Tau pathology-dependent remodelling of cerebral arteries precedes Alzheimer’s disease-related microvascular cerebral amyloid angiopathy
- First Online:
- Cite this article as:
- Merlini, M., Wanner, D. & Nitsch, R.M. Acta Neuropathol (2016) 131: 737. doi:10.1007/s00401-016-1560-2
- 1.5k Downloads
Alzheimer’s disease (AD) is characterised by pathologic cerebrovascular remodelling. Whether this occurs already before disease onset, as may be indicated by early Braak tau-related cerebral hypoperfusion and blood–brain barrier (BBB) impairment found in previous studies, remains unknown. Therefore, we systematically quantified Braak tau stage- and cerebral amyloid angiopathy (CAA)-dependent alterations in the alpha-smooth muscle actin (α-SMA), collagen, and elastin content of leptomeningeal arterioles, small arteries, and medium-sized arteries surrounding the gyrus frontalis medialis (GFM) and hippocampus (HIPP), including the sulci, of 17 clinically and pathologically diagnosed AD subjects (Braak stage IV–VI) and 28 non-demented control subjects (Braak stage I–IV). GFM and HIPP paraffin sections were stained for general collagen and elastin with the Verhoeff–van Gieson stain; α-SMA and CAA/amyloid β (Aβ) were detected using immunohistochemistry. Significant arterial elastin degradation was observed from Braak stage III onward and correlated with Braak tau pathology (ρ = 0.909, 95 % CI 0.370 to 0.990, p < 0.05). This was accompanied by an increase in neutrophil elastase expression by α-SMA-positive cells in the vessel wall. Small and medium-sized arteries exhibited significant CAA-independent α-SMA loss starting between Braak stage I and II–III, along with accumulation of phosphorylated paired helical filament (PHF) tau in the perivascular space of intraparenchymal vessels. α-SMA remained at the decreased level throughout the later Braak stages. In contrast, arterioles exhibited significant α-SMA loss only at Braak stage V and VI/in AD subjects, which was CAA-dependent/correlated with CAA burden (ρ = −0.422, 95 % CI −0.557 to −0.265, p < 0.0001). Collagen content was only significantly changed in small arteries. Our data indicate that vessel wall remodelling of leptomeningeal arteries is an early-onset, Braak tau pathology-dependent process unrelated to CAA and AD, which potentially may contribute to downstream CAA-dependent microvascular pathology in AD.
KeywordsHyperphosphorylated tau Collagen Internal elastic lamina Neutrophil elastase Vascular smooth muscle Cerebrovascular pathology
Pathologic cerebrovascular alterations are prominent in Alzheimer’s disease (AD) and include smooth muscle loss, vasculitis, cerebral amyloid angiopathy (CAA) with or without microhaemorrhaging, blood–brain barrier (BBB) leakage, altered vessel wall collagen content, and alterations affecting neurovascular coupling [16, 17, 25, 39, 40]. They contribute to the genesis of the neuron-hostile environment characteristic in AD. The question remains whether these vascular alterations are merely AD-related or have other, earlier origins. For example, vessel wall pathology is observed in normal ageing [25, 26] and may also play a role in primary age-related tauopathy (PART) [13, 61]. In the latter, hyperphosphorylation of tau protein causing neurofibrillary tangle (NFT) pathology occurs in the absence of amyloid β (Aβ) accumulation . NFTs are also present at the earliest neurodegenerative stages that may progress to incipient AD or fully developed AD with or without concomitant Aβ pathology , which is interpreted by some that PART could be a mere pre-AD stage rather than a separate neurodegenerative disorder . Clear is, though, that Braak tau pathology—whether as PART, related to AD, or underlying both—is accompanied by decreased cerebral blood flow [3, 9, 41]. Further, decreased cerebral blood flow [22, 32, 36, 50] along with increased vascular resistance and impaired cerebral oxygen uptake [23, 36] and overall arterial stiffness  is observed in subjects with mild cognitive impairment, suggesting the existence of early vessel wall remodelling because changes in the vessel wall anatomy can affect blood flow. Arterial wall remodelling is of special interest herein since an important aspect of especially large and medium-sized muscular arteries, besides functioning as a conduit, is to cushion the cardiac pulsatile blood flow waves to such extent that the pulsations experienced by the relatively fragile arteriolar and capillary wall are largely proportional to the wall’s distension capacities. This prevents microvascular wall damage . Furthermore, the cerebral arterial wall is instrumental in adequate functioning of perivascular drainage, one of the main cerebral clearance routes [60, 62]. Collagen, elastin, and vascular smooth muscle form the arterial wall triad that contributes to the regulation of these blood flow dynamics, with arterial elastin playing an important role in arterial wall recoil, affecting shape during and after vascular smooth muscle-mediated vessel constriction and dilatation.
