TYPE Original Research PUBLISHED 11 April 2023 DOI 10.3389/fnagi.2023.1134399 Association of candidate genetic OPEN ACCESS variants and circulating levels of EDITED BY Taher Darreh-Shori, Karolinska Institutet (KI), Sweden ApoE/ApoJ with common REVIEWED BY Satoshi Saito, neuroimaging features of cerebral National Cerebral and Cardiovascular Center, Japan Lieke Jäkel, amyloid angiopathy Radboud University Medical Center, Netherlands Anna Bonaterra-Pastra1 , Sònia Benítez2,3 , Olalla Pancorbo4 , *CORRESPONDENCE Mar Hernández-Guillamon David Rodríguez-Luna4 , Carla Vert5 , Alex Rovira5 , M. Mar Freijo6 , mar.hernandez.guillamon@vhir.org Silvia Tur7 , Maite Martínez-Zabaleta8 , Pere Cardona Portela9 , Jose Luis Sánchez-Quesada jsanchezq@santpau.cat Rocío Vera10 , Lucia Lebrato-Hernández11 , Juan F. Arenillas12,13 , SPECIALTY SECTION Soledad Pérez-Sánchez14 , Ana Domínguez-Mayoral14 , This article was submitted to Alzheimer’s Disease and Related Dementias, Joan Martí Fàbregas15 , Gerard Mauri16 , Joan Montaner1,17,18 , a section of the journal Jose Luis Sánchez-Quesada2,3* and Frontiers in Aging Neuroscience RECEIVED 30 December 2022 Mar Hernández-Guillamon1* ACCEPTED 27 March 2023 1 Neurovascular Research Laboratory, Vall d’Hebron Research Institute, Universitat Autònoma PUBLISHED 11 April 2023 de Barcelona, Barcelona, Spain, 2 Cardiovascular Biochemistry Group, Research Institute of the Hospital CITATION de Sant Pau (IIB Sant Pau), Barcelona, Spain, 3 Center for Biomedical Research Network on Diabetes Bonaterra-Pastra A, Benítez S, Pancorbo O, and Associated Metabolic Diseases (CIBERDEM), Instituto de Salud Carlos III (ISCIII), Madrid, Spain, 4 Rodríguez-Luna D, Vert C, Rovira A, Freijo MM, Stroke Research Group, Vall d’Hebron Research Institute, Barcelona, Spain, 5 Section of Neuroradiology, Tur S, Martínez-Zabaleta M, Cardona Portela P, Department of Radiology, Vall d’Hebron University Hospital, Vall d’Hebron Research Institute, Universitat Vera R, Lebrato-Hernández L, Arenillas JF, Autònoma de Barcelona, Barcelona, Spain, 6 Neurovascular Group, BioCruces Health Research Institute, Pérez-Sánchez S, Domínguez-Mayoral A, Barakaldo, Spain, 7 Department of Neurology, Son Espases University Hospital, Balearic Islands, Spain, 8 Fàbregas JM, Mauri G, Montaner J, Department of Neurology, Donostia University Hospital, San Sebastián, Spain, 9 Department Sánchez-Quesada JL and of Neurology, Bellvitge University Hospital, L’Hospitalet de Llobregat, Spain, 10 Stroke Unit, Department Hernández-Guillamon M (2023) Association of Neurology, Ramón y Cajal University Hospital, Madrid, Spain, 11 Stroke Unit, Department of Neurology of candidate genetic variants and circulating and Neurophysiology, Virgen del Rocío University Hospital, Seville, Spain, 12 Stroke Program, Department levels of ApoE/ApoJ with common of Neurology, Hospital Clínico Universitario, Valladolid, Spain, 13 Clinical Neurosciences Research Group, neuroimaging features of cerebral amyloid Department of Medicine, University of Valladolid, Valladolid, Spain, 14 Department of Neurology, Virgen angiopathy. Macarena University Hospital, Seville, Spain, 15 Stroke Unit, Department of Neurology, Hospital de la Front. Aging Neurosci. 15:1134399. Santa Creu i Sant Pau, Barcelona, Spain, 16 Stroke Unit, Department of Neurology, Hospital Universitari doi: 10.3389/fnagi.2023.1134399 Arnau de Vilanova de Lleida, Lleida, Spain, 17 Stroke Research Program, Institute of Biomedicine of Seville (IBiS), Virgen del Rocío University Hospital, University of Seville, Seville, Spain, 18 Department COPYRIGHT of Neurology, Virgen Macarena University Hospital, Seville, Spain © 2023 Bonaterra-Pastra, Benítez, Pancorbo, Rodríguez-Luna, Vert, Rovira, Freijo, Tur, Martínez-Zabaleta, Cardona Portela, Vera, Lebrato-Hernández, Arenillas, Pérez-Sánchez, Introduction: Cerebral amyloid angiopathy (CAA) is characterized by the Domínguez-Mayoral, Fàbregas, Mauri, accumulation of amyloid-β (Aβ) in brain vessels and is a main cause of lobar Montaner, Sánchez-Quesada and Hernández-Guillamon. This is an open-access intracerebral hemorrhage (ICH) in the elderly. CAA is associated with magnetic article distributed under the terms of the resonance imaging (MRI) markers of small vessel disease (SVD). Since Aβ is also Creative Commons Attribution License accumulated in Alzheimer’s disease (AD) in the brain parenchyma, we aimed to (CC BY). The use, distribution or reproduction in other forums is permitted, provided the study if several single nucleotide polymorphisms (SNPs) previously associated original author(s) and the copyright owner(s) with AD were also associated with CAA pathology. Furthermore, we also studied are credited and that the original publication in this journal is cited, in accordance with the influence of APOE and CLU genetic variants in apolipoprotein E (ApoE) and accepted academic practice. No use, clusterin/apolipoprotein J (ApoJ) circulating levels and their distribution among distribution or reproduction is permitted which does not comply with these terms. lipoproteins. Methods: The study was carried out in a multicentric cohort of 126 patients with lobar ICH and clinical suspicion of CAA. Results: We observed several SNPs associated with CAA neuroimaging MRI markers [cortical superficial siderosis (cSS), enlarged perivascular spaces in the centrum semiovale (CSO-EPVS), lobar cerebral microbleeds (CMB), white matter hyperintensities (WMH), corticosubcortical atrophy and CAA-SVD burden Frontiers in Aging Neuroscience 01 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 score]. Concretely, ABCA7 (rs3764650), CLU (rs9331896 and rs933188), EPHA1 (rs11767557), and TREML2 (rs3747742) were significantly associated with a CAA- SVD burden score. Regarding circulating levels of apolipoproteins, protective AD SNPs of CLU [rs11136000 (T) and rs9331896 (C)] were significantly associated with higher HDL ApoJ content in the lobar ICH cohort. APOEε2 carriers presented higher plasma and LDL-associated ApoE levels whereas APOEε4 carriers presented lower plasma ApoE levels. Additionally, we observed that lower circulating ApoJ and ApoE levels were significantly associated with CAA-related MRI markers. More specifically, lower LDL-associated ApoJ and plasma and HDL- associated ApoE levels were significantly associated with CSO-EPVS, lower ApoJ content in HDL with brain atrophy and lower ApoE content in LDL with the extent of cSS. Discussion: This study reinforces the relevance of lipid metabolism in CAA and cerebrovascular functionality. We propose that ApoJ and ApoE distribution among lipoproteins may be associated with pathological features related to CAA with higher ApoE and ApoJ levels in HDL possibly enhancing atheroprotective, antioxidative, and anti-inflammatory responses in cerebral β-amyloidosis. KEYWORDS ApoE, ApoJ, CAA, MRI, lipoproteins, EPVS 1. Introduction grouped as cerebral β-amyloidosis diseases. In addition, AD brains present neurofibrillary tangles (NFT) formed by insoluble Cerebral amyloid angiopathy (CAA) is a degenerative small hyperphosphorylated Tau, that together with the presence of Aβ vessel disease (SVD) caused by the deposition of amyloid, plaques induce neurodegenerative processes causing dementia commonly amyloid-β (Aβ), in the wall of the blood vessels of the (Scheltens et al., 2021). Besides the biological overlap between central nervous system, affecting leptomeningeal vessels, arteries AD and CAA, based on Aβ production and clearance processes, and cortical arterioles of medium or small caliber, as well as different factors are described to favor the Aβ accumulation in cerebral capillaries (Charidimou et al., 2017a). The most relevant different brain localizations (Greenberg et al., 2020). For example, pathological consequence of CAA is the presence of intracerebral the length of the peptide is clearly related to its deposition: hemorrhage (ICH) in cortical and subcortical localization with Aβ42 is the main component of amyloid plaques in AD, whereas ICH recurrence being a common complication of this disease Aβ40 is predominantly aggregated in blood vessels (Yamada, (Charidimou et al., 2017d). Cognitive deterioration, dementia and 2015). Apolipoprotein E (ApoE), which has the main function transient focal neurologic episodes (TFNE) are also relevant clinical of regulating lipids transport and distribution in the brain (Bu, manifestations of CAA (Arvanitakis et al., 2011; Smith et al., 2021). 