"VSports在线直播" VEGF-A₁₆₅b Is an Endogenous Neuroprotective Splice Isoform of Vascular Endothelial Growth Factor A in Vivo and in Vitro

Materials and Methods

All reagents were sourced from Sigma-Aldrich (Dorset, UK) unless otherwise stated. Antibodies were sourced from R&D Systems (Carlsbad, CA) [αVEGF-A₁₆₅b (monoclonal antibody [mAb] 3045), α-caspase-3, αTrkA, and αNF-200] or Cell Signaling Technology (Danver, MA) [phospho-p44/42 mitogen-activated protein kinase (MAPK) (9106) and p44/42 MAPK (9102)]. Recombinant human (rh) VEGF-A₁₆₅b was provided by Philogene (New York, NY) or from R&D Systems. Computer-aided analysis of immunohistochemistry (IHC), retinal Fluorogold staining, and cytotoxicity was performed using Macintosh computers running public domain ImageJ version 1.46 plus Cell Counter Plugin (NIH, Bethesda, MD; available at http://rsb.info.nih.gov/nih-image).

Results

IHC of human brain samples demonstrated that VEGF-A₁₆₅b was strongly expressed in the human hippocampal region (Figure 1A), throughout all three pyramidal regions (CA1, CA2, and CA3) (Figure 1B), and the dentate gyrus (Figure 1C), with neurons also being expressed in occasional neurons scattered through the cortex. The use of a mouse IgG at the same concentration or preincubation of the primary antibody with recombinant VEGF₁₆₅b demonstrated no significant staining (Figure 1, D and E). Preincubation with recombinant VEGF₁₆₅a, however, did not block the staining (Figure 1F). To quantify expression, protein was extracted from rat brains and VEGF-A_xxxb and total VEGF-A levels measured by ELISA. VEGF-A_xxxb levels averaged 45% and 41% of total VEGF-A in the cortex and hippocampus, respectively (Figure 1, G and H).

Because VEGF-A_xxxa splice variants are known to be neuroprotective, we assessed the effectiveness of VEGF-A_xxxb splice variants as neuroprotective agents in hippocampal neurons, central nervous system neurons that are particularly vulnerable to damage.

RhVEGF-A₁₆₅b reversed glutamate-induced hippocampal neuronal death in a concentration-dependent manner (Figure 2, A and B) and was no less potent than galanin, which has previously been described as a potent hippocampal neuroprotective agent¹⁶ (Figure 2B). At 10 nmol/L, the maximum VEGF-A₁₆₅b concentration used, cell death was the same as untreated cells, revealing complete inhibition of excitotoxicity. VEGF-A₁₆₅b is a weak partial agonist at VEGF-A receptor 2 (VEGFR2, also known as Flk1) in vitro.²⁴ In hippocampal neurons, the neuroprotective action of VEGF-A₁₆₅b was dependent on VEGFR2 activation (Figure 2C), as has been described for VEGF-A₁₆₅a.²⁵ The partial reversal of excitotoxic cell death seen in the presence of VEGF-A₁₆₅b was blocked by the 100 nmol/L VEGFR blockers PTK787 (PTK, blocks both VEGFR1 and VEGFR2²⁶) and 10 nmol/L ZM323881 (ZM, blocks VEGFR2²⁷) but was unaffected by SU5146 (SU, at 100 nmol/L has specificity for VEGFR1¹²). Hippocampal neurons in culture expressed VEGFR2 (Figure 2D). Hippocampal neuroprotection elicited by VEGF-A₁₆₅b was dependent on downstream signaling of the VEGFR2 through the p44/42 MAPK pathway (Figure 3A), as it was also blocked by the MEK2 inhibitor, 10 μmol/L PD98059, as is also the case for VEGF-A₁₆₅a.²⁸ Neuroprotection was not affected by blockade of either p38 MAPK by SB203580 or phosphatidylinositol 3–kinase (PI3K) with LY294002 (Figure 3, A and B). Treatment of 50B11 neurons with VEGF-A₁₆₅b significantly increased p44/42 MAPK phosphorylation (Figure 3C), which was blocked by the inhibitor PD98059 (Figure 3D).

VEGF-A₁₆₅b also exerted a neuroprotective action on retinal neurons in vivo (Figure 4). In retinal ischemia a reduction in retrograde transport of fluorescent tracer was seen (Figure 4A). This ischemic neuronal loss was reversed by prior intraocular VEGF-A₁₆₅b treatment (Figure 4A). To determine whether this was due to apoptosis, retinae were stained for active caspase-3 (Figure 4B). Uninjected and HBSS-injected ischemic eyes revealed significant retinal neuronal loss through apoptosis, as assessed by Fluorogold labeling of live neurons (Figure 4, C and D) and caspase-3 staining respectively (Figure 4E). The reduction in neuronal retrograde transport was significantly reversed by VEGF-A₁₆₅b (Figure 4, A, C, and D), when compared with the uninjected contralateral eye and reduced apoptosis in both RGCs and inner nuclear layer cells (Figure 4, B and D).

