Astrocyte-derived VEGF increases cerebral microvascular permeability under high salt conditions

High salt diet increases the cerebral microvascular permeability by down-regulating expression of tight junction (TJ) proteins

Increased cerebral microvascular permeability is the key to many peripheral inflammatory factors entering the CNS, leading to a variety of diseases. Inspired by previous epidemiological data, we studied the effect of HSD on the CNS. First, we analyzed the permeability of the mice cortical microvessels with a diameter of 20-40um in vivo by two-photon imaging. We found that intravascular dye leaked into extravascular space in the HSD group (Figure 1A). Next, we used the Evans blue dye (EBD) to assess the effect of HSD rats on the permeability of whole cerebral vessels. As shown in Figure 1B, the EBD leakage was significantly increased in the HSD group compared to the normal diet group (Day-180: P<0.05). Thus, we analyzed the expression of TJ-associated proteins in the brain tissues from rats fed HSD or normal salt diet (NSD), by using immunofluorescence staining and western blotting (Figure 1C, 1D). No significant changes were observed in the expression of ZO-1, occluding, and claudin-5 between the HSD and NSD groups, when high salt was fed for 30 days (Figure 1D, 1E). However, HSD dramatically decreased the expression of ZO-1, occludin, and claudin-5 after 180 days of high salt feeding (Figure 1D, 1E). Immunofluorescence double staining confirmed that the ZO-1 expression was down-regulated in CD31-positive microvascular in HSD rats fed for 180 days (Figure 1C).

Figure 1. HSD increases cerebral microvascular permeability. (A) Representative images of the mice cerebral cortical microvessels and dye leakage at 10 min after injection of Rhodamine B isothiocyanate-dextran (n=3 per group). (B) Evans blue leakage analyzed in rats fed with selected diets for 180 days (n=3 per group). (C) ZO-1 (green) co-stained with CD31 (red), a microvascular endothelia marker, and DAPI (blue; n=6 per group) in the rats brain slices. (D, E) Expression of Occludin, Claudin-5 and ZO-1 in the brain tissues from rats analyzed by immunoblotting. (F) Systolic blood pressure (SBP) of rats fed with NSD or HSD for 180 days. *P<0.05, **P<0.01, compared with NSD; at least three separate experiments were conducted, means ± SD.

Moreover, systolic blood pressure (SBP) was measured using a tail-cuff system in rats fed with normal or high-salt diet for 180 days. As shown in Figure 1F, there was no significant difference in the SBP between the normal and HSD groups.

High salt activates astrocytes and increases expression of VEGF

Due to the strong effect on increased vascular permeability, VEGF is also called vascular permeability factor, which is mainly secreted by astrocytes [20]. Therefore, we examined the protein expression of VEGF and the astrocyte marker GFAP in the brain. Interestingly, immunofluorescence double staining showed that HSD up-regulated the expression of VEGF and GFAP in the rat brain (Figure 2A); this was also confirmed by western blotting (Figure 2B, 2C).

Figure 2. High salt activates astrocytes and up-regulates VEGF. (A) Representative images of double immunofluorescence staining for GFAP(red) and VEGF(green) in the brain specimens of rats; DAPI is stained in blue(n=5 per group). (B, C) Expression of GFAP and VEGF in the brain tissues from rats analyzed by western blotting and densitometry (n=5 per group). (D) Immunofluorescence staining of GFAP and AQP4 in red. (E, F) Expression of AQP4 and GFAP analyzed by western blotting in NaCl-treated astrocytes; *P<0.05, **P<0.01, compared with NS. At least three separate experiments were conducted; means ± SD.

To investigate the effect of a high salt environment on astrocytes, primary astrocytes isolated from rat brain were treated with 40mM NaCl for 24h. Immunofluorescence staining and western blotting demonstrated an enhanced expression of GFAP (Figure 2D–2F). Aquaporin-4 (AQP-4) is a cell membrane protein that is highly expressed on the surface of astrocytes’ end-feet [21]. As shown in Figure 2D, the AQP4 expression on the end-feet changed from flat and uniform to slender and multipolar. Western blotting revealed that high salt treatment decreased the expression of AQP-4 (Figure 2E, 2F).

