A Novel Nonradioactive Method to Evaluate Vascular Barrier Breakdown and Leakage

Abstract

Purpose.: To identify a novel, sensitive, nonradioactive leakage assay that can be used in the assessment of retinal vascular permeability in rats and mice.

Methods.: Breakdown of the vascular barrier was induced by vascular endothelial growth factor (VEGF), lipopolysaccharide (LPS), or diabetes. Biotinylated bovine serum albumin (bBSA) was administered as a tracer. After perfusion with lactated Ringer's solution, extravasated bBSA was detected with immunoprecipitation and Western blot analysis or sandwich ELISA. The results were then normalized against the final bBSA plasma concentration, the circulation time, and the protein concentration of the tissue.

Results.: Six hours after VEGF injection, BRB breakdown was quantified in the injected eye and was 2.5-fold higher than in the contralateral phosphate-buffered saline (PBS)–injected eye (n = 6 rats, P < 0.01). Intravitreal LPS injection induced severe inflammation in the directly injected eye and moderate inflammation in the contralateral untreated eye. Leakage was six- and threefold higher, respectively, compared with that in the untreated control animals (n = 5 rats, P < 0.01). Nine-month diabetic rats had a threefold increase in vascular leakage compared with age-matched control animals (n = 6 retinas, P < 0.05). Twenty-four hours after intraperitoneal administration of LPS in mice, the animals showed increased vascular leakage in all tissue organs examined (retina, 1.7-fold; brain, 1.5-fold; and kidney, 1.3-fold).

Conclusions.: bBSA can serve as an effective alternative to the current methods used for quantitating vascular leakage and especially the blood–retinal barrier breakdown. It is reasonably easy to perform, low in cost, and adaptable to experiments in mice.

Breakdown of the blood–tissue barrier occurs in many eye disorders, such as diabetic retinopathy ¹ and uveitis. ² The breakdown of the blood–retinal barrier (BRB) is an early sign of vascular dysfunction and contributes to macular edema, an important cause of visual loss. The quantification of BRB leakage remains difficult in animal models. Radioactive isotope dilution techniques are sensitive, ³ but carry a high cost and safety concerns. Fluorescence-conjugated macromolecules such as FITC-dextran ^4–6 have been used, but their use is limited in tissues such as the retina, because of endogenous autofluorescence. Evans blue, ⁷ a dye that binds irreversibly to large macromolecules such as albumin, has been used to measure plasma albumin extravasation in the retina. ⁸ However, the sensitivity of the assay is limited by the small amount of retinal tissue being measured. Investigators have used higher doses of Evans blue ⁹ to increase its sensitivity but high doses lead to saturation of albumin–Evans blue complexes, with an excess of free Evans blue in the plasma. ¹⁰  

Our goal was to develop an alternative, easy to perform, robust, and reliable nonradioactive microleakage assay utilizing biotinylated albumin for quantitative studies of vascular permeability in rats and mice. Breakdown of the vascular barrier induced by VEGF, lipopolysaccharide (LPS), or diabetes ^11–13 was quantified by measuring the extravasation of intravenously administered biotinylated (b)BSA. After proving the specificity and sensitivity of the assay with immunoprecipitation and Western blot analysis, we modified the ELISA sandwich technique, to obtain a simpler and more streamlined process. 

Materials and Methods

Animals

All animal experiments adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and the protocols were approved by the Animal Care Committee of the Massachusetts Eye and Ear Infirmary. Male Brown Norway rats weighing 200 to 300 g each and C57Bl mice weighing 20 to 25 g each were used for the experiments. The animals were fed standard laboratory chow and allowed free access to water in an air-conditioned room with a 12-hour light/12-hour dark cycle until they were used for the experiments. 