Ageing negatively affects arterial elastin, thereby altering perfusion patterns [19, 64]. To our knowledge, it is not known whether Braak tau and/or AD pathology affects elastin in cerebral arteries, similar to the finding that perivascular tau accumulation as observed in sporadic AD subjects and in a mouse model of tauopathy can instigate pathologic vessel wall changes [5, 48, 57]. Furthermore, although previous studies have shown CAA-dependent alpha-smooth muscle actin (α-SMA) and collagen loss in cerebral arterioles of AD subjects [54, 55], quantitative analyses of such vessel wall remodelling in cerebral arteries according to Braak tau stage (called “Braak stage” henceforth) in parallel to analysis of the arterioles in the same subjects have, to our knowledge, not been extensively performed. Therefore, we systematically quantified Braak stage- and CAA-dependent and -independent changes in collagen, elastin, and α-SMA and luminal diameters of the leptomeningeal arterioles, small arteries, and medium-sized arteries surrounding the gyrus frontalis medialis (GFM) and hippocampus (HIPP), including the sulci, of non-demented control (NDCTRL) and AD subjects. Concomitantly, we investigated the potential role of neutrophil elastase in affecting elastin integrity, and perivascular tau accumulation in affecting the α-SMA content of the arterial wall at the different Braak stages. The three lead questions of this study are whether (1) remodelling of the vessel wall occurs already at early Braak stages, (2) the type of remodelling is distinct between the arterioles and arteries, and (3) the type of remodelling is Braak stage and/or CAA dependent.
Materials and methods
Subjects and brain material
Demographics, neuropathological diagnoses, and clinical history of the subjects analysed
Age at death (mean ± SD, years)
Cause of deathe
82.1 ± 9.6
Angina pectoris, COPD, CVA, DM2, hypertension
Acute myocardial infarction, cachexia, dehydration, pneumonia, sepsis
81.4 ± 5.7
Cardiac decompensation, COPD, DM2, emphysema, hypertension
Cachexia, cardiac arrest, CVA (medium-sized), dehydration
84.0 ± 7.5
Angina pectoris, atrial fibrillation, cardiac decompensation, hypertension
Cardiac arrest, CVA, sudden death
91.2 ± 4.4
Cardiac decompensation, DM2, hypertension, pneumonia
Cachexia, cardiac arrest, dehydration, uncontrolled anti-coagulation therapy
89.0 ± 3.1
Atrial fibrillation, cardiac decompensation, hypercholesterolaemia, hypertension
Cardiac arrest, dehydration, pneumonia
89.2 ± 3.4
Angina pectoris, CVA, hypertension, myocardial infarction, vascular dementia (1 subject)
Acute myocardial infarction, cachexia, dehydration, pneumonia
All reagents listed were purchased from Sigma-Aldrich (Basel, Switzerland) unless otherwise specified. The HIPP and GFM tissue blocks obtained from the NBB were cut into 5-µm-thick sections and were mounted on SuperFrost® Plus microscope slides (VWR, Dietikon, Switzerland). Detailed histological staining procedures can be found in the Supplementary Materials and Methods. Briefly, following standard deparaffinisation and rehydration steps, sections for immunohistochemical staining were treated with antigen retrieval buffer followed by co-incubation with primary goat anti-α-SMA antibody (ab21027; Abcam, Cambridge, UK) and primary mouse anti-Aβ (6E10, purified; Lucerna-Chem, Luzern, Switzerland) for 1 h at room temperature (RT). Subsequently, the sections were co-incubated with donkey anti-mouse-Alexa488 and donkey anti-goat-Cy3 antibody (both from Jackson ImmunoResearch, Suffolk, UK). HIPP and GFM sections adjacent to the ones immunostained for α-SMA and Aβ were stained for collagen and elastin using the Verhoeff–van Gieson (VVG) stain. Another series of adjacent HIPP and GFM sections was stained for neutrophil elastase (primary rabbit anti-neutrophil elastase antibody, ab21595; Abcam, Cambridge, UK) using the VectaStain® Elite staining kit (ReactoLab, Servion, Switzerland) in combination with the Vector® SG substrate per the manufacturer’s instructions. To compare collagen stained by the Van Gieson stain with collagen stained with a collagen IV-specific antibody, a subset of sections was incubated with a mouse monoclonal antibody against human collagen IV (M 0785; DAKO, Gløstrup, Denmark) followed by staining with a donkey anti-mouse-Cy5 antibody (Jackson ImmunoResearch, Suffolk, UK). Accumulation of phosphorylated paired helical filament tau (PHF-tau) in the perivascular space of intraparenchymal vessels was detected with an antibody against phosphorylated tau (mouse anti-phospho-PHF-tau (AT8), MN1020; ThermoFisher Scientific, Reinach, Switzerland) followed by staining with a donkey anti-mouse-Cy3 antibody (Jackson ImmunoResearch, Suffolk, UK).