2010), is also a crucial element associated with Aβ brain deposition. Symptomatic CAA is characteristically associated with magnetic This last assumption is based on the fact that the ε4 variant resonance imaging (MRI) markers of small vessel brain injury, of the APOE gene is the main genetic risk factor for AD including hemorrhagic markers as lobar cerebral microbleeds (Schmechel et al., 1993), but also for CAA (Greenberg et al., 1996). (CMB) and cortical superficial siderosis (cSS) (Charidimou et al., Indeed, APOEε4 is related not only to CAA but also to lobar 2017c,d), but also non-hemorrhagic features of SVD, such as ICH and lobar microhemorrhages, characteristic traits in CAA enlarged perivascular spaces in the centrum semiovale (CSO- patients (Inoue et al., 2021). On the other hand, the APOEε2 EPVS) (Charidimou et al., 2017b) and periventricular white matter variant is more frequent in brains with more severe vascular hyperintensities (WMH) (Chojdak-Łukasiewicz et al., 2021). These changes and it is associated with the presence and recurrence imaging features together with clinical characteristics are currently of ICH attributed to CAA (Greenberg et al., 1998). However, used to diagnose patients as possible or probable CAA according while both APOEε2 and APOEε4 alleles are considered risk to the modified Boston criteria (Linn et al., 2010), which have factors for ICH-CAA (Giralt-Steinhauer et al., 2022), APOEε2 is recently been validated and updated (Charidimou et al., 2022). a protective factor for AD (Serrano-Pozo et al., 2021). On the In parallel, a SVD score obtained from some of these common other hand, apolipoprotein J (ApoJ) or clusterin (CLU) coded by CAA MRI features, previously validated and associated with the CLU gene, is the second most expressed apolipoprotein in pathological CAA, has been defined to predict the vascular the brain after ApoE. Both ApoE and ApoJ are important factors affectation degree in this disease (Charidimou et al., 2016; Boulouis in lipid metabolism that can also modulate Aβ aggregation and et al., 2017). deposition in the brain (Huynh et al., 2017; Foster et al., 2019). Pathologically, CAA frequently coexists with Alzheimer’s The distribution of these apolipoproteins among different plasma disease (AD), where Aβ is accumulated in the brain parenchyma lipoproteins can play a role in their functionality (Koch and Jensen, forming insoluble plaques. Therefore, both CAA and AD are 2016; Lanfranco et al., 2020). In this regard, we have previously Frontiers in Aging Neuroscience 02 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 described how the content of ApoE and ApoJ in lipoproteins 2.2. Magnetic resonance imaging differs between CAA-ICH and AD patients (Bonaterra-Pastra et al., evaluation 2021). To further understand the rationale underlying cerebral β- Brain MRI was acquired using a 1.5-T whole-body scanner amyloidosis and the biological relationship between CAA and for most patients (n = 118, 96%) for diagnostic and clinical AD, our study was based on the determination of the frequency purposes in each center at 6.70 ± 16.79 months from the last lobar of different genetic variants previously associated with AD ICH. The following sequences were obtained: axial T2-weighted pathology in a cohort of patients with lobar ICH associated with turbo spin-echo, axial T1-weighted spin-echo, axial T2-weighted CAA. In this sense, candidate single nucleotide polymorphisms turbo fluid-attenuated inversion recovery (FLAIR), and axial T2∗ - (SNPs) were selected for being previously associated with AD weighted echo-planar gradient-echo sequence. All MRI images occurrence and progression in different GWAs studies, including were evaluated in Vall d’Hebron University Hospital (VHUH) genetic variants in APOE, CLU, ABCA7, PICALM, BIN1, SORL1, by the same neuroradiologist, who was blinded to clinical and PTK2B, EPHA1, CD33, CD2AP, MS4A6A, HLA-DRB5/HLA- biological information. DRB1, TREM1, TREM2, and TREML2 genes (Lambert, 2013; Intracerebral hemorrhages were defined as hypointense foci Replogle et al., 2015) which are assumed to be associated with AD >5 mm in diameter on the T2∗ -weighted images, and the number through different pathways (Replogle et al., 2015; Giri et al., 2016). and location were recorded. cSS was defined as the presence of thin In particular, we focused on the potential associations of candidate hypointensity on T2∗ -weighted in a curvilinear pattern following SNPs frequency with the most common MRI markers associated the gyral cortical surface and was evaluated and classified as focal with CAA in the lobar ICH cohort. (restricted to ≤3 sulci) or disseminated (≥4 sulci) according to Furthermore, because of the particularly important role of its distribution and severity as previously described (Linn et al., ApoE and ApoJ in cerebral β-amyloidosis, we analyzed the 2010). cSS contiguous or potentially anatomically connected with association of the levels of total plasma ApoE and ApoJ proteins, any lobar ICH were not included in the aforementioned categories. as well as their content in different circulating lipoproteins, with Perivascular spaces (PVS) also known as Virchow–Robin spaces the genetic variants in APOE and CLU genes related to AD. Finally, were evaluated on axial T2-weighted images. Enlarged perivascular with the aim to improve our understanding of ApoE and ApoJ in spaces (EPVS) in the basal ganglia (BG) or centrum semiovale cerebral β-amyloidosis, the associations between ApoE and ApoJ (CSO) were rated using a previously described validated 4-point levels in plasma and lipoproteins with the most common MRI visual rating scale (0 = no EPVS, 1 ≤ 10 EPVS, 2 = 11–20 EPVS, markers associated with CAA were also analyzed. 3 = 21–40 EPVS, and 4 > 40 EPVS) (Wardlaw et al., 2013). EPVS were also classified as low (≤20 EPVS) or high (≥21 EPVS) degree. The presence of CMB (diameter <5 mm), together with their 2. Materials and methods number and location, was recorded following the Brain Observer MicroBleed Scale (BOMBS) (Cordonnier et al., 2009). Depending 2.1. Study population on the number of lobar microbleeds, patients were classified as having less than 5 lobar CMB or more than 5 (Lobar CMB >5). The population studied in this cohort consisted of 126 WMH were defined as hyperintense signal lesions in T2-FLAIR or participants who had suffered at least one lobar ICH and presented T2-weighted images and were classified as deep or periventricular a clinical suspicion of CAA. Patients were recruited during a follow- depending on their location. Deep and periventricular WMH up visit in 11 different Spanish centers (Bonaterra-Pastra et al., were assessed according to the four-point Fazekas rating scale 2021; Marazuela et al., 2021). All patients were >55 years old (Fazekas et al., 1987). For periventricular WMH: 0 = absence, and those presenting with symptomatic deep ICHs or treated with 1 = “caps” or pencil-thin lining, 2 = smooth “halo,” 3 = irregular anticoagulant therapy were excluded. WMH extending into the deep white matter; and for deep WMH: The data obtained from the cohort included the recruitment 0 = absence, 1 = punctate foci, 2 = beginning confluence of foci, date, date of stroke events, demographic characteristics (sex and 3 = large confluent areas. We further classified WMH burden as age), relevant vascular risk factors [hypertension (HT), diabetes mild (Fazekas: 0, 1) or severe (Fazekas: 2, 3). The WMH score was mellitus (DM), and dyslipidemia (DL)], medication, and cognitive evaluated in the hemisphere not affected by hemorrhage, except in status. Hypertension was defined as systolic BP ≥140 mmHg, cases when both hemispheres were involved. Global brain atrophy diastolic BP ≥90 mmHg, or use of antihypertensive medication. was visually assessed and classified as mild, moderate or severe. DM was defined as fasting glucose levels >7 mmol/L or the The MRI analysis allowed the classification of cases according to use of antidiabetic drugs or insulin. Dyslipidemia was based Boston criteria 1.5 (Linn et al., 2010); 106 patients accomplished on total cholesterol levels above 6.20 mmol/L or the use of a CAA diagnosis, whereas 13 patients presented mixed-pathology lipid-lowering medication. Patients that presented mild cognitive (presenting lobar and deep microbleeds), 7 patients were excluded impairment or dementia were all considered as cognitive impaired from the Boston criteria classification because of missing imaging patients. The study was approved by the Clinical Investigation information, and 1 patient was diagnosed as probable CAA with Ethical Committee of the Vall d’Hebron University Hospital, supporting pathology. In addition, all patients were classified Barcelona, Spain (PR(AG)326/2014) and had the approval of the according to the CAA-SVD burden score (Charidimou et al., 2016). Ethical Committees of all the participating centers. The study was It ranges from 0 to 6 points and it is based on the presence and conducted in accordance with the Helsinki Declaration. grade of MRI markers associated with CAA including CMBs, cSS, Frontiers in Aging Neuroscience 03 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 CSO-EPVS, and WMH. In our study, the CAA-SVD burden was the appropriate quantity of KBr, according to the equation of classified as low (≤3) or high degree (≥4). Radding and Steinberg (1960). Each sequential step consisted of ultracentrifugation at the aforementioned density at 100,000 g for 24 h, at 4◦ C, using 3.2 ml open-top Thickwall Polycarbonate 2.3. Blood collection tubes (Ref. No. 