Biotinylation of BSA

The biotinylation of BSA (Santa-Cruz Biotechnology, Santa Cruz, CA) was performed per the manufacturer's instructions (EZ-Link Sulfo NHS-LC-Biotin product manual; Pierce, Thermo Scientific, Rockford, IL). It was then dialyzed for 24 hours against a 1000× volume of PBS without potassium. 

bBSA Immunoprecipitation from Retina Lysates

An equal amount of retinal lysates (300 μg) was incubated with 5 μL of 1 mg/mL anti-BSA antibody (Bethyl Laboratories, Montgomery, TX) and 20 μL of protein A/G agarose beads (Pierce, Thermo Scientific) from 2 to 24 hours at 4°C. The beads were washed five to seven times with lysis buffer and PBS. The immunopellets were then run on an SDS-PAGE system and transferred onto PVDF membranes, where they were probed with streptavidin-HRP (GE HealthCare, Piscataway, NJ). 

Induction of Microvascular Leakage

After general anesthesia was induced in the rats with ketamine and xylazine, the pupils were dilated with 0.5% tropicamide (Bausch & Lomb Pharmaceuticals, Tampa, FL). Leakage was induced either by intravitreal injection of 250 ng recombinant murine VEGF 164 (R&D Systems Inc., Minneapolis, MN) or 5 μg LPS from Salmonella typhimurium (Sigma-Aldrich, St. Louis, MO), through a 33-gauge syringe (Hamilton Company, Reno, NV) in 5 μL PBS. The contralateral eye received an equal volume of solvent and served as a paired control. In separate experiments, 5 mg/kg of LPS in PBS or PBS alone was given intraperitoneally in mice. Approximately 6 hours after VEGF and 18 hours after LPS was administered, vascular barrier breakdown was measured as described, in the appropriate section. 

Retina Leakage in Long-Term Diabetes

Brown Norway Rats were made diabetic by injection of streptozotocin (50 mg/kg; Sigma-Aldrich) at 6 to 8 weeks of life. After 9 months of hyperglycemia (>350 mg/dL), microvascular leakage was measured. 

Measurement of BRB Breakdown Tested with bBSA

After microvascular leakage was induced, 0.2 mL of 30 mg/mL bBSA was injected through the femoral vein. After 60 minutes, the chest cavity was opened, and blood was drawn from the right ventricle. The rats were then perfused via the left ventricle at 37°C with lactated Ringer's solution mixed with heparin. The perfusion lasted 6 minutes at a physiological pressure of 120 mm Hg (height of apparatus adjusted to allow an approximate flow rate of 260 mL/min before insertion of the catheter and start of perfusion). The perfusion solution was warmed to 37°C to prevent vasoconstriction. Immediately after perfusion, both eyes were enucleated and placed in cold PBS. The retinas were then carefully dissected under an operating microscope, and each was placed in 0.5 mL chilled lysis buffer. ¹⁴ The retinas were then sonicated five times for 2 seconds each in a 20% to 30% duty cycle and centrifuged at 15,000 rpm for 10 minutes. The supernatant was collected and stored at −80°C. Protein concentrations were determined by protein assay (Bio-Rad, Hercules, CA). Equal amounts (∼350 μg) of the retinal lysates were incubated overnight with 5 μL of anti-BSA antibody, 20 μL of protein agarose beads (Protein A/G Plus; Pierce, Thermo Scientific). The beads were washed five times with PBS containing 1% Triton X-100. The samples were run on 4% to 12% polyacrylamide gels on an SDS-PAGE system. After electrophoresis, the proteins were transferred onto PVDF membranes (Millipore, Billerica, MA) at 35 V for 45 minutes and then at 100 V for 1 hour in Towbin buffer containing 10% methanol and 0.01% SDS. The membranes were blocked with 0.5 mg/mL polyvinyl alcohol (10–30 kDa) for 2 seconds and 2% gelatin in PBST 0.4% for 30 minutes. They were then probed with streptavidin-horseradish peroxidase for 20 minutes (1:3000 in phosphate-buffered saline with 0.4% Tween-20). The membranes were washed three times for 5 minutes with PBS containing 0.4% Tween-20. They were then developed with enhanced chemiluminescence (ECL) by incubation for 60 seconds and exposure to autoradiograph film (RX; Fuji, Tokyo, Japan) for approximately 1 minute. 