Image analysis: measurement of luminal diameter and Aβ, collagen, α-SMA, and neutrophil elastase vessel wall ratios
A detailed description of the analysis method can be found in the Supplementary Materials and Methods and in Supplementary Fig. 1. The green fluorescent (Aβ) and red fluorescent (α-SMA) image channels were merged in ImageJ (National Institutes of Health, Bethesda, ML, USA). Stained cells and/or cellular material, e.g. red blood cells and collagen deposits, in the lumen of both the merged fluorescent and bright field VVG- and neutrophil elastase-stained vessels were removed using ImageJ to avoid false-positive measurement hits in the subsequent quantification steps. The vessel wall fractions of Aβ (CAA), collagen, α-SMA, and neutrophil elastase and the luminal diameters were quantified using an in-house established semi-automatic quantification script written in MATLAB™ (The MathWorks, Inc., Natick, MA, USA).
Image analysis: elastin degradation
The semi-automatic MATLAB™ quantification method mentioned above could not consistently detect local, small breakages in the arterial elastin layer (Fig. 1b, mild elastin degradation: arrowhead and inset) that were part of the elastin degradation process. Therefore, the degree of overall arterial elastin degradation was assessed in a semi-quantitative, blinded manner (see Supplementary Materials and Methods). Since no significant differences in the elastin fraction were found between CAA-affected (CAA+) and non-CAA-affected (CAA−) arteries—only in rare cases of arteries with CAA fractions >1.0 and in which CAA was present in the media (Fig. 1c: c2), which were excluded from the analysis—and between vessels in the HIPP and the GFM, the number of all small arteries and all medium-sized arteries was pooled per Braak stage.
Alpha-smooth muscle actin is decreased in leptomeningeal arterioles and arteries of clinically and pathologically defined Alzheimer’s disease subjects
Previous studies have reported that cerebral microvessels of subjects with clinically and pathologically defined AD exhibited vascular smooth muscle loss and collagen increase [54, 55]. To confirm this finding in our study population and to understand whether leptomeningeal arterioles, small arteries, and medium-sized arteries in the same subjects might show different pathology, we analysed the α-SMA and collagen fraction in the vessel wall of pooled (i.e. leptomeningeal CAA+ and CAA−) arterioles, small arteries, and medium-sized arteries. We then grouped them according to the brain region they surrounded (HIPP or GFM) and the clinical, antemortem cognitive status of the subjects (AD or NDCTRL) (Fig. 2). All three vessel types surrounding the HIPP showed a significant decrease in the α-SMA fraction in the AD compared to NDCTRL subjects (~30–50 % decrease, p < 0.05) (Fig. 2a, left). The α-SMA fraction of the small and of the medium-sized arteries surrounding the GFM of AD subjects was significantly lower than that of their NDCTRL counterparts; the α-SMA fraction of arterioles surrounding the GFM showed a trend of being lower in the AD subjects than in the NDCTRL subjects (Fig. 2a, right). The collagen fractions were not significantly different between the AD and NDCTRL subjects in any of the vessel types and brain regions (Fig. 2b), although the collagen fraction of all three vessel types surrounding the GFM of both the AD and NDCTRL subjects tended to be lower than those surrounding the HIPP. Thus, overall, leptomeningeal arterioles, small arteries, and medium-sized arteries exhibited significant α-SMA loss in AD subjects whereas their vessel wall collagen content was similar between the AD and NDCTRL subjects.