362305, Beckman Coulter). VLDL was collected in 400 µl, and LDL and HDL in 600 µl fractions of the Ten milliliters of blood in EDTA-containing Vacutainer tubes upper phase. ApoE determination was determined in a Cobas (Becton Dickinson, Franklin Lakes, NJ, USA) were collected from 6000/c501 autoanalyzer using reagents from Kamiya Biomedicals each participant at a follow-up visit. Blood was centrifuged at 4◦ C (Cat. No. KAI-007, Seattle, WA, USA), using the parameters for for 15 min at 2,500 rpm and plasma was immediately aliquoted and automatic analyzer provided by the manufacturer. ApoE in both frozen at −80◦ C. plasma and lipoprotein fractions were measured undiluted. ApoJ quantification, in plasma and lipoproteins, was analyzed by ELISA (#3713-1HP-1, Mabtech, Stockholm, Sweden) at VHIR. Before 2.4. Genetic determinations ELISA quantification, lipoprotein samples were diluted to the same cholesterol concentration (VLDL: 0.1 mmol cholesterol/L; LDL: Single nucleotide polymorphisms (SNPs) candidates were 0.4 mmol cholesterol/L; HDL: 0.3 mmol cholesterol/L). Samples genotyped from blood using SNPlexTM at the Spanish National with null values for a particular quantification were discarded from Genotyping Centre (CEGEN) and included rs6656401 and such analysis. rs6701713 of CR1, rs4311 of ACE, rs11136000, rs7012010, rs9331896, and rs9331888 of CLU, rs4147929 and rs3764650 of ABCA7, rs3851179 and rs10792832 of PICALM, rs744373 and rs6733839 of BIN1 gene, rs11767557 and rs11771145 of 2.6. Statistical analysis EPHA1, rs3865444 of CD33, rs9349407 and rs10948363 of CD2AP, rs11218343 of SORL1, rs9271192 of HLA-DRB5-DRB1 region, To study allelic associations, each SNP was dissected into its rs28834970 of PTK2B, rs6910730 of TREM1, rs75932628 of two alleles to perform an allelic test. The association of all the TREM2, and rs3747742 of TREML2, and rs4938933 and rs983392 SNPs and demographic variables with CAA neuroimaging features of MS4A6A and rs429358 and rs7412 of the APOE gene to was studied by raw analysis using contingency tables and a χ2 determine APOEε genotype. The minor allele frequency (MAF) of test using the Pearson or Fisher p-value as needed. To adjust the the SNPs analyzed in our cohort are reported in Supplementary results, a forward LR binary logistic regression using sex, age, and Table 1. The information about MAF in the global population was the significant SNPs for each radiological feature was assessed. The obtained from 1,000 Genome Project phase 3 (Auton et al., 2015). odds ratios (ORs) and 95% confidence intervals (CIs) for the effect on radiological characteristics were estimated using binary logistic regression analysis. This regression was performed as a common 2.5. Plasma ApoE and ApoJ levels dominant model using SNPs as binary categorical variables [not having the minor allele (MA) vs. having the minor allele] or as Apolipoproteins levels determination in plasma and in an additive model (comparing the three different genotypes). For lipoprotein fractions was done in a representative subcohort of the binary logistic regression of the SVD burden, a simplified score 60 lobar ICH patients with clinical suspicion of CAA randomly of low-high SVD burden score was used. In addition, p-values selected. The lipoprotein fractions determined were very low- were adjusted for multiple corrections by Benjamini–Hochberg density lipoproteins (VLDL), low-density lipoproteins (LDL) and (BH) false discovery rate (FDR). The distribution of the continuous high-density lipoproteins (HDL). Blood samples from this group variables was tested using the Kolmogorov–Smirnov test. If the were obtained 10.42 ± 16.77 months after the last ICH and distribution was normal, an independent samples t-test was processed as previously explained. The standard lipid profile was quantified in the Clinical Biochemistry Unit of the VHUH performed, and if the distribution was not normal an independent using a direct HDL-cholesterol method (HDL-C plus) or by samples Mann–Whitney U test. When analyzing the association ultracentrifugation when the triglycerides concentration was higher between apolipoprotein levels and age or ordinal variables with than 3 mmol/L, in agreement with the National Cholesterol more than 2 categories, Spearman correlations were performed. Education Program (National Cholesterol Education Program Data are expressed as the mean ± standard deviation (SD) for [NCEP] Expert Panel on Detection, Evaluation, and Treatment normal distributions or as the median [interquartile range] for of High Blood Cholesterol in Adults [Adult Treatment Panel non-normal distributions. To adjust our results, backward linear III]., 2002). These determinations were performed in an AU regressions were performed for ApoJ and ApoE circulating levels 5800 autoanalyzer (Beckman Coulter, Pasadena, CA, USA) using using the significant variables for each apolipoprotein fraction and reagents from Beckman Coulter. To study the apolipoprotein adjusting by sex and age. The slope of the regression (B) with content in the lipoprotein fractions, lipoproteins were isolated a 95% CI was estimated by linear regressions. A p-value below by flotation sequential ultracentrifugation, using 1 ml of plasma, 0.05 was considered statistically significant. Statistical analyses were according to density: VLDL (1.006–1.019 g/ml), LDL (1.019– conducted using SPSS Statistics version 21 (IBM Corporation, 1.063 g/ml), and HDL (1.063–1.210 g/ml) in the Research Institute Armonk, NY, USA) and R (R Foundation for Statistical Computing, of the Hospital de Sant Pau. Density was adjusted by adding Vienna, Austria). Frontiers in Aging Neuroscience 04 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 3. Results of ABCA7 [rs3764650 (G)], EPHA1 [rs11767557 (C)], and CLU [rs9331896 (C)] resulted to act as potential protective factors as The demographic characteristics of the study population are their frequency significantly decreased at higher CAA-SVD score. displayed in Table 1. All patients presented at least one lobar ICH, 30.2% of patients presented 2 or more lobar ICH and 50.4% TABLE 2 Magnetic resonance imaging characteristics of a lobar ICH presented cognitive impairment. The most common MRI features cohort with clinical suspicion of CAA (N = 118). previously associated with CAA were analyzed and resulted well represented in our cohort (Table 2). From them, 42.4% of patients cSS 50 (42.4%) presented cSS, 70% of whom displayed disseminated cSS. Almost Focal 15 (30%) all patients (92.4%) presented MRI-visible EPVS; 97.25% of them Disseminated 35 (70%) in the BG, mostly with low degree (80%), whereas 80.7% presented EPVS in the CSO, among them 48.9% with high degree. Also, EPVS 109 (92.4%) 61.9% of the population presented CMB. Among them, most of BG-EPVS 106 (97.25%) the patients showed lobar CMBs (94.52%), including 36.23% that Low degree (1–20) 85 (80.2%) presented more than 20 lobar CMBs. Also, nearly the whole cohort High degree (21 to >40) 21 (19.8%) (91.5%) presented WMH, both periventricular (89.8%) and deep (90.7%). Additionally, 37.3% of patients presented brain atrophy. CSO-EPVS 88 (80.7%) A total of 11.1% of the cohort, and 10% of the subcohort for Low degree (1–20) 45 (51.1%) biomarker analysis in plasma, presented mixed pathology. High degree (21 to >40) 43 (48.9%) The main aim of this study was to evaluate whether genetic CMB 73 (61.9%) variants described as AD risk factors were also related to CAA, especially focusing on imaging features associated with this Lobar CMB 69 (94.52%) pathology in patients with lobar ICH. To do so, associations 1–5 25 (36.23%) between the frequency of the candidate SNPs and MRI features 6–10 12 (17.39%) and demographic characteristics were considered through allelic 11–20 7 (10.14%) univariant analysis (Supplementary Table 2). After applying a binary logistic regression adjusted by age and sex, only the >20 25 (36.23%) candidate SNPs that remained significantly associated with specific Deep CMB 14 (19.18%) MRI features are presented in Table 3. We detected that several 1–5 12 (85.7%) CLU SNPs were associated with MRI features associated with 6–10 1 (7.1%) CAA, such as the presence of cSS, CSO-EPVS, lobar CMB, and both deep and periventricular WMH. In addition, specific SNPs in 11–20 1 (7.1%) ABCA7 (rs3764650; p = 0.009), CLU (rs9331888; p = 0.014), EPHA1 >20 0 (0%) (rs11767557; p = 0.010), and TREML2 (rs3747742; p = 0.019) were Cerebellum CMB 15 (20.5%) independently associated with the CAA-SVD burden score when 1–5 12 (80%) the common dominant model was applied (Figure 1). Moreover, the variants in ABCA7 (rs3764650; p = 0.013), CLU (rs9331896; 6–10 3 (20%) p = 0.018), EPHA1 (rs11767557; p = 0.006), and TREML2 11–20 0 (0%) (rs3747742; p = 0.033) were also independently associated with >20 0 (0%) the CAA-SVD burden score when the additive model was applied WMH 108 (91.5%) (Supplementary Figure 1). The minor alleles in the specific SNPs Periventricular 97 (89.8%) TABLE 1 Demographic and clinical characteristics. Caps or