Measurement of Vascular Leakage in the Brain and Kidneys with bBSA

After the induction of systemic leakage by LPS in mice, 0.1 mL of 30 mg/mL bBSA was injected through the femoral vein. After 60 minutes, the chest cavity was opened, blood was drawn from the right ventricle, and the mice were perfused via the left ventricle at 37°C with lactated Ringer's solution. The perfusion lasted 6 minutes at a physiological pressure of 120 mm Hg (height of apparatus adjusted to allow approximate flow rate of 260 mL/min before insertion of catheter and start of perfusion). The perfusion solution was warmed to 37°C to prevent vasoconstriction. Immediately after perfusion, both the kidneys and the brain were isolated and placed in cold PBS. They were then homogenized separately and each homogenate placed in 10 mL chilled lysis buffer. ¹⁴ They were centrifuged at 15,000 rpm for 10 minutes, and the supernatants were collected and stored in −80°C. Protein concentrations were determined by protein assay (Bio-Rad), and the amount of bBSA was measured by sandwich ELISA. 

Sandwich ELISA of the Retina bBSA Levels

Flat-bottomed, high-binding, 96-well polystyrene microtiter plates (R&D Biosystems) were coated for 16 to 24 hours at 4°C with 100 μL/well (5 μg/mL) of the rabbit anti-BSA antibody (Bethyl Laboratories) in a coating buffer containing PBS (pH 7.4). Nonspecific sites were then blocked with 0.5 mg/mL polyvinyl alcohol for 2 seconds and 0.5 mg/mL casein in PBST (Tween-20 0.05%) for 1 hour at room temperature. One-hundred-microliter retinal samples (∼150 μg protein) were added and incubated for 2 hours at room temperature. After the samples were washed with PBS with 0.05% Tween, 100 μL of streptavidin-HRP (1:2000) in PBS-Tween-20 0.4% was added, and the samples were incubated for 20 minutes at room temperature and subsequently washed five times for 1 minute each time with 3× PBS containing Tween-20 0.4%. A final wash with 1× PBS was performed before incubation with tetramethylbenzidine (100 μL; Sigma-Aldrich) for 5 minutes at room temperature, after which 100 μL of stop solution (2 M H₂SO₄) was added. Optical density was read at 450 nm with a spectrophotometer (model Lambda Bio 40; Perkin Elmer, Boston, MA). 

BRB breakdown, expressed as the vascular leakage rate, was calculated with the following equation, with results expressed in nanoliters of plasma/amount of protein in the retina milligrams/hour.  

 

 

Results

Polyclonal Anti-BSA Antibody Immunoprecipitation of bBSA from Rodent Retina Lysate and Blood Serum

Since all tissues contain enzymes that are endogenously biotinylated (carboxylases), the identification of a biotinylated tracer such as bBSA may be difficult. To overcome this problem, we immunoprecipitated bBSA with a polyclonal anti-BSA antibody from retinal tissue spiked with 7.5 and 15 ng of the tracer followed by streptavidin-HRP Western blot analysis. The antibody immunoprecipitated >95% of the added bBSA (Fig. 1). Measurement of the bBSA in the blood is not hindered by the endogenous carboxylases, given the almost million-fold excess bBSA over endogenous biotin-containing enzymes in the circulating blood (Supplementary Fig. S1. 

Figure 1.

 

The polyclonal anti-BSA antibody precipitated, in a quantitative manner, bBSA from rat retina lysate and blood serum that was free of contaminating endogenous biotinylated enzymes (carboxylases). Three hundred micrograms of retinal lysates was spiked with 0, 7.5, and 15 ng bBSA and incubated with rabbit nonimmune serum, or rabbit polyclonal anti-BSA antibody and protein A/G agarose beads for 2 hours. After extensive washing, the lysates, immunoprecipitates, and lysates after immunoprecipitation were run on a 4% to 12% SDS-PAGE system, transferred onto PVDF membranes, and probed with streptavidin-HRP. Extracts after anti-BSA immunoprecipitation demonstrate almost complete immunodepletion of bBSA from the tissue extract.