The van Gieson component of the VVG stain detects collagen (bright-red stain) but lacks specificity for the different collagen types, e.g. collagen IV present in the basal lamina. Nevertheless, comparison of VVG-stained brain sections with adjacent brain sections stained with a mouse monoclonal antibody against human collagen IV showed similarity in the collagen patterns/distributions within the vessel wall between the two staining methods (Supplementary Fig. 2). The relatively high age of our study population may explain the increased collagen IV content in the basal lamina/thickening of the basal lamina compared to that in young subjects . All VVG-stained vessels were carefully isolated from the acquired images and any non-vessel-related collagen was removed from the images to avoid quantification of collagen present in the leptomeninges and vessel lumen. That the van Gieson component of the VVG stain could also detect vessel wall collagen present in the medial layer can be appreciated from Supplementary Fig. 1, “Collagen fraction”.
Analysis of CAA dependence shows CAA-dependent alpha-smooth muscle actin loss in leptomeningeal arterioles, but not in leptomeningeal arteries
To understand whether the decrease in the α-SMA fraction of the arterioles, small arteries, and medium-sized arteries in AD subjects (Fig. 2a) was related to CAA pathology and if the apparent lack of changes in the collagen fraction was due to the pooling of CAA+ and CAA− vessels (dilution effect), the three vessel types were analysed according to CAA status (CAA+ or CAA−) and antemortem cognitive status of the subjects (AD or NDCTRL) (Fig. 2c). None of the NDCTRL leptomeningeal vessels showed CAA. The α-SMA fraction of both AD CAA+ and CAA− arterioles was significantly lower than that of NDCTRL arterioles (AD CAA+ arterioles: ~70 % decrease, p < 0.001; AD CAA− arterioles: ~40 % decrease, p < 0.01) (Fig. 2c, left). Moreover, the α-SMA fraction of AD CAA+ arterioles was ~50 % lower than that of the AD CAA− arterioles, indicating that CAA exacerbated arteriolar α-SMA loss. Similar to the findings of Perry et al. , the α-SMA fraction of AD CAA− and CAA+ small and medium-sized arteries was significantly and equally decreased compared to that of NDCTRL small arteries and medium-sized arteries (~35–50 % decrease, p < 0.01; Fig. 2c, left), suggestive of a CAA-independent decrease in the arterial α-SMA fraction in the AD subjects. None of the vessels showed significant changes in their collagen fraction (Fig. 2c, right).
Analysis of Braak stage dependence shows that leptomeningeal arteries, but not leptomeningeal arterioles, exhibit early and Braak stage-dependent alpha-smooth muscle actin loss
The collagen fraction of small arteries of Braak stage II–III and VI subjects was significantly decreased compared to that of their Braak stage I counterparts (Braak stage II–III: ~21 % decrease, p < 0.05; Braak stage VI: ~43 % decrease, p < 0.01) (Fig. 4b); Braak stage and the observed decreased collagen fraction of the small arteries were, however, only weakly correlated (ρs = −0.256, 95 % CI −0.390 to −0.112, p < 0.001). Although not significant, medium-sized arteries tended to show an increase in the collagen fraction between Braak stage I and II–III and I and VI (~25 % increase, p = 0.07 and p = 0.09) (Fig. 4b). Similar to the lack of Braak stage-dependent arteriolar α-SMA loss, the arterioles in these CAA-negative samples did not show a significant change or trend in the collagen fraction with increasing Braak stage (Fig. 4b), further confirming the Braak tau-independent origins of the microvascular pathology found in this and a previous study .