High salt up-regulates expression of VEGF in astrocytes by activating NFκB/MMP-9 pathway

Nuclear factor-κB (NF-κB) is a nuclear transcription factor that regulates the expression of pro-inflammatory genes [22]. The activation of NF-κB is considered to be part of a stress response as it is activated by a variety of stress stimuli. Phosphorylation of p65 NF-κB subunit and its nuclear translocation reflect the degree of NF-κB activation [22]. Our results show that the ratio of p-p65/p65 increases and phosphorylated p65 translocates from the cytoplasm to the nucleus in NaCl-treated astrocytes (Figure 3A–3C).

Figure 3. Increased expression of VEGF is mediated by NFκB/MMP-9 pathway. (A, B) HS increases protein levels of VEGF, MMP9 and p-p65. (C) Immunofluorescence staining of NaCl (40mM, 24 h)-induced p65 phosphorylation and nuclear translocation; (D) HS upregulates mRNA expression levels of VEGF. (E) SN50 antagonized NaCl-induced up-regulation of VEGF and MMP-9 protein expression. *P<0.05, **P<0.01, compared with NS, n=3, means ± SD.

Matrix metalloprotease-9 (MMP-9) is involved in the homeostasis of the extracellular matrix, and participates in angiogenesis by releasing VEGF [23]. We found that the MMP-9 and VEGF expression was significantly increased under high-salt conditions (Figure 3A, 3B, 3D). By using the SN50 inhibitor of p65 nuclear translocation, we found that SN50 decreased the protein expression of VEGF and MMP9 (Figure 3E), indicating that HSD induces VEGF and MMP-9 through the p65 NF-κB-dependent pathway.

Astrocyte-derived VEGF is the main factor causing tight junction destruction in endothelial cells under high salt condition

To determine whether the astrocyte-derived VEGF is involved in HS-induced tight junction destruction, mice brain microvascular endothelial cells, bEnd.3, were co-cultured with primary rat astrocytes in a transwell culture plate divided into 6 groups: 1. Endothelial cells cultured alone in normal medium (NS+EC group); 2. Endothelial cells co-cultured with astrocytes in normal medium (NS+EC+AS group); 3. Endothelial cells cultured in high-salt medium (40mM NaCl for 24h) alone (HS+EC group); 4. Endothelial cells co-cultured with astrocytes in high-salt medium (HS+EC+AS group); 5. Astrocytes pre-treated with 40mM NaCl(24h), and cultured with endothelial cells in conditioned medium (EC+CM group); 6. VEGF neutralizing antibodies (NA; 0.1μg/mL) added into the HS+EC+AS group, and cultured 24h (HS+EC+AS+NA group). Immunofluorescence staining was performed to analyze the morphology of ZO-1, one of the TJ proteins. As shown in Figure 4A, the structure of ZO-1 was intact and normal when endothelial cells were exposed to NaCl directly. However, the expression of ZO-1 was destroyed when endothelial cells were co-cultured with astrocytes in high salt conditions, or treated with conditioned medium from high salt-cultured astrocytes. Western blotting showed that the expression of occludin, claudin-5, and ZO-1 decreased in the HS+EC+AS and EC+CM groups (Figure 4B, 4C). Analysis of the permeability of tight junctions using NaF showed similar results to western blotting (Figure 4D).

Figure 4. Astrocyte-derived VEGF mediates HS-induced BBB breakdown. (A) Representative double immunofluorescence staining of ZO-1⁺ bEnd.3 endothelium. Endothelial cells were cultured alone, or co-cultured with primary rats′ astrocytes, and treated with NS, HS, conditioned medium, or VEGF neutralizing antibody (NA). (B, C) Western blotting analysis of Occludin, Claudin-5, and ZO-1 in endothelial cells. (D) Permeability of tight junctions measured using NaF; *P<0.05, **P<0.01, compared with EC+NS. At least three separate experiments were conducted; means ± SD.

Effect of VEGF on tight junction is mediated via ERK and eNOS pathway

Next, we analyzed the molecular mechanism involved in the NaCl-mediated down-regulation of TJ protein expression. It has been reported that high-salt conditions increase eNOS phosphorylation, and activate thep38/MAPK pathway [24, 25]. Moreover, VEGF activates PLCγ, which stimulates MEK-MAPK and p38in endothelium, and induces expression of eNOS via PLCγ and PI3K [26]. Thus, we examined the phosphorylation status of ERK, eNOS, and p38, which represents the degree of activation, by western blotting. As shown in Figure 5, a dramatic up-regulation of p-ERK/ERK and p-eNOS/eNOS levels was observed in endothelial cells treated with conditioned medium from high salt-cultured astrocytes, or with VEGF recombinant rat protein. Anti-VEGF neutralizing antibody (NA) partially reversed the up-regulation of p-ERK/ERK and p-eNOS/eNOS levels.

