Astrocyte-derived VEGF-A drives blood-brain barrier disruption in CNS inflammatory disease

Abstract

In inflammatory CNS conditions such as multiple sclerosis (MS), current options to treat clinical relapse are limited, and more selective agents are needed. Disruption of the blood-brain barrier (BBB) is an early feature of lesion formation that correlates with clinical exacerbation, leading to edema, excitotoxicity, and entry of serum proteins and inflammatory cells. Here, we identify astrocytic expression of VEGF-A as a key driver of BBB permeability in mice. Inactivation of astrocytic Vegfa expression reduced BBB breakdown, decreased lymphocyte infiltration and neuropathology in inflammatory and demyelinating lesions, and reduced paralysis in a mouse model of MS. Knockdown studies in CNS endothelium indicated activation of the downstream effector eNOS as the principal mechanism underlying the effects of VEGF-A on the BBB. Systemic administration of the selective eNOS inhibitor cavtratin in mice abrogated VEGF-A-induced BBB disruption and pathology and protected against neurologic deficit in the MS model system. Collectively, these data identify blockade of VEGF-A signaling as a protective strategy to treat inflammatory CNS disease.

Figure 1. Efficient inactivation of VEGF-A in the inflamed CNS in GfapCre:Vegfa^fl/fl mice.

(A) Mice with loxP sites flanking Vegfa exon 3 were bred with an mGfapCre line. (B) Final matings generated GfapCre:Vegfa^fl/fl mice and GfapCre:Vegfa^+/fl, Vegfa^fl/fl, and Vegfa^+/fl controls, with experimental pups representing 25% of total. (C and D) Immunoreactivity for the endothelial protein F8R and BBB junction components CLN-5 and OCLN in cerebral cortices of 12-week-old GfapCre:Vegfa^fl/fl mice and littermates (n = 3 per genotype). (E) ELISA of primary human astrocytes showing VEGF-A induction by vehicle or 10 ng/ml IL-1β, TGF-β1, or IFN-γ at 24 hours. Results are typical of cultures from 3 separate brains. (F and G) AdIL-1 (10⁷ pfu) was microinjected into cortical gray matter of 12-week-old GfapCre:Vegfa^fl/fl mice and littermates (n = 21, at least 4 per genotype); animals were sacrificed at 7 dpi, based on time course studies (see Supplemental Figures 1 and 2); and lesions were stained and examined to determine the number of GFAP⁺ and VEGF-A⁺ cells. Data are representative of 3 independent experiments. In G, individual channels from sections of the merged images are shown at right. See Supplemental Figure 1D for higher-magnification images of VEGF-A⁺GFAP⁺ cells. Scale bars: 50 μm (C and G). *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA plus Bonferroni test.

Figure 2. GfapCre:Vegfa^fl/fl animals display reduced BBB breakdown.

(A) 3-dimensionally rendered projection from cortex of a 12-week-old C57BL/6 mouse, illustrating intimate association of GFAP⁺ astrocytic endfeet with CLN-5⁺ endothelium. (B) Cortical section from the same mouse immunostained for VEGFR2, demonstrating localization to endothelium. (C–E) MVEC cultures from mouse cortex (C and E) and spinal cord (D) were treated with vehicle or 10 ng/ml VEGF-A and harvested at 24 hours, followed by immunostaining (C and D) and immunoblotting (E). Results are typical of data from 3 separate cultures. (F–I) AdIL-1–injected cerebral cortices from 12-week-old GfapCre:Vegfa^fl/fl mice and littermates sacrificed at 7 dpi (n = 21, at least 4 per genotype, as in Figure 1). (F and G) Immunostaining for CLN-5 (F) and OCLN (G). Individual channels from sections of the merged images are shown below, enlarged 1.5-fold. (H) Morphometry of OCLN and CLN-5. (I) Morphometry of albumin and fibrinogen extravasation (measures of BBB breakdown). Data are representative of 3 independent experiments. Scale bars: 20 μm (A and C); 40 μm (B); 10 μm (D); 50 μm (F and G). *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA plus Bonferroni test.

Figure 3. Reduced lymphocyte infiltration in GfapCre:Vegfa^fl/fl mice.

