | NDCNTRL |
III | 4 | 87 ± 3.7 | M (3) F (1) | + (2), − (2) | A(1), B(2), C(3) | NDCNTRL |
IV | 6 | 89.3 ± 5.4 | M (2) F (4) | + (3), − (3) | B(2), C(4) | NDCNTRL (2), AD (4) |
V | 3 | 84.3 ± 13.2 | M (0) F (3) | + (2), − (1) | C | AD |
VI | 5 | 89.2 ± 3.4 | M (0) F (5) | + (4), − (1) | C | AD |
Brain Region: Cortex (frozen) |
Braak Stage | Number of Subjects | Age at Death | Gender | APOE4 Status | CERAD Score | Diagnosis |
I – II | 4 | 88.7 ± 7.9 | M (0) F (4) | + (0), − (2) | O(1), A(1), B(1), C(1) | NDCNTRL |
III – IV | 5 | 91.4 ± 2.9 | M (1) F (4) | + (1), − (2) | O(1), B(2), C(2) | NDCNTRL (3), AD (2) |
V – VI | 5 | 75.2 ± 8.1 | M (2) F (3) | + (2), −(2) | B(1), C(4) | AD |
Immunofluorescence staining
Sections of paraffin-embedded hippocampus and cortex samples with 5 μm thickness were deparaffinized in xylene and then rehydrated by immersing the slides in order in 100% ethanol, 95% ethanol, 70% ethanol and water. Sections were pretreated by boiling in 0.1 M sodium citrate buffer for antigen retrieval as previously described [
20]. Only for amyloid plaque staining, sections were also incubated in 95% formic acid for 5 min prior to citrate buffer pre-treatment. Unspecific binding sites were blocked with 10% horse serum in phosphate-buffered saline (PBS) with 0.2% Triton X-100 (PBS-T). After blocking, sections were incubated overnight at 4 °C with primary antibodies diluted in 5% horse serum in PBS-T. All primary antibodies and dilutions used in immunohistochemistry are listed in Additional file
1. After washing with PBS, sections were incubated for 2 h at room temperature with secondary antibodies (1:200 dilution in 5% horse serum in PBS-T, donkey anti-mouse/rabbit/goat conjugated to Alexa488, Cy3, Alexa647 and streptavidin conjugated to Alexa488) purchased from Jackson Immunoresearch (Pennsylvania, USA). Following the secondary antibody step, sections were incubated for 10 min with 2 mg/ml DAPI (4′,6-diamidino-2-phenylindole) (Sigma Aldrich, Missouri, USA) in PBS for nucleus staining and for 5 min in 0.2% Sudan Black (Sigma Aldrich) in 70% ethanol to quench auto-fluorescence [
38]. Coverslips were mounted with water based mounting medium and slides were stored at 4 °C. Pictures were taken with Leica DM4000B microscope or Leica TCS SP8 confocal microscope. For peptide blocking/competition assay, 1 μg anti-Aβ34 and 1 μg anti-PDGFR-β antibodies were pre-incubated with 10 μg recombinant human Aβ34 peptide (Anaspec, California, USA) in 5% horse serum in PBS-T for 1 h. The immunohistochemistry protocol was then followed as described above.
Quantification of immunofluorescence staining
For each subject, 2 sections, which are at least 50 μm apart from each other, were used and 10 pictures with 20X magnification were randomly taken per section. For Aβ34 and PDGFR-β quantitative analyses, sections were also stained with anti-Collagen IV antibody to allow a quantification of the total number of vessels in each visual field. Vessels with Aβ34 or PDGFR-β immunoreactivity were counted manually and divided by the total number of vessels in the visual field assessed by Collagen IV immunostaining. In order to quantify only capillaries, vessels with a diameter > 10 μm were excluded. Mean of 20 visual fields were calculated for each subject and results were reported as percentage of Aβ34 or PDGFR-β positive vessels. For tau tangle and amyloid plaque quantifications, pictures were converted to 8-bit images and constant thresholds were applied to each picture. Intra-neuronal APP signal was also eliminated by using this threshold in amyloid plaque quantification. The area for each immunostained picture was calculated with ImageJ and results were reported as percentage of amyloid or tau positive area. Mean of 20 visual fields were calculated for each subject and reported in the figures.
