ORIGINAL RESEARCH published: 13 March 2019 doi: 10.3389/fneur.2019.00187 Brain ApoA-I, ApoJ and ApoE Immunodetection in Cerebral Amyloid Angiopathy Jessica Camacho 1 , Teresa Moliné 1 , Anna Bonaterra-Pastra 2 , Santiago Ramón y Cajal 1 , Elena Martínez-Sáez 1* and Mar Hernández-Guillamon 2* 1 Pathology Department, Vall d’Hebron University Hospital, Universitat Autònoma de Barcelona, Barcelona, Spain, 2 Neurovascular Research Laboratory, Vall d’Hebron Research Institute, Universitat Autònoma de Barcelona, Barcelona, Spain Cerebral amyloid angiopathy (CAA) is a common cause of lobar intracerebral hemorrhage (ICH) in elderly individuals and it is the result of the cerebrovascular deposition Edited by: of beta-amyloid (Aβ) protein. CAA is frequently found in patients with Alzheimer’s Aurel Popa-Wagner, University Hospital Essen, Germany disease (AD), although it has an independent contribution to the cognitive deterioration Reviewed by: associated with age. Specific apolipoproteins (Apo) have been associated with Aβ Paul O’Callaghan, fibrillization and clearance from the brain. In this regard, in the present study, we Uppsala University, Sweden analyzed the brain levels of ApoE, ApoA-I, and ApoJ/clusterin in autopsy brains from Bo Zhou, Stanford University, United States 20 post-mortem cases with CAA type I, CAA type II, with parenchymal Aβ deposits *Correspondence: or without Aβ deposits. Our objective was to find a possible differential pattern of Elena Martínez-Sáez apolipoproteins distribution in the brain depending on the CAA pathological presentation. eamartinez@vhebron.net Mar Hernández-Guillamon The protein expression levels were adjusted by the APOE genotype of the patients Mar.hernandez.guillamon@vhir.org; included in the study. We found that ApoE and ApoJ were abundantly present in marguillamon@gmail.com meningeal, cortical, and capillary vessels of the brains with vascular Aβ accumulation. Specialty section: ApoE and ApoJ also deposited extracellularly in the parenchyma, especially in cases This article was submitted to presenting Aβ diffuse and neuritic parenchymal deposits. In contrast, ApoA-I staining Neurodegeneration, was only relevant in capillary walls in CAA type I cases. On the other hand, ICH was the a section of the journal Frontiers in Neurology principal cause of death among CAA patients in our cohort. We found that CAA patients Received: 06 September 2018 with ICH more commonly had APOEε2 compared with CAA patients without ICH. In Accepted: 14 February 2019 addition, patients who suffered an ICH presented higher vascular ApoE levels in brain. Published: 13 March 2019 However, higher ApoE presence in cortical arteries was the only independent predictor of Citation: Camacho J, Moliné T, suffering an ICH in our cohort after adjusting by age and APOE genotype. In conclusion, Bonaterra-Pastra A, Ramón y Cajal S, while ApoE and ApoJ appear to be involved in both vascular and parenchymal Aβ Martínez-Sáez E and pathology, ApoA-I seems to be mainly associated with CAA, especially in CAA type I Hernández-Guillamon M (2019) Brain ApoA-I, ApoJ and ApoE pathology. We consider that our study helps to molecularly characterize the distribution Immunodetection in Cerebral Amyloid subtypes of Aβ deposition within the brain. Angiopathy. Front. Neurol. 10:187. doi: 10.3389/fneur.2019.00187 Keywords: ApoE, ApoA-I, ApoJ, clusterin, β-amyloid, cerebral amyloid angiopathy, intracerebral hemorrhage Frontiers in Neurology | www.frontiersin.org 1 March 2019 | Volume 10 | Article 187 Camacho et al. Apolipoproteins in Cerebral Amyloid Angiopathy INTRODUCTION emerged as a natural chaperone able to bind Aβ and prevent its aggregation and toxicity in vitro (16–18). Cerebral beta-amyloidosis is defined as the anomalous deposition With this background, the aim of the present study was of β-amyloid (Aβ) protein in the brain. Aβ originates by to immunohistochemically investigate the presence of ApoE, the sequential processing of the amyloid precursor protein ApoJ, and ApoA-I in brains classified as CAA type I, CAA (APP), primarily generating peptides of 40 and 42 amino type II, brains with Aβ parenchymal deposits without vascular acids. The abnormal folding of amyloid material produces deposits and controls. Our hypothesis was that the expression a β-pleated sheet secondary structure, which forms highly and distribution of this family of apolipoproteins in the brain insoluble amyloid fibrils. In the most common form of Cerebral was modulated by the accumulation of Aβ as insoluble plaques, Amyloid Angiopathy (CAA), insoluble Aβ accumulates in the deposits in cortical/meningeal vessels or capillaries. Furthermore, wall of the cortical and meningeal vessels, replacing the smooth we examined whether the immunopositivity of ApoE, ApoJ, muscle cells, and inducing vascular degeneration. However, Aβ and ApoA-I in brain vessels was associated with vascular Aβ can also accumulate in brain capillaries. In this regard, CAA deposition and symptomatic ICH appearance in patients affected can be neuropathologically classified as CAA type I, which by CAA. is characterized by Aβ deposition in cerebral capillaries, or CAA type II, where cerebral capillaries are not involved (1). The most relevant clinical consequence of CAA is the risk MATERIALS AND METHODS of lobar symptomatic intracerebral hemorrhages (ICH) and microscopic cortical infarcts (2). CAA patients also present Selection of Patients cognitive impairment, lower perceptual speed and episodic Twenty patients with a postmortem neuropathological study memory impairment, separately from the effect of Alzheimer’s performed at the Pathology Department of Vall d’Hebron disease (AD) pathology (3). Indeed, AD is the most frequent University Hospital were selected for the study. A written and neurodegenerative disease, morphologically characterized by the informed consent was obtained from all patients (presenting presence of hyperphosphorylated tau protein in the cytoplasm Aß brain deposits), or their representatives, to participate in of neurons and extracellular Aβ deposits forming parenchymal this research protocol before the tissue donation. For control plaques. Although CAA is observed in more than 90% of cases, we received the exemption of the Ethic Committee to AD patients (4), only a minority of AD subjects demonstrate analyze them as anonymized samples in the context of this study. hemorrhages suggestive of advanced CAA (5). In fact, the The project was approved by the Clinical Investigation Ethical combination of factors that determine whether Aβ accumulation Committee of the Vall d’Hebron University Hospital, Barcelona, occurs in parenchyma, developing into AD, or exclusively in Spain