Free
Retina  |   August 2013
Polychromatic Angiography for the Assessment of VEGF-Induced BRB Dysfunction in the Rabbit Retina
Author Affiliations & Notes
  • Samir R. Tari
    PCAsso Diagnostics LLC, North Brunswick, New Jersey
  • Maher Youssif
    PCAsso Diagnostics LLC, North Brunswick, New Jersey
  • C. Michael Samson
    New York Eye and Ear Infirmary, New York, New York
  • Robert L. Harris
    Rutgers University, New Brunswick, New Jersey
  • Cheng-Mao Lin
    Kellogg Eye Center, University of Michigan, Ann Arbor, Michigan
  • Uday B. Kompella
    University of Colorado Denver, Denver, Colorado
  • David A. Antonetti
    Kellogg Eye Center, University of Michigan, Ann Arbor, Michigan
  • Gaetano R. Barile
    Manhattan Eye Ear and Throat Hospital, New York, New York
  • Correspondence: Samir R. Tari, PCAsso Diagnostics LLC, 675 US Highway 1, North Brunswick, NJ 08902;samir.tari@pcasso.org
Investigative Ophthalmology & Visual Science August 2013, Vol.54, 5550-5558. doi:10.1167/iovs.13-12144
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Samir R. Tari, Maher Youssif, C. Michael Samson, Robert L. Harris, Cheng-Mao Lin, Uday B. Kompella, David A. Antonetti, Gaetano R. Barile; Polychromatic Angiography for the Assessment of VEGF-Induced BRB Dysfunction in the Rabbit Retina. Invest. Ophthalmol. Vis. Sci. 2013;54(8):5550-5558. doi: 10.1167/iovs.13-12144.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose.: To determine the utility of polychromatic angiography (PCA) in the assessment of VEGF-induced blood retinal barrier (BRB) dysfunction in rabbits.

Methods.: Twenty-six eyes of 24 Dutch Belted rabbits were injected intravitreally with 1.25 μg (group A, n = 5), 10 μg (group C, n = 7), or 4 μg (group B, n = 6; group D, n = 4; and group E, n = 4) of VEGF on day 0. Groups D and E were also injected intravitreally with 1.25 μg and 12.5 μg bevacizumab, respectively, on day 2. On days 0, 2, 4, 7, 11, and 14, PCA was performed using a contrast agent mixture composed of fluorescein sodium, indocyanine green, PCM102, and PCM107 and imaged with a modified fundus camera. PCA scores were based on detected leaking fluorophores.

Results.: On day 7, there was a statistically significant difference between PCA scores of group A (0.6 ± 0.89) and both groups B (2.67 ± 1.37, P = 0.0154) and C (3.33 ± 0.52, P = 0.00085). There was also a statistically significant difference between groups B and E (PCA score 0.75 ± 0.96, P = 0.032) on day 7. On day 11, there was statistically significant difference between group C (1.80 ± 1.1) and both groups A (0, P = 0.021) and B (0.33 ± 0.52, P = 0.037).

Conclusions.: A differential response to both increasing VEGF dose and administration of bevacizumab could be discerned using the PCA. PCA allowed stratification of VEGF-induced BRB dysfunction and inhibitory effects of bevacizumab therapy in the rabbit retina.