pencil/thin lining (1) 24 (24.7%) Smooth halo or irregular 73 (75.3%) Lobar ICH (n = 126) periventricular (2–3) Age, years 76.18 ± 7.11 Deep 98 (90.7%) Sex (females) 68 (54%) Punctate foci (1) 43 (43.9%) HT 73 (60.3%) Beginning or large confluent areas 55 (56.1%) (2–3) DM 18 (15.5%) DL 41 (36.9%) Cortico-subcortical 44 (37.3%) atrophy Cognitive impairment 58 (50.4%) Mild 29 (65.9%) Corticosubcortical ICH 126 (100%) Moderate 13 (29.5%) ≥2 38 (30.2%) Severe 2 (4.5%) ≥3 13 (10.3%) The percentages displayed are within each category. Data are expressed as n (%). cSS, cortical Data are expressed as n (%). ICH, intracerebral hemorrhage; HT, hypertension; DM, diabetes superficial siderosis; EPVS, enlarged perivascular spaces; BG, basal ganglia; CSO, centrum mellitus; DL, dyslipidemia. semiovale; CMB, cerebral microbleeds; WMH, white matter hyperintensity. Frontiers in Aging Neuroscience 05 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 TABLE 3 Binary logistic regression for CAA MRI markers with SNPs minor alleles adjusting by sex and age. Genetic and clinical Common dominant model Additive model variables OR (95% CI) p-Value BH OR (95% CI) p-Value BH p-Value p-Value cSS CLU rs9331888 2.490 0.017 0.029 1.889 0.039 0.041 (G allele, MA) (1.175–5.278) (1.033–3.454) cSS extent CD2AP rs10948363 0.188 0.028 0.040 – – – (G allele, MA) (0.042–0.836) CD2AP rs9349407 – – – 0.150 0.006 0.024 (C allele, MA) (0.039–0.581) CLU rs11136000 6.722 0.013 0.028 8.264 0.009 0.024 (T allele, MA) (1.500–30.127) (1.698–40.000) CSO-EPVS – – – – – – – High degree CLU rs7012010 2.370 0.033 0.043 1.935 0.026 0.033 CSO-EPVS (C allele, MA) (1.073–5.237) (1.082–3.462) BG-EPVS – – – – – – – High degree CD33 rs3865444 – – – 3.333 0.005 0.024 BG-EPVS (A allele, MA) (1.437–7.752) Lobar CMB ABCA7 rs3764650 0.313 0.012 0.028 0.290 0.008 0.024 (G allele, MA) (0.126–0.778) (0.115–0.728) CLU rs7012010 2.185 0.048 0.048 1.853 0.044 0.044 (C allele, MA) (1.006–4.749) (1.016–3.378) Lobar CMB >5 ABCA7 rs3764650 0.323 0.039 0.044 – – – (G allele, MA) (0.110–0.945) CLU rs11136000 – – – 0.503 0.023 0.031 (T allele, MA) (0.278–0.911) TREML2 rs3747742 2.306 0.039 0.044 1.918 0.023 0.031 (C allele, MA) (1.44–5.092) (1.095–3.362) High deep WMH CR1 rs6701713 0.420 0.048 0.048 0.419 0.030 0.036 burden (A allele, MA) (0.177–0.994) (0.191–0.919) CLU rs9331888 3.054 0.005 0.028 2.351 0.009 0.024 (G allele, MA) (1.404–6.642) (1.234–4.478) High periventricular CLU rs9331888 2.673 0.015 0.028 – – – WMH burden (G allele, MA) (1.211–5.897) CLU rs9331896 – – – 0.478 0.012 0.027 (C allele, MA) (0.270–0.848) Atrophy BIN1 rs6733839 0.307 0.002 0.028 0.457 0.019 0.031 (T allele, MA) (0.132–0.715) (0.237–0.881) CD2AP rs10948363 0.257 0.006 0.028 0.359 0.007 0.024 (G allele, MA) (0.111–0.595) (0.170–0.755) CLU rs7012010 – – – 2.101 0.020 0.031 (C allele, MA) (1.125–3.924) CAA-small vessel ABCA7 rs3764650 0.261 0.009 0.028 0.284 0.013 0.027 disease burden score (G allele, MA) (0.096–0.711) (0.105–0.764) CLU rs9331896 – – – 0.470 0.018 0.031 (C allele, MA) (0.251–0.879) CLU rs9331888 2.921 0.014 0.028 – – – (G allele, MA) (1.246–6.851) EPHA1 rs11767557 0.285 0.010 0.028 0.288 0.006 0.024 (C allele, MA) (0.109–0.744) (0.118–0.698) TREML2 rs3747742 2.787 0.019 0.029 1.963 0.033 0.037 (C allele, MA) (1.184–6.650) (1.055–3.652) BH p-value, adjusted Benjamini–Hochberg p-value; cSS, cortical superficial siderosis; EPVS, enlarged perivascular spaces; CSO, centrum semiovale; CMB, cerebral microbleeds; WMH, white matter hyperintensity; OR, odds ratio; CI, confidence interval; MA, minor allele. Frontiers in Aging Neuroscience 06 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 On the other hand, the minor alleles in the specific SNPs of CLU plasma ApoE in patients carrying the ε2 allele (non- [rs9331888 (G)] and TREML2 [rs3747742 (C)] were considered a APOEε2 carriers = 43.16 ± 14.02 mg/L vs. APOEε2 risk factor, since their frequency was significantly correlated with a carriers = 67.04 ± 21.56 mg/L; p < 0.001) and significantly higher CAA-SVD score (Table 3). Interestingly, carriers of APOEε2 lower total plasma ApoE levels associated with the presence of the showed a statistical tendency to present a higher CAA-SVD burden ε4 allele (non-APOEε4 carriers = 48.59 ± 17.17 mg/L vs. APOEε4 score (Supplementary Table 2), however, the association was lost carriers = 36.82 ± 11.73 mg/L; p = 0.024) (Supplementary after the logistic regression. Table 7), as previously described (Montañola et al., 2016; From these results and considering the known role of Marais, 2019). Furthermore, the APOEε2 genotype was also ApoJ/CLU in cerebral β-amyloidosis (Howlett et al., 2013; Beeg significantly associated with higher levels of ApoE in LDL et al., 2016), we next focused on the study of the potential genotypic (non-APOEε2 carriers = 35.07 [17.64–51.61] µmol/mol chol association of the CLU SNPs (rs11136000, rs7012010, rs9331888, vs. APOEε2 carriers = 67.48 [58.79–171.32] µmol/mol chol; and rs9331896) with ApoJ levels in plasma, as well as its distribution p = 0.001) (Supplementary Table 7). The associations between among lipoproteins in samples from a subgroup of 60 patients APOE genotype and ApoE levels in plasma and lipoproteins with clinical suspicion of CAA. First, none of the demographic remained significant when analyzing the subcohort of patients characteristics and main vascular risk factors evaluated were with blood samples obtained chronically (>90 days after ICH) related with ApoJ levels in plasma or its content in VLDL, LDL, (Supplementary Table 8). or HDL lipoproteins (Supplementary Table 3). However, the Interestingly, when analyzing the possible associations between variant rs1113600 was significantly associated with ApoJ content circulating ApoE levels and the presence of the different MRI in HDL [T allele (MA) = 795.87 [586.48–1,065.01] µg/mmol markers, we detected lower content of ApoE in LDL significantly cholesterol (chol) vs. C allele = 614.68 [0–950.47] µg/mmol chol; associated with a higher extent of cSS (focal cSS = 76.19 [48.16– p = 0.012], whereas the association between rs9331896 and HDL- 168.08] vs. disseminated cSS = 27.73 [9.85–60.51] µmol/mol ApoJ content showed a statistical tendency [C allele (MA) = 795.87 chol; p = 0.003) (Figure 4A). This association was maintained [549.54–1,065.01] µg/mmol chol vs. T allele = 644.44 [13.22– when chronic plasma samples (>90 days after ICH) were 961.01] µg/mmol chol; p = 0.056]. In both cases, the minor allele exclusively analyzed (focal cSS = 124.93 [58.22–176.51] vs. [rs1113600 (T) and rs9331896 (C)] was associated with higher ApoJ disseminated cSS = 27.73 [11.32–60.30] µmol/mol chol; p = 0.010). content in HDL (Figure 2). No other associations were detected Also, as observed with the ApoJ protein study, we found between CLU SNPs and circulating ApoJ levels (Supplementary that ApoE content in lipoproteins was associated with the Table 4). These associations remained significant when including presence of visible CSO-EPVS and their degree score. Lower only patients with blood samples obtained >90 days after ICH total and HDL ApoE levels were significantly correlated with (N = 40) (Supplementary Table 5). the presence of CSO-EPVS (absence = 55.61 ± 21.15 mg/L and On the other hand, whereas total plasma ApoJ levels 441.36 ± 267.94 µmol/mol chol vs. presence = 42.92 ± 14.48 mg/L resulted not linked to CAA-MRI markers, several associations and 306.72 ± 192.85 µmol/mol chol; p = 0.014, and p = 0.044, were identified regarding ApoJ distribution among lipoproteins respectively) and with the corresponding degree score (p = 0.045 (Supplementary Table 2). Remarkably, lower ApoJ content in and p = 0.032, respectively) (Figure 4B). Association between LDL was significantly associated with the presence of CSO- ApoE total and HDL levels with CSO-EPVS presence and EPVS (absence = 104.06 [66.49–149.33] µg/mmol chol vs. severity remained significant for chronically obtained samples presence = 60.78 [9.57–117.16] µg/mmol chol; p = 0.033), and (>90 days after ICH) (absence = 57.10 [50.45–69.45] mg/L and with the corresponding EPVS rating scale (p = 0.041) (Figure 3A). 