No replicating evidence for anti-amyloid-β autoantibodies in cerebral amyloid angiopathy-related inflammation

Abstract

Objective: Elevated levels of anti-amyloid-β (anti-Aβ) autoantibodies in cerebrospinal fluid (CSF) have been proposed as a diagnostic biomarker for cerebral amyloid angiopathy-related inflammation (CAA-RI). We aimed to independently validate the immunoassay for quantifying these antibodies and evaluate its diagnostic value for CAA-RI.

Methods: We replicated the immunoassay to detect CSF anti-Aβ autoantibodies using CSF from CAA-RI patients and non-CAA controls with unrelated disorders and further characterized its performance. Moreover, we conducted a literature review of CAA-RI case reports to investigate neuropathological and CSF evidence of the nature of the inflammatory reaction in CAA-RI.

Results: The assay demonstrated a high background signal in CSF, which increased and corresponded with higher total immunoglobulin G (IgG) concentration in CSF (r_sp = 0.51, p = 0.02). Assay levels were not elevated in CAA-RI patients (n = 6) compared to non-CAA controls (n = 20; p = 0.64). Literature review indicated only occasional presence of B-lymphocytes and plasma cells (i.e., antibody-producing cells), alongside the abundant presence of activated microglial cells, T-cells, and other monocyte lineage cells. CSF analysis did not convincingly indicate intrathecal IgG production.

Interpretation: We were unable to reproduce the reported elevation of anti-Aβ autoantibody concentration in CSF of CAA-RI patients. Our findings instead support nonspecific detection of IgG levels in CSF by the assay. Reviewed CAA-RI case reports suggested a widespread cerebral inflammatory reaction. In conclusion, our findings do not support anti-Aβ autoantibodies as a diagnostic biomarker for CAA-RI.

Figure 1

Quantified immunoassay concentration in control subjects and patients. Displayed are the quantified immunoassay concentrations (ng/mL) in the cerebrospinal fluid (CSF) of control subjects (n = 20), cerebral amyloid angiopathy‐related inflammation (CAA‐RI) patients (n = 6), and subjects that participated in the AN1792 active amyloid‐β (Aβ) immunization trial (n = 4). CSF from one subject who received placebo treatment (open triangle) was not included in the analysis of variance (ANOVA). Concentrations were similar for all three groups (p = 0.64; ANOVA). Mean values and standard deviations are indicated. The legend indicates center of origin for CAA‐RI samples and the treatment type for AN1792 trial participants. Sample numbers for the latter are specified in the figure. The anti‐Aβ plasma titers for these samples were: 1707 for sample #1, 142 for samples #2 and #3, 0 for sample #4.

Figure 2

Relationship between quantified immunoassay concentrations and total immunoglobulin G levels in cerebrospinal fluid of control subjects. (A) Quantified immunoassay concentrations (ng/mL) in the cerebrospinal fluid (CSF) of control subjects with normal (<30 mg/L; n = 11) or elevated (≥30 mg/L; n = 10) total immunoglobulin G (IgG) CSF levels. Quantified immunoassay concentrations were increased in subjects with elevated total IgG levels (p = 0.04; t‐test). Asterisk indicates a significant p value. Mean values and standard deviations are indicated. (B) Spearman's correlation (r _sp) between quantified immunoassay concentrations and total IgG levels in the CSF of control subjects. A positive correlation was observed (r _sp = 0.51, p = 0.02).