Blockage of VEGF function by bevacizumab alleviates early-stage cerebrovascular dysfunction and improves cognitive function in a mouse model of Alzheimer's disease

Abstract

Background: Alzheimer's disease (AD) is a neurodegenerative disorder and the predominant type of dementia worldwide. It is characterized by the progressive and irreversible decline of cognitive functions. In addition to the pathological beta-amyloid (Aβ) deposition, glial activation, and neuronal injury in the postmortem brains of AD patients, increasing evidence suggests that the often overlooked vascular dysfunction is an important early event in AD pathophysiology. Vascular endothelial growth factor (VEGF) plays a critical role in regulating physiological functions and pathological changes in blood vessels, but whether VEGF is involved in the early stage of vascular pathology in AD remains unclear.

Methods: We used an antiangiogenic agent for clinical cancer treatment, the humanized monoclonal anti-VEGF antibody bevacizumab, to block VEGF binding to its receptors in the 5×FAD mouse model at an early age. After treatment, memory performance was evaluated by a novel object recognition test, and cerebral vascular permeability and perfusion were examined by an Evans blue assay and blood flow scanning imaging analysis. Immunofluorescence staining was used to measure glial activation and Aβ deposits. VEGF and its receptors were analyzed by enzyme-linked immunosorbent assay and immunoblotting. RNA sequencing was performed to elucidate bevacizumab-associated transcriptional signatures in the hippocampus of 5×FAD mice.

Results: Bevacizumab treatment administered from 4 months of age dramatically improved cerebrovascular functions, reduced glial activation, and restored long-term memory in both sexes of 5×FAD mice. Notably, a sex-specific change in different VEGF receptors was identified in the cortex and hippocampus of 5×FAD mice. Soluble VEGFR1 was decreased in female mice, while full-length VEGFR2 was increased in male mice. Bevacizumab treatment reversed the altered expression of receptors to be comparable to the level in the wild-type mice. Gene Set Enrichment Analysis of transcriptomic changes revealed that bevacizumab effectively reversed the changes in the gene sets associated with blood-brain barrier integrity and vascular smooth muscle contraction in 5×FAD mice.

Conclusions: Our study demonstrated the mechanistic roles of VEGF at the early stage of amyloidopathy and the protective effects of bevacizumab on cerebrovascular function and memory performance in 5×FAD mice. These findings also suggest the therapeutic potential of bevacizumab for the early intervention of AD.

Keywords: Alzheimer’s disease; Bevacizumab; Cerebrovascular function; Vascular endothelial growth factor.

Fig. 1

Bevacizumab treatment improves long-term memory in 5×FAD mice. a The experimental timeline of bevacizumab treatment. b Schematic of the novel object recognition (NOR) task. A and B indicate different objects. c Representative heatmaps of the traveling paths of female mice during the test session of the NOR task. d The discrimination index (DI%) and the total time spent exploring both objects (total exploration time) during the training and test sessions in the NOR task (n = 6–10 female mice per group). e Representative heatmaps of the traveling paths of male mice during the test session of the NOR task. f The discrimination index (DI%) and the total time spent exploring both objects (total exploration time) during the training and test sessions in the NOR task (n = 6–9 male mice per group). Data are presented as the mean ± SEM and were analyzed by one-way ANOVA followed by Fisher’s LSD test, except Test DI (%) data in panel f, which were analyzed by the Kruskal–Wallis test followed by Dunn’s test. WT (Veh): wild-type littermates receiving sham treatment, AD (Veh): 5×FAD mice receiving sham treatment, AD (Bev): 5×FAD mice receiving bevacizumab treatment