PR(IR)93/2013). Patients’ characteristics and causes of blood brain vessels, triggering CAA-associated ICHs, still needs death are presented in Table 1. All samples were obtained from to be elucidated. It is plausible that the transport routes of 2 to 20 h after decease and post-mortem specimens were fixed in Aβ will be crucial in this process. Evidence for that is the 10% buffered formalin for 3–4 weeks. Macroscopic examination presence of vascular amyloid and the occurrence of cerebral was then performed, and 27 cortical and subcortical areas, microhemorrhages and vasogenic edema when parenchymal Aβ including midbrain, and cerebellum, were selected for routine is mobilized by active or passive immunization in AD (6). neuropathological diagnosis. Paraffin-embedded blocks were cut Specific apolipoproteins, beyond their natural role in into 4 µm sections and stained with hematoxylin & eosin. cholesterol metabolism, have been proposed to be involved in Aβ Aß immunohistochemistry was performed in selected areas (as transport and clearance from the brain (7, 8). Apolipoproteins described at the end of this section). All cases were reviewed represent the protein part of lipoproteins, which classically by two neuropathologists at a multiheaded microscope. Patients transport cholesterol from tissues to the liver. However, their were divided into four groups according to Aß localization in the presence and function in the brain are not well-described. The brain (Figure 1): CAA type I patients (group 1: 6/20 cases); CAA main apolipoproteins present in brain are apolipoprotein E type II patients (group 2: 8/20 cases); patients with parenchymal (ApoE) and apolipoprotein J (ApoJ), both of them genetically Aß deposits without vascular involvement (group 3: 3/20 cases) associated with cerebral β-amyloidosis. On one hand, ApoE and control cases with neither vascular nor parenchymal Aß has a determining role in the progression of Aβ deposition, deposits (group 4: 3/20 control cases). since having the APOEε4 allele is the major risk factor for AD CAA lesions were graded following the criteria described by (9), as well as for CAA (10). The role of ApoJ [also known as Vonsattel et al. (19) and modified by Greenberg and Vonsattel clusterin, (CLU)] in Aβ pathology is highlighted by genome-wide (20). Grade 0 CAA was given when no amyloid could be found in association studies (GWAS) that found a statistical association the vessel wall. Grade 1 CAA lesions showed amyloid deposits in between a single nucleotide polymorphism within the CLU gene otherwise normal vessels. The thickening of the vessel wall with and the risk of suffering AD (11). In addition, more recently, a stiff appearance defined a grade 2 CAA lesion, where the media the association between the CLU genotype and Aβ deposition, was replaced by amyloid. Grade 3 CAA showed cracking of the determined by PET imaging, has been described (12). Indeed, vessel wall, creating a “vessel-within-vessel” appearance affecting both ApoE and ApoJ have been shown to influence the amyloid at least 50% of the circumference of the vessel. In grade 4 CAA fibrillogenesis process and clearance in experimental models lesions, scarring of the wall was observed, with fibrinoid necrosis (13–15). In this regard, apolipoprotein A-I (ApoA-I) has also and traces of intermingled amyloid deposits. Classification into Frontiers in Neurology | www.frontiersin.org 2 March 2019 | Volume 10 | Article 187 Camacho et al. Apolipoproteins in Cerebral Amyloid Angiopathy TABLE 1 | Cause of death and patients’ characteristics. Case Group Age Sex APOE Genotype Braak Cerad CAA type Vonsattel Cause of death 1 1 85–90 F E4/E3 V–VI C I II Intracerebral hemorrhage 2 1 75–80 F E3/E3 IV C I III Intracerebral hemorrhage 3 1 75–80 F E3/E3 I A I II Metastatic colon adenocarcinoma 4 1 70–75 M E4/E3 II B I II Intracerebral hemorrhage 5 1 80–85 M E4/E2 – – I II Intracerebral hemorrhage 6 1 75–80 M E4/E3 III C I II Heart failure 7 2 80–85 M E2/E3 IV B II II Intracerebral hemorrhage 8 2 85–90 F E2/E3 V–VI C II IV Intracerebral hemorrhage 9 2 65–70 M E3/E3 II C II II Pericarditis 10 2 90–95 F E3/E3 III C II I Mesenteric thrombosis 11 2 75–80 M E4/E3 I B II II Metastatic prostate adenocarcinoma 12 2 80–85 F E2/E3 III A II II Intracerebral hemorrhage 13 2 80–85 F E3/E3 – – II I Secondary amyloidosis 14 2 70–75 F E3/E3 – – II I Metastatic breast carcinoma 15 3 75–80 F E3/E3 I A – – Acute meningitis 16 3 55–60 F E3/E3 II C – – Ischemic stroke 17 3 75–80 F E3/E3 I B – – Acute pneumonia 18 4 50–55 M E4/E3 – – – – Pulmonary hypertension 19 4 60–65 F E3/E3 – – – – Large B cell lymphoma 20 4 70–75 M E4/E3 – – – – Multiple Myeloma F, female; M, male. FIGURE 1 | Classification of cases into 4 groups according to the parenchymal and vascular Aß deposition by immunostaining. Representative images of Aß immunohistochemistry. (A) Group 1: CAA type I, defined by the presence of capillary and arterial Aß deposits. (B) Group 2: CAA type II, defined by the presence of arterial Aß deposits without capillary involvement. (C) Group 3, characterized by the presence of parenchymal Aß deposits without vascular involvement. (D) Group 4, defined by the absence of Aß deposits (control cases). 100x magnification. CAA type I and CAA type II was made according to the criteria pathology was evaluated and staged according to Braak (21) described by Thal et al. (presence and absence of capillary Aß and CERAD’s criteria (22) through immunohistochemistry using deposits in types I and II, respectively) (1). Alzheimer disease the mouse anti-human Amyloid-ß Clone 6F/3D 1:500 (Dako, Frontiers in Neurology | www.frontiersin.org 3 March 2019 | Volume 10 | Article 187 Camacho et al. Apolipoproteins in Cerebral Amyloid Angiopathy Glostrup, Denmark) and the mouse anti-human Tau protein of deposits (5–10/ HPF); and score 3 indicated frequent deposits Clone AT8 1:20 (Thermo Scientific, Rockford, USA). (>10/ HPF). Immunodetection of ApoE, ApoJ, and ApoA-I in the cytoplasm of subpial astrocytes and pyramidal neurons Immunohistochemistry was semi-quantitatively rated as: absence, mild, moderate, or To describe a representative distribution of Aβ deposits, up to intense expression. 