548.65 ± 229.02 µmol/mol chol vs. presence = 38.90 [32.90–49.55] In addition, lower ApoJ content in HDL was significantly mg/L and 301.62 ± 188.00 µmol/mol chol; p = 0.004 and p = 0.002, associated with corticosubcortical atrophy (absence = 756.79 respectively; degree score p = 0.001 and p = 0.001, respectively). [466.97–1,028.20] µg/mmol chol vs. presence = 304.47 [0.00– After a linear regression adjustment considering age and sex to 795.87] µg/mmol chol; p = 0.031) and its severity (p = 0.020) underscore those variables independently associated with ApoE (Figure 3B). After applying a linear regression adjusting by age circulating levels, we found that APOEε2 and ε4 genotypes and sex to underscore those variables independently associated resulted independently associated with total plasma protein levels. with the protein circulating levels, we only found that certain MRI In addition, APOEε2 genotype and cSS severity maintained the features remained significantly associated with ApoJ distribution independent association with LDL ApoE levels. And finally, the in lipoproteins whereas the association with CLU SNPs lost presence of visible CSO-EPVS remained significantly associated signification. On the other hand, CSO-EPVS remained associated with HDL ApoE levels (Figure 4C). with circulating LDL ApoJ content and corticosubcortical atrophy remained associated with HDL ApoJ, independently of CLU SNPs (Figure 3C). If only chronic samples (>90 days after ICH) were analyzed, no associations were found, probably due to a reduced 4. Discussion sample size (n = 40). We next explored the potential alteration of the circulating Detecting the cerebrovascular Aβ pathology is a crucial ApoE levels in the lobar ICH cohort. Although none of the endpoint with diagnostic value and potential for monitoring CAA demographic characteristics or clinical variables analyzed progression. In recent years, the development of positron emission were associated with ApoE concentration in plasma or tomography (PET) radiotracers for Aβ imaging (Farid et al., 2017; with its distribution among lipoproteins (Supplementary Chang et al., 2021) and cerebrospinal fluid (CSF) biomarkers Table 6), we found significantly higher levels of total (Kuiperij et al., 2020) have been an intensive focus of research in Frontiers in Aging Neuroscience 07 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 FIGURE 1 Allelic distribution of the SNPs associated with the CAA-SVD burden score using the common dominant model. The genetic risk factor (presenting or not the minor allele) associated with a higher score is represented in dark gray, whereas the protective factor is represented in light gray. Allelic distribution is expressed as a percentage in each category of the score (0–6). the CAA field, although high costs and the invasiveness of these patients. For this purpose, we tested the MAF of candidate AD- tests make them still limited for the clinical practice. The evolution linked SNPs identified in previous GWAS (Lambert, 2013; Replogle of the vascular pathology could be potentially monitored through et al., 2015) in a multicenter cohort of patients that suffered a lobar the progression of SVD neuroimaging markers strongly associated ICH with a clinical suspicion of CAA. with CAA pathology, although most of these features have Initially, we found that the frequency of minor alleles in certain little specificity (Chojdak-Łukasiewicz et al., 2021). Furthermore, candidate genetic variants in the ICH-CAA cohort did not follow although these MRI markers reflect distinct pathophysiologic the same direction as the results previously described for AD aspects of the disease, biological mechanisms underlying their pathology. This phenomenon is well established for the presence occurrence and topography are still under investigation. In this of the APOEε2 allele, which is a protective genetic variant for AD, regard, the hypothesis of the present study was that candidate whereas it is associated with a higher hemorrhagic load in CAA genes previously associated with different pathological aspects in (Lambert, 2013; Greenberg et al., 2020). In this sense, for instance, AD could also be related, in accordance or in opposition, to the ABCA7 rs3764650 (G) has been described as a risk factor for AD, deposition of Aβ in blood vessels and may shed light on the without a direct influence on CAA (Beecham et al., 2014). However, biological impact of the MRI features commonly found in CAA other ABCA7 SNPs have been associated with both CAA and small Frontiers in Aging Neuroscience 08 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 FIGURE 2 Association of CLU SNPs [rs1136000 (T), rs9331896 (C)] with ApoJ content in HDL. Boxplot representation of ApoJ content in HDL expressed as µg of ApoJ per mmol of cholesterol in HDL according to CLU SNPs alleles. *p < 0.05; # p < 0.1. vessel ischemic disease (Mäkelä et al., 2018; Blumenau et al., 2020). causing a lower burden of neuritic plaques together with a major In our cohort, we observed that ABCA7 SNP rs3764650 (G) acted accumulation of Aβ in the cerebral vessels, which could explain as a protective factor regarding the presence and number of lobar how protective factors for AD could cause risk for CAA and vice CMBs and also resulted to be less frequent in those cases with a versa. However, these genes are also involved in many biological higher CAA-SVD burden score. Also, TREML2 rs3747742 (C), a mechanisms that may differently influence both pathologies, such protective factor for AD (Benitez et al., 2014), acted as a risk factor as Tau pathology or BBB integrity maintenance (Obermeier et al., for the presence of a high number of lobar CMB (>5) and resulted 2013; Shulman et al., 2013; Beecham et al., 2014; Benitez et al., more frequent when increasing the CAA-SVD burden score. 2014). Finally, EPHA1 rs11767557 (C) was significantly associated with Interestingly, we observed that the four CLU SNPs studied lower punctuation in the neuroimaging CAA-SVD burden score. were independently and significantly associated with several This SNP was previously related to a higher risk of AD; although CAA neuroimaging markers. In more detail, rs7012010 (C) and other EPHA1 genetic polymorphisms have also been associated rs9331888 (G) were previously widely described as risk markers with small vessel ischemic disease (Wei et al., 2019; Blumenau for AD (Lambert et al., 2009; Yu and Tan, 2012; Tan et al., 2016; et al., 2020). All these results add evidence to the complexity of Foster et al., 2019). In the present study, we observed significant pathophysiological processes that lead to the development of CAA independent associations of rs7012010 (C) with the degree of CSO- or AD. ABCA7 has been previously associated with Aβ metabolism EPVS and the presence of lobar CMB and brain atrophy; and and neuritic plaque burden in AD (Shulman et al., 2013; Beecham rs9331888 (G) with the presence of cSS, WMH burden and a et al., 2014; Mäkelä et al., 2018). Furthermore, ABCA7, in addition global higher CAA-SVD burden, as features presumably associated to TREML2 and EPHA1, has been proposed to influence Aβ with a higher CAA severity in the lobar ICH population. On clearance (Benitez et al., 2014; Karch and Goate, 2015; Mäkelä the contrary, rs11136000 (T) and rs9331896 (C) were previously et al., 2018). Alterations in Aβ production and clearance may cause described as protective factors for AD (Foster et al., 2019). In a redistribution of Aβ from parenchyma to vasculature, possibly this sense, we observed rs9331896 (C) to be independently and Frontiers in Aging Neuroscience 09 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 FIGURE 3 Association of ApoJ levels with MRI features associated with CAA. ApoJ levels are expressed as µg of ApoJ per mmol of cholesterol in each lipoprotein. (A) Boxplot representation of ApoJ levels in LDL according to EPVS in centrum semiovale (CSO). (B) ApoJ levels in HDL according to the presence and severity of cortico-subcortical atrophy. (C) Linear regression for circulating ApoJ levels using the significant variables for each lipoprotein fraction (CLU SNPs and MRI markers) and adjusting by sex and age. *p < 0.05. significantly associated with a lower burden of periventricular (Montañola et al., 2016). Altogether, in the lobar-ICH cohort, the WMH and CAA-SVD score and rs11136000 (T) with a reduced allelic frequencies of CLU SNPs rs9331896 (C) and rs9331888 (G) number of CMB, in line with a previous study of our group where followed the pattern of potential risk/protective factors previously rs11136000 (T) was associated with the absence of lobar CMBs described in AD. A recent study found several different CLU Frontiers in Aging Neuroscience 10 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 FIGURE 4 Association of ApoE levels with MRI features. ApoE levels are expressed as mg/L in plasma and µmol of ApoE per mol of cholesterol in each lipoprotein. (A) Boxplot distribution of ApoE levels in LDL according to cortical superficial siderosis severity and distribution. (B) Total and HDL ApoE levels according to EPVS in centrum semiovale (CSO). (C) Linear regression for circulating ApoE levels using the significant variables for each lipoprotein fraction (APOE genotypes and MRI markers) and adjusting by sex and age. *p < 0.05. Frontiers in Aging Neuroscience 11 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 haplotypes associated with lobar ICH, suggesting that different compared to controls, but not in the ICH-CAA cohort (Bonaterra- CLU SNPs are clearly involved in ICH and that different SNPs Pastra et al., 2021), suggesting that the content of ApoJ in LDL within the same gene could cause the same phenotypic endpoint could influence the Aβ traffic within the brain. Moreover, it has (Sawyer et al., 2022). Regarding APOE genotype, the presence of been previously described that ApoJ treatment