Fig. 2

Bevacizumab treatment reduces capillary stalling of neutrophils and improves cerebrovascular responses to stress hormones in both sexes of 5×FAD mice. a Flow chart of the experimental design of in vivo two-photon imaging. Imaging was conducted through a cranial window in the cortex of anesthetized mice. b Representative two-photon projection images showing anti-Ly6G-488 antibody-labeled neutrophils trapped in capillaries. The vascular network was labeled with 70 kDa Texas Red dextran. Two-photon imaging was performed 15 min after the injection of both the anti-Ly6G-488 antibody and Texas Red dextran. Scale bars: 100 μm. c Enlarged images showing flowing or stalled neutrophils in cortical capillaries (red: Texas Red dextran-labeled blood vessels; Green: neutrophils labeled with anti-Ly6G-488). Scale bars: 25 μm. d The percentage of neutrophil stalls in capillaries was measured in both WT and 5×FAD mice (n = 4–5 mice per group, male and female combined). e Flow chart of the experimental design for Evans blue injection to determine blood–brain barrier integrity in 5×FAD mice. f, g Quantitative analysis of Evans blue dye leakage in the (f) cortex and (g) hippocampus of both male and female 5×FAD mice. Evans blue content was normalized by the weight of the cortex or hippocampus, followed by normalization to the plasma concentration of Evans blue (n = 7–8 mice per group, male and female mice combined). Data in d, f and g were analyzed by one-way ANOVA followed by Fisher’s LSD test. h Flow chart of the experimental design for the CBF test in 5×FAD mice. i Representative pseudocolor laser speckle flowmetry maps of CBF before and after norepinephrine injection into 5×FAD mice. j The curve graph shows the dynamic CBF changes (relative to the baseline CBF, ΔCBF) before and after norepinephrine injection (n = 7–11 mice per group). CBF changes in the WT (Veh), AD (Veh) and AD (Bev) groups at all time points were analyzed by two-way ANOVA followed by Tukey’s multiple comparison test. k Quantitative analysis of CBF changes relative to baseline CBF in response to norepinephrine injection in 5×FAD mice (n = 7–11 mice per group). The sum of CBF changes from the 48 s time point (immediately before norepinephrine injection) to the 248 s time point was averaged to represent the CBF changes in each animal, and then the results were analyzed by one-way ANOVA followed by Fisher’s LSD test. Data in (d), (f, g), and (j, k) showed similar trends of changes and effects by bevacizumab treatment in both female and male 5×FAD mice, and the data of both sexes were combined for statistical analysis. Data are presented as the mean ± SEM. CBF: cerebral blood flow, PU: units of blood perfusion. WT (Veh): wild-type littermates receiving sham treatment, AD (Veh): 5×FAD mice receiving sham treatment, AD (Bev): 5×FAD mice receiving bevacizumab treatment

Fig. 3

Bevacizumab treatment decreases Aβ levels in the brains of 5×FAD mice. a Representative images of Aβ immunofluorescence staining and quantitative analyses of Aβ fluorescence intensity and % area in the hippocampus of female mice (n = 6–8 mice per group). Scale bar: 200 µm. b Representative images of Aβ immunofluorescence staining and quantitative analyses of Aβ fluorescence intensity and % area in the sensory cortex of female mice (n = 6–8 mice per group). Scale bar: 50 µm. c Representative images of Aβ immunofluorescence staining and quantitative analyses of Aβ fluorescence intensity and % area in the hippocampus of male mice (n = 6–7 mice per group). Scale bar: 200 µm. d Representative images of Aβ immunofluorescence staining and quantitative analyses of Aβ fluorescence intensity and % area in the sensory cortex of male mice (n = 6–7 mice per group). Scale bar: 50 µm. All data are presented as the mean ± SEM and were analyzed by unpaired two-tailed Student’s t test. WT (Veh): wild-type littermates receiving sham treatment, AD (Veh): 5×FAD mice receiving sham treatment, AD (Bev): 5×FAD mice receiving bevacizumab treatment

Fig. 4

Reduced activation of glial cells in the brains of 5×FAD mice after bevacizumab treatment. a Representative images of GFAP immunofluorescence staining in the hippocampus and sensory cortex of female mice. Scale bar of the hippocampus: 200 µm. Scale bar of the sensory cortex: 50 µm. b Quantitative analyses of GFAP fluorescence % area in the hippocampus and sensory cortex of female mice (n = 6–10 mice per group). c Representative images of IBA1 immunofluorescence staining in the hippocampus and sensory cortex of female mice. Scale bar of the hippocampus: 200 µm. Scale bar of the sensory cortex: 100 µm. d Quantitative analyses of IBA1 fluorescence % area in the hippocampus and sensory cortex of female mice (n = 6–10 mice per group). e Representative images of GFAP immunofluorescence staining in the hippocampus and sensory cortex of male mice. Scale bar of the hippocampus: 200 µm. Scale bar of the sensory cortex: 50 µm. f Quantitative analyses of GFAP fluorescence % area in the hippocampus and sensory cortex of male mice (n = 6–9 mice per group). g Representative images of IBA1 immunofluorescence staining in the hippocampus and sensory cortex of male mice. Scale bar of the hippocampus: 200 µm. Scale bar of the sensory cortex: 100 µm. h Quantitative analyses of IBA1 fluorescence % area in the hippocampus and sensory cortex of male mice (n = 6–9 mice per group). All data are presented as the mean ± SEM and were analyzed by one-way ANOVA followed by Fisher’s LSD test. WT (Veh): wild-type littermates receiving sham treatment, AD (Veh): 5×FAD mice receiving sham treatment, AD (Bev): 5×FAD mice receiving bevacizumab treatment