Figure 5. Effect of VEGF on tight junction is mediated via ERK/eNOS. (A, B) Protein levels of ERK/pERK, eNOS/pSer¹¹⁷⁷, eNOS, and p38/p-p38 in endothelial cells; *P<0.05, **P<0.01, compared with vehicle. At least three separate experiments were conducted; means ± SD.

VEGF^hi/+ mice exhibit increased cerebral microvascular permeability

To corroborate the above results, we constructed VEGF^hi/+ mice with up-regulated astrocyte-derived VEGF by transferring a plasmid with GFAP promoter and mVegfa open reading frame into a fertilized egg (Figure 6A). The VEGF^hi/+ mice showed no significant changes in the expression of ZO-1, but a down-regulation of occludin and claudin-5 (Figure 6B). In addition, in the VEGF^hi/+ mice cortical, the fluorescent dye solution leaked from microvessels significantly more than in the wildtype mice (Figure 6C).

Figure 6. Increased cerebral microvascular permeability in VEGF^hi/+ mice. (A) Schematic diagram of VEGF^hi/+ mice generation. (B) Immunoblotting showing VEGF, Occludin, Claudin-5 and ZO-1expression in VEGF^hi/+ and wildtype mice. (C) Cerebral cortical micro-vessel and dye leakage at 5 and 10 min after injection of Rhodamine B isothiocyanate-dextran; n=5, *P<0.05, **P<0.01, compared with NSD. At least three separate experiments were conducted; means ± SD.

Blocking VEGF attenuates HSD-induced disruption of tight junction

Next, we tested if blocking VEGF using the bevacizumab antibody could reduce the destruction of tight junctions induced by HSD. Rats in the HSD group were intraperitoneally injected with10 mg/kg bevacizumab or an equal volume of normal saline (continuous administration for 30days). We found that the HSD-induced disruption of tight junction proteins in the rat brain could be attenuated by blocking VEGF with bevacizumab (Figure 7A–7C).

Figure 7. Blocking VEGF attenuates disruption of tight junctions induced by HSD. (A) Representative images showing double immunofluorescence staining of ZO-1 in green and DAPI in the brain specimens of rats. (B, C) Expression of Occludin, Claudin-5 and ZO-1 determined by western blotting; n=5, *P<0.05, **P<0.01, compared with NSD. At least three separate experiments were conducted; means ± SD.

Animal experiments

8-week-old specific pathogen-free male Sprague-Dawley rats (weighing 250-280g), 1-day-oldspecific pathogen-free Sprague-Dawley rats, and 8-week-old male C57BL/6 mice weighing 20–25 g) were obtained from the experimental animal center of Sun Yat-sen University, and kept in a pathogen-free room. Transgenic male mice overexpressing VEGF against a C57BL/6 background were obtained from Cyagen Biosciences, Inc (Suzhou, China). The animal use protocols were reviewed and approved by the Institutional Animal Care and Use Committee of Sun Yat-sen University (Guangzhou, China).

In the rat experiment, we randomly assigned 40 male Sprague-Dawley rats aged 8 weeks to 2 groups. Group 1 received normal chow and tap water (normal salt diet, NSD) for 30 and 60 days. Group 2 received a sodium-rich chow containing 8% NaCl and 1% saline (high salt diet, HSD) for 30and 180 days. Group 2 was divided into two sub-groups: the intervention group receiving 10 mg/kg bevacizumab (Avastin, 25mg/ml, Roche, USA) at day 150, or the control group receiving an equal volume of saline solution intraperitoneally injected for the remaining 30 days.

In the mice experiment, we randomly assigned 10 C57BL/6 aged 8 week-old mice to LSD (< 0.1% NaCl) and tap water, or HSD (4% NaCl) and 1% saline. After mice had been on their specified diets for 90 days, they were used in two-photon imaging experiment.

Transgenic vector

We inserted the Glial fibrillary acidic protein (GFAP) promoter and mVegfa open reading frame in a regular plasmid to construct a gene expression vector. By pronuclear microinjection, we injected the constructed vector into the fertilized egg, which was then implanted into a surrogate mother to obtain the desired offspring through germline transmission.