VEGF-A₁₆₅a has been found to be cytoprotective against injury induced by a variety of cellular insults to neurons.²⁹ To determine whether neuroprotection was confined to central neurons or also affected peripheral neurons, cultured primary DRG neurons were investigated. The addition of 1 nmol/L VEGF-A₁₆₅b significantly increased the length of neurite outgrowth during 24 hours from 41.7 ± 2.2 μm to 73.8 ± 20 μm (P < 0.05, U-test) but did not affect the number of DRG neurons in treated and control cultures, indicating a neurotrophic action. A total of 2.5 nmol/L VEGF-A₁₆₅b significantly prevented an increase in activated caspase-3 in primary DRG neurons induced by 24 hours of treatment with the chemotherapeutic oxaliplatin. Treatment with oxaliplatin alone (5, 10, or 20 μg/mL) significantly increased the percentage of caspase-3–positive DRG neurons compared with untreated neurons, and VEGF-A₁₆₅b prevented this (Figure 5, A and B).

To determine whether VEGF₁₆₅b was acting as an endogenous neuroprotective factor for DRG, we treated cultured primary rat DRG cells with an antibody to VEGF₁₆₅b with or without the toxic chemotherapeutic oxaliplatin. In the absence of oxaliplatin, neither anti-VEGF₁₆₅b antibody (means ± SEM, 98% ± 14%, 103% ± 12%, and 95% ± 6.6% of control, respectively) nor anti–pan VEGF antibody (93% ± 10%, 98% ± 16%, and 95% ± 12%) had any effect at 1, 10, or 100 μg/mL, respectively. In contrast, when combined with oxaliplatin anti-VEGF₁₆₅b treatment resulted in a significant increase in the proportion of dead cells after 24 hours compared with mouse IgG (Figure 5C). As a more sensitive measure we also used the PrestoBlue cell viability assay where absorption is proportional to mitochondrial metabolic activity and can be taken as a measure of cell viability, in which oxaliplatin also resulted in significant reduction in cell viability, which was further reduced by the addition of anti-VEGF₁₆₅b (Figure 5D).

Staining of cultured DRG neurons for VEGFR2 revealed strong expression on the cell membrane (Figure 6A). To investigate the receptor mediating the neuroprotective response to VEGF-A₁₆₅b, DRG cultures were treated with 10 nmol/L specific VEGFR2 inhibitor ZM323881 or vehicle during oxaliplatin treatment and stained for NeuN and activated caspase-3 (Figure 6B). Treatment with 20 μg/mL of oxaliplatin, either alone or with ZM323881, induced a significant increase in the percentage of activated caspase-3–positive, NeuN-positive cells compared with media (51.4% ± 3.0% and 55.7% ± 1.9% versus 37.8% ± 1.3% and 42.5% ± 2.6%, respectively). Concurrent treatment with 2.5 nmol/L recombinant human VEGF-A₁₆₅b without ZM323881 significantly reduced the percentage of activated caspase-3–positive, NeuN-positive cells compared with vehicle (35.4% ± 3.1%) (Figure 6B), and this effect was blocked by treatment with the VEGFR2 inhibitor ZM323881 (54.3% ± 3.3%). To further rule out a role for VEGFR1, cells were treated with placenta growth factor (PlGF), which activates only VEGFR1. This neither affected the oxaliplatin-induced cell death nor prevented its amelioration by VEGF₁₆₅b (Figure 6B). As an additional control to determine whether VEGF₁₆₅b was acting through VEGFR2, cells were treated with oxaliplatin and VEGF₁₆₅b in the presence of either mouse IgG as a control or the VEGFR2 neutralizing antibody DC101. Although VEGFR2 blockade by itself did not increase neuronal caspase activity or affect the oxaliplatin-induced effect, it prevented the VEGF₁₆₅b-mediated cytoprotection (Figure 6C).

To determine whether systemic administration of VEGF-A₁₆₅b could be neuroprotective, we used a mouse model of peripheral traumatic nerve injury that results in activation of injury-response genes in DRG neurons, such as galanin,³⁰ and ATF3 (Figure 7, A and B). Biweekly treatment with rhVEGF-A₁₆₅b reduced the intensity of neuronal ATF3 expression in L3/4 DRG compared with sham surgery controls after 10 days (Figure 7C).

Discussion

We report that, like VEGF-A₁₆₅a, the splice isoform VEGF-A₁₆₅b is neuroprotective for central and peripheral neurons, in vitro and in vivo, and mediates this neuroprotective effect through VEGFR2 and MEK1/2 activation. In addition to its neuroprotective effects, VEGF-A₁₆₅b also has neurotrophic actions on neurons.