(A–G) AdIL-1–microinjected cortices were harvested at 7 dpi from 12-week-old GfapCre:Vegfa^fl/fl mice and littermates (at least 4 per genotype, n = 21). Shown are immunostaining and morphometry of (A and B) CD11b, (B and C) CD4, (C) CD19, (D and E) VCAM-1 and CXCL12, and (F and G) MMP-9 and MMP-3. Data are representative of findings from 3 independent experiments. (H and I) Human CNS MVECs were treated with 10 ng/ml IL-1β or 10 or 100 ng/ml VEGF-A for 24 hours, and (H) expression of VCAM-1 and ICAM-1 were determined by immunoblotting, and (I) concentrations of CC and CXC chemokines and cytokines were quantified. Data are representative of 3 experiments in separate cultures. (J and K) 12-week-old C57BL/6 mice (n = 4 per group) received 50 mg/ml MP or vehicle i.p., then intracerebral microinjection of AdIL-1 24 hours later, and were sacrificed at 7 dpi and examined to determine (J) the proportion of GFAP⁺, VEGF-A⁺, and CD4⁺ cells (measures of immunoreactivity and lymphocyte infiltration) and (K) the expression of fibrinogen and VCAM (measures of BBB breakdown). Data are representative of findings from 3 independent experiments. Scale bars: 50 μm (A, C, E, G, and all insets). *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA plus Bonferroni test (B, D, F, and I) or Student’s t test (J and K).

Figure 4. Reduced neuropathology in GfapCre:Vegfa^fl/fl animals.

(A–C) AdIL-1–injected cortices from 12-week-old GfapCre:Vegfa^fl/fl mice and littermates (7 dpi, n = 3 per genotype) were (A) stained for NeuN (also shown separately) and fibrinogen, and neuronal loss was quantified and plotted (B) per group and (C) against fibrinogen area (a measure of BBB breakdown). (D and E) Cells from A–C were (D) stained for cleaved caspase-3, and (E) the proportion of cleaved caspase-3–positive cells was determined (a measure of apoptosis). (F–I) GfapCre:Vegfa^fl/fl mice and littermates (12 weeks old, n = 3 per genotype) received a stereotactic microinjection of lysolecithin into the corpus callosum and sacrificed at 7 dpi. (F and G) Immunostaining and quantification for MBP (a measure of demyelination). Arrowheads, extent of demyelination; arrows, injection tracks. (H and I) Staining and morphometry for Olig2 (a marker of oligodendrocyte loss). Scale bars: 150 μm (A and F); 15 μm (D and H). *P < 0.05, ANOVA plus Bonferroni test. Results are representative of data from 3 independent studies.

Figure 5. Restricted clinical deficit and tissue damage in EAE in GfapCre:Vegfa^fl/fl mice.

GfapCre:Vegfa^fl/fl mice and Vegfa^fl/fl littermates (8-week-old males, at least 10 per group) were sensitized with encephalitogenic MOG_35–55/CFA. (A and B) Disease was scored using a 5-point paradigm (35) and plotted (A) as a function of time and (B) by peak score. (C–F) EAE pathology, assessing for (C) astrogliosis, (D) BBB disruption, (E) inflammatory cell infiltration, and (F) demyelination and oligodendrocyte loss. (G–K) Respective morphometry of (G) VEGF-A and GFAP, as in C; (H and I) CLN-5, Ig, and fibrinogen, as in D; (J) CD45, as in E; and (K) Olig2 and demyelination (assessed by fluoromyelin), as in F. Scale bars: 50 μm (C–E); 150 μm (F); 40 μm (D and F, insets). *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA plus Bonferroni test (A) or Student’s t test (B and G–K). Data are representative of 3 independent experiments.

Figure 6. GfapCre:Hif1a^fl/fl mice show normal VEGF-A expression and BBB opening in inflammatory lesions.