Demonstration of a Reliable Standard Curve of Different Amounts of bBSA with Immunoprecipitation and Western Blot Analysis

The sensitivity and quantification ability of our assay was examined by establishing a standard curve. Known amounts of bBSA in 350 μL of lysis buffer were immunoprecipitated with a constant amount of anti-BSA antibody (2 μg), and the immunopellets were run on 4% to 12% polyacrylamide gels by SDS-PAGE and blotted with streptavidin-HRP. Figure 2 shows that by using this technique and by increasing the exposure time in the Western analysis to 10 minutes, we detected as little as 200 pg of bBSA (Fig 2A). With a short exposure time (1 minute), the sensitivity deceased to 1 ng, and the signal became saturated, reaching a plateau at approximately 50 to 60 ng of bBSA. Within each exposure time, 1 log unit of linear relationship was established (Figs. 2A, 2B, insets). 

Figure 2.

 

Demonstration of a standard curve of increasing amounts of bBSA by immunoprecipitation and Western blot. Western blot with streptavidin-HRP of increasing amounts of bBSA in 350 mL of lysis buffer immunoprecipitated with 2 μg of anti-BSA antibody. Image-analysis software was used to convert the signal from Western blot to numerical data (pixels). (A) Ten-minute and (B) 1-minute exposure times.

bBSA Clearance in the Rat Plasma

bBSA measurements 10 minutes after injection and 60 minutes later revealed that it was minimally metabolized, ¹⁵ with final blood levels 96% of initial values (Supplementary Fig. S1). 

Quantification of VEGF-Induced BRB Breakdown

Figure 3 demonstrates retinal vasculature breakdown as measured by bBSA with anti-BSA immunoprecipitation and streptavidin Western analysis after intravitreal VEGF injection. mVEGF (250 ng) was injected intravitreally in one eye. The contralateral eye received an equal volume of sterile PBS (n = 5 for each treatment). Six hours after VEGF injection, BRB breakdown was quantified. The mean ± SD leakage was measured to be 36 ± 6 and 92 ± 24 nL/mg protein/h for the PBS and VEGF injection, respectively (n = 5, P < 0.01). 

Figure 3.

 

VEGF-induced breakdown of the BRB as measured by bBSA leakage 6 hours after intravitreal injection of 250 ng VEGF or PBS. Top: streptavidin-HRP Western blot of anti-BSA immunopellets from retinal extracts of eyes treated with VEGF or PBS. Bottom: quantification of the leakage. Data are expressed as the mean ± SD (n = 5; P < 0.01).

bBSA Measurements of BRB Leakage in Chronically Diabetic Rats

Figure 4 shows the BRB breakdown as measured by bBSA in chronically diabetic rats 9 months after a single streptozocin injection (60 mg/kg IP). Nondiabetic control data are from animals that received an equal volume of citrate buffer only. Diabetic animals had a threefold increase in leakage, compared with the age-matched control animals (n = 6 eyes for each treatment group). 

Figure 4.

 

Quantification of the diabetes-induced breakdown of the BRB by the bBSA method. Rats were injected with a single dose of streptozotocin or citrate buffer control, and 9 months later, BRB breakdown was assessed by the bBSA method. Top: streptavidin-HRP Western blot of anti-BSA immunopellets. Bottom: quantification of the leakage. Results are the mean ± SD (n = 6 eyes per group).

Quantification of VEGF-Induced BRB Breakdown by Sandwich ELISA

To simplify the new method of measuring vascular permeability, we sought to adapt it for use in an ELISA format. The sensitivity and quantification ability of ELISA were examined by establishing a standard curve with known amounts of bBSA in 100 μL of lysis buffer. As shown in Figure 5 we achieved a linear relationship over 1 log unit, with casein as the blocking agent or when the known amount of bBSA was dissolved in retinal extracts instead of simple lysis buffer. 