Microvessel isolation and preparation of vessel enriched brain lysates
Previously published microvessel isolation protocols for mouse and rat brains were adapted for frozen human brains [
6,
10,
65]. Briefly, 1 g frozen human cortex was placed in 5 ml Hank’s balanced salt solution (HBSS) with protease and phosphatase inhibitors and homogenized with glass Teflon homogenizer (20 strokes at 400 rpm). Homogenates were centrifuged at 2000 g for 10 min at 4 °C. The pellet was resuspended in 17.5% Dextran (Sigma Aldrich) in PBS and centrifuged at 4400 g for 15 min at 4 °C. The pellet was again resuspended in 1 ml HBSS and passed through filters with 100 μm and 20 μm pore size (Pluriselect Life Science, Leipzig, Germany) in the respective order. The material retained on the 20 μm filter was collected by gentle washing with HBSS and centrifuged at 2000 g for 10 min at 4 °C. The resulting pellet (corresponding to microvessels) was resuspended in 4% paraformaldehyde (PFA) in PBS for fixation. Similar to the immunofluorescence protocol, microvessels were blocked with 10% horse serum in PBS-T and incubated 2 h with primary antibodies and 2 h with secondary antibodies in PCR tubes at room temperature. Microvessels were then placed on glass slides and coverslips were mounted with water-based mounting medium. Images were taken with Leica TCS SP8 confocal microscope. For vessel enriched brain lysates, material retained on both filters was collected and combined by washing with HBSS and centrifugation at 2000 g for 10 min at 4 °C. After centrifugation, the pellet was incubated in RIPA buffer at 4 °C with rotation for 1 h, sonicated for 5 min in ultrasonic bath and then centrifuged at 10000 g for 15 min at 4 °C. Supernatants were collected and stored at -80 °C.
Immunoassays
Custom-made 4-plex (Aβ34, Aβ38, Aβ40, and Aβ42) meso scale discovery (MSD) plates were purchased from Meso Scale Diagnostics (Maryland, USA). For 1-plex Aβ34 assay, high binding MSD plates were spot-coated with monoclonal anti-Aβ34 antibody and incubated overnight. The assay was performed according to manufacturer’s instructions as previously reported [
30]. Briefly, diluent 35 was added into each well and incubated for 1 h at room temperature for blocking. PBS with 0.2% Tween-20 was used for washing. 25 μl sample or standard and 25 μl sulfo-tag anti-amyloid beta (6E10 or 4G8) detection antibody were added. Plates were incubated at 4 °C overnight with rigorous shaking. On the following day, 2X read buffer was added after washing and plates were read with SECTOR Imager 600 plate reader (Meso Scale Diagnostics, USA) and analyzed with MSD Workbench software (Meso Scale Diagnostics, USA).
The human total PDGFR-β enzyme-linked immunosorbent assay (ELISA) and ELISA ancillary reagent kits were purchased from R&D Systems (Minnesota, USA). The assay was performed according manufacturer’s instructions. Briefly, plates were incubated with capture antibody diluted in PBS overnight at 4 °C. PBS with 0.2% Tween-20 was used for washing. Aliquots of 100 μl sample or standard were diluted in diluent 12 and incubated for 2 h at room temperature. After washing, biotinylated detection antibody, diluted in diluent 14, was added into wells and incubated for 2 h at room temperature. Following a wash step, streptavidin-horseradish peroxidase (HRP) substrate diluted in diluent 14 was added. Color reagents were added and incubated for 20 min; the reaction was stopped by adding stop solution (2 N sulfuric acid) and absorbance at 520 nm was measured with Tecan Infinite M Nano plate reader.
Cell culture
Human primary pericytes were purchased from ScienCell (California, USA). Cells were maintained in Pericyte Medium (ScienCell) with addition of 2% fetal bovine serum (FBS) (ScienCell), penicillin/streptomycin (ScienCell) and pericyte growth supplement (ScienCell) in poly-L-lysine (ScienceCell or SigmaAldrich) coated T75 flasks. Cells from each passage were frozen in 10% dimethyl sulfoxide (DMSO) in FBS and stored at -150 °C. Cells from passage 3 to 6 were used in experiments.
