Immunolocalization of platelet‐derived growth factor receptor‐β (PDGFR‐β) and pericytes in cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL)

Aims Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is identified by aggregates of NOTCH3 extracellular domain (N3ECD) along capillaries and the deposition of granular osmiophilic material (GOM). We assessed the pattern of distribution of pericytes in relation to N3ECD deposits in cerebral microvessels of CADASIL subjects. Methods We assessed post mortem brains from (n = 50) subjects with CADASIL, cerebral small vessel disease, and similar‐age cognitively normal and older controls. Immunohistochemical and immunofluorescent staining methods were used to study the distribution and quantify immunoreactivities of the platelet‐derived growth factor receptor‐β (PDGFR‐β) (for pericytes) and microvascular markers in the frontal cortex and white matter. Results PDGFR‐β antibody stained cells typical of pericytes in capillaries and small arterioles in both the grey and white matter. PDGFR‐β reactive pericytes adopted ‘crescent’ morphology wrapped closely around capillary walls readily evident in cross‐sections. We noted considerable overlap between PDGFR‐β and N3ECD imunoreactivities in capillaries. Quantitative analysis of PDGFR‐β immunoreactivity revealed significant differences in PDGFR‐β %A in CADASIL compared with young controls (P < 0.05). PDGFR‐β %A was further positively correlated with the basement membrane marker collagen IV (r = 0.529, P = 0.009), but was not associated with GLUT‐1, the marker for endothelial cells. Conclusions Our results suggest increased expression of PDGFR‐β immunoreactive pericytes in cerebral microvessels in CADASIL compared with similar age controls. While we cannot confirm whether PDGFR‐β‐expressing pericytes produce N3ECD and hence GOM, our findings demonstrate that up‐regulation of pericyte‐like cells is associated with microvascular changes, including loss of vascular smooth muscle cells in CADASIL.

600 families worldwide [1][2][3]. It is caused by mutations in the NOTCH3 gene and results in aggregation of NOTCH3 extracellular domain in granular osmiophilic material (GOM) within blood vessel walls, degeneration of vascular smooth muscle cells (VSMCs) and a small vessel disease (SVD) type dementia [1,2,4]. We have previously shown that capillary beds are also associated with an accumulation of N3ECD deposits, which mostly represent GOM in CADASIL. It has been previously postulated that pericyte-like cells, owing to the similar developmental lineage as VSMCs, may produce GOM. Previous electron microscopy (EM) studies have also shown that amorphous GOM is located at the cell membranes of pericytelike cells [5][6][7].
Pericytes are located abluminal to endothelial cells within the basement membrane of capillaries, precapillary arterioles and post-capillary venules [8][9][10]. They have multiple functions beyond contractile control of blood flow, providing paracrine signals during angiogenesis and some phagocytic activity. Pericytes are further specialized as a vital component of the bloodbrain barrier (BBB) and central nervous system (CNS) capillaries have around 22-32% pericyte coverage, second in density only to capillaries of the retina [11,12]. Pericyte processes wrap around the length of capillaries and form close contacts with endothelial cells through peg and socket junctions, gap junctions, and adhesion plaques. They are thought to contract the capillary in response to neuronal activity and during ischaemia [13]. Three major signalling pathways have been identified between endothelial and pericyte cell populations during angiogenesis: transforming growth factor-β, angiopoietin and platelet-derived growth factor (PDGF) [14]. In particular, PDGF subunit B expressed by endothelial cells recruits pericytes expressing PDGF receptor-β (PDGFR-β) to the capillary wall during vessel maturation in angiogenesis. Blockage of this signal results in fewer recruited pericytes to the vessel, causing vessel leakage, tortuosity, formation of microaneurysms and microbleeds [15]. NOTCH3 is also expressed in pericytes and perturbed NOTCH3 signalling during development results in reduced pericyte recruitment to capillaries [16]. In accord with angiogenic factors, PDGFR-β expression is activated following ischaemic injury, where PDGFR-β signalling surprisingly promotes transformation of VSMC to a different phenotype and loss of contractile function [17,18]. Notch1 expression has similarly been shown to be activated by ischaemia with a similar effect to promote differentiation of VSMCs to different phenotypes [19]. In contrast, the expression of NOTCH3 is reduced following vascular injury [20]. A study exploring cultured VSMCs derived from a CADASIL patient reported that there was a negative feedback mechanism between the expression of NOTCH3 and PDGFR-β [21], suggesting that NOTCH3 and PDGFR-β signalling may act antagonistically to control VSMC differentiation state. Another recent study has further identified signature genes downstream of Notch activity in retinal pericytes and suggests that tight regulation of Notch signalling is crucial for pericyte survival [22]. In an effort to explore some of these mechanisms [2], we primarily aimed to localize and quantify the expression of PDGFR-β in CADASIL in relation to N3ECD and compare against similar age relevant controls. Table 1 provides demographic details and diagnoses in the subjects. The mean ages of the CADASIL and young control subjects were not different. Available case notes and radiological reports indicated that CADASIL subjects showed extensive white matter (WM) changes consistent with small vessel disease (SVD) of the brain and met the minimum criteria for cognitive impairment used in our post-stroke survivors study [24]. CADASIL diagnosis was confirmed by the presence of NOTCH3 gene mutations or the presence of granular osmiophilic material (GOM) in arteries within skin biopsies [25]. None of the controls had neurological or pathological evidence for cerebrovascular disease or neurodegenerative disorder. Tissue blocks from the frontal lobe of the CADASIL subjects and agematched controls were collected from four sources. In addition to the Newcastle Brain Tissue Resource (NBTR), Newcastle University, Campus for Ageing and Vitality, we obtained cases from the MRC London Brain Bank for Neurodegenerative Diseases, the MRC Sudden Death Brain and Tissue Bank, University of Edinburgh and Neurology Department, Ludwig Maximilians University, Germany. Tissue from SVD subjects and older controls were obtained from the NBTR. Use of brain tissue was approved by the local research ethics committee of the Newcastle upon Tyne Hospitals NHS Foundation Trust, the NBTR committee, and the ethics committees overseeing the Brain Banks at the other respective sites.