Figure 5.

 

Demonstration of a standard curve from data obtained by anti-BSA and streptavidin-HRP sandwich ELISA. Background-subtracted absorbance (450–570 nm) of bBSA standards in lysis buffer. Plates were blocked with PBS-Tween (0.05%), gelatin 0.25%, or casein (1 mg/mL) or retinal lysate.

To further verify that the ELISA gives results similar to those of the immunoprecipitation and Western blot analysis, the VEGF-induced BRB breakdown was measured by ELISA. The mean ± SD of VEGF-induced leakage measured 62 ± 5 nL/mg protein/h versus 27 ± 3 nL/mg protein/h in the control (n = 6 for each group, P < 0.01). These results demonstrate a 2.3-fold increase in vascular leakage, which is very close to the results obtained by Western blot (2.5-fold increase in leakage). 

Quantification of LPS-Induced BRB Breakdown in Rats

Intravitreal injection of LPS from Salmonella typhimurium was used as a model of acute retinal vascular inflammation. To achieve various amounts of inflammation, we injected only one eye of treated animals with 5 μg of intravitreal LPS. Sixteen hours later, the directly injected eye had severe inflammation, whereas the contralateral eye had moderate inflammation. Additional animals that did not receive LPS in either eye served as the control. Figure 6 demonstrates that increasing leakage correlated with the severity of inflammation. Data (mean ± SD) from eyes with severe inflammation showed leakage levels of 124.1 ± 28.3 nL/mg protein/h compared with 65.5 ±2.6 nL/mg protein/h in eyes with moderate inflammation and 26.1 ± 2.9 nL/mg protein/h in the control group (n = 5 for each group, P < 0.01). 

Figure 6.

 

Intravitreal LPS induced breakdown of the BRB. Sixteen hours after a 5-μL intravitreal injection of 5 μg LPS or PBS leakage was measured by bBSA perfusion and ELISA assay. Results are expressed as the mean ± SD. Bottom: clinical images of representative eyes. (A) Control eyes. (B) Eyes with moderate inflammation. (C) Eyes with severe inflammation.

Quantification of the Systemic Effect of LPS on Vascular Leakage in Mice

To further establish the potential of the bBSA vascular leakage assay in other tissues, we examined the effects of 5 mg/kg intraperitoneal LPS in mice on vascular leakage in several organ tissues (retina, brain, and kidneys). Twenty-four hours after LPS administration, the animals showed increased vascular leakage in all organs measured. Vascular leakage was increased 1.37-fold in the kidney (P < 0.05) by LPS treatment and 1.5-fold in the brain (P < 0.05). Intraperitoneal LPS caused in no visible inflammation in the eye. Nevertheless, vascular leakage had increased 1.7-fold compared with the control (P < 0.05; Fig. 7). 

Figure 7.

 

Tissue vasculature leakage induced by intraperitoneal injection of LPS in mice, as measured by bBSA. LPS or vehicle control were administered intraperitoneally at a dose of 5 mg/kg, and 24 hours later vascular permeability was measured by bBSA perfusion and ELISA quantification. Results are expressed as the increase in LPS- versus control-treated animals. (n = 5 for each group, P < 0.01).

Discussion

Several assays have been used for measuring vascular leakage. Radioactivity has been the most sensitive technique, ¹⁶ but its cost and the need for a special license limit its use. Evans blue ^8,17–19 or fluorescent dyes ⁶ have been used with variable success in different tissues, but their sensitivity is inferior to the radioactivity method, and investigators working to determine BRB permeability in rodents are forced to work close to the detection limits of the instruments. Xu et al. ⁸ have an interesting discourse in the second to last paragraph of their paper, explaining the issues of Evans blue, related to the limited amount of rat retinal tissue and instrument sensitivity, necessitating triplicate measurements of the same cuvette. For this reason, we performed a noncomparative study demonstrating an alternative technique that quantifies BRB breakdown. 