Braak tau pathology does not significantly affect the wall-to-lumen ratio and the luminal diameter of leptomeningeal vessels
As described by others, CAA-affected vessels show a decrease in their wall-to-lumen ratio [29, 54]. Contrary to this, herein, we found that Braak tau pathology did not significantly alter the wall-to-lumen ratio of the arterioles, small arteries, and medium-sized arteries at all Braak stages analysed (Supplementary Fig. 3a). The average wall-to-lumen ratio at all Braak stages of the arterioles, small arteries, and medium-sized arteries was 0.37, 0.29, and 0.25, respectively, in line with the reduction in this ratio with increasing vessel size . The relatively high values found for all three vessel types may be expected for vessels from aged subjects [2, 56]. The lack of a change in the arteriolar wall-to-lumen ratio with increasing Braak stage agrees with previous findings  and with our other results of Braak stage-independent changes in the arterioles described herein. We ascribe the unchanged wall-to-lumen ratio of the small and medium-sized arteries to the fact that we measured the overall, total thickness of the vessel wall (intima to outer layer of the adventitia) to correct for the arterial Braak stage-dependent α-SMA loss described above, which would otherwise skew these results. Thus, arterial α-SMA loss did not affect the overall, total vessel wall thickness of the arteries.
Analysis of the luminal diameter per Braak stage, however, revealed that the percentage of arterioles within the smallest luminal diameter range measured (50–75 µm) was increased at Braak stage V and VI (~20–25 % increase) although this was not significant (p > 0.05) (Supplementary Fig. 3b). A less pronounced shift toward decreased luminal diameters was observed for small arteries (Braak stage IV and higher) and medium-sized arteries (Braak stage II/III and higher) (Supplementary Fig. 3b). Given the relatively high α-SMA but low collagen content of specifically arterioles—i.e. low arteriolar wall stiffness—compared to arteries, arteriolar loss of α-SMA-positive cells/vascular smooth muscle cells as we found at Braak stage V and VI (Fig. 4) may explain the more profound decrease in arteriolar luminal diameter compared to that of the stiffer small and medium-sized arteries. These reductions in arteriolar luminal diameter may play a role in cerebral hypoperfusion at progressive Braak tau stages as found in previous studies [3, 9, 41]. On the other hand, the lack of a profound effect of Braak stage on luminal diameter of the leptomeningeal arteries, which are upstream of the arterioles, suggests that cerebral hypoperfusion—at least at early Braak tau stages—may be more related to altered arterial wall compliance causing alterations in perivascular drainage and blood flow regulation rather than to actual arterial luminal diameter changes.
Accumulation of phosphorylated paired helical filament tau in the perivascular space of intraparenchymal vessels is accompanied by intraparenchymal alpha-smooth muscle actin loss and increases with Braak stage
Given that accumulation of PHF-tau has been observed in the perivascular space of both CAA-affected and -unaffected intraparenchymal vessels in subjects with sporadic AD [48, 57], we hypothesised that the CAA-independent but Braak stage-dependent loss of α-SMA in the small and medium-sized leptomeningeal arteries (Figs. 3b, c, 4a) might have been related to the perivascular drainage of toxic PHF-tau species from intraparenchymal vessels to the leptomeningeal arteries. Staining of HIPP sections of all 45 subjects for PHF-tau revealed the presence of PHF-tau in the perivascular space of intraparenchymal vessels in especially the CA1 region bordering the subiculum  in subjects with Braak stage II pathology and higher (Fig. 4c, Supplementary Fig. 4). The number of subjects affected by PHF-tau increased with progression of Braak tau pathology (Fig. 4d); however, the number of vessels showing this phenomenon highly varied between the subjects (on average, 5–20 vessels out of 50 vessels analysed were affected in the subjects). Perivascular PHF-tau accumulation was observed around intraparenchymal veins, arterioles, and small arteries, and was accompanied by α-SMA loss (Fig. 4c, Supplementary Fig. 4). None of the leptomeningeal vessels analysed in this study showed such perivascular PHF-tau accumulation, indicating that potential PHF-tau-related α-SMA loss in the leptomeningeal arteries might only have been caused by perivascular drainage of soluble PHF-tau species.
Leptomeningeal arteries exhibit Braak stage-dependent elastin degradation starting at early Braak tau pathology
Correlation analysis between the number of small and medium-sized leptomeningeal arteries in the respective elastin degradation category and Braak stage
CI (95 %)
CI (95 %)
Braaka vs. “0”
−0.965 to 0.246
−0.941 to 0.476
Braaka vs. “1”
−0.995 to −0.638
−0.535 to 0.931
Braaka vs. “2”
0.370 to 0.990
−0.247 to 0.965
Braaka vs. “3”
−0.056 to 0.976
−0.825 to 0.797
We postulate that the significant collagen loss with progressing Braak stage pathology specifically in the small arteries causes higher vessel wall stress and thus more pressure-induced elastin loss in these vessels than in the medium-sized arteries. This may explain the difference in elastin degradation severity between these otherwise similar artery types.