3 blocks from cortical areas, including perihippocampal cortex in five cases, were selected in order to evaluate an equivalent APOE Genotype number of meningeal and cortical vessels (20–30 per case). Genomic DNA was extracted from formalin-fixed, paraffin- Brainstem and cerebellar areas were not used in this study. embedded tissue using EZ1 DNA Tissue kit (48) and Bio Robot Consecutive sections of the selected blocks were performed for EZ1 (Qiagen, Hilden, Germany). APOE genotypes (rs429358 the detection of ApoE, ApoJ, and ApoA-I. In those samples and rs7412) were determined by allelic discrimination using from patients who suffered an ICH, the immunopositivity of the the SNP genotyping mixes C-3084793 and C-904973 (Applied different apolipoproteins were evaluated outside the lesion. Biosystems, Foster City, CA, USA) and the TaqMan_Genotyping Paraffin was removed, endogenous peroxidases were blocked Master Mix (Applied Biosystems) in a Rotor Gene 6000 Real- with methanol-H2 O2 , and non-specific binding sites were Time PCR analyzer (Corbett Life Sciences, Valencia, CA, USA). blocked with 10% normal goat serum for 1 h. Sections were incubated with mouse anti-human ApoE 1:200 (Clone E6D7, Calbiochem—NE1004, Madrid, Spain), mouse anti-human ApoJ Statistical Analysis 1:1,000 (Clone 5 BD Pharmingen −552886, Madrid, Spain) and Statistical analysis was conducted using SPSS package 17.0. rabbit anti-human ApoA-I 1:200 (Clone EP1368Y, Abcam— Clinical or pathological associations with age (continuous ab52945, Cambridge, UK) for 40 min at RT. Secondary antibody variable following a normal distribution after applying the (goat anti-mouse horseradish peroxidase, 1:500) was applied for Kolmogorov–Smirnov test) were tested using the Student’s t-test 8 min. Immunoreactive sites were developed with DAB solution or One-way ANOVA test, as appropriate. The rest of the and serial sections were stained with hematoxylin & eosin variables analyzed were considered categorical variables and the (H&E) to assess morphology. Meningeal and cortical vessels, significance of frequency differences were calculated using the neurons, subpial astrocytes, and parenchymal deposits in cortical Chi Square (χ2 ) test or the exact test of Fisher for 2 × 2 and perihippocampal areas were semi-quantitatively evaluated. contingency tables. Logistic regression analysis were performed Blind to the group distribution, two neuropathologists assessed to identify independent associations with the apolipoproteins the three antibodies under a multiheaded microscope until expression. A p < 0.05 was considered statistically significant. agreement was reached. RESULTS ApoE, ApoJ, and ApoA-I Immunopositivity Scores ApoE, ApoJ, and ApoA-I cerebral levels and localization were Supplemental Table 1 summarizes the scores used for the analyzed in brains from the four different groups of patients vascular and parenchymal immunodetection. The presence of categorized according to the parenchymal and vascular Aß ApoE, ApoJ, and ApoA-I immunostaining in cortical and deposition determined by immunostaining (Figure 1). Principal meningeal arteries was graded according to the percentage of demographic characteristics, AD, and CAA pathological scores arteries showing immunopositivity: score 0 was given when no and APOE genotype of all patients included in the study can positive arteries could be found; score 1 indicated <30% of be found in Table 1. CAA cases (encompassing groups 1 and positive arteries; score 2 corresponded to 30–75% of positive 2) were significantly older than non-CAA cases (encompassing arteries; and score 3 indicated immunopositivity in more than groups 3 and 4) (81 ± 7 vs. 66 ± 12, p = 0.026). As expected, 75% of the evaluated arteries. ApoE, ApoJ, and ApoA-I presence ICH was the cause of death in CAA patients more often than in cortical capillaries was evaluated in 10 high-power field (HPF) in non-CAA cases [50% (7/14) vs. 0% (0/6), p = 0.036]. Among and scored as follows: score 0 when no signal was detected CAA patients, no differences were found in age (p = 0.983), in cortical capillaries; score 1 when the signal was focal (in sex (p = 0.774), or cause of death between CAA type I (group fewer than 50% of the capillaries); score 2 when the signal 2) and type II (group 1) cases (p = 0.298). However, APOEε4 was focal and linear (in fewer than 50% of the capillaries); genotype was more represented in CAA type I patients than score 3 when the signal was diffuse (in more than 50% of in those lacking capillary CAA (type II) [67% (4/6) vs. 13% capillaries); and score 4 when the signal was strong and thick in (1/8), p = 0.036]. CAA type II cases had a slightly higher ε2 the capillaries. ApoE, ApoJ, and ApoA-I parenchymal deposits allele frequency than CAA type I, although differences were were evaluated in 10 HPF, after detecting the hotspot for each not significant [38% (3/8) vs. 17% (1/6), p = 0.393]. APOE block, and graded according to the number of parenchymal genotype was not statistically associated with sex, age or other deposits, divided into neuritic-like plaques (with a dense core pathological scores. However, APOE genotype was associated surrounded by a clear halo and, sometimes, a neuritic corona) with having suffered an ICH (p = 0.017). Interestingly, CAA and diffuse-like plaques (without any central core): score 0 patients with ICH had a significantly higher APOEε2 genotype indicated absence of parenchymal deposits; score 1 indicated frequency compared with CAA patients without ICH [57% (4/7) sparse deposits (<5/HPF); score 2 indicated a moderate number vs. 0% (0/7), p = 0.018]. Frontiers in Neurology | www.frontiersin.org 4 March 2019 | Volume 10 | Article 187 Camacho et al. Apolipoproteins in Cerebral Amyloid Angiopathy The presence of ApoE, ApoJ, and ApoA-I in cortical 100% (6/6) of CAA type I cases presented the highest score and meningeal arteries, capillaries, neuritic-like, and diffuse in ApoJ levels (score 4), whereas the mean score was lower in parenchymal deposits was evaluated by immunohistochemistry. the other groups (p = 0.001). ApoJ presence was also stronger Examples of immunopositivity for each case are shown in in microvessels affected by Aß accumulation, so levels were Figure 2. Parenchymal ApoE, ApoJ, and ApoA-I presence was higher in brain capillaries from CAA type I cases than from also analyzed within brain cells (Supplemental Figure 1). CAA type II cases (p = 0.002). Interestingly, ApoA-I positivity in the capillaries was more pronounced globally and more ApoE, ApoJ, and ApoA-I Presence in homogenous between groups, although microvessels from CAA type I samples also presented the highest score in 83% (5/6) Cerebral Arteries of the cases, which was more often than the other groups In meningeal vessels, ApoE positivity was not significantly (p = 0.006). Indeed, ApoA-I levels in brain capillaries were also different among groups (p = 0.307): score 3 was given to 67% significantly higher in CAA type I cases than in CAA type II (4/6) of CAA type I, 50% (4/8) of CAA type II, and 33% (1/3) of cases (p = 0.003) (Figure 4, Supplemental Table 2). In addition, control cases. ApoJ levels in meningeal vessels were higher than potential associations between ApoE, ApoJ, and