Image acquisition and analysis
Images of capillary beds or regions of interest (ROIs) within the tissue sections were captured on a Zeiss Axioplan 2.0 microscope and Image capture software (Infinity Capture V4.6.0, Lumenera Corporation, Ottawa, Ontario, Canada), taking care to avoid larger arterioles >50 μm external diameter. Immunohistochemical staining was quantified using Image Pro Plus (V.4.0; Media Cybernetics, Silver Spring, MD, USA). We assessed the percent area (%A) for each case from at least 10 ROI images representing the vascular area stained with PDGFR-β (as PDGFR-β %A). To test the quality of the immunoreactvities between individual sections and cases, and we ascertained the integrated optical density (IOD).There were no significant differences in IOD values between CADASIL and control samples, however. We also found that there were no relationships between the immunohistochemical staining of PDGFR-β and length of fixation, or post mortem interval, between the groups. All of the histopathological analyses were performed blind to the operator.

Immunofluorescence (IF)
Six-μm thick formalin-fixed paraffin-embedded tissue sections were cut serial to the sections used for immunohistochemistry, and underwent antigen retrieval in either 3 mM EDTA ethylenediaminetetra acetic acid buffer (pH 8.0) and blocked with 10% normal horse serum/PBS 'phosphate buff-   . Images were captured using Leica TCS SP2 (upright) and Zeiss Spinning Disk (Invert) confocal microscopes (Leica Microsystems, Wetzlar, Germany and Zeiss Microscopy, Oberkochen, Germany). In situ co-localization of two antibodies was assessed using the JaCoP Plugin for Image J [45]. The Manders co-localization coefficient used to compare the extent of co-localization between antibodies.The overlap of green and red (co-localization) antigens was set to be yellow.

Sclerotic index (SI)
Ten-μm sections were stained with H&E haematoxylin and eosin' for assessment of the SI of microvessels in the WM. Briefly, all arterioles with external diameter between 35 and 350 μm within the region of interest were imaged using a Zeiss Axioplan 2 microscope and Image capture software (Infinity Capture V4.6.0, Lumenera Corporation). SI was measured as described previously [25,26], and SI values of greater than 0.3 are considered to be a state of mild SVD, and those greater than 0.5 are considered to be severe [29].