In this study we used bBSA as a tracer. ^20,21 Since all tissues contain endogenous biotin-containing proteins (such as carboxylases) which could interfere with our measurements, we used immunoprecipitation of bBSA before its quantification. The anti-BSA antibody used in these studies can efficiently immunoprecipitate more than 95% of the exogenous bBSA administered with high specificity. Detection and quantification performed by streptavidin-HRP in Western blot analysis or ELISA allowed us to detect nanograms or even picograms of the extravasated tracer. These results make this assay sensitive enough to detect as little as 1 nL of plasma leakage from the vascular bed, allowing for detection of small differences in vascular leakage. 

To further reduce variability in the data, we measured the protein concentration of the samples instead of the wet weight of the tissue. Wet-weight measurements can be a source of variability in tissues such as retina, where the sticky vitreous can contaminate the samples with its high levels of water content and minimal protein. 

We used vascular endothelial growth factor (VEGF) and streptozotocin-induced diabetes to disrupt the BRB in both an acute and chronic model of vascular barrier breakdown. ^22,23 VEGF increased leakage 2.5-fold as quickly as 6 hours after injection, whereas 9-month-old diabetic rats exhibited a threefold increase in vascular leakage. These results compare favorably with those in previously published reports. ^24,25 Sixteen hours after intravitreal administration of a more potent inducer of inflammation (LPS) eyes had severe inflammation and a sixfold increase in leakage. Contralateral eyes that did not receive intravitreal LPS still exhibited moderate inflammation due to the systemic action of LPS and exhibited a threefold increase in leakage. 

Most of the use of nonradioactive assays to detect intraocular leakage has been in rats, since their sensitivity limits their effectiveness in smaller animals such as mice. After modifying our method for the ELISA, we used the LPS model of systemic inflammation, by injecting 5 mg/kg LPS intraperitoneally in mice and examined the breakdown of the BRB, breakdown of the blood–brain barrier, and kidney leakage. Twenty-four hours after the injection, there was a significant increase in vascular leakage in all organ tissues examined (retina, 1.7-fold; brain, 1.5-fold; and kidney, 1.3-fold), indicating that the ELISA is suitable for use in mice. 

Although bBSA immunoprecipitation followed by streptavidin-HRP Western blot analysis and ELISA quantification were equally sensitive, there are obvious advantages that favor the use of ELISA, including simplicity, speed, and ease of quantification. 

In most BRB leakage assays, including ours, the intravascular tracer has to be cleared from the circulation postmortem. This necessity can create problems that can affect the measurement of leakage. After the death of the animal and ensuing hypoxia, vascular tone changes, endothelial integrity is affected, and clotting of blood occurs. For this reason, timely perfusion with warm, heparinized physiologic solution is essential to minimize artifacts produced by postmortem changes. 

In conclusion, we have shown that biotin BSA can serve as an effective alternative to the current methods used for quantitating vascular leakage and especially BRB breakdown. It is reasonably easy to perform, can be performed in mice, is low in cost, and can be used in a high-throughput ELISA and 96-well plate format. This method can be applied to different disorders for measuring vascular barrier breakdown and help in the evaluation of new treatments. 

Supplementary Materials

Supplementary Figure S1 - (PDF) 

Footnotes

 Supported by the Boston area Diabetes and Endocrinology Research Center, Fight for Sight, Research to Prevent Blindness, the Lions Onassis Foundation, and the Bacardi Fund.

Footnotes

 Disclosure: G. Trichonas, None; A. Manola, None; Y. Morizane, None; A. Thanos, None; X. Koufomichali, None; T.D. Papakostas, None; S. Montezuma, None; L. Young, None; J.W. Miller, None; E. Gragoudas, None; D. Vavvas, None