ApoA-I presence ApoE but was equally represented in all groups (p = 0.448). The in brain capillaries and APOE genotype were determined. In this highest ApoJ level score (score 3) was given to 83% (5/6) of CAA regard, only the expression of ApoE protein in capillaries was type I cases, 87% (7/8) of CAA type II cases, 67% (2/3) of brains related to APOEε4 genotype (p = 0.035), although the association with parenchymal Aß deposits without Aß vascular involvement did not achieve statistical significance after the adjustment by age and 67% (2/3) of control cases. ApoA-I showed a lower presence and CAA subtype (odds ratio (OR) 2.32; 95% confidence interval in meningeal vessels than ApoE or ApoJ and did not differ among (CI) [0.97–5.87], p = 0.073). groups (p = 0.905); most of the samples received scores of 1 or 0. Expression of ApoE, ApoJ, and ApoA-I in the meningeal arteries was not significantly different between CAA type I and CAA ApoE, ApoJ, and ApoA-I type II cases or between CAA and non-CAA cases (Figure 3A, Supplemental Table 2). APOE genotype was not associated with Parenchymal Accumulation Frequent ApoE neuritic-like plaques (score 3) were found in the expression of ApoE, ApoJ, and ApoA-I in meningeal arteries. 50% (3/6) of CAA type I cases, 63% (5/8) of CAA type II cases Regarding the presence of these apolipoproteins in cortical and 67% (2/3) of cases with parenchymal Aß deposits without arteries, ApoE showed a score of 3 in 50% (3/6) of CAA type I vascular involvement. No ApoE neuritic-like plaques were found and 63% (5/8) of CAA type II cases. Interestingly, cortical vessels in control cases (differences among the 4 groups: p = 0.002). from brains with parenchymal Aß deposits without vascular Aß Similarly, 33% (2/6) of CAA type I and 63% (5/8) of CAA type involvement were not positive for ApoE (all cases scored 0), II cases showed frequent ApoJ neuritic-like plaques (score 3), but the differences among the 4 groups did not reach statistical whereas there were lower scores in cases with parenchymal Aß significance (p = 0.114). ApoJ vascular cortical positivity was deposits without vascular involvement and no ApoJ neuritic- higher than ApoE, showing a score of 3 in all (100%, 6/6) CAA like plaques in control cases (p = 0.002). ApoA-I deposits type I cases, in 63% (2/3) of CAA type II cases and in 33% were almost absent in all of the analyzed samples, and the (1/3) of brains with parenchymal Aß deposits without vascular distribution was not different among groups (p = 0.490). No involvement. ApoJ expression in the cortical vessels of controls differences were found in the number of ApoE, ApoJ, or ApoA- was lower than in those cases with cerebral Aß accumulation I neuritic-like plaques between CAA type I and type II cases (p = 0.048). Presence of ApoA-I was also less frequent in cortical (Figure 5A, Supplemental Table 2). APOE genotype was not vessels and was similarly distributed among groups (p = 0.441). associated with the presence of ApoE, ApoJ, and ApoA-I forming Expression of ApoE, ApoJ, and ApoA-I in cortical arteries was neuritic or diffuse-like parenchymal deposits not significantly different between CAA type I and CAA type Frequent ApoE diffuse-like plaques (score 3) were found in II cases. However, CAA brains presented a higher percentage of 50% (3/6) of CAA type I cases, in 63% (5/8) of CAA type II ApoE-positive cortical vessels than non-CAA cases (p = 0.099), cases and 67% (2/3) of cases with parenchymal Aß deposits and this was especially relevant for ApoJ (p = 0.018) (Figure 3B, without vascular Aß involvement. Brain samples from controls Supplemental Table 2). In addition, the presence of ApoE in presented fewer ApoE diffuse-like plaques, although differences cortical arteries was nearly significantly related to the APOEε2 among the 4 groups were not statistically significant (p = 0.389). genotype (p = 0.058). The presence of ApoJ forming diffuse-like plaques was frequent (score 3) in 33% (2/6) of CAA type I and 63% (5/8) of CAA ApoE, ApoJ, and ApoA-I Presence in Brain type II cases, in comparison with a lower detection in the rest Cortical Capillaries of the groups (p = 0.017). Although ApoE and ApoJ diffuse- ApoE staining in the capillaries scored 4 in 83% (5/6) of CAA like plaques were found in 67% (2/3) of controls, groups 3 and type I cases (group 1), in which the presence was higher than 4 presented lower scores compared with CAA groups 1 and in the other groups (p = 0.002). In fact, ApoE expression 2 in both cases. ApoA-I diffuse-like plaques were globally less in brain capillaries was significantly more abundant in CAA frequent than ApoE and ApoJ, without significant differences type I cases than in CAA type II cases (p = 0.004). ApoJ among groups (p = 0.234). Presence of ApoE, ApoJ, or ApoA- positivity in the capillaries followed the same pattern as ApoE: I in diffuse-like plaques was also not different between CAA Frontiers in Neurology | www.frontiersin.org 5 March 2019 | Volume 10 | Article 187 Camacho et al. Apolipoproteins in Cerebral Amyloid Angiopathy FIGURE 2 | Representative images of brain ApoE, ApoJ, and ApoA-I immunopositivity. Serial sections were stained with anti-ApoE (A,D,G,J), anti-ApoJ (B,E,H,K), and anti-ApoA-1 antibodies (C,F,I,L). Detection of these proteins were found in meningeal vessels (A–C), cortical vessels (D–F), cortical capillaries (G–I), and diffuse-like (asterisks) and neuritic-like plaques (arrows) (J–L). Meningeal and cortical vessels: 100x magnification. Neuritic plaques and capillaries: 400x magnification. type I and type II cases (Figure 5B, Supplemental Table 2). presence was evaluated far from the acute hematoma in ICH APOE genotype was not associated with the presence of ApoE, cases. In this regard, ApoE was significantly higher in capillaries ApoJ, and ApoA-I forming neuritic or diffuse-like parenchymal and in meningeal and cortical arteries in those patients who deposits in our cohort. suffered an ICH (p = 0.035, p = 0.048, and p = 0.015, ApoE, ApoJ, and ApoA-I were expressed in the cytoplasm respectively) (Figure 6), although ApoE immunodetection was of pyramidal neurons and subpial astrocytes from different not associated with age. In contrast, ApoE levels in diffuse or cortical and perihippocampal areas analyzed. No differences in neuritic-like plaques were not different in patients who suffered the number of positive cells nor in the intensity of the expression an ICH. ApoA-I levels in meningeal arteries were higher in ICH were detected for ApoE, ApoJ, or ApoA-I among the 4 groups patients than in the other subjects (p = 0.006). Finally, ApoJ analyzed or between CAA type I and CAA type II cases. ApoE, parenchymal presence was not statistically associated