For Aβ uptake assays, human recombinant Aβ40 (rPeptide, Watkinsville, USA) and mouse Aβ40 synthetic peptides (Anaspec, California, USA) were prepared according to published protocols [
52]. Briefly, lyophilized peptides were dissolved in hexafluoro-2-propanol (HFIP) on ice and aliquoted into 1.5 ml Eppendorf tubes. Aliquots were lyophilized and stored at -80 °C. Before the experiment, lyophilized peptide was dissolved in DMSO and concentration adjusted with pericyte medium. Pericytes were cultured in either 60-mm dishes or 6-well plates and treated with different concentrations of Aβ40 for varying incubation times. At the end of the experiment cell lysates were prepared in RIPA buffer, cell media were collected and stored for further analysis. In order to block Aβ uptake, cells were treated with 10 μM endocytosis blocker IPA3 (1,1′-Dithiodi-2-naphthtol) (Abcam, Cambridge, United Kingdom) or 500 nM low density lipoprotein receptor-related protein 1 (LRP1) blocker RAP (receptor associated protein) (R&D Systems) for 2 h prior to Aβ40 (2 μM) addition. Solutions of 10 μM IPA3 and 500 nM RAP were used since these concentrations were reported to reduce Aβ40 uptake without affecting cell viability [
61]. Cell lysates were prepared with RIPA buffer, cell media were collected and stored for further analysis. In order to study Aβ uptake, pericytes were seeded on coverslips in a 24-well plate and incubated with Hilyte Fluor 488-labeled human Aβ40 (Anaspec) after RAP or IPA3 treatment. After fixation and DAPI staining, cells were imaged with Leica TCS SP8 confocal microscope. For 96-well assay, pericytes were plated in 96-wells and incubated with Hilyte Fluor 488-labeled human Aβ40 after RAP or IPA3 treatment. After washing, extracellular Aβ was quenched with 0.2% trypan blue fluorescence and measured at 485 nm excitation and 535 nm emission wavelength for detection of Hilyte Fluor-488 with Tecan Spark plate reader. Subsequently, cells were incubated with DAPI and fluorescence was measured with 360 nm excitation and 485 nm emission wavelength to normalize Hilyte Fluor 488 fluorescence for cell count in each well. For BACE1 inhibition assays, pericytes were treated with varying concentrations (10 μM, 5 μM, 2 μM, 1 μM) of BACE1 Inhibitor IV (Sigma Aldrich) for 12 h prior to Aβ40 (2 μM) addition. As in previous experiments, lysates were prepared with RIPA buffer, cell media were collected and stored for further analysis.
Immunocytochemistry
Coverslips were coated with poly-L-lysine and placed in 24-well plate. Cells were seeded in 24-well plates, incubated overnight and then fixed with 4% PFA in PBS. Unspecific binding was blocked with 10% horse serum in PBS. Cells were incubated with primary antibody for 2 h, followed by secondary antibody for 90 min at room temperature (Additional file
1). For nucleus staining, cells were incubated with 2 mg/ml DAPI solution in distilled water for 10 min. Then, coverslips were transferred on slides with water-based mounting medium. Cells were imaged with Leica DM4000B microscope or Leica TCS SP8 confocal microscope.
Western blot and immunoprecipitation
Pericyte lysates were mixed with 2X Novex SDS sample buffer (Thermo Fisher Scientific, Massachusetts, USA) containing 10% β-mercaptoethanol 1:1, incubated at 95 °C for 5 min and loaded on Novex 10–20% Tricine gel (Thermo Fisher Scientific). Electrophoresis was performed at 100 V for 2 h in Tricine running buffer. After transfer, nitrocellulose membranes were blocked with 5% milk blocking solution and incubated with primary antibodies at 4 °C overnight. On the next day, membranes were incubated with HRP-coupled secondary antibodies (Jackson Immunoresearch) diluted in 5% milk blocking solution for 2 h at room temperature and developed with SuperSignal West Femto Maximum Sensitivity Substrate (ThermoFisher). Images were taken with ImageQuant LAS 400 (GE Healthcare, Illinois, USA). BACE1 immunoprecipitation was performed using magnetic separation with Dynabead M-280 Sheep Anti-Rabbit IgG (Invitrogen) according to the manufacturer’s protocol. Briefly, cell lysates were incubated with rabbit anti-BACE1 antibody (ab2077, abcam) overnight at 4 °C with gentle rotation. After addition of magnetic beads, samples were incubated for 1 h with gentle rotation at room temperature. After washing with PBS, beads were magnetically separated and incubated in 2X Novex SDS sample buffer containing 10% β-mercaptoethanol at 95 °C for 5 min. Beads were magnetically separated and proteins loaded on Novex 10–20% Tricine gel. Electrophoresis and Western blotting were performed as described above.