Severe leptomeningeal arterial elastin degradation tends to be accompanied by increased leptomeningeal arterial smooth muscle neutrophil elastase expression
The enzymatic degradation of elastin referred to above is prominent in chronic obstructive pulmonary disease (COPD), where neutrophil-derived neutrophil elastase compromises the vessel wall integrity of pulmonary arteries resulting in emphysema . Inhibition of neutrophil elastase alleviates elastin degradation and emphysema in COPD [1, 31], as well as vasogenic edema and neuronal loss in acute cerebral ischemia and traumatic brain injury [28, 53]. Here, severe elastin degradation in small and medium-sized arteries (Fig. 5b) tended to be accompanied by increased vessel wall neutrophil elastase fractions (Fig. 5b, c). Confocal microscopy revealed that the neutrophil elastase observed within the vessel wall was expressed by α-SMA-positive cells/smooth muscle cells (Fig. 5d, arrows in right panel) and most likely not by neutrophils (neutrophil elastase is also a specific neutrophil marker) as these were only observed attached to the luminal side of the intimal layer (Fig. 5d, asterisk and arrowheads) and between the leptomeningeal layer and the adventitia (Fig. 5d, number sign). However, extravasation of neutrophils into the vessel wall could not be ruled out, as some neutrophils attached to the luminal side of the intimal layer exhibited a flattened morphology suggestive of diapedesis (Fig. 5d, arrowheads). Thus, these results suggest a role for increased and/or aberrant neutrophil elastase expression by smooth muscle cells in severe elastin degradation in leptomeningeal arteries.
Herein, we found that leptomeningeal arteries show CAA-independent pathologic vessel wall remodelling that is related to early Braak tau pathology, while leptomeningeal arterioles only show significant vessel wall remodelling at late/AD Braak stages and which is CAA dependent. This suggests that early, pre-AD cerebral artery pathology unrelated to Aβ burden may contribute to the impairment of downstream arterioles in AD. Nonetheless, the following two major limitations of our study need to be considered: the relatively small population size and the exclusion of intraparenchymal vessels in the majority of our analyses. (1) Although our population size and Braak stage range were similar to those of previous studies [17, 54, 55], the majority of subjects had vascular-related morbidities, including chronic hypertension, diabetes type 2, and cardiac impairment. These factors were not biased toward a specific Braak stage (Table 1) but all of them could impact vascular health. (2) Our goal was to determine the effects of Braak tau pathology and CAA on the cerebrovasculature. Therefore, to avoid measuring these along with the (combined) effects of parenchymal Aβ plaques, neurofibrillary tangles, glial cells, and the accompanying pathologic changes in the parenchymal brain milieu on vessel wall remodelling, we had analysed only leptomeningeal vessels. Consequently, potential Braak stage- and CAA-dependent pathology of intraparenchymal vessels, except for perivascular PHF-tau accumulation, was not determined.