Fig. 5

Sex-specific changes in the levels of VEGF receptors in the 5×FAD mouse brain. (a) Representative images and (b) quantitative analysis of full-length VEGFR1 (VEGFR1^FL) and soluble VEGFR1 (sVEGFR1) protein levels in the hippocampus of female 5×FAD mice (n = 7–10 mice per group). (c) Representative images and (d) quantitative analysis of VEGFR1^FL and sVEGFR1 protein levels in the cortex of female 5×FAD mice (n = 7–10 mice per group). (e) Representative images and (f) quantitative analysis of VEGFR2 protein levels in the hippocampus of female 5×FAD mice (n = 7–9 mice per group). (g) Representative images and (h) quantitative analysis of VEGFR2 protein levels in the cortex of female 5×FAD mice (n = 7–9 mice per group). (i) Representative images and (j) quantitative analysis of VEGFR1^FL and sVEGFR1 protein levels in the hippocampus of male 5×FAD mice (n = 4–5 mice per group). (k) Representative images and (l) quantitative analysis of VEGFR1^FL and sVEGFR1 protein levels in the cortex of male 5×FAD mice (n = 4–5 mice per group). (m) Representative images and (n) quantitative analysis of VEGFR2 protein levels in the hippocampus of male 5×FAD mice (n = 8–10 mice per group). (o) Representative images and (p) quantitative analysis of VEGFR2 protein levels in the cortex of male 5×FAD mice (n = 8–10 mice per group). All data are presented as the mean ± SEM and were analyzed by one-way ANOVA followed by Fisher’s LSD test. WT (Veh): wild-type littermates receiving sham treatment, AD (Veh): 5×FAD mice receiving sham treatment, AD (Bev): 5×FAD mice receiving bevacizumab treatment

Fig. 6

Reversal of transcriptomic profiles enriched in cerebrovascular and mitochondrial functions in bevacizumab-treated 5×FAD mice. a Experimental design and the workflow of the RNA-seq analysis. b Principal component analysis (PCA) plot of the RNA-seq results (n = 3 pools per group, 2–3 animals of the same sex per pool). c Volcano plots of the DEGs of the female groups, with horizontal lines at −log10 (P value) = −log10 (0.05) and vertical lines at log2 (fold change, FC) = log2 (1.5). d Volcano plots of the DEGs of the male groups. e Venn diagram depicting limited overlaps of the DEGs of the female groups. f Venn diagram depicting limited overlaps of the DEGs of the male groups. g Heatmaps of the DEGs affected by bevacizumab treatment in female 5×FAD mice. Twenty-three DEGs were upregulated in 5×FAD mice compared with wild-type mice but downregulated after bevacizumab treatment. Thirty-two DEGs were downregulated in 5×FAD mice but upregulated after bevacizumab treatment. h Heatmaps of the DEGs affected by bevacizumab treatment in male 5×FAD mice. Five DEGs were upregulated in 5×FAD mice but downregulated after bevacizumab treatment. Nineteen DEGs were downregulated in 5×FAD mice but upregulated after bevacizumab treatment. i, j Cell-type enrichment analysis of DEGs in the (i) female groups and (j) male groups. k, l Quantitative PCR validation of DEGs identified from 5×FAD versus wild-type groups, which were reversed by bevacizumab treatment in the (k) female (n = 6–8 mice per group) and (l) male (n = 7–9 mice per group) groups. All data are presented as the mean ± SEM and were analyzed by one-way ANOVA followed by Fisher’s LSD test. m KEGG pathway analysis based on GSEA was performed for RNA-seq data of the female and male 5×FAD mice. WT (Veh): wild-type littermates receiving sham treatment, AD (Veh): 5×FAD mice receiving sham treatment, AD (Bev): 5×FAD mice receiving bevacizumab treatment

Fig. 7

The levels of proteins associated with the blood–brain barrier in the hippocampus of 5×FAD mice. (a) Representative images and (b) quantitative analysis of Cgn, ZO-1, and claudin 5 protein levels in the hippocampus of female 5×FAD mice (n = 7–10 mice per group). (c) Representative images and (d) quantitative analysis of Cgn, ZO-1, and claudin 5 protein levels in the hippocampus of male 5×FAD mice (n = 8–10 mice per group). All data are presented as the mean ± SEM and were analyzed by one-way ANOVA followed by Fisher’s LSD test. WT (Veh): wild-type littermates receiving sham treatment, AD (Veh): 5×FAD mice receiving sham treatment, AD (Bev): 5×FAD mice receiving bevacizumab treatment. Cgn: cingulin, ZO-1: zonula occludens-1