Statistical analysis
Graphpad Prism 8 software was used for all statistical analyses. Assumption of normal (Gaussian) distribution was tested with Shapiro-Wilk normality test (α = 0.05). If the assumption of normality was true, the significance of the differences between two groups was tested with a two-tailed unpaired Student’s t test. The significance of the differences between more than two groups was tested with 1-way analysis of variance (ANOVA) followed by Tukey’s multiple comparison test. Correlations were calculated with Pearson’s correlation (α = 0.05, confidence interval 95%, two tailed). In case of a deviation from normal distribution, the significance of the differences between two groups was tested with a two-tailed unpaired Mann-Whitney test. Correlations were calculated with Spearman’s correlation (α = 0.05, confidence interval 95%, two tailed). Correlation coefficients (r) and p-values were reported in the figures.
Discussion
The C-terminally truncated Aβ isoform, Aβ34, has been identified as an important intermediate product of enzymatic Aβ degradation clearance. Although, in principle, Aβ34 can be generated via various enzymatic pathways, BACE1 was recently shown to be a major enzyme involved in Aβ34 generation in vivo [
30]. In our recently published study, we showed elevated CSF Aβ34 levels in early clinical stages of AD and a correlation of CSF Aβ34 levels with Aβ clearance rates in subjects with evidence of cerebral amyloid deposition [
30]. In the present study, we analyzed the distribution of Aβ34 in post-mortem human brain samples to further investigate its role in AD pathology and neurodegenerative disease progression. Our comprehensive histological analysis revealed a unique association of Aβ34 immunoreactivity with pericytes in brain capillaries and a loss of this capillary associated Aβ34 immunoreactivity with advancing disease stages in AD patients. Interestingly, we observed a peak in capillary Aβ34 immunoreactivity in the earlier Braak stages (stage 2–3) which then dramatically decreased between Braak stages 3 and 4. The correlation of Braak stages with cognitive scores used in clinical practice had been investigated by several previous studies suggesting that individuals with MCI were most likely distributed between Braak stages 2 and 4 [
3,
21,
63]. Based on our current data, however, a definitive conclusion whether the observed increase in capillary Aβ34 immunoreactivity in the earlier Braak stages indeed represents a histopathological correlate of elevated CSF Aβ34 levels detected in the recently published prodromal AD clinical sample [
30] cannot be drawn at this point. On the one hand, we were not able to assess CSF Aβ34 levels in our post-mortem cohort. On the other hand, the post-mortem neuropathological sample was clinically divided only into two diagnostic categories, i.e. AD dementia and non-demented controls, the latter including both healthy elderly individuals and non-demented controls meeting the clinical criteria for a mild cognitive impairment. Another limitation of our study sample was the lack of younger individuals and age-matched controls of even earlier Braak stages (Braak stage 0) mainly due to the low prevalence and unavailability of this control group [
37]. Analysis of such additional samples may be particularly helpful in clarifying the question whether pericyte mediated Aβ degradation occurs already at a very early stage in the absence of overt AD pathology, and whether it is upregulated in an age-dependent manner.
In our study, we provided substantial histological and biochemical evidence for an association of Aβ34 with brain pericytes. In vitro mechanistic assessments moreover revealed that pericytes can generate Aβ34 by uptake and degradation of extracellular Aβ40. Nevertheless, the results of these experiments do not exclude the possibility that Aβ34 is generated elsewhere, e.g. in neurons, and ultimately only taken up by pericytes for further degradation. As a major Aβ source and the primary cell type expressing BACE1 in the brain [
5], neurons highly likely represent an alternative source of Aβ34 generation in the central nervous system. This assumption is also supported by a previous study reporting intraneuronal Aβ34 immunoreactivity [
12]. The detection of capillary Aβ34 immunoreactivity, however, may in our opinion be better explained by an in situ production of Aβ34 in pericytes. Due to its short half-life and its susceptibility to further enzymatic degradation [
11,
30], it appears rather unlikely that neuronal Aβ34 would reach the microvasculature in sufficient amounts in order to be taken up by brain pericytes. We rather propose that the more stable and more abundant Aβ isoforms like Aβ40 and possibly also Aβ42 which were previously shown to be drained along vascular clearance routes [
59,
60] are taken up by pericytes and locally degraded to Aβ34 at the neurovascular unit. In line with this assumption, our mechanistic studies in human primary pericytes demonstrated that pericytes are fully capable of taking up and degrading Aβ40 to Aβ34. We did not specifically assess whether Aβ42 can also be degraded to Aβ34 by pericytes in vitro since the previously reported relative abundance of Aβ40 at the vasculature compared to Aβ42 [
26,
53,
59] and our correlative analyses and findings in the human microvessels would rather be in line with Aβ40 being the major Aβ34 precursor in vivo. However, earlier studies illustrated that Aβ42 is also a substrate of BACE1 [
19,
30] hence pericyte-associated Aβ34 could be generated via BACE1-mediated degradation of Aβ42. Both Aβ40 and Aβ42 fibrils were reported to induce pericyte death in vitro; nevertheless Aβ34 was shown not to aggregate even in high concentrations [
11] and did not show any deleterious effects in our pericyte culture experiments (data not shown). Previously, Caillava et al. even suggested a protective role for this isoform since Aβ34 treatment lowered the apoptotic cell death in HEK cells in vitro [
12]. Considering susceptibility of this isoform to further enzymatic degradation by metalloproteases [
12,
30], we suggest that BACE-1 mediated degradation of Aβ40 or Aβ42 would enhance both enzymatic and vascular clearance by generating more accessible, benign and likely protective against apoptosis, isoform Aβ34.