Statistics
Data were analysed using SPSS (V19.0, IBM Corporation, Armonk, NY, USA); data were tested for normality using Shapiro-Wilk test. Differences between means of groups were first tested using anova and where appropriate using Tukey's post hoc test . Linear correlations between PDGFR-β immunoreactivity and other vascular markers were performed using the Pearson's correlation, as described previously [26]. The points represent %A for 23 pairs of values. Significant difference between means was considered when the P value was less than 0.05.

Immunohistochemical staining of PDGFR-β within capillaries
PDGFR-β antibody immunostained pericyte-like cells associated with capillaries of the smallest internal diameters (∼6 μm) and some pre-capillaries in both grey matter (GM) and WM (Figure 1). We also noted some PDGFR-β immunostaining of pyramidal neurons within the cortex. Pericytes were identified by their morphology and in crosssections evident as typical 'crescent' shaped cell bodies, wrapped closely around capillary walls ( Figure 1B). They were also observed around the transverse length of capillaries ( Figure 1A,C). Dual IF labelling with GLUT-1 antibodies revealed the close association between PDGFR-β and endothelial cells, but these markers labelled different cell types ( Figure 1C,D).
To further differentiate and validate PDGFR-β immunoreactivity, we performed triple labelling of capillaries with PDGFR-β, SMA and GLUT-1 ( Figure 2). This confirmed PDGFR-β immunoreactivity was largely associated with pericytes in capillaries in the general absence of SMA reactivity. However, we did observe both PDGFR-β and SMA immunoreactivities in some larger capillaries with diameters greater than 10 μm ( Figure 2D). In CADASIL cases, there were often severely degenerating capillaries, where expression of the endothelial protein GLUT-1 was perturbed and some perivascular cells stained positive (possibly leaked erythrocytes) for GLUT-1 with evidence of endothelial blebbing (Figure 2A,B). This was not frequent in SVD and less so in control.
To assess the relationship between pericytes and N3ECD deposits or GOM in CADASIL, we first replicated our previous findings to show that N3ECD immunoreactivity was localized as granular material along capillaries (Figure 3). We then used dual IF labelling to assess co-localization of PDGFR-β positive pericytes and N3ECD deposits. We observed N3ECDstained granular deposits along capillary walls, indicating an abundance of N3ECD and likely GOM deposits within the capillary beds, but these did not always overlap with PDGFR-βimmunostained cells ( Figure 3D). However, upon overall screening, PDGFR-β positive pericytes were co-localized with N3ECD in ∼70% of the CADASIL cases. Using dual IF labelling to assess co-localisation between PDGFR-β positive pericytes and N3ECD deposits we found strong association between the two antibodies with the following statistics: Pearson's r 2 = 0.905; Manders' coefficients M = 0.969 for fraction of N3ECD overlapping PDGFR-β, and M = 0.920 for fraction of PDGFRβ overlapping N3ECD (P < 0.001). This indicated a close relationship between pericytes with N3ECD deposits or GOM. media and more diffusely in the adventitia of small arterioles (internal diameter approximately 30 μm) in CADASIL, SVD cases and less so in controls ( Figure 4). However, diffuse and granular PDGFR-β staining within the intima was observed in CADASIL subjects where vessels had undergone hyalinosis (SI >0.5), but to a lesser extent in SVD (Figure 4). PDGFR-β immunostained smaller arterioles more frequently, concomitant with degenerated smooth muscle cells and N3ECD deposits ( Figure 4B,C). Using multiple IF staining for PDGFR-β, SMA and GLUT-1, we assessed the distribution of the markers in CADASIL, SVD and controls. We observed that in both CADASIL and SVD arterioles, there was upregulation or enhanced PDGFR-β immunoreactivity in endothelial cells compared with controls with positive GLUT-1 but little PDGFR-β immunostaining ( Figure  4G-I). In some large arterioles in CADASIL cases, there was near lack of GLUT-1 reactivity, but strong PDGFR-β reactivity ( Figure 5). We also observed PDGFR-β staining within the tunica media and adventitia of arterioles more frequently in CADASIL compared with SVD, and found concomitant immunostaining for N3ECD alongside