VEGF-A₁₆₅a has been found to play a key role in neuronal protection, in addition to its actions on the vasculature, directly protecting motoneurons under conditions of hypoxia, oxidative stress, and serum deprivation.³¹ Disturbance of VEGF-A expression contributes to the development of amyotrophic lateral sclerosis in humans.²⁹ VEGF-A and its receptors are also expressed in central and peripheral nervous system support cells, such as astrocytes and Schwann cells, thereby also contributing to neuronal survival and growth.^10,32

In the central nervous system, VEGF-A₁₆₅a protects hippocampal, cortical, and cerebellar granule neurons against numerous insults^33–36 through VEGFR2, signaling through activation of multiple intracellular pathways, including phospholipase C, PI3K, and MEK1/2,³² whereas effects on supporting cells, such as Schwann cells, and astrocytes, are generally mediated through VEGFR1.¹⁰ In contrast, our data indicate that neuroprotection by VEGF-A₁₆₅b in hippocampal neurons, although also mediated through VEGFR2, involves MEK1/2 but not either PI3K or p38 MAPK. The VEGF-A_xxx bisoforms compete for and inhibit VEGF-A_xxxa binding at VEGFR2,^5,7,37 but the VEGF-A_xxxb isoforms are not simply competitive inhibitors of the VEGFR2 because binding of VEGF-A₁₆₅b results in differential tyrosine residue phosphorylation of VEGFR2.²⁴ Neuroprotective and neurotrophic actions by VEGF-A₁₆₅a may also involve the VEGF co-receptor neuropilin-1 (NP-1)³²; VEGF-A₁₆₅b binds weakly to NP-1 and does not require NP-1 binding to phosphorylate and activate VEGFR2.²⁴ Our data support the conclusion that the differential VEGFR2 phosphorylation and MEK1/2 activation exerted by VEGF-A₁₆₅b is sufficient to protect central and peripheral neurons, without NP-1 binding. Our data on both central nervous system and peripheral sensory neurons indicate that the neuroprotective mechanism of VEGF-A₁₆₅b is like that ascribed to VEGF-A₁₆₅a in that neuroprotection occurs through prevention of caspase-3 induction.³²

In the eye, endogenous VEGF-A is a survival factor for RGCs, protecting against ischemia-reperfusion injury³⁸ and preventing neuronal apoptosis without the necessity for NP-1 binding because VEGF-A₁₂₀a (which lacks the NP-1 binding domain³⁹) also exerted neuroprotective effects. Our findings indicate that exogenous VEGF-A₁₆₅b can reduce both RGC and inner nuclear cell loss through VEGFR2 activation. The loss of neuronal retrograde transport is seen clinically in diabetic retinopathy as cotton wool spots under fundus examination, and the loss of ganglion cells contributes to vision loss in glaucoma.⁴⁰ We have previously found that VEGF-A₁₆₅b is cytoprotective for endothelial and retinal pigmented epithelial cells,⁴¹ whereas it is antiangiogenic in the eye.⁴² Prevention of the ischemia-induced damage to neurons in the retina by VEGF-A₁₆₅b would therefore also be a substantial advantage in therapeutic approaches for diabetic retinopathy and glaucoma.

In hippocampal and peripheral sensory cultured neurons, VEGF-A₁₆₅b also enhances neurite outgrowth, demonstrating that VEGF-A₁₆₅b exerts neurotrophic effects, also through VEGFR2. In peripheral sensory and autonomic neurons, VEGF-A₁₆₅a enhances neuronal neurite outgrowth in culture.^11,32,43 VEGFR2 is expressed in both the central and peripheral nervous systems but often in different neuronal populations from VEGF-A proteins.^44,45 These different distributions suggest that VEGF-A isoforms exert paracrine actions on neurons through which neuroprotective and neurotrophic effects may be mediated. We demonstrate that VEGF-A₁₆₅b is a major VEGF-A splice variant commonly found in central and peripheral neurons, where it is well placed to exert such paracrine effects. VEGF-A₁₆₅b is also expressed in human skin, prostate, and kidney, among other tissues,^5,46 where it could also exert paracrine effects on peripheral sensory neurons.

VEGF-A₁₆₅b has clear inhibitory effects on tumor growth^5,14,47 but does not result in hypotension, angiogenesis, or proteinuria.^5,6,48 Our data indicate that VEGF-A₁₆₅b has neuroprotective effects similar to those ascribed to VEGF-A₁₆₅a, on central and peripheral neurons. VEGF-A–dependent neovascularization is key in the pathophysiology of many conditions, and anti–VEGF-A therapies have entered clinical practice in oncology⁹ and ophthalmology.⁸ After preclinical studies of VEGF-A administration in neurodegenerative disease,^{11,29,49–51} VEGF-A supplementation is now under trial for treatment of neuronal degenerative diseases, for instance in diabetic neuropathy.⁵² Although effective in some patients, there has to be some concern about the safety profile of these strategies in relation to the potential compromise of nonendothelial tissue and cell function in which VEGF-A has been found to have protective properties, particularly neurons¹⁰ and podocytes.⁵³ We report that VEGF-A_xxxb can exert similar neuroprotective and neurotropic effects with VEGF-A_xxxa both centrally (eg, hippocampal neurons in culture, and retinal neurons in vivo) and peripherally.

With the requirement that therapy for pathological conditions in which VEGF-A has been implicated should not adversely affect the function of the normal vasculature or other cell types, we suggest that VEGF-A₁₆₅b may be an alternative therapeutic agent in neurodegenerative conditions with fewer adverse vascular effects.