could improve HDL ε2 and ε4 alleles had no impact on the common CAA MRI features, functionality regarding atheroprotective, anti-oxidative and anti- although APOEε2 carriers presented a statistical tendency to higher inflammatory properties (Rivas-Urbina et al., 2020). Also, ApoE CAA-SVD burden score. Previous studies found an association of in HDL could act as a protective factor since AD patients present APOEε4 with strictly lobar CMB (Shams et al., 2022) and APOEε2 HDL with lower ApoE content (Pedrini et al., 2022), and in our with cSS (Charidimou et al., 2015) although these results have previous study, we observed that subjects from an ICH-CAA cohort not been confirmed in other studies (Shoamanesh et al., 2014; presented a higher ratio of ApoE/ApoC-III in HDL than controls Montañola et al., 2016; Shams et al., 2022). which we hypothesized to be a defensive response against vascular Because we found a relevant association of different CLU SNPs Aβ deposition (Bonaterra-Pastra et al., 2021). Higher circulating with the presence of several MRI features associated with CAA, ApoE levels are known to be atheroprotective, and ApoE-enriched we further explored their potential linked with ApoJ peripheral HDL are able to reduce CAA in an in vitro model (Morton et al., levels. In fact, it has been extensively described that certain CLU 2019; Robert et al., 2020). Altogether we propose that higher SNPs may affect ApoJ expression and levels (Szymanski et al., content of ApoJ and ApoE in LDL and HDL may enhance the 2011; Xing et al., 2012; Cai et al., 2016; Tan et al., 2016). None drainage of Aβ from brain vessels, displaying protective properties. of the SNPs analyzed was associated with changes in ApoJ plasma Both ApoE and ApoJ bind Aβ and are established ligands for concentration, but some associations with the distribution of ApoJ different receptors of the LDL receptor family such as LDL receptor, in lipoproteins were observed. In this sense, we detected that the ApoE receptor, VLDL receptor, LRP1, and megalin. Therefore, rs11136000 (T) and rs9331896 (C) SNPs were specifically related lower content of ApoJ and ApoE in lipoproteins may be related with higher content of ApoJ in HDLs. Indeed, rs11136000 (T) with higher vascular Aβ deposition, which could be translated to has been described to increase cerebral ApoJ expression in AD a more prominent occurrence and severity of different CAA-SVD patients and controls (Allen et al., 2012; Ling et al., 2012), although markers. studies regarding its impact on plasma levels have found different results (Schürmann et al., 2011; Mullan et al., 2013). Therefore, we observed that only the AD-protective CLU SNPs [rs11136000 (T) 5. Conclusion and rs9331896 (C)] were associated with higher ApoJ content in HDL, the circulating lipoprotein with higher content of ApoJ, in an In summary, we observed several genetic variants previously allelic-dependent manner. associated with AD (SNPs on ABCA7, BIN1, CD2AP, CLU, CR1, On the other hand, we used the same strategy regarding EPHA1, and TREML2) to be associated with MRI markers of CAA. associations between genetic APOE variants and ApoE plasma In more detail, the CLU SNPs studied were independently and concentration and its distribution among lipoproteins. Unlike what significantly associated with several CAA markers and followed the is observed for ApoJ, total plasma levels of ApoE are modulated same risk/protective direction as previously described for AD. We by the APOE alleles. In this sense, we detected that carrying observed CLU SNPs and APOE genotype described as protective the ε2 allele significantly increased total plasma ApoE levels, factors for AD to be associated with higher ApoJ and ApoE content whereas ε4 significantly decreased its levels as has been extensively in lipoproteins. Furthermore, ApoJ and ApoE distribution among described in different cohorts (Montañola et al., 2016; Rasmussen, lipoproteins was associated with CAA-MRI features, such as the 2016). Regarding ApoE distribution among plasma lipoproteins, presence of CSO-EPVS and their degree score. We suggest that we observed that APOEε2 carriers presented also significantly higher ApoE and ApoJ levels in HDL and LDL might enhance increased ApoE content in LDL. It is important to consider that atheroprotective and anti-inflammatory responses in cerebral apart from altering ApoE levels, APOE genotype also influences amyloidosis. This study reinforces the relevance of peripheral ApoE structure and functionality, and thus, affects processes such lipid metabolism and cerebrovascular functionality in CAA and as the clearance of parenchymal Aβ through LDL receptor-related could open the path for new therapeutic strategies. However, the protein 1 (LRP1) (Yin and Wang, 2018). meaning of our findings regarding ApoJ and ApoE distribution in In order to understand the biological impact of the MRI lipoproteins and their link to Aβ pathology needs further research markers analyzed, we studied potential associations with ApoJ and and confirmation in larger cohorts. An important limitation of ApoE circulating levels in the CAA-ICH cohort. ApoE and ApoJ our study is that all the patients from our cohort presented a content in both HDL and LDL are negatively associated with several symptomatic ICH and some of the results observed could be due to MRI features. Specifically, we found that lower ApoJ content in pathological changes associated with this hemorrhagic phenotype. LDL and lower ApoE levels in plasma and HDL were associated We tried to overcome this issue by collecting the blood samples with CSO-EPVS presence and degree. EPVS in CSO have been of the participants for protein determinations in a non-acute related to high Aβ burden in CAA and AD patients in post-mortem phase of the disease to avoid the massive inflammatory cascade histopathological studies (Roher et al., 2003; van Veluw et al., produced after the stroke. Also, as an observational study, we 2016; Perosa et al., 2022), suggesting that these pathological space cannot confirm the causality of the differences observed in ApoJ enlargements may be the consequence of the impairment in the and ApoE circulating levels. Finally, although we are in front of a perivascular drainage pathways (Hawkes et al., 2011; Arbel-Ornath multicenter study, the sample size is still small and larger studies et al., 2013; Perosa et al., 2022). Interestingly, in a previous study, we with a higher number of patients should be conducted, especially found ApoJ LDL content to be significantly elevated in AD patients regarding the genetic association’s analysis. Frontiers in Aging Neuroscience 12 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 Data availability statement and co-financed by the European Regional Development Fund (FEDER). The Neurovascular Research Laboratory was part of the The raw data supporting the conclusions of this article will be RICORS-ICTUS-Enfermedades Vasculares Cerebrales Network, made available by the authors, without undue reservation. ISCIII, Spain (RD21/0006/0007). CIBERDEM (CB07/08/0016) was an ISCIII Project. Ethics statement Conflict of interest The studies involving human participants were reviewed and approved by the Clinical Investigation Ethical Committee The authors declare that the research was conducted in the of the Vall d’Hebron University Hospital, Barcelona, Spain absence of any commercial or financial relationships that could be (PR(AG)326/2014). The patients/participants provided their construed as a potential conflict of interest. written informed consent to participate in this study. Author contributions Publisher’s note All claims expressed in this article are solely those of the MH-G: conceptualization. AB-P, CV, AR, and SB: methodology. authors and do not necessarily represent those of their affiliated OP, DR-L, MF, ST, MM-Z, PC, RV, LL-H, JA, SP-S, AD-M, organizations, or those of the publisher, the editors and the JF, GM, and JM: recruitment of patients. AB-P and MH-G: reviewers. Any product that may be evaluated in this article, or formal analysis. AB-P: writing—original draft preparation. MH-G claim that may be made by its manufacturer, is not guaranteed or and JS-Q: writing—review and editing, supervision, and funding endorsed by the publisher. acquisition. All authors read and agreed to the published version of the manuscript. Supplementary material Funding The Supplementary Material for this article can be found This work was funded by the Instituto de Salud Carlos III online at: https://www.frontiersin.org/articles/10.3389/fnagi.2023. (ISCIII), (PI17/00275, PI019/00421, PI20/00465, and PI20/00334) 1134399/full#supplementary-material References Allen, M., Zou, F., Chai, H. S., Younkin, C. S., Crook, J., Shane Pankratz, Bonaterra-Pastra, A., Fernández-De-retana, S., Rivas-Urbina, A., Puig, N., Benítez, V., et al. (2012). Novel late-onset Alzheimer disease loci variants associate S., Pancorbo, O., et al. (2021). Comparison of plasma lipoprotein composition and with brain gene expression. Neurology 79, 221–228. doi:10.1212/WNL.0b013e3182 function in cerebral amyloid angiopathy and Alzheimer’s disease. Biomedicines 9:72. 