Quantitative analysis of PDGFR-β and the relationship with vascular markers
Quantitative analysis in capillaries in the frontal WM immunostained for PDGFR-β revealed a trend for differences in the %A immunoreactivity between groups (anova, P = 0.084). However, post hoc tests revealed significant differences between CADASIL and young control, and between old and young controls (Tukey's test, P < 0.05; Figure 6A).
To assess if PDGFR-β immunostaining was related to changes in vascular cell markers, we correlated the PDGFR-β %A determined in the frontal WM against the endothelial cell marker, GLUT-1 and the basement membrane marker, COL4. We found no association between the %A immunostaining of PDGFR-β and GLUT-1 (Pearson's r = 0.312, P = 0.148, n = 23, Figure 6B). However, we observed a significant correlation between %A immunostaining of PDGFR-β and COL4 (Pearson's r = 0.529, P = 0.009, n = 23, Figure 6C). We further calculated the PDGFR-β %A: COL4 %A ratio of the CADASIL cases and observed that there were no significant differences between the young control and CADASIL groups, indicating lack of changes in overall PDGFR-β immunoreacvity per unit amount of COL4 or basement membrane protein (P > 0.05). To also relate whether arteriolar sclerosis in general had any impact on capillary density, we correlated PDGFR-β immunoreactivity as %A with SI and found no relationship between %A staining of PDGFR-β and SI (data not shown).