with ICH ApoJ, and ApoA-I immunopositivity in pyramidal neurons and in this cohort (Figure 6). Nevertheless, after adjusting for age and subpial astrocytes are presented in Supplemental Figure 1. APOE genotype, higher ApoE presence in cortical arteries was the only independent predictor of suffering an ICH (OR, 9.40, Vascular ApoE, ApoJ, and ApoA-I Levels 95% CI [1.05–84.41], p = 0.046). Associated With ICH In the cohort of analyzed postmortem brains, lobar ICH was the DISCUSSION principal cause of death among CAA patients (N = 7, Table 1). Sex was not associated with ICH presence, but subjects who It is still not completely clear why Aβ accumulates in the brain suffered an ICH were older than the others (83 ± 6 vs. 73 ± in certain neuropathological conditions, such as AD and CAA. 11, p = 0.040). We evaluated the possible association between Although certain inherited forms of CAA presenting cerebral the levels of ApoE, ApoJ, and ApoA-I in brain vessels and the hemorrhage are associated with autosomal dominant mutations occurrence of a parenchymal bleeding, although apolipoprotein in APP and other genes [including the cystatin C (CST3), integral Frontiers in Neurology | www.frontiersin.org 6 March 2019 | Volume 10 | Article 187 Camacho et al. Apolipoproteins in Cerebral Amyloid Angiopathy FIGURE 3 | Distributions of ApoE, ApoJ and ApoA-I in meningeal (A) and cortical (B) arteries. 0: No staining; 1: staining in <30% of the vessels; 2: staining in 30–75% of the vessels; 3: staining in >75% of the vessels. Statistical differences were calculated using the Chi Square (χ2) test. FIGURE 4 | Distributions of ApoE, ApoJ, and ApoA-I in cortical capillaries. 0: No staining; 1: focal and granular wall staining (<50% of the capillaries); 2: focal and linear wall staining (<50% of the capillaries); 3: diffuse wall staining (>50% of the capillaries); 4: strong staining. Statistical differences were calculated using the Chi Square (χ2) test. transmembrane protein 2B (ITM2B), transthyretin and gelsolin that tries to explain the sporadic cerebrovascular Aβ deposition genes] in most cases, sporadic CAA does not clearly associate refers to “protein elimination failure” due to a deficit in protein with any genetic risk factor other than the association described aggregate clearance mechanisms (24). The cerebrovasculature is for APOEε4 (10, 23). One of the most accepted hypotheses the main system responsible for the clearance of cerebral Aβ, Frontiers in Neurology | www.frontiersin.org 7 March 2019 | Volume 10 | Article 187 Camacho et al. Apolipoproteins in Cerebral Amyloid Angiopathy FIGURE 5 | Distribution of ApoE, ApoJ and ApoA-I in neuritic plaques (A) or diffuse parenchymal deposits (B). 0: No staining; 1: 1-5/HPF; 2: 5-10/HPF; 3: >10/HPF. Statistical differences were calculated using the Chi Square (χ2) test. which is evacuated through perivascular drainage routes and neurogenesis, mitochondrial function, tau phosphorylation, excluded through the blood-brain barrier (BBB). The fibrillar and neuronal atrophy, neuroinflammation, and the metabolism and compact Aβ deposition in the brain vessel walls has been related aggregation of Aβ (26). Previous immunohistochemical studies to abnormalities in the vasculature, which can subsequently in AD brains revealed the presence of ApoE in parenchymal compromise the integrity of the BBB. Indeed, the most relevant deposits and in the wall of brain vessels, where the ApoE level pathological consequence of CAA is the development of ICHs correlated with Aβ deposits (8, 27, 28). Our results confirm some affecting cortical and subcortical areas. Furthermore, recurrence of these previous findings, as we detected higher parenchymal is a common complication in CAA patients, which results ApoE levels in brains where a previous presence of Aβ, forming in elevated mortality and disability rates (25). Nevertheless, neuritic plaques, was confirmed. However, no significant the mechanisms of vessel rupture triggering ICH due to Aβ differences were found in the ApoE level in meningeal arteries deposition have not been clarified either. In this regard, our affected or not by CAA, although a tendency toward a higher aim was to study the presence of factors that codeposit with ApoE presence was observed in cortical vessels from CAA brains Aβ in brains showing different Aβ localization patterns. In than those from non-CAA brains. particular, we analyzed specific apolipoproteins, ApoE, ApoJ, and Several evidences have demonstrated that APOE genotype ApoA-I, which may influence the insoluble deposit aggregation modifies the distribution of cerebrovascular Aβ in AD individuals and clearance (13–16) and therefore can be relevant in the (23, 29). In this regard, we studied whether the APOE genotype distribution of Aβ in the brain. could also influence the expression of ApoE, ApoJ, and ApoA-I Brain ApoE is produced by glia (astrocytes and microglia), in CAA cases. Although APOEε4 is a well-known risk factor for pericytes, smooth muscle cells, and, to a much lesser extent, CAA (10, 23), we could not show this association in our study, by neurons under certain stress conditions and, potentially, in probably because of the small sample size of the control group brain endothelial cells. ApoE modulates multiple mechanistic (cases without Aβ pathology). However, among CAA cases, we pathways that affect cognition, including cholesterol/lipid found that APOEε4 genotype was more represented in CAA homeostasis, synaptic function, glucose metabolism, type I patients, whereas CAA type II patients seemed to present Frontiers in Neurology | www.frontiersin.org 8 March 2019 | Volume 10 | Article 187 Camacho et al. Apolipoproteins in Cerebral Amyloid Angiopathy FIGURE 6 | Distribution of ApoE, ApoJ, and ApoA-I in capillaries and meningeal and cortical arteries in brains from subjects who had an intracerebral hemorrhage (ICH) or not. Graphs represent the percentage of samples in each group (presenting ICH or not). The presence in brain capillaries was classified as 0: no staining; 1: focal and granular wall staining (<50% of the capillaries); 2: focal and linear wall staining (<50% of the capillaries); 3: diffuse wall staining (>50% of the capillaries); 4: strong staining. The presence in brain arteries was classified as 0: no staining; 1: staining in <30% of the vessels; 2: staining in 30–75% of the vessels; 3: staining in >75% of the vessels. Statistical differences were calculated using the Chi Square (χ2) test. higher APOEε2 frequency, in accordance with previous studies between APOE ε2 genotype and higher presence of ApoE in (1, 30). No significant association could be found between the cortical arteries. In fact, in CAA-associated