605801 doi: 10.3390/biomedicines9010072 Arbel-Ornath, M., Hudry, E., Eikermann-Haerter, K., Hou, S., Gregory, J. L., Boulouis, G., Charidimou, A., Pasi, M., Roongpiboonsopit, D., Xiong, L., Auriel, Zhao, L., et al. (2013). Interstitial fluid drainage is impaired in ischemic stroke and E., et al. (2017). Hemorrhage recurrence risk factors in cerebral amyloid angiopathy: Alzheimer’s disease mouse models. Acta Neuropathol. 126, 353–364. doi: 10.1007/ Comparative analysis of the overall small vessel disease severity score versus individual s00401-013-1145-2 neuroimaging markers. J. Neurol. Sci. 380, 64–67. doi: 10.1016/j.jns.2017.07.015 Arvanitakis, Z., Leurgans, S. E., Wang, Z., Wilson, R. S., Bennett, D. A., and Bu, G. (2010). Apolipoprotein E and its receptors in Alzheimer’s disease: Pathways, Schneider, J. A. (2011). Cerebral amyloid angiopathy pathology and cognitive domains pathogenesis and therapy. Ratio (Oxf) 36, 490–499. doi: 10.1124/dmd.107.016501. in older persons. Ann. Neurol. 69, 320–327. doi: 10.1038/jid.2014.371 CYP3A4-Mediated Auton, A., Abecasis, G. R., Altshuler, D. M., Durbin, R. M., Bentley, D. R., Cai, R., Han, J., Sun, J., Huang, R., Tian, S., Shen, Y., et al. (2016). Plasma clusterin Chakravarti, A., et al. (2015). A global reference for human genetic variation. Nature and the CLU gene rs11136000 variant are associated with mild cognitive impairment 526, 68–74. doi: 10.1038/nature15393 in type 2 diabetic patients. Front. Aging Neurosci. 8:179. doi: 10.3389/fnagi.2016.00179 Beecham, G. W., Hamilton, K., Naj, A. C., Martin, E. R., Huentelman, M., Myers, Chang, Y., Liu, J., Wang, L., Li, X., Wang, Z., Lin, M., et al. (2021). Diagnostic utility A. J., et al. (2014). Genome-wide association meta-analysis of neuropathologic features of integrated11c -Pittsburgh compound B positron emission tomography/magnetic of Alzheimer’s disease and related dementias. PLoS Genet. 10:1004606. doi: 10.1371/ resonance for cerebral amyloid angiopathy: A pilot study. Front. Aging Neurosci. journal.pgen.1004606 13:721780. doi: 10.3389/fnagi.2021.721780 Beeg, M., Stravalaci, M., Romeo, M., Carrá, A. D., Cagnotto, A., Rossi, A., et al. Charidimou, A., Boulouis, G., Frosch, M. P., Baron, J.-C., Pasi, M., Albucher, (2016). Clusterin binds to Aβ1-42 oligomers with high affinity and interferes with J. F., et al. (2022). The Boston criteria version 2.0 for cerebral amyloid angiopathy: peptide aggregation by inhibiting primary and secondary nucleation. J. Biol. Chem. A multicentre, retrospective, MRI-neuropathology diagnostic accuracy study. Lancet 291, 6958–6966. doi: 10.1074/jbc.M115.689539 Neurol. 21, 714–725. doi: 10.1016/S1474-4422(22)00208-3 Benitez, B. A., Jin, S. C., Guerreiro, R., Graham, R., Lord, J., Harold, D., et al. (2014). Charidimou, A., Boulouis, G., Gurol, M. E., Ayata, C., Bacskai, B. J., Frosch, M. P., Missense variant in TREML2 protects against Alzheimer’s disease. Neurobiol. Aging et al. (2017a). Emerging concepts in sporadic cerebral amyloid angiopathy. Brain 140, 35, 1510.e19–1510.e26. doi: 10.1016/j.neurobiolaging.2013.12.010 1829–1850. doi: 10.1093/brain/awx047 Blumenau, S., Foddis, M., Müller, S., Holtgrewe, M., Bentele, K., Berchtold, D., Charidimou, A., Imaizumi, T., Moulin, S., Biffi, A., Samarasekera, N., Yakushiji, et al. (2020). Investigating APOE, APP-Aβ metabolism genes and Alzheimer’s disease Y., et al. (2017d). Brain hemorrhage recurrence, small vessel disease type, and GWAS hits in brain small vessel ischemic disease. Sci. Rep. 10:7103. doi: 10.1038/ cerebral microbleeds: A meta-analysis. Neurology 89, 820–829. doi: 10.1212/WNL. s41598-020-63183-5 0000000000004259 Frontiers in Aging Neuroscience 13 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 Charidimou, A., Boulouis, G., Roongpiboonsopit, D., Auriel, E., Pasi, M., Haley, Lambert, J.-C., Heath, S., Even, G., Campion, D., Sleegers, K., Hiltunen, M., et al. K., et al. (2017c). Cortical superficial siderosis multifocality in cerebral amyloid (2009). Genome-wide association study identifies variants at CLU and CR1 associated angiopathy: A prospective study. Neurology 89, 2128–2135. doi: 10.1212/WNL. with Alzheimer’s disease. Nat. Genet. 41, 1094–1099. doi: 10.1038/ng.439 0000000000004665 Lanfranco, M. F., Ng, C. A., and Rebeck, G. W. (2020). ApoE lipidation as a Charidimou, A., Boulouis, G., Pasi, M., Auriel, E., van Etten, E. S., Haley, therapeutic target in Alzheimer’s disease. Int. J. Mol. Sci. 21, 1–19. doi: 10.3390/ K., et al. (2017b). MRI-visible perivascular spaces in cerebral amyloid angiopathy ijms21176336 and hypertensive arteriopathy. Neurology 88, 1157–1164. doi: 10.1212/WNL. Ling, I. F., Bhongsatiern, J., Simpson, J. F., Fardo, D. W., and Estus, S. (2012). 0000000000003746 Genetics of clusterin isoform expression and Alzheimer’s disease risk. PLoS One Charidimou, A., Martinez-Ramirez, S., Reijmer, Y. D., Oliveira-Filho, J., Lauer, A., 7:e33923. doi: 10.1371/journal.pone.0033923 Roongpiboonsopit, D., et al. (2016). Total magnetic resonance imaging burden of small Linn, J., Halpin, A., Demaerel, P., Ruhland, J., Giese, A., Dichgans, M., et al. (2010). vessel disease in cerebral amyloid angiopathy an imaging-pathologic study of concept Prevalence of superficial siderosis in patients with cerebral amyloid angiopathy. validation. JAMA Neurol. 73, 994–1001. doi: 10.1001/jamaneurol.2016.0832 Neurology 74, 1346–1350. doi: 10.1212/WNL.0b013e3181f002c1 Charidimou, A., Martinez-Ramirez, S., Shoamanesh, A., Oliveira-Filho, J., Frosch, Mäkelä, M., Kaivola, K., Valori, M., Paetau, A., Polvikoski, T., Singleton, A. B., et al. M., Vashkevich, A., et al. (2015). Cerebral amyloid angiopathy with and without (2018). Alzheimer risk loci and associated neuropathology in a population-based study hemorrhage: Evidence for different disease phenotypes. Neurology 84:1206. doi: 10. (Vantaa 85+). Neurol. Genet. 4:e211. doi: 10.1212/NXG.0000000000000211 1212/WNL.0000000000001398 Marais, A. D. (2019). Apolipoprotein E in lipoprotein metabolism, health and Chojdak-Łukasiewicz, J., Dziadkowiak, E., Zimny, A., and Paradowski, B. (2021). cardiovascular disease. Pathology 51, 165–176. doi: 10.1016/j.pathol.2018.11.002 Cerebral small vessel disease: A review. Adv. Clin. Exp. Med. 30, 349–356. doi: 10. 17219/ACEM/131216 Marazuela, P., Bonaterra-Pastra, A., Faura, J., Penalba, A., Pizarro, J., Pancorbo, O., et al. (2021). Circulating AQP4 levels in patients with cerebral amyloid angiopathy- Cordonnier, C., Potter, G. M., Jackson, C. A., Doubal, F., Keir, S., Sudlow, C. L. M., associated intracerebral hemorrhage. J. Clin. Med. 10:989. doi: 10.3390/jcm10050989 et al. (2009). Improving interrater agreement about brain microbleeds: Development of the brain observer microbleed scale (BOMBS). Stroke 40, 94–99. doi: 10.1161/ Montañola, A., de Retana, S. F., López-Rueda, A., Merino-Zamorano, C., Penalba, STROKEAHA.108.526996 A., Fernández-Álvarez, P., et al. (2016). ApoA1, ApoJ and ApoE plasma levels and genotype frequencies in cerebral amyloid angiopathy. Neuromolecular Med. 18, 99– Farid, K., Charidimou, A., and Baron, J. C. (2017). Amyloid positron emission 108. doi: 10.1007/s12017-015-8381-7 tomography in sporadic cerebral amyloid angiopathy: A systematic critical update. Neuroimage Clin. 15, 247–263. doi: 10.1016/j.nicl.2017.05.002 Morton, A. M., Furtado, J. D., Mendivil, C. O., and Sacks, F. M. (2019). Dietary unsaturated fat increases HDL metabolic pathways involving apoE favorable to reverse Fazekas, F., Chawluk, J. B., Alavi, A., Hurtig, H. I., and Zimmerman, R. A. (1987). cholesterol transport. JCI Insight 4, 1–14. doi: 10.1172/jci.insight.124620 MR signal abnormalities at 1.5-T in Alzheimer dementia and normal aging. Am. J. Roentgenol. 149, 351–356. doi: 10.2214/ajr.149.2.351 Mullan, G. M., McEneny, J., Fuchs, M., McMaster, C., Todd, S., McGuinness, B., et al. (2013). Plasma clusterin levels and the rs11136000 genotype in individuals with Foster, E. M., Dangla-Valls, A., Lovestone, S., Ribe, E. M., and Buckley, mild cognitive impairment and Alzheimer’s disease. Curr. Alzheimer Res. 10, 973–978. N. J. (2019). Clusterin in Alzheimer’s disease: Mechanisms, genetics, and doi: 10.2174/15672050113106660162 lessons from other pathologies. Front. Neurosci. 13:164. doi:10.3389/fnins.201 9.00164 National Cholesterol Education Program [NCEP] Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults [Adult Treatment Giralt-Steinhauer, E., Jiménez-Balado, J., Fernández-Pérez, I., Rey Álvarez, L., Panel III],. (2002). Third report of the national cholesterol education program (NCEP) Rodríguez-Campello, A., Ois, Á, et al. (2022). Genetics and epigenetics of expert panel on detection, evaluation, and treatment of high blood cholesterol in adults spontaneous intracerebral hemorrhage. Int. J. Mol. Sci. 23:6479. doi:10.3390/ijms23 (adult treatment panel III) final report. Circulation 106, 3143–3421. 