The increase in Aβ34 immunoreactivity which we observed in our post-mortem sample may reflect a compensatory upregulation of vascular Aβ degradation in the early (preclinical or prodromal) stages of AD. As mentioned above, such an upregulation of Aβ34-related degradation clearance could provide an explanation for the observed increase in CSF Aβ34 levels in patients with prodromal AD [
30]. As an alternative, elevated CSF Aβ34 levels might also be explained by an unspecific release of Aβ34 by degenerating cells, which in our case would be pericytes. We could not fully address with our cross-sectional correlative postmortem study design, whether the loss of Aβ34 immunoreactivity is caused by the physical loss of pericytes during disease progression. The decrease in Aβ34+ vessels appeared to be more pronounced than the decrease in the percentage of PDGFR-β + capillaries with advancing disease stages, suggesting in addition to the loss of pericytes, a functional impairment might have contributed to the observed findings. Nevertheless, the robust positive correlation between Aβ34 and PDGFR-β immunoreactivity in the brain and levels in our microvessel preparations support that at least partially an early loss of pericytes could be an important factor. In that regard, Aβ34 could also represent a sensitive candidate marker reflecting impairments in cerebrovascular pericyte function.
In the present study, we occasionally observed Aβ34 immunoreactivity in association with CAA in arterioles and large arterial vessels. However, only a limited number of individuals in our sample set showed CAA pathology, and capillary CAA (capCAA) was completely absent (Additional file
4). Wisniewski et al. previously reported presence of Aβ fibrils in cerebral vessels associated with pericytes [
64]; due to a lack of capillary CAA cases we were unfortunately not able to assess the association between Aβ34 and aggregated Aβ forms in capillaries in our study. In case of arterial CAA, we speculate that Aβ34 immunoreactivity in these arteries might be the result of binding of this isoform to deposited vascular amyloid in arterial walls, e.g. upon failed vascular drainage. Alternatively, Aβ34 could also be locally generated in CAA-laden vessels, e.g. through the proteolytic cleavage of longer Aβ isoforms by BACE1 or MMPs [
19,
25]. In support of this hypothesis, elevated levels of vascular BACE1 were previously described in CAA-laden arteries [
7,
13,
16]. Further studies focusing on CAA and vascular amyloidosis in samples including capCAA will be required to obtain a better understanding of the origin of Aβ34 in CAA-affected arteries.
Pericytes have vital roles in vascular development, formation and maintenance of the blood brain barrier (BBB), blood flow regulation as well as in modulating immune reactions in the brain [
1,
2,
4,
15,
29]. In addition, it has recently been shown that pericytes can take up extracellular Aβ42 via a LRP1/ APOE dependent mechanism in vitro [
32]. Pericyte deficiency was shown to worsen AD pathology in transgenic animals with enhanced Aβ levels in the brain; conversely, pericyte implantation improved clearance of Aβ and reduced Aβ deposition [
44,
55]. Given their location in the intramural periarterial drainage pathways and their ability to take up and degrade extracellular Aβ, pericytes have been implicated to have a role in Aβ clearance [
17,
24]. We have shown, for the first time, an association of these cells with an important intermediate product of Aβ clearance. Our results are thus in line with the emerging role of pericytes in enzymatic Aβ degradation in human brain [
32]. A failure of the pericyte-mediated Aβ degradation pathway could result in a progressive loss of vascular Aβ clearance, enhanced Aβ40 accumulation in the vasculature and consecutive impact on parenchymal Aβ clearance and deposition. The strong inverse association between Aβ34 immunoreactivity and tau pathology which we detected in our post-mortem sample suggests that this clearance pathway may have tremendous impact on neurodegenerative processes associated with AD.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.