Introduction
Presently, there are no clinical diagnostic tools to quantify blood-retinal barrier (BRB) dysfunction. Fluorescein angiography (FA) merely demonstrates the presence or absence of retinal vascular leakage. So, although it may provide a qualitative/descriptive assessment of the nature of leakage, it is not quantitative. Several efforts have been made to quantify fluorescein leakage either by measuring its accumulation in the vitreous cavity, 1,2 calculating the area of leakage, 3,4 or by estimating brightness intensity from photographs. 5 However, none of these techniques is widely used. Optical coherence tomography (OCT) can quantify retinal thickening, and reveal cyst formation and other gross morphological alterations resulting from imbalance in fluid ingress and egress from the retinal tissues, 6 but it does not directly quantify BRB dysfunction. 
The feasibility of using fluorescinated dextrans of different molecular weights for angiographic evaluation of BRB dysfunction is well established in animals 713 and tested with promising results in humans. 14 The correlation between progression or regression of model pathologies and the size of leaking particles has also been demonstrated. 12,13 McNaught et al. 12 have shown that as laser-induced BRB dysfunction in rabbit healed, less of the larger molecular weight fluorescein-dextran particles leaked. In a monkey model of uveitis, Lightman et al. 13 found a gradual increase in the molecular weight of leaking fluorescein-dextran as the disease progressed. Using electron microscopy, these authors also found that vascular segments that leaked 70-KDa FITC-dextran had more severe pathological changes to the endothelial cell junctions compared with vascular segments that leaked 20-KDa FITC-dextran or 376-Da fluorescein, which showed no abnormalities. 
Despite the correlation of molecular weight leakage to pathological processes and the potentially important diagnostic information it may provide, this diagnostic approach has not been widely used due to limitations, such as the poor signal intensity of fluorescinated dextrans, the low sensitivity of the older fundus cameras, and, most importantly, the need for a separate imaging session for each molecular weight. 
With advances in a camera's fluorescence sensitivity and improved loading of dextrans with fluorophores of higher quantum yields, it has become feasible to overcome many of these limitations. Furthermore, simultaneous administration of multiple fluorophores with different excitation and emission spectra attached to dextrans of various sizes could provide a more practical clinical assessment of retinal vascular function. 
In this study, we report a new angiographic technique to assess the size selective BRB permeability and to stratify dysfunction by using multiple fluorophores (conjugated and unconjugated) of different effective molecular weights. We term this new technique polychromatic angiography (PCA). 
Materials and Methods
Contrast Agent
The PCA contrast agent was composed of a mixture four components: fluorescein sodium (Altaire Pharmaceuticals, Inc., Aquebogue, NY); indocyanine green (ICG) (Pulsion Medical, Inc., Irving, TX); and PCM102 (Tetramethylrhodamine-Dextran 10) and PCM107 (Cy5-Dextran 40) (both from PCAsso Diagnostics, LLC, North Brunswick, NJ). Table 1 summarizes the characteristics and the dose of each component of the PCA contrast agent. The mixture was diluted to a total volume of 4.0 mL using normal saline and administered to rabbits intravenously through the auricular vein using a 24-gauge intravenous cannula (Terumo Medical Corporation, Somerset, NJ). 
Table 1. 
 
PCA Filters and Contrast Agent Components
Table 1. 
 
PCA Filters and Contrast Agent Components
Imaging Channel Filters Contrast Agent MW Dose, mg/kg
Excitation Emission Name Conjugation
Ch 1 480/20 525/20 Na fluorescein Partially free 376 Da 5
Ch 2 561/10 600/40 PCM 102 10 KDa dextran ≈10 KDa 15
Ch 3 645/35 690/50 PCM 107 40 KDa dextran ≈40 KDa 10
Ch 4 775/50 845/55 Indocyanine green Plasma components >70 KDa* 1
Bovine Retinal Endothelial Cell Permeability Assay
Bovine retinal endothelial cells (BRECs) were grown to confluence on fibronectin-coated filters with 0.4-μm pores (Transwell; Corning Costar, Acton, MA). VEGF165 (PeproTech, Inc., Rocky Hill, NJ), at 50 ng/mL, was applied to both the apical and basolateral sides of the membrane for 30 minutes before the addition of 8.3 μM of PCM107 to the apical chamber. After the addition of PCM107, aliquots were removed at 30-minute intervals (up to 210 minutes) from the basolateral chamber and placed in 96-well polystyrene plates (black with clear bottoms; Corning Costar). A sample was taken from the apical chamber at the last time point and placed in the 96-well plate. The fluorescence of the aliquots was quantified with FLUOstar Omega microplate reader (BMG LABTECH GmbH, Ortenberg, Germany) and the rate of diffusive flux (Po) in centimeters per second was calculated by the following formula: Po = ([F At]V A)/(F L A), where F A is basolateral fluorescence, FL is apical fluorescence, Δt is change in time, A is the surface area of the filter (cm2), and V A is the volume of the basolateral chamber (in mL). 
WTS-1 Toxicity Assay
BRECs grown in MDCB-131 medium (Sigma-Aldrich, St. Louis, MO) supplemented with 10% fetal calf serum, EGF 10 nm/mL, EndoGro 0.2 mg/mL (VEC Technologies, Inc., Rensselaer, NY), heparin 0.09 mg/mL, 1% antibiotic/antimycotic, and 0.1% tylosin were seeded in a 96-well plate (10,000 cells per well) and incubated at 37°C with 5% CO2 for 40 hours before assessed for cytotoxicity. The bottom row of the plate was not seeded and used as blank. After the medium was removed, 100 μL of each dye solution diluted in MDCB-131 medium and a mixture of all four dyes (1× and 10× of the concentrations used in vivo) were added to the wells. Ethanol (10% and 20%) was used as a positive control. After 4 hours incubation at 37°C with 5% CO2, the dye solutions and ethanol were removed and 100 μL of medium containing 10% of WST-1 assay reagent (Roche Diagnostics GmbH, Mannheim, Germany) was added to each well. After another 4 hours of incubation at 37°C with 5% CO2, the absorbance of the plate was measured at 440 nm with reference at 740 nm (FLUOstar Omega; BMG LABTECH GmbH). Data were obtained from triplicated samples and background absorbance from the same treatment without WST-1 reagent was subtracted. 
Animals
All animal experiments were conducted in accordance with ARVO statement for the use of animals in ophthalmic and vision research and were approved by the institutional animal care and use committee of Rutgers University. Twenty-six eyes of 24 male Dutch Belted rabbits, weighing between 1.7 and 2.1 kg, were assigned to one of five groups as described in Table 2
Table 2
 