ICH patients, our immunodetection of ApoE, ApoJ, or ApoA-I and the APOE group previously showed higher plasma ApoE protein levels in genotype, except for a statistical trend pointing toward a link APOEε2 carriers (31). On the other hand, the presence of either Frontiers in Neurology | www.frontiersin.org 9 March 2019 | Volume 10 | Article 187 Camacho et al. Apolipoproteins in Cerebral Amyloid Angiopathy APOEε4 or APOEε2 have been described to increase the risk of Another important aim of our study was to determine whether lobar ICH (32). Previous pathological studies demonstrated that these codeposition factors were present when cases were classified APOEε2 may promote the CAA-related vessel breakdown (33), according to Aβ capillary involvement. Indeed, the clinical whereas APOEε4 was associated with the amount of vascular relevance of the CAA neuropathological classification is under amyloid (10). Interestingly, in our cohort, CAA patients with investigation. For instance, one study found that CAA type ICH more commonly were APOEε2 carriers compared with CAA I seems to be associated with the severity of CAA, dementia patients without ICH, validating the results described previously and severe AD-type neuropathology (46), while another study (34). In addition to this, the presence of ApoE in cortical vessels reported that the associations of CAA with cognitive outcomes was significantly higher in patients who suffered an ICH, after were not driven by the presence of capillary involvement (47). adjusting for APOE genotype and age. This data suggests the In our study, CAA capillary involvement was not associated possible role of this apolipoprotein in the destabilization and with the CAA or AD pathological severity. CAA subtype was disruption of brain vessels, as proposed in different experimental not related to ICH occurrence either. On the other hand, models (35, 36). the biological pattern of both CAA entities has not been ApoJ is a multifunctional heterodimeric protein that acts as studied extensively (48). In this context, the expression of these a natural chaperone. ApoJ binds to Aβ in CSF and plasma apolipoproteins in cerebral vessels affected by Aβ deposition was (37), and increased plasma ApoJ has been associated with a previously described (27, 28), although no specific reports have higher prevalence and severity of AD (38). In AD brains, described a possible differential distribution pattern depending ApoJ codeposits with Aβ in cerebrovascular and parenchymal on the CAA pathological classification. Therefore, we determined lesions (27, 39, 40). Once again, the results from our study the levels and localizations of ApoE, ApoJ, and ApoA-I in are in accordance with the published data, since we clearly brains classified as CAA type I or type II. We found that the confirmed higher ApoJ levels in brain vessels from CAA cases presence of the three apolipoproteins in capillaries from CAA and in parenchymal neuritic-like and diffuse-like deposits in cases was more evident than in those from non-CAA cases. cases presenting Aβ plaques. However, it is worth mentioning Furthermore, the endothelial expression levels of ApoE and that ApoJ and ApoE diffuse-like deposits were also found in ApoJ, as well as those of ApoA-I, were higher in the CAA some control cases, where no deposits of Aβ were detected type I capillaries than in the CAA type II capillaries. In fact, in parenchyma. The use of more sensitive immunodetection endothelial expression with strong and thick wall staining and, methods (for example, using Aβ antibodies that recognize sometimes, perivascular deposits was found in CAA type I different epitopes) would be required to confirm these results. capillaries, while only endothelial expression without significant The increased levels of ApoJ has been postulated as a protective wall staining was found in CAA type II capillaries. Interestingly, response against the aberrant accumulation of Aβ in the brain. our results highlight the relevant specific association between However, there are conflicting results in the literature regarding ApoA-I presence and capillary Aβ, although whether this ApoA- the protective role of ApoJ. Knocking out CLU in a transgenic I has a peripheral origin and is attracted by the microvascular mouse model of AD resulted in an exacerbation of vascular Aβ Aβ or this is a specific expression mechanism in endothelial deposition, whereas parenchymal Aβ deposition was remarkably cells needs to be elucidated. Furthermore, higher levels of ApoE reduced (41). In another study, intraventricular administration of and ApoA-I were also detected in capillaries from patients who an ApoJ-mimetic peptide promoted a reduction in parenchymal suffered an ICH, independently of the CAA subtype, suggesting and vascular Aβ accumulation and improved cognitive function the possible involvement of the accumulation of these two in an experimental AD model (42). Thus, the impact of the apolipoproteins in the alteration of some properties associated to alteration of the ApoJ level in the brain still needs to be clarified. the BBB integrity. In our study, in contrast to ApoE, no significant association was In summary, ApoE and ApoJ appear to be involved in both found between ApoJ cerebrovascular levels and the occurrence vascular and parenchymal Aβ pathology, while ApoA-I seems of ICH. to be mainly associated with CAA, especially in CAA type I. Finally, ApoA-I is a major constituent of high-density In addition, ApoE presence in cortical arteries appears as an lipoprotein (HDL) in plasma and plays a crucial role in reverse independent factor for the occurrence of ICH in CAA. Our study cholesterol transport. It is produced mainly in the liver and presents important limitations, such as the small sample size of intestine, but ApoA-I is considered to be able to cross the BBB the subgroups (principally evident for the non-CAA groups) or (43) because it has been found in CSF and throughout the brain the analysis of different cortical areas, which may have influenced (44), although the mechanisms behind this trafficking are not the apolipoprotein immunodetection rating. Further studies are well-described. In fact, in AD brains, ApoA-I has been associated needed to evaluate the impact of the co-deposition of these with senile plaques (8, 27), although it is not clear whether apolipoproteins in Aβ-related pathologies and in the occurrence this plaque-associated ApoA-I comes from the periphery or of CAA-associated ICH. whether it is expressed by specific brain cells, such as endothelial capillary cells (45). In the present study, ApoA-I was also found AUTHOR CONTRIBUTIONS in occasional brain parenchymal deposits. However, ApoA-I positivity