Fig 2: Chronic anti Aβ-immunization with 350 μg/wk of 3D6 mAb for 15 weeks was associated with mAb binding to vascular amyloid, complement, and immune cell activation and increased brain C3 levels but had no effect on blood C3 levels.a, Schematic of 15-week passive immunization in 16-month-old APP/PS1dE9; hAPOE4 mice (n = 8–9/group) with 3D6 or IgG2a isotype control (350 μg/week). b, Immunostaining in the cerebellum and meningeal layers showing anti-Ms IgG2a (to detect mAb) binding, vascular amyloid (Amylo-Glo), and C1q. Arrows indicate co-localization of mAb, vascular amyloid, and C1q. c, Multiple immunolabelling with Iba1, CD68 showing amoeboid-like activated microglia/macrophages (arrows) near CAA-laden vessels (Amylo-Glo) and C1q staining in the cerebellum and meninges. Arrowheads indicate C1q accumulation in cerebellar and meningeal vessels. d–e, Soluble C3 levels measured in d, Tper brain extracts and e, plasma. Mann-Whitney test, **p < 0.005. Data are expressed as mean ± SEM. Scale bars: b–c, 25 μm.

Fig 3: Passive anti-Aβ immunization for 7 weeks was associated with differential gene expression in the brain, particularly upregulation of inflammatory genes, including complement genes.a, Schematic for anti-Aβ immunotherapy in the 20-month-old APP/PS1dE9;hApoE4 mice immunized for 7-weeks, 500 μg/week, followed by RNA extraction from cerebellar tissue and Smart-seq2 (n=3, 2M and 1F). b, Volcano plot showing differentially expressed genes with p-values < 0.0001. c, Pheatmap plot showing significantly expressed genes after multiple correction, p-adjusted value < 0.05. d, List of selected differentially expressed upregulated genes with respective functions. e, Bar plots with the top 20 up-regulated and down-regulated biological pathways.

Fig 4: Chronic anti-Aβ passive immunization with 3D6 mAb was associated with elevated C3 levels in the brain and a corresponding lowering in the serum, in aged 20-mo-old male APP/PS1dE9:hAPOE4 mice following 13 weeks of passive immunization.a, Immunostaining for fibrillar amyloid (Amylo-Glo, AG), endothelial marker CD31, and complement C1q in cortical and meningeal vessels. Asterisks indicate plaque-associated C1q immunoreactivity (IR); arrows indicate C1q accumulation along vessels. b–d, Quantification of (b), C1q IR (c), AG+C1q+ colocalized area and (d), CD31+C1q+ colocalized area. e, Representative images from the cortex showing complement C3 immunoreactivity in GFAP-positive astrocytes, including perivascular localization (arrows). f–g, Quantification of % area for C3, GFAP, and C3+GFAP+ colocalization in the (f) isocortex and (g) cerebellum. h–i, Soluble C1q levels in h, T-per brain homogenates and i, serum. j–k, Soluble C3 levels in j, T-per brain homogenates and k, serum measured by ELISA. l, Linear regression between brain soluble C3 levels and % Prussian blue–stained area (hemosiderin/microhemorrhages) across IgG2a and 3D6–treated groups. Unpaired t-test: *p < 0.05, **p < 0.005. Data represent mean ± SEM; n = 5 per group. Scale bars: a, e = 50 μm.

Fig 5: MRI and histopathology-based detection of immunotherapy-associated microbleeds and CAA-associated classical complement activation in mice chronically treated for 13 weeks.a, Schematic of anti-Aβ immunization with 500 μg/wk 3D6 IgG2a mAb and an IgG2a isotype control for 13 weeks from 17 to 20 months in aged male APP/PS1dE9;hAPOE4 AD-like mouse model. For baseline data, baseline control mice were used. For spontaneous hemorrhage detection, 1x PBS-injected aged controls were used. b, T2*-weighted FLASH MRI of 20-month-old male APP/PS1dE9;hApoE4 mice after 13 weeks of passive immunization with 3D6 mAb (500 μg/week), showing cerebral microbleeds as hypointense spots (arrowheads), compared to 1× PBS–treated aged controls. c, Frequency of cerebral microbleeds detected by T2* MRI in 3D6–treated (n = 3) and PBS-treated (n = 4) mice. d, Quantification of Prussian blue–positive (% hemosiderin-stained) area in IgG2a and 3D6–treated mice. Unpaired t-test: **p < 0.005, ***p < 0.0005. Data represent mean ± SEM. e, Representative Prussian blue–stained sections showing hemosiderin deposits (arrowheads, with selected examples in insets) in cortex, hippocampus, thalamus, and cerebellum from 1× PBS (n=4), IgG2a isotype control (n = 5), and 3D6 IgG2a–treated mice (n = 5). Scale bar: 200 μm. f, Immunofluorescence staining in the cerebellum showing 3D6 mAb localization (anti-Ms IgG2a) and C1q deposition at sites of vascular and parenchymal amyloid labeled with Amylo-Glo (AG). g, Detection of activated complement fragments C3b/iC3b/C3c at vascular amyloid sites associated with 3D6 mAb. h, Localization of membrane attack complex (MAC/C5b-9) along vascular amyloid (arrows) and plaques (asterisk) in 3D6–treated mice. Minimal complement staining is observed in IgG2a-treated controls. Scale bars: f–h, 25 μm.