The presence of circulating human apolipoprotein J reduces the occurrence of cerebral microbleeds in a transgenic mouse model with cerebral amyloid angiopathy

Abstract

Background: Cerebral amyloid angiopathy (CAA) is characterized by amyloid-β (Aβ) deposition in cerebral vessels, leading to lobar cerebral microbleeds (CMB) and intracerebral hemorrhages (ICH). Apolipoprotein J (ApoJ) is a multifunctional chaperone related to Aβ aggregation and clearance. Our study investigated the vascular impact of chronic recombinant human Apolipoprotein J (rhApoJ) treatment in a transgenic mouse model of β-amyloidosis with prominent CAA.

Methods: Twenty-month-old APP23 C57BL/6 mice received 25 doses of rhApoJ (1 mg/kg) (n = 9) or saline (n = 8) intraperitoneally for 13 weeks, while Wild-type (WT) mice received saline (n = 13). Postmortem brains underwent T2*-weighted magnetic resonance imaging (MRI) to detect hemorrhagic lesions. Aβ levels and distribution, cerebral fibrinogen leakage, brain smooth muscle actin (sma), and plasma matrix metalloproteinases and inflammatory markers were analyzed after treatments. Additionally, plasma samples from 22 patients with lobar ICH were examined to determine the clinical relevance of the preclinical findings.

Results: rhApoJ-treated APP23 presented fewer cortical CMBs (50-300 μm diameter) (p = 0.012) and cortical larger hemorrhages (> 300 μm) (p = 0.002) than saline-treated mice, independently of Aβ brain levels. MRI-detected hemorrhagic lesions correlated with fibrinogen cerebral extravasation (p = 0.011). Additionally, rhApoJ-treated mice presented higher number of sma-positive vessels than saline-treated mice (p = 0.038). In rhApoJ-treated mice, human ApoJ was detected in plasma and in occasional leptomeningeal vessels, but not in the parenchyma, suggesting that its mechanism of action operates through the periphery. The administration of rhApoJ induced an increase in plasma Groα (p = 0.035) and MIP-1α (p = 0.035) levels, while lower MMP-12 (p = 0.046) levels, compared to the saline-treated group. In acute lobar ICH patients, MMP-12 plasma levels correlated with larger hemorrhage volume (p = 0.040) and irregular ICH shape (p = 0.036).

Conclusions: Chronic rhApoJ treatment in aged APP23 mice ameliorated CAA-related neurovascular damage by reducing the occurrence of CMB. We propose that rhApoJ may prevent blood-brain barrier (BBB) leakage and CMB appearance partly through circulating MMP-12 modulation.

Keywords: Amyloid-β; Apolipoprotein J; Cerebral amyloid Angiopathy; Cerebral microbleeds; MMP-12.

Fig. 1

(A) Representative T2*-weighted cerebral MRI sections of APP23 mice chronically treated with saline or rhApoJ and WT mice treated with saline. CMB are indicated with red arrows. Graphical representation of the total number of hemorrhagic lesions. (B) Graphical representation of the number of hemorrhagic lesions, CMB (50–300 μm diameter) and large hemorrhagic lesions (> 300 μm) in the cortex, and (C) in deep brain regions (thalamus and basal ganglia). (D) Comparison of cerebral hemorrhagic lesions in T2*-MRI and Prussian blue staining showing iron hemosiderin deposits. The scale bar represents 20 μm. Data are presented as the mean + SD. #: count; *:p < 0.05; **:p < 0.01; ***:p < 0.001

Fig. 2

(A) Representative images of immunofluorescence staining of fibrinogen in green from brain sections of APP23 mice chronically treated with saline or rhApoJ and WT mice. (B) Graphical quantification of fibrinogen-positive cerebral vessels. Correlation between the number of fibrinogen-positive vessels and the number of cerebral hemorrhagic lesions detected by T2*-MRI. (C) Representative images of immunohistochemical staining of sma in brown from brain sections of APP23 mice chronically treated with saline or rhApoJ and WT mice. (D) Graphical quantification of sma-positive cerebral vessels per mm². Correlation between the number of sma-positive vessels and the number of hemorrhagic lesions. The scale bar represents 50 μm. Data are presented as boxplots. #: count; *:p < 0.05; **:p < 0.01; ***:p < 0.001

Fig. 3

Detection of fibrillar and vascular Aβ in chronically treated APP23 mice. Representative brain sections from APP23 mice chronically treated with saline or rhApoJ and the corresponding graphical quantification in both groups. Vascular Aβ stained with resorufin in red. Fibrillary Aβ stained with thioflavin S in green. Immunohistochemistry staining for Aβ₄₀ in brown. Scale bars represent 1000 μm. Data are presented as boxplots. #: count

Fig. 4

(A) Graphical quantification of plasma human ApoJ levels (hApoJ) (ng/mL) and representative image of human ApoJ immunodetection in a brain cortex section from a rhApoJ-treated APP23 mouse. A consecutive brain slice was stained with ThS to confirm the presence of CAA in the vessel. The scale bar represents 20 μm. (B) Graphical representation of plasma levels of Groα (pg/mL), MIP-1α (pg/mL), and MMP-12 (ng/mL) in mice from Group 2. Data are presented as boxplots. *:p < 0.05; ***:p < 0.001. (C) Correlation between plasma MMP-12 levels (ng/mL) and the number of large hemorrhagic lesions in the brain cortex (A) and the volume (mm³) of cortical hemorrhagic lesions

Fig. 5

(A) Correlation between MMP-12 plasma levels and ICH volume (cm³) and (B) the rate of ICH shape from a human cohort presenting with acute lobar ICH.