was less diffuse than ApoE and ApoJ in meningeal and MH-G and EM-S designed and coordinated the study. JC cortical arteries, and there was no association between ApoA-I participated in the design of the study. SR advised on levels and Aβ arterial deposition. the theoretical aspects of the study. TM performed the Frontiers in Neurology | www.frontiersin.org 10 March 2019 | Volume 10 | Article 187 Camacho et al. Apolipoproteins in Cerebral Amyloid Angiopathy immunohistochemistry analysis. JC and EM-S selected brain co-financed by the European Regional Development Fund samples and revised all the immunohistochemistry analysis until (FEDER). The Neurovascular Research Laboratory is part of agreement was reached. AB-P performed the APOE genotype the INVICTUS+ network, ISCIII, Spain [RD16/0019/0021]. analysis from paraffin-embedded sections. JC, EM-S, and MH-G MH-G is supported by the Miguel Servet programme, ISCIII, wrote the manuscript. All authors read and approved the Spain [CPII17/00010]. final manuscript. SUPPLEMENTARY MATERIAL FUNDING The Supplementary Material for this article can be found This work has been funded by Instituto de Salud Carlos online at: https://www.frontiersin.org/articles/10.3389/fneur. III (ISCIII) [CP12/03259, PI14/01134, and PI17/00275], 2019.00187/full#supplementary-material REFERENCES and apolipoproteins E and J in the mouse central nervous system. J Cereb Blood Flow Metab. (2007). 27:909–918. doi: 10.1038/sj.jcbfm.9600419 1. Thal DR, Ghebremedhin E, Rüb U, Yamaguchi H, Del Tredici K, Braak H. 16. Koldamova RP, Lefterov IM, Lefterova MI, Lazo JS. Apolipoprotein Two types of sporadic cerebral amyloid angiopathy. J Neuropathol Exp Neurol. A-I directly interacts with amyloid precursor protein and inhibits (2002) 61:282–93. doi: 10.1093/jnen/61.3.282 A beta aggregation and toxicity. Biochemistry. (2001) 40:3553–60. 2. Charidimou A, Boulouis G, Gurol ME, Ayata C, Bacskai BJ, Frosch MP, et al. doi: 10.1021/bi002186k Emerging concepts in sporadic cerebral amyloid angiopathy. Brain. (2017) 17. Paula-Lima AC, Tricerri MA, Brito-Moreira J, Bomfim TR, Oliveira FF, 140:1829–50. doi: 10.1093/brain/awx047 Magdesian MH, et al. Human apolipoprotein A-I binds amyloid-beta 3. Arvanitakis Z, Leurgans SE, Wang Z, Wilson RS, Bennett DA, Schneider and prevents Abeta-induced neurotoxicity. Int J Biochem Cell Biol. (2009) JA. Cerebral amyloid angiopathy pathology and cognitive domains in older 41:1361–70. doi: 10.1016/j.biocel.2008.12.003 persons. Ann Neurol. (2011) 69:320–7. doi: 10.1002/ana.22112 18. Merino-Zamorano C, Fernández-de Retana S, Montañola A, Batlle 4. Charidimou A, Gang Q, Werring DJ. Sporadic cerebral amyloid angiopathy A, Saint-Pol J, Mysiorek C, et al. Modulation of Amyloid-β1-40 revisited: recent insights into pathophysiology and clinical spectrum. J Neurol Transport by ApoA1 and ApoJ across an in vitro model of the blood- Neurosurg Psychiatry. (2012) 83:124–37. doi: 10.1136/jnnp-2011-301308 brain barrier. J Alzheimers Dis. (2016) 53:677–91. doi: 10.3233/JAD- 5. Zhang-Nunes SX, Maat-Schieman MLC, van Duinen SG, Roos 150976 RAC, Frosch MP, Greenberg SM. The cerebral beta-amyloid 19. Vonsattel JP, Myers RH, Hedley-Whyte ET, Ropper AH, Bird ED, angiopathies: hereditary and sporadic. Brain Pathol. (2006) 16:30–9. Richardson EP. Cerebral amyloid angiopathy without and with cerebral doi: 10.1111/j.1750-3639.2006.tb00559.x hemorrhages: a comparative histological study. Ann Neurol. (1991) 30:637–49. 6. Uro-Coste E, Russano de Paiva G, Guilbeau-Frugier C, Sastre N, Ousset PJ, doi: 10.1002/ana.410300503 da Silva NA, et al. Cerebral amyloid angiopathy and microhemorrhages after 20. Greenberg SM, Vonsattel JP. Diagnosis of cerebral amyloid angiopathy. amyloid beta vaccination: case report and brief review. Clin Neuropathol. Sensitivity and specificity of cortical biopsy. Stroke. (1997) 28:1418–22. (2010) 29:209–16. doi: 10.5414/NPP29209 doi: 10.1161/01.STR.28.7.1418 7. Cho HS, Hyman BT, Greenberg SM, Rebeck GW. Quantitation of 21. Braak H, Alafuzoff I, Arzberger T, Kretzschmar H, Del Tredici K. Staging apoE domains in Alzheimer disease brain suggests a role for apoE of Alzheimer disease-associated neurofibrillary pathology using paraffin in Abeta aggregation. J Neuropathol Exp Neurol. (2001) 60:342–9. sections and immunocytochemistry. Acta Neuropathol. (2006) 112:389–340. doi: 10.1093/jnen/60.4.342 doi: 10.1007/s00401-006-0127-z 8. Wisniewski T, Golabek AA, Kida E, Wisniewski KE, Frangione B. 22. Mirra SS, Heyman A, McKeel D, Sumi SM, Crain BJ, Brownlee LM, et al. Conformational mimicry in Alzheimer’s disease. Role of apolipoproteins in The consortium to establish a registry for Alzheimer’s disease (CERAD). Part amyloidogenesis. Am J Pathol. (1995) 147:238–44. II Standardization of the neuropathologic assessment of Alzheimer’s disease. 9. Saunders AM, Schmader K, Breitner JC, Benson MD, Brown WT, Goldfarb L, Neurology. (1991) 41:479–86. et al. Apolipoprotein E epsilon 4 allele distributions in late-onset Alzheimer’s 23. Trembath D, Ervin JF, Broom L, Szymanski M, Welsh-Bohmer K, Pieper disease and in other amyloid-forming diseases. Lancet. (1993) 342:710–1. C, et al. The distribution of cerebrovascular amyloid in Alzheimer’s doi: 10.1016/0140-6736%2893%2991709-U disease varies with ApoE genotype. Acta Neuropathol. (2007) 113:23–31. 10. Greenberg SM, Briggs ME, Hyman BT, Kokoris GJ, Takis C, Kanter DS, et al. doi: 10.1007/s00401-006-0162-9 Apolipoprotein E epsilon 4 is associated with the presence and earlier onset 24. Carare RO, Hawkes CA, Jeffrey M, Kalaria RN, Weller RO. Review: cerebral of hemorrhage in cerebral amyloid angiopathy. Stroke. (1996) 27:1333–7. amyloid angiopathy, prion angiopathy, CADASIL and the spectrum of doi: 10.1161/01.STR.27.8.1333 protein elimination failure angiopathies (PEFA) in neurodegenerative disease 11. Lambert JC, Heath S, Even G, Campion D, Sleegers K, Hiltunen M, et al. with a focus on therapy. Neuropathol Appl Neurobiol. (2013) 39:593–611. Genome-wide association study identifies variants at CLU and CR1 associated doi: 10.1111/nan.12042 with Alzheimer’s disease. Nat Genet. (2009) 41:1094–9. doi: 10.1038/ng.439 25. Greenberg SM, Gurol ME, Rosand J, Smith EE. Amyloid angiopathy- 12. Tan L, Wang HF, Tan MS, Tan CC, Zhu XC, Miao D, et al. Effect of CLU related vascular cognitive impairment. Stroke. (2004) 35(Suppl. 1):2616–9. genetic variants on cerebrospinal fluid and neuroimaging markers in healthy, doi: 10.1161/01.STR.0000143224.36527.44 mild cognitive impairment and Alzheimer’s disease cohorts. Sci Rep. (2016) 26. Tai LM, Thomas R, Marottoli FM, Koster KP, Kanekiyo T, Morris AWJ, et al. 6:26027. doi: 10.1038/srep26027 The role of APOE in cerebrovascular dysfunction. Acta Neuropathol. (2016) 13. Matsubara E, Frangione B, Ghiso J. Characterization of apolipoprotein J- 131:709–23. doi: 10.1007/s00401-016-1547-z Alzheimer’s A beta interaction. J Biol Chem. (1995) 270:7563–7. 