Cerebral Amyloid Angiopathy-Related Inflammation Biomarkers: Where are we Now?

Cerebral amyloid angiopathy (CAA) is a common age-related microangiopathy affecting the small cortical and leptomeningeal vessels by progressive amyloid-β deposition [1]. It is most often recognized clinically by spontaneous lobar intracerebral hemorrhage, transient focal neurological episodes, or as a contributor to vascular cognitive impairment [1, 2]. CAA (of varying severity) is almost invariably found in Alzheimer’s disease brains, although it also occurs independently [2]. CAA– related inflammation (CAA-ri) is an aggressive disease subtype of CAA [1]. This spontaneous meningoencephalitis syndrome is characterized by rapidly progressive (or subacute) mental status and behavioral-cognitive changes, headache, seizures, and occasionally focal neurological deficits, with neuropathological evidence of CAA-associated vascular or perivascular inflammatory infiltrates [3, 4]. Characteristic brain MRI findings include unifocal or multifocal white matter hyperintensities, which are typically asymmetric and extend to the immediately adjacent subcortical white matter (U-fibers). Blood-sensitive MRI sequences might also reveal hemorrhagic signatures of CAA: multiple lobar microbleeds and cortical superficial siderosis [1, 3]. CAA-ri is an important diagnosis not to be missed in clinical practice, as patients typically respond to prompt immunosuppressive treatment, i.e., high dose corticosteroids or cyclophosphamide, with only rare relapses [5, 6]. However, a definitive diagnosis of CAA-ri still requires brain biopsy, and hence developing non-invasive diagnostic criteria and biomarkers for this syndrome are key priorities.

CAA-ri has generated additional interest for its notable clinical and radiological similarities to amyloid-related imaging abnormalities (ARIA), a complication of immunotherapy treatments in Alzheimer’s disease patients (e.g., bapineuzumab, an experimental humanized monoclonal antibody to amyloid-β) [7–10]. Although the pathogenesis of ARIA remains undefined, it occurs more commonly in trial subjects receiving higher doses of the administered antibody and APOE ɛ4 allele carriers; it also seems to be related with the severity of parenchymal amyloid deposition and the degree of CAA [11]. These raise the intriguing possibility that ARIA is the result of an immune response directed toward cerebrovascular amyloid deposits (i.e., CAA), in combination to rapid shifts of amyloid from brain parenchyma to the perivascular spaces surrounding an already impaired microvasculature [7, 11]. An important advance in the field is the recent demonstration of autologous anti-amyloid-β antibodies in the cerebrospinal fluid (CSF) of patients with active CAA-ri [12, 13]. Anti-amyloid-beta autoantibodies levels showed a temporal relationship with patients’ clinical and radiologic improvement, in addition to specifically discriminating CAA-ri from sporadic CAA without inflammation, other non-CAA inflammatory and autoimmune disorders or healthy controls [5, 11– 13]. Thus, CSF amyloid-β autoantibodies not only point to a plausible aetiopathogenic process (i.e., inflammatory response to vascular amyloid deposits), but importantly pave the way for novel potential biomarkers for non-invasive CAA-ri diagnosis and monitoring.

The pilot study by Carmona-Iragui et al. in this issue of the Journal of Alzheimer’s Disease extends these findings and provides a possible framework for future CAA-ri characterization based on potential biomarkers. The authors evaluated pre- and post-treatment CSF markers and Florbetapir PET scans in a case series of four patients presenting with typical clinical and imaging manifestations of CAA-ri. They reported increased anti-amyloid autoantibodies, t-Tau, and p-Tau and decreased amyloid-β 40 and 42 levels; after treatment, anti-amyloid autoantibodies, but not other CSF markers, decreased. Florbetapir-PET (available only in two cases) performed after corticosteroid treatment, showed cortical amyloid-β deposition with lower retention in areas with white matter hyperintensities and edema during the active phase of the disease [14]. The report is novel, in that CSF markers and amyloid PET, although increasingly recognized as useful biomarker tools for sporadic CAA without inflammation [15], have not been systematically studied in CAA-ri [11, 17]. There are important reasons for treating these new data with caution and consider them hypothesis-generating. First, this is a convenient small patient sample, making replication of the result in other larger cohorts critical. Second, the patients’ diagnosis of CAA-ri in this report was not established on histopathological examination, but was rather based on characteristic clinical-neuroimaging findings and natural history.

The report by Carmona-Iragui and colleagues coincides with the validation of the first standardized clinico-radiological criteria for possible or probable CAA-ri diagnosis (Table 1) [18]. These non-invasive criteria were recently tested in patients with histologically proven CAA-ri (n = 17) and non-inflammatory CAA (n = 37) yielding good sensitivity and excellent specificity (82% – 67.6% and 82% – 97.3% , respectively) [18]. Similar to the Boston criteria for sporadic CAA [19, 20], the validation of CAA-ri criteria will allow a reliable diagnosis of the disease to be reached from basic clinical and MRI information alone, sparing at least some CAA-ri patients from the morbidity of unnecessary brain biopsies [18]. This new development will also facilitate the better characterization of CAA-ri through other biomarkers, e.g., CSF anti-amyloid autoantibodies level and amyloid PET. Future studies should define and validate clinical diagnostic cut-offs [11] for the biomarkers suggested by Carmona-Iragui et al. [14] and further explore how these can inform diagnosis and management of CAA-ri in the context of the new diagnostic criteria [18]. These findings might have a direct application for ARIA and future CAA immunotherapy trials aimed to safely clear amyloid from vessels and possibly reduce the risk of future CAA-related consequences.