Cell culture

Primary astrocytes were isolated from 1-day-old Sprague-Dawley rats as described before [16, 40]. The immortalized murine brain microvascular endothelial cell line bEnd.3 (ATCC, Manassas, VA, USA) was used for the in vitro assay. Cells were cultured alone or co-cultured in transwell plates. All experiments were conducted using 80%–85% confluent cells. In the high salt group, the plated cells were incubated with serum-free DMEM medium for1 h, followed by a 24 h incubation in serum-free DMEM containing 40 mM NaCl (Sigma-Aldrich). Recombinant VEGF (CST, #5211), anti-VEGF neutralization antibody (R&D, #AF564) or SN50 (MCE, #213546-53-3) were added into the medium as indicated.

In vivo two-photon imaging

Mice were anesthetized, operated on to construct a thin cranial window (3 mm in diameter), fixed on a custom-fabricated metal frame and placed under a two-photon laser scanning microscope (Leica, Germany)equipped with a water-immersion objective lens (25×). Data acquisition and laser scanning were performed using Leica Application Suite Advanced Fluorescence 2.5 software, at a wave length of 860 nm. To monitor the cerebral microvascular permeability using detection of leaked dyes, Rhodamine B isothiocyanate-dextran (1.4% in saline, 70 kDa molecular weight, Sigma-Aldrich) was injected intravenously to visualize the brain vasculature. We selected the red fluorescence channel for detection, and calculated the average fluorescence intensity in the extravascular compartment. Images of the XYZ stacks (512 × 512 pixels) were collected to a depth of 200 μm (2-μm step size) below the cortical surface, at 5and 10 min after the injection. We defined the vessels with a diameter of 20-40um as microvascular.

Evans blue dye (EBD) extravasation

In brief, a 4% solution of EBD (4 ml/kg of body weight) was injected intraperitoneally and allowed to circulate for 2 hours at day 180 before execution. Under deep anesthesia, rats were perfused with saline until colorless fluid outflowed from the right atrium. Then, ischemic cerebral hemispheres were collected after decapitation. The brain specimens were weighed (wet weight of each sample was 50 mg), homogenized in 1 ml of 50% trichloroacetic acid, and centrifuged at 15,000× g for 20 minutes. Then, 0.5 ml of the resultant supernatant was added to 1.5 ml of anhydrous ethanol for a colorimetric assay using a fluorescence spectrophotometer (Ex620 nm, Em680 nm) to determine the EBD concentration. The EBD content (per mg of wet weight) within the brain tissue was used to determine the BBB permeability rate of EBD.

Noninvasive blood pressure measurement

The computerized tail-cuff system (BP2010A, sofron, China) was used to perform the non-invasive systolic blood pressure measurement. Measurements were conducted at day 180before execution.

Immunofluorescence staining

Frozen sections were incubated with control 1.5% goat serum for 30 min, and then stained using mouse monoclonal anti-CD31 (1:100, Thermo Fisher, #MA1-81051), rabbit polyclonal anti-ZO1 (1:100, Thermo Fisher, #40-2200), mouse monoclonal anti-GFAP (1:1000, CST, #3670), rabbit polyclonal anti-VEGF (1:100; Millipore, #07-1420), mouse monoclonal anti-p65 (1:1000, CST, #6956), or rabbit monoclonal anti-p-p65 (1:1000, CST, #3033) primary antibodies, followed by fluorescent conjugated secondary antibody Alexa Fluor 647goat anti-mouse IgG (1:400, Thermo Fisher, A32728) or Alexa Fluor 488-conjugated goat anti-rabbit IgG (1:400, Thermo Fisher, A32731).

Cells were plated on poly-L-lysine (PLL)-treated cover slips, fixed in 4% paraformaldehyde for 20 min, and blocked with 10% goat serum in PBS. Slides were incubated overnight in a humidified chamber at 4 °C with the following primary antibodies: rabbit polyclonal anti-AQP4 (1:200, CST, Thermo Fisher, #PA5-77716) or rabbit polyclonal anti-ZO1 (1:100, Thermo Fisher, #40-2200). After primary antibody incubation, samples were washed, and incubated with fluorescent secondary antibody Alexa Fluor 647-conjugated goat anti-rabbit IgG (1:400, Thermo Fisher, A32733)for 1 hour. The cells were counterstained by DAPI. Images were captured with a fluorescence microscope (IX71, Olympus).