(A) Primary human astrocytes were treated with 10 ng/ml IL-1β, TGF-β1, or LPS for the indicated times, and HIF-1α induction was examined by immunoblot. (B) Human astrocytes were treated with 10 ng/ml IL-1β or vehicle for 24 hours. In IL-1β–treated cultures, HIF-1α localized to astrocytic nuclei (inset; enlarged 2-fold). Results in A and B are typical of data from 3 separate cultures of human astrocytes from different brains. (C–G) AdIL-1 was microinjected into cortical gray matter of 12-week-old GfapCre:Hif1a^fl/fl mice and littermate controls (n = 4 per genotype), and animals were sacrificed at 7 dpi. Immunostaining and morphometry were performed for (C and D) HIF-1α, (C and E) VEGF-A, and (F and G) fibrinogen (a measure of BBB breakdown) and CLN-5. In D, representative cells (arrows) are shown enlarged 2-fold in the insets (arrowheads). Results in C–G are representative of 3 independent experiments. Scale bars: 30 μm (B, D, and E); 40 μm (G). **P < 0.01, ANOVA plus Bonferroni test.

Figure 7. Effects of VEGF-A on CLN-5 are mediated via eNOS.

(A) VEGF-A signaling pathways. (B) Human CNS MVECs were pretreated with 10 ng/ml anti-VEGFR2 or IgG for 2 hours, then VEGF-A or vehicle for 24 hours, and CLN-5 and OCLN levels were determined by immunoblot. (C and D) 12-week-old C57BL/6 mice (3 per group) received cortical coinjection of 20 ng VEGF-A or vehicle, plus 2 μg anti-VEGFR2 or IgG, and were sacrificed at 24 hours for determination of CLN-5 and fibrinogen area. (E and F) Immunoblot for CLN-5 in VEGF-A–treated human CNS MVECs also (E) treated with 5 μM U73122, 10 ng/ml VEGF-A, 5 μM U73343, or 3 μM PP2 or (F) pretreated with 2 μM cavtratin. (G) Immunoblot for CLN-5, eNOS, and p38 (representative of other signaling components) in human MVECs nucleofected with sham, nontargeting (NT), or eNOS siRNA. CLN-5/p38 ratio is shown below. (H–M) 12-week-old C57BL/6 mice (3 per group) received 2.5 mg/kg cavtratin or vehicle i.p., then 30 minutes later received cortical injection of (H–K) 60 ng VEGF-A, followed by sacrifice at 24 hours, or (L and M) 10⁷ pfu AdIL-1, followed by daily cavtratin until sacrifice at 7 dpi. Immunostaining and morphometry for NeuN, fibrinogen, CLN-5, and CD3 are shown. Scale bars: 20 μm (C); 150 μm (H, I, and L). *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA plus Bonferroni test (D) or Student’s t test (J, K, and M). Results are representative of at least 3 separate cultures (B and E–G) or 3 independent experiments (C, D, and H–M).

Figure 8. Reduced clinical severity and neuropathology of EAE in cavtratin-treated mice.

Mice (8-week-old male C57BL/6, 10 per group) were sensitized with MOG_35–55/CFA, then from onset of weight loss (7 dpi) treated for 7 days with 2.5 mg/kg/d i.p. cavtratin or vehicle. (A and B) Disease was scored using a 5-point paradigm (35) and plotted (A) as a function of time and (B) by peak score. (C) Lumbar spinal cord sections were subjected to immunostaining at 21 dpi to assess BBB breakdown (top) as well as demyelination and oligodendrocyte loss (bottom). (D–F) Morphometry of (D) fibrinogen (a measure of BBB breakdown) and CLN-5, (E) CD45 (a measure of inflammatory cell infiltration), and (F) Olig2 and demyelination (assessed by fluoromyelin). These findings resembled the phenotype of EAE in GfapCre:Vegfa^fl/fl mice (see Figure 5). Scale bars: 50 μm (C, top); 150 μm (C, bottom); 40 μm (C, insets). *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA plus Bonferroni test (A) or Student’s t test (B and D–F). Data are representative of 3 independent experiments.

Figure 9. Proposed mechanism underlying BBB disruption in CNS inflammatory lesions.

The inflammatory cytokine IL-1β is produced early in lesion pathogenesis by mononuclear phagocytes and binds to its receptor on astrocytes adjacent to CNS microvessels, leading to Vegfa induction. VEGF-A released from the astrocyte binds to VEGFR2 on CNS MVECs, activating eNOS-dependent downregulation of Cln5 and Ocln and leading to disruption of endothelial tight junctions and BBB breakdown. These events facilitate parenchymal entry of autoantibodies, plasma proteins, and lymphocytes, leading to inflammation, edema, and excitotoxicity and exacerbation of neuropathology. How these astrocyte-driven effects interact with pericyte effects at the BBB is not known.