27. Harr SD, Uint L, Hollister R, Hyman BT, Mendez AJ. Brain expression 14. Martel CL, Mackic JB, Matsubara E, Governale S, Miguel C, Miao W, et al. of apolipoproteins E, J, and A-I in Alzheimer’s disease. J Neurochem. Isoform-specific effects of apolipoproteins E2, E3, and E4 on cerebral capillary (1996) 66:2429–35. sequestration and blood-brain barrier transport of circulating Alzheimer’s 28. Namba Y, Tomonaga M, Kawasaki H, Otomo E, Ikeda K. Apolipoprotein E amyloid beta. J Neurochem. (1997) 69:1995–2004. immunoreactivity in cerebral amyloid deposits and neurofibrillary tangles in 15. Bell RD, Sagare AP, Friedman AE, Bedi GS, Holtzman DM, Deane R, et al. Alzheimer’s disease and kuru plaque amyloid in Creutzfeldt-Jakob disease. Transport pathways for clearance of human Alzheimer’s amyloid beta-peptide Brain Res. (1991) 541:163–6. Frontiers in Neurology | www.frontiersin.org 11 March 2019 | Volume 10 | Article 187 Camacho et al. Apolipoproteins in Cerebral Amyloid Angiopathy 29. Love S, Chalmers K, Ince P, Esiri M, Attems J, Jellinger K, et al. Development, of metalloproteinases 3 increase in leptomeningeal arteries affected by appraisal, validation and implementation of a consensus protocol for the cerebral amyloid angiopathy. Neuropathol Appl Neurobiol. (2017) 43:492–504. assessment of cerebral amyloid angiopathy in post-mortem brain tissue. Am J doi: 10.1111/nan.12342 Neurodegener Dis. (2014) 3:19–32. 41. Wojtas AM, Kang SS, Olley BM, Gatherer M, Shinohara M, 30. Thal DR, Papassotiropoulos A, Saido TC, Griffin WS, Mrak RE, Kölsch Lozano PA, et al. Loss of clusterin shifts amyloid deposition to the H, et al. Capillary cerebral amyloid angiopathy identifies a distinct cerebrovasculature via disruption of perivascular drainage pathways. APOE epsilon4-associated subtype of sporadic Alzheimer’s disease. Acta Proc Natl Acad Sci USA. (2017) 114:E6962–71. doi: 10.1073/pnas.17011 Neuropathol. (2010) 120:169–83. doi: 10.1007/s00401-010-0707-9 37114 31. Montañola A, de Retana SF, López-Rueda A, Merino-Zamorano C, Penalba 42. Qi XM, Wang C, Chu XK, Li G, Ma JF. Intraventricular infusion of clusterin A, Fernández-Álvarez P, et al. ApoA1, ApoJ and ApoE plasma levels and ameliorated cognition and pathology in Tg(6799). BMC Neurosci. (2018) 19:2. genotype frequencies in cerebral amyloid angiopathy. Neuromol Med. (2016) doi: 10.1186/s12868-018-0402-7 18:99–108. doi: 10.1007/s12017-015-8381-7 43. Balazs Z, Panzenboeck U, Hammer A, Sovic A, Quehenberger O, Malle 32. Woo D, Sauerbeck LR, Kissela BM, Khoury JC, Szaflarski JP, Gebel J, E, et al. Uptake and transport of high-density lipoprotein (HDL) and et al. Genetic and environmental risk factors for intracerebral hemorrhage: HDL-associated alpha-tocopherol by an in vitro blood-brain barrier preliminary results of a population-based study. Stroke. (2002) 33:1190–5. model. J Neurochem. (2004) 89:939–50. doi: 10.1111/j.1471-4159.2004. doi: 10.1161/01.STR.0000014774.88027.22 02373.x 33. Nicoll JA, Burnett C, Love S, Graham DI, Dewar D, Ironside JW, 44. Dietschy JM, Turley SD. Cholesterol metabolism in the brain. Curr Opin et al. High frequency of apolipoprotein E epsilon 2 allele in hemorrhage Lipidol. (2001) 12:105–12. doi: 10.1097/00041433-200104000-00003 due to cerebral amyloid angiopathy. Ann Neurol. (1997) 41:716–21. 45. Möckel B, Zinke H, Flach R, Weiss B, Weiler-Güttler H, Gassen doi: 10.1002/ana.410410607 HG. Expression of apolipoprotein A-I in porcine brain endothelium 34. Charidimou A, Martinez-Ramirez S, Shoamanesh A, Oliveira-Filho J, Frosch in vitro. J Neurochem. (1994) 62:788–98. doi: 10.1046/j.1471-4159.1994.6202 M, Vashkevich A, et al. Cerebral amyloid angiopathy with and without 0788.x hemorrhage: evidence for different disease phenotypes. Neurology. (2015) 46. Mäkelä M, Paetau A, Polvikoski T, Myllykangas L, Tanskanen M. Capillary 84:1206–12. doi: 10.1212/WNL.0000000000001398 amyloid-β protein deposition in a population-based study (Vantaa 85+). J 35. Teng Z, Guo Z, Zhong J, Cheng C, Huang Z, Wu Y, et al. ApoE influences the Alzheimers Dis. (2016) 43:492–504. doi: 10.3233/JAD-150241 blood-brain barrier through the NF-κB/MMP-9 pathway after traumatic brain 47. Boyle PA, Yu L, Nag S, Leurgans S, Wilson RS, Bennett DA, et al. Cerebral injury. Sci Rep. (2017) 7:6649. doi: 10.1038/s41598-017-06932-3 amyloid angiopathy and cognitive outcomes in community-based older 36. Pang J, Chen Y, Kuai L, Yang P, Peng J, Wu Y, et al. Inhibition of blood-brain persons. Neurology. (2015) 85:1930–6. doi: 10.1212/WNL.0000000000002175 barrier disruption by an apolipoprotein E-mimetic peptide ameliorates early 48. Matsuo K, Shindo A, Niwa A, Tabei, KI, Akatsu H, Hashizume Y, et al. brain injury in experimental subarachnoid hemorrhage. Transl Stroke Res. Complement activation in capillary cerebral amyloid angiopathy. Dement (2017) 8:257–72. doi: 10.1007/s12975-016-0507-1 Geriatr Cogn Disord. (2017) 44:343–53. doi: 10.1159/000486091 37. Ghiso J, Matsubara E, Koudinov A, Choi-Miura NH, Tomita M, Wisniewski T, et al. The cerebrospinal-fluid soluble form of Alzheimer’s amyloid beta is Conflict of Interest Statement: The authors declare that the research was complexed to SP-40,40 (apolipoprotein J), an inhibitor of the complement conducted in the absence of any commercial or financial relationships that could membrane-attack complex. Biochem J. (1993) 293(Pt 1):27–30. be construed as a potential conflict of interest. 38. Schrijvers EMC, Koudstaal PJ, Hofman A, Breteler MMB. Plasma clusterin and the risk of Alzheimer disease. JAMA. (2011) 305:1322–6. Copyright © 2019 Camacho, Moliné, Bonaterra-Pastra, Ramón y Cajal, Martínez- doi: 10.1001/jama.2011.381 Sáez and Hernández-Guillamon. This is an open-access article distributed under the 39. McGeer PL, Kawamata T, Walker DG. Distribution of clusterin terms of the Creative Commons Attribution License (CC BY). The use, distribution in Alzheimer brain tissue. Brain Res. (1992) 579:337–41. or reproduction in other forums is permitted, provided the original author(s) and doi: 10.1016/0006-8993(92)90071-G the copyright owner(s) are credited and that the original publication in this journal 40. Manousopoulou A, Gatherer M, Smith C, Nicoll JAR, Woelk CH, Johnson is cited, in accordance with accepted academic practice. No use, distribution or M, et al. Systems proteomic analysis reveals that clusterin and tissue inhibitor reproduction is permitted which does not comply with these terms. Frontiers in Neurology | www.frontiersin.org 12 March 2019 | Volume 10 | Article 187