Discussion
Using immunohistochemical methods, we localized PDGFR-β immunoreactivity in brains of subjects with CADASIL and SVD. Concentrating on the WM, we confirmed that PDGFR-β positive cells are pericyte-like cells frequently found around capillaries. These cells are also associated with N3ECD immunoreactivity, which is a reliable marker of GOM in CADASIL [4]. From these static studies, it is not clear whether pericytes produce or uptake N3ECD or GOM as a measure for perivascular clearance [30]. However, our findings are consistent with the postulation that pericytes, identified by electron microscopy, are associated with GOM [5][6][7] and by extension N3ECD deposits.
We also showed that there is an increase in PDGFR-β immunoreactivity associated with pericytes in CADASIL compared with similar age controls, with a similar trend in SVD compared with respective older controls. These results suggest changes in the capillary population of PDGFR-β-expressing pericytes in the WM in CADASIL. The findings further suggest that pericyte coverage of capillaries is robust in CADASIL despite the presence of a significant burden of N3ECD deposits or GOM in their proximity. This possibly indicates that the integrity of the BBB is not completely diminished because it has been previously shown that the loss of pericytes in PDGF-null mice impairs the barrier [9]. We also observed that PDGFR-β immunoreactivity was associated with COL4 immunoreactivity, but not GLUT1 in capillaries. This suggests that immunoreactivity to PDGFR-β or density of pericytes was not changed in relation to changes in COL4 reactivity, but does confirm the known relationship between pericytes and capillary basement membranes [9]. The lack of a significant relationship between PDGFR-β and GLUT-1 may partly reflect abnormalities in endothelial cells, which are also in close proximity to pericytes [9].
PDGFR-β is known to be up-regulated in response to hypoxia [19,31] and its expression can be induced in different cell types under different conditions [32]. A recent reported a significant reduction in the number of pericytes following acute stroke, followed by proliferation of PDGFR-β+ cells of neurovascular origin during scar formation [31]. However, this study utilized a model of acute large artery occlusion which produces a large ischaemic lesion [31], a process not common to CADASIL or SVD [2,3]. Instead, CADASIL is known to undergo slow insidious progression with multiple small strokes (and transient ischaemic attacks), resulting in microinfarcts and an overall loss of axonal myelin without related WM changes [2,3,26]. We therefore postulate a differential involvement of PDGFR-β-expressing pericytes in CADASIL and possibly SVD, where a chronic hypoxic state [33] or low levels of transient hypoxia may up-regulate various factors such as bone morphogenic protein-4 (Maiko Uemura and Masafumi Ihara, pers. comm.), which co-localizes with PDGFR-β and may in turn affect its expression in pericytes. This, however, does not result in large-scale increases in PDGFR-β-cells migrating into the parenchyma to form scar tissue.
PDGF is also expressed by endothelial cells during angiogenesis to attract PDGFR-β-expressing pericytes to capillary walls, where PDGF acts as a downstream of angiogengic signal, which in turn increases in response to hypoxic signals [15,34]. Further to involvement in pericyte recruitment in angiogenesis, PDGF expression interacts with Notch signalling, where both Notch1 and Notch3 ligands promoted PDGFR-β expression in cultured VSMC through independent mechanisms, but both recep- Figure 6. Quantification of PDGFR-β immunoreactivity in the frontal white matter in CADASIL. (A) Quantitative analysis of PDGFR-β percent area (%A) revealed a trend for differences between the groups (anova, P = 0.084, n = 6-7 samples per group). However, post hoc tests revealed significant differences between CADASIL subjects and young controls, and the old control group compared with young controls (Tukey's test, *P < 0.05). (B) Lack of relationship between PDGFR-β %A and GLUT-1 immunoreactivity (r = 0.312, P = 0.148). (C) PDGFR-β %A was significantly correlated with the basement membrane marker COL4 (r = 0.529, P = 0.009). CADASIL, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy; PDGFR-β, platelet-derived growth factor receptor-β. tors appear to instigate increased functional PDGFR-β expression [21]. This mechanism may serve to transiently attenuate VSMC phenotype in response to extracellular cues. A muted response of PDGFR-β up-regulation in cultured VSMC from a CADASIL patient with the R133C NOTCH3 mutation [21], a version of NOTCH3 receptor thought to be hypomorphic, although the functional phenotype of CADASIL associated NOTCH3 mutations, is still under debate [35,36]. These findings may explain the observed increase in PDGFR-β staining in capillaries of CADASIL, where dysfunctional NOTCH3 mutant proteins present in CADASIL result in an exacerbated response of PDGF and PDGFR-β expression in states of hypoxia to promote angiogenesis.
We observed PDGFR-β reactivity in endothelial cells of small arterioles, where the staining was more intense frequently in CADASIL. We further observed (qualitative) diffuse PDGFR-β immunostaining within the vessel wall, specifically in the small arterioles, which had undergone sclerosis of the vessel wall. PDGF signalling has been associated with other vascular pathologies, such as atherosclerosis [37,38], kidney fibrosis [39], lung fibrosis [32], pulmonary arterial hypertension (PAH) [40,41] and diabetes [42]. Some of these vascular disorders exhibit similar pathophysiology to those observed in arteriolosclerosis and hyalinosis of arterioles commonly observed in dementing cerebrovascular disorders, where vessels exhibit significant vessel wall thickening, increased collagen deposition and changes in VSMC morphology [26]. Similarly, in atherosclerosis and arterial restenosis, VSMCs de-differentiate from their contractile state towards a less specialized cell type, switching to a synthetic phenotype and migrating into the intima where they undergo proliferation, a mechanism regulated by PDGFR-β signalling [32,37,42]. We previously reported that SVD and CADASIL cases exhibit increased arteriolosclerosis in the frontal WM [26]. Taken together, our results suggest that increased PDGFR-β in endothelial cells of arterioles may be linked to degeneration of VSMC and vessel wall hyalinosis.
Our results suggest that there is a similar mechanism responsible for the vessel sclerosis observed in small vessel diseases of the brain as that which has been reported in other systemic disorders, where an increase in PDGF signalling is initiated by a primary vascular disorder such as hypertension or diabetes. This PDGF signal results in a phenotype conversion of VSMCs from contractile to a synthetic phenotype, which would function to reduce the contractile ability of the blood vessel and therefore a loss of autoregulation of cerebral blood flow. Once the VSMCs begin to dedifferentiate, the control of local blood flow may be attenuated, resulting in further states of hypoxia, which in turn up-regulates the Notch and PDGF pathways, and the vessel is further impeded in a vicious cycle. Although the consequences of CADASIL-causing mutations are debated [35,36], NOTCH3 mutations might be considered the primary factor in the disease process, and the similarities between CADASIL and SVD could be explained by NOTCH3 polymorphisms associated with late-onset SVD [35,43], where presence of a hypomorphic NOTCH3 phenotype may render a patient more vulnerable to ischaemic processes occurring as a result of other environmental damage to blood vessel walls, such as hypertension.
Our study is not without limitations. The availability of a larger pool of brain samples would have permitted more in-depth characterization of the PDGFR-β-expressing pericytes in the WM regions. The availability of more specific markers of pericytes would also have been useful to verify our findings. However, our initial study is sufficiently novel to suggest that the persistent hypoxic state within the deep WM [44] may reflect activation of capillary-associated pericytes in CADASIL. The significance of these findings is important for both CADASIL and SVD subjects, where agents, which modulate PDGF signalling, may provide therapeutic strategies to reduce the morbidity in cerebral SVDs.