Western blot analysis

Cells and brain tissues were homogenized with RIPA buffer containing protease inhibitor cocktail (Sigma). Supernatants were collected by centrifugation at 12,000 g for 20 min at 4°C. The concentration of total protein in each sample was quantitated by BCA method (Thermo). Equal amounts of proteins were separated by SDS-PAGE, and then transferred to membranes (Immobilon-P; Millipore) for 1 hour. The membranes were blocked for 1 hour and incubated with the following primary antibodies: mouse monoclonal anti-GFAP (1:1000, CST, #3670), rabbit polyclonal anti-VEGF (1:1000; Millipore, #07-1420), rabbit polyclonal anti-ZO1 (1:1000, Thermo Fisher, #40-2200), mouse monoclonal anti-Occludin (1:1000, Thermo Fisher, #33-1520), rabbit polyclonal anti-Claudin5 (1:1000, Thermo Fisher, #PA5-37527), mouse monoclonal anti-MMP9 (1:500, Thermo Fisher, #MA5-14228), mouse monoclonal anti-p65 (1:1000, CST, #6956), rabbit monoclonal anti-p-p65 (1:1000, CST, #3033), rabbit monoclonal anti-p38 (1:1000, CST, #8690), rabbit monoclonal anti-p-p38 (1:1000, CST, #4511), rabbit polyclonal anti-ERK (1:1000, CST, #9102), rabbit polyclonal anti-p-ERK (1:1000, CST, #9101), rabbit polyclonal anti-eNOS (1:1000, CST, #9572), or rabbit polyclonal anti-p-eNOS^thr495 (1:1000, CST, #9574) antibody in blocking solution at 4°C overnight. After triple washing, the membranes were incubated with the corresponding secondary antibody (1: 5000) at room temperature for 1 hour. In order to confirm equal protein loading, the blots were also incubated with antibodies against GAPDH (1:5000, CST, #8884). The target protein was detected using the ECL-Plus kit (Millipore), and ImageJ software was used to analyze the relative protein densities.

Measuring the permeability of co-cultured cells by using NaF

NaF was used for permeability studies as described before [41]. Astrocyte/endothelial cells co-cultured in the transwell plate were used to measureapical to basolateral (A to B) transport of NaF. About 200mL of fresh pre-warmed transport buffer (141 mM NaCl, 4 mM KCl, 2.8 mM CaCl2, 1 mM MgSO4, 10 mM HEPES, and 10 mM D-glucose, pH 7.4) containing 250 mM NaF were added to the apical chamber and 800 mL of fresh transport buffer were added to the basolateral chamber. Cells were incubated in a humidified atmosphere (5%CO₂/95% air) at 37°C for 1 hour. At the end of incubation period, the buffer from both chambers was separately collected. NaF concentrations in apical and basolateral compartments were determined by measuring their fluorescence intensities at excitation/emission wavelengths of 720 and 529 nm, respectively (multi-plate reader BIO-TEK, Winooski, VT, USA). Apparent permeation coefficient (Pc, cm/sec) was calculated from the following equation: P_c(cm/sec)=V_b×C_b/ C_a×A×T. V_b is the volume of basolateral side (800 mL), C_b is the concentration of NaF (mM) in the basolateral side, C_a is the concentration of NaF (mM) in the apical side, A is the membrane area (0.33 cm²), and T is the time of transport (3,600 sec).

Quantitative real-time PCR

Total RNA was extracted using TRIzol (Life Technologies), and reverse transcription was performed from 3 μg total RNA using oligo (dT) and RevertAid Reverse Transcriptase (Thermo Scientific) according to the supplier’s instructions. Quantitative PCR was performed with SuperReal PreMix SYBR Green (TIANGEN) using an Applied Biosystems 7500 Fast Real-Time PCR System (Life Technologies). Relative cDNA level was calculated by the comparative CT (cycle threshold) method. PCR primers were the following (5′ to 3′): VEGF Sense: GTGAGCCAGGCTGCAGGAAG, Antisense: GAATGCGTCTGCCGGAGTCT, GAPDH Sense: ATGATTCTACCCACGGCAAG, Antisense: CTGGAAGATGGTGATGGGTT.

Statistical analysis

Data are shown as mean ± SD. One-way analysis of variance (ANOVA) was performed on datasets using SPSS 13.0. For each end point, the treatment means were compared using the Bonferroni least significant difference procedure. A P value <0.05 was considered significant.

Ethics statement

This study was approved by the ethics committee of the Third Affiliated Hospital of Sun Yat-sen University, and all participants signed informed consent forms.

The animal study was carried out in accordance with the recommendations of the animal use protocol, which was approved by the Institutional Animal Care and Use Committee of Sun Yat-Sen University. No potential conflicts of interest were disclosed.