Our findings of Braak stage-dependent arterial elastin degradation offer insight into an understudied pathology of the cerebral arteries. Elastin loss is known to contribute to vessel tortuosity and, thereby, aberrant perfusion [34, 47]. Thus, the arterial elastin degradation we have observed may play a role in previously described impaired perivascular drainage [25, 26, 27, 59, 62] and cerebral hypoperfusion that worsens with advancing Braak stages [3, 9, 41]. Especially, reduced brain tissue oxygenation due to cerebral hypoperfusion becomes significant at Braak stage III–IV , similar to the moderate and severe arterial elastin degradation we found. Such arterial elastin degradation is, to our knowledge, not known to be present in relatively healthy aged individuals . Given that the majority of the Braak stage IV and all of the Braak stage V and VI subjects had clinically and pathologically defined AD and that moderate and severe elastin degradation were prominent at these advanced Braak stages, elastin degradation might be an important degenerative process contributing to vessel wall pathology affecting vascular tone in AD. The trend for an increased expression of neutrophil elastase by α-SMA-positive cells in the arteries with severe elastin degradation we found points to one potential mechanism underlying the observed arterial elastin degradation, which we are currently studying in detail. Other potential mechanisms are (1) the upregulated expression and/or activity of matrix metalloproteinases (MMPs), particularly MMP-2 and MMP-9, in the vessel wall as they are known to degrade arterial elastin, are linked to mild cognitive impairment and AD, and are activated by Aβ oligomers [10, 24, 35]; and (2) altered production of vascular smooth muscle-derived arterial elastin due to the change of a contractile to a proliferative vascular smooth muscle phenotype [30, 47]. In light of the latter, elastin is known to promote the physiological contractile, non-proliferative vascular smooth muscle phenotype [30, 47]; thus, elastin loss/degradation in itself may contribute to changes in vascular smooth muscle behaviour. Important to note here is that neutrophil elastase induces both vascular smooth muscle cell apoptosis and increased vascular smooth muscle expression of pro-inflammatory chemokines causing vessel wall remodelling [12, 46]. These findings suggest that the loss of arterial α-SMA-positive cells/smooth muscle cells we describe may also be related to the cells’ increased neutrophil elastase expression.
The CAA-independent but Braak stage-dependent α-SMA loss in both artery types from Braak stage II/III onward we observed was supported by previous findings of CAA-independent vascular smooth muscle loss in the posterior cerebral artery in AD , and was different from the CAA dependence of the arteriolar/microvascular α-SMA loss we and others [17, 54, 55] had found. CAA causes pathologic remodelling of the vessel wall that is linked to CAA-induced hypoperfusion  and impaired perivascular drainage . Conversely, chronic, CAA-/Aβ-independent hypoperfusion instigates microvascular CAA formation, cortical microinfarcts, and neurovascular dysfunction [16, 45], and progresses with Braak tau pathology [9, 41]. Further, while Aβ42-induced endothelin-1 upregulation causing vasoconstriction/hypoperfusion becomes significant only at Braak stage V–VI/in AD patients , as mentioned, significantly reduced cerebral tissue oxygenation can already be observed at Braak stage III–IV . The finding of PHF-tau accumulation in the perivascular space of intraparenchymal vessels starting at early Braak stages we describe here and is present in AD as has been reported by others [48, 57] may suggest a tau pathology-driven mechanism underlying this hypoperfusion. Given that the intraparenchymal vessels with perivascular PHF-tau accumulation showed α-SMA loss, and that perivascular molecules have access to perivascular drainage pathways and thus leptomeningeal arteries [26, 60, 62], it is likely that toxic (oligomeric) PHF-tau species contributed to the early Braak stage arterial α-SMA loss. This notion is supported by the fact that arterial α-SMA loss, arterial elastin degradation, and intraparenchymal perivascular PHF-tau accumulation all started around Braak stage II–III, suggesting a link between the loss of vessel wall constituents important for cerebral blood flow/perivascular drainage and perivascular PHF-tau.
Yet, a potential involvement of Aβ in the early arterial wall changes observed herein cannot be ruled out. Soluble Aβ oligomeric species, for example, are already present at Braak stage I–II . Even though Aβ oligomer-mediated vasoconstriction via endothelin-1 upregulation may be unlikely to occur at early, pre-AD Braak stages as mentioned above, (soluble) Aβ oligomers are cytotoxic and can be generated by vascular smooth muscle cells [20, 21]. Neuron-derived Aβ oligomers, similar to tau oligomers, are cleared along and may accumulate around the leptomeningeal arteries through the workings of the perivascular drainage system [26, 60]. Furthermore, although a clear effect of CAA on arterial elastin was not observed in this study, amyloidogenic proteins have been reported to bind and interact with elastin [18, 63] and, as mentioned, can activate MMPs, which degrade elastin.
We thank the Netherlands Brain Bank, Netherlands Institute for Neuroscience, Amsterdam, for providing us with the brain material used herein and the related clinical and neuropathological reports. We thank and are grateful to Dr. Kim Baeten, University of California, Berkeley, USA, for careful proofreading and helpful comments.
Compliance with ethical standards
Conflict of interest
RMN owns stock in Neurimmune Therapeutics AG, Switzerland, and is co-founder and member of the board of directors of said company.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.