126479 Obermeier, B., Daneman, R., and Ransohoff, R. M. (2013). Development, Giri, M., Zhang, M., and Lü, Y. (2016). Genes associated with Alzheimer’s disease: maintenance and disruption of the blood-brain-barrier. Nat. Med. 19, 1584–1596. An overview and current status. Clin. Interv. Aging 11, 665–681. doi: 10.2147/CIA. doi: 10.1038/nm.3407.Development S105769 Pedrini, S., Doecke, J. D., Hone, E., Wang, P., Thota, R., Bush, A. I., et al. Greenberg, S. M., Bacskai, B. J., Hernandez-Guillamon, M., Pruzin, J., Sperling, R., (2022). Plasma high-density lipoprotein cargo is altered in Alzheimer’s disease and and van Veluw, S. J. (2020). Cerebral amyloid angiopathy and Alzheimer disease—one is associated with regional brain volume. J. Neurochem. 163, 53–67. doi: 10.1111/jnc. peptide, two pathways. Nat. Rev. Neurol. 16, 30–42. doi: 10.1038/s41582-019-0281-2 15681 Greenberg, S. M., Briggs, M. E., Hyman, B. T., Kokoris, G. J., Takis, C., Kanter, D. S., Perosa, V., Oltmer, J., Munting, L. P., Freeze, W. M., Auger, C. A., Scherlek, et al. (1996). Apolipoprotein E ε4 is associated with the presence and earlier onset of A. A., et al. (2022). Perivascular space dilation is associated with vascular amyloid-β hemorrhage in cerebral amyloid angiopathy. Stroke 27, 1333–1337. doi: 10.1161/01. accumulation in the overlying cortex. Acta Neuropathol. 143, 331–348. doi: 10.1007/ STR.27.8.1333 s00401-021-02393-1 Greenberg, S. M., Vonsattel, J.-P. G., Segal, A. Z., Chiu, R. I., Clatworthy, A. E., Radding, C. M., and Steinberg, D. (1960). Studies on the synthesis and secretion Liao, A., et al. (1998). Association of apolipoprotein E and ε2 vasculopathy in cerebral of serum lipoproteins by rat liver slices. J. Clin. Invest. 39, 1560–1569. doi: 10.1172/ amyloid angiopathy. Neurology 50, 638–643. doi: 10.1212/wnl.50.4.961 JCI104177 Hawkes, C. A., Härtig, W., Kacza, J., Schliebs, R., Weller, R. O., Nicoll, J. A., et al. Rasmussen, K. L. (2016). Plasma levels of apolipoprotein E, APOE genotype and (2011). Perivascular drainage of solutes is impaired in the ageing mouse brain and risk of dementia and ischemic heart disease: A review. Atherosclerosis 255, 145–155. in the presence of cerebral amyloid angiopathy. Acta Neuropathol. 121, 431–443. doi: 10.1016/j.atherosclerosis.2016.10.037 doi: 10.1007/s00401-011-0801-7 Replogle, J., Chan, G., White, C., Raj, T., Winn, P., Evans, D., et al. (2015). A TREM1 Howlett, D. R., Hortobágyi, T., and Francis, P. T. (2013). Clusterin associates variant alters the accumulation of Alzheimer-related amyloid pathology. Ann. Neurol. specifically with Aβ40 in Alzheimer’s disease brain tissue. Brain Pathol. 23, 623–632. 67, 558–569. doi: 10.1111/j.1365-2958.2007.06085.x.Compensation doi: 10.1111/bpa.12057 Rivas-Urbina, A., Rull, A., Aldana-Ramos, J., Santos, D., Puig, N., Farre- Huynh, T., Davis, A., Ulrich, J., and Holtzman, D. (2017). Apolipoprotein E and Cabrerizo, N. et al. (2020). Subcutaneous administration of apolipoprotein J-derived Alzheimer’s disease: The influence of apolipoprotein E on amyloid-β and other mimetic peptide D-[113-122]apoJ improves LDL and HDL function and prevents amyloidogenic proteins. J. Lipid Res. 58, 824–836. doi: 10.1194/jlr.R075481 atherosclerosis in LDLR-KO mice. Biomolecules 10:829. doi: 10.3390/biom10060829 Inoue, Y., Ando, Y., Misumi, Y., and Ueda, M. (2021). Current management and Robert, J., Button, E. B., Martin, E. M., McAlary, L., Gidden, Z., Gilmour, M., et al. therapeutic strategies for cerebral amyloid angiopathy. Int. J. Mol. Sci. 22:3869. doi: (2020). Cerebrovascular amyloid angiopathy in bioengineered vessels is reduced by 10.3390/ijms22083869 high-density lipoprotein particles enriched in apolipoprotein E. Mol. Neurodegener. 15, 1–21. doi: 10.1186/s13024-020-00366-8 Karch, C. M., and Goate, A. M. (2015). Alzheimer’s disease risk genes and mechanisms of disease pathogenesis. Biol. Psychiatry 77, 43–51. doi: 10.1016/j. Roher, A. E., Kuo, Y. M., Esh, C., Knebel, C., Weiss, N., Kalback, W., et al. (2003). biopsych.2014.05.006.Alzheimer Cortical and leptomeningeal cerebrovascular amyloid and white matter pathology in Alzheimer’s disease. Mol. Med. 9, 112–122. doi: 10.1007/bf03402043 Koch, M., and Jensen, M. K. (2016). HDL-cholesterol and apolipoproteins in relation to dementia. Curr. Opin. Lipidol. 27, 76–87. doi: 10.1097/MOL. Sawyer, R. P., Demel, S. L., Comeau, M. E., Marion, M., Rosand, J., Langefeld, 0000000000000257 C. D., et al. (2022). Alzheimer’s disease related single nucleotide polymorphisms and correlation with intracerebral hemorrhage incidence. Medicine 101:e30782. doi: Kuiperij, H. B., Hondius, D. C., Kersten, I., Versleijen, A. A. M., Rozemuller, A. J. M., 10.1097/md.0000000000030782 Greenberg, S. M., et al. (2020). Apolipoprotein D: A potential biomarker for cerebral amyloid angiopathy. Neuropathol. Appl. Neurobiol. 46, 431–440. doi: 10.1111/nan. Scheltens, P., de Strooper, B., Kivipelto, M., Holstege, H., Chételat, G., Teunissen, 12595 C. E., et al. (2021). Alzheimer’s disease. Lancet 397, 1577–1590. doi: 10.1016/S0140- 6736(20)32205-4 Lambert, J.-C. (2013). Meta-Analysis of 74,046 individuals identifies 11 new susceptibility loci for Alzheimer’s disease. Nat. Genet. 45, 1452–1458. doi: 10.1038/ng. Schmechel, D. E., Saunders, A. M., Strittmatter, W. J., Crain, B. J., Hulette, C. M., 2802.Meta-analysis Joo, S. H., et al. (1993). Increased amyloid β-peptide deposition in cerebral cortex as a Frontiers in Aging Neuroscience 14 frontiersin.org Bonaterra-Pastra et al. 10.3389/fnagi.2023.1134399 consequence of apolipoprotein E genotype in late-onset Alzheimer disease. Proc. Natl. Tan, L., Wang, H., Tan, M., Tan, C., Zhu, X., Miao, D., et al. (2016). Effect of Acad. Sci. U.S.A. 90, 9649–9653. doi: 10.1073/pnas.90.20.9649 CLU genetic variants on cerebrospinal fluid and neuroimaging markers in healthy, mild cognitive impairment and Alzheimer’ s disease cohorts. Berlin: Nature Publishing Schürmann, B., Wiese, B., Bickel, H., Weyerer, S., Riedel-Heller, S. G., Pentzek, Group, 1–12. doi: 10.1038/srep26027 M., et al. (2011). Association of the Alzheimer’s disease clusterin risk allele with plasma clusterin concentration. J. Alzheimers Dis. 25, 421–424. doi: 10.3233/JAD- van Veluw, S. J., Biessels, G. J., Bouvy, W. H., Spliet, W. G., Zwanenburg, 2011-110251 J. J., Luijten, P. R., et al. (2016). Cerebral amyloid angiopathy severity is linked to dilation of juxtacortical perivascular spaces. JCBFM 36, 576–580. doi: 10.1177/ Serrano-Pozo, A., Das, S., and Hyman, B. T. (2021). APOE and Alzheimer’s disease: 0271678X15620434 Advances in genetics, pathophysiology, and therapeutic approaches. Lancet Neurol. 20, 68–80. doi: 10.1016/S1474-4422(20)30412-9 Wardlaw, J. M., Smith, E. E., Biessels, G. J., Cordonnier, C., Fazekas, F., Frayne, R., et al. (2013). Neuroimaging standards for research into small vessel disease and Shams, M., Shams, S., Martola, J., Cavallin, L., Granberg, T., Kaijser, M., et al. (2022). its contribution to ageing and neurodegeneration. Lancet Neurol. 12, 822–838. doi: MRI markers of small vessel disease and the APOE allele in cognitive impairment. 10.1016/S1474-4422(13)70124-8 Front. Aging Neurosci. 14:897674. doi: 10.3389/fnagi.2022.897674 Wei, C. J., Cui, P., Li, H., Lang, W. J., Liu, G. Y., and Ma, X. F. (2019). Shared genes Shoamanesh, A., Martinez-Ramirez, S., Oliveira-Filho, J., Reijmer, Y., Falcone, G. J., between Alzheimer’s disease and ischemic stroke. CNS Neurosci. Ther. 25, 855–864. Ayres, A., et al. (2014). Interrelationship of superficial siderosis and microbleeds doi: 10.1111/cns.13117 in cerebral amyloid angiopathy. Neurology 83, 1838–1843. doi: 10.1212/WNL. 0000000000000984 Xing, Y. Y., Yu, J. T., Cui, W. Z., Zhong, X. L., Wu, Z. C., Zhang, Q., et al. (2012). Blood clusterin levels, rs9331888 polymorphism, and the risk of Alzheimer’s disease. Shulman, J. M., Chen, K., Keenan, B. T., Chibnik, L. B., Fleisher, A., Thiyyagura, P., J. Alzheimers Dis. 29, 515–519. doi: 10.3233/JAD-2011-111844 et al. (2013). Genetic susceptibility for Alzheimer disease neuritic plaque pathology. JAMA Neurol. 70, 1150–1157. doi: 10.1001/jamaneurol.2013.2815 Yamada, M. (2015). Cerebral amyloid angiopathy: Emerging concepts. J. Stroke 17, 17–30. doi: 10.5853/jos.2015.17.1.17 Smith, E. E., Charidimou, A., Ayata, C., Werring, D. J., and Greenberg, S. M. (2021). Cerebral amyloid angiopathy-related transient focal neurologic episodes. Neurology Yin, Y., and Wang, Z. (2018). ApoE and neurodegenerative diseases in aging. Adv. 97, 231–238. doi: 10.1212/WNL.0000000000012234 Exp. Med. Biol. 1086, 77–92. doi: 10.1007/978-981-13-1117-8_5 Szymanski, M., Wang, R., Bassett, S. S., and Avramopoulos, D. (2011). Alzheimer’s Yu, J. T., and Tan, L. (2012). The role of clusterin in Alzheimer’s disease: Pathways, risk variants in the clusterin gene are associated with alternative splicing. Transl. pathogenesis, and therapy. Mol. Neurobiol. 45, 314–326. doi: 10.1007/s12035-012- Psychiatry 1:e18. doi: 10.1038/tp.2011.17 8237-1 Frontiers in Aging Neuroscience 15 frontiersin.org