Study Design
Table 2
 
Study Design
n Day 0 Day 2 Day 4 Day 7 Day 11 Day 14
PCA VEGF165 PCA Bevacizumab PCA PCA PCA PCA
Group A
 5 + 1.25 μg + + + + +
Group B
 6 + 4 μg + + + + +
Group C
 7 + 10 μg + + + + +
Group D
 4 + 4 μg + 1.25 μg + + + +
Group E
 4 + 4 μg + 12.5 μg + + + +
Anesthesia
The rabbits were anesthetized with a subcutaneous injection of a mixture of ketamine hydrochloride (35 mg/kg) (Butler Shein Animal Health, Dublin, OH) and xylazine hydrochloride (5 mg/kg) (Lloyd, Inc., Shenandoah, IA). The pupils were dilated with a topical application of phenylephrine hydrochloride 2.5% (Akorn, Inc., Buffalo Grove, IL) and tropicamide 0.5% (Akorn, Inc.) eye drops. 
Intravitreal Injection
The intravitreal injections were administered according to previously published methods. 15 Two drops of 5% povidine iodine (Butler Shein Animal Health) were instilled. Topical tetracaine (Bausch & Lomb, Inc., Tampa, FL) was applied for additional topical anesthesia. A pediatric lid speculum was used to keep the eyelids open. Injections were performed using a 28-gauge needle attached to a 0.5-mL syringe. The needle was introduced transconjunctivally into the vitreous cavity at the superior temporal quadrant, 1.5 mm behind the limbus. 
All five groups of animals (Table 2) received intravitreal injections of human recombinant VEGF165 (PeproTech, Inc.) on day 0. Groups D and E received bevacizumab (Genentech, Inc., South San Francisco, CA) on day 2. The injection volume of both VEGF165 and bevacizumab was 50 μL for all doses. Both intravitreal injections were performed immediately after taking PCA images. At the end of the procedure, topical antibiotic ointment (Neo-polycin; Fera Pharmaceuticals LLC, Locust Valley, NY) was applied. 
PCA
Imaging.
Images were taken with a TRC 50 VT fundus camera (Topcon, Tokyo, Japan) retrofitted with a Canon 5D Mark II CMOS camera (Canon, Tokyo, Japan). To isolate and prevent crosstalk between channels, the original filters were removed and custom excitation and emission filter pairs were inserted to image each fluorophore (as described in Table 1). All band-pass optical filters were purchased from Chroma Technologies (Bellows Falls, VT). Channel 3 images were pseudo colored (from red to cyan) using Photoshop CS 4 (Adobe Systems, Inc., San Jose, CA) to increase contrast and differentiation from channel 2. 
Reading and Scoring PCA Images.
Images were taken immediately after the intravenous injection of the contrast agents and intermittently for 5 to 6 minutes. One image from each channel, taken between 1.5 and 3.0 minutes after contrast agent administration, was used for PCA scoring. Each of the chosen four images was then assessed as being either leaking or not leaking. In this study, leakage was defined as the presence of fluorophore extravasation from the retina vessels (Fig. 1). The PCA score was based on the channel showing the largest molecular weight leakage; for example, leakage on channels 1 and 2 but not channels 3 and 4 would be scored as PCA 2 (Fig. 2). 
Figure 1. 
 
Example of PCA channel 2 images before (A) and 4 days after intravitreal injection of 4 μg VEGF (B). In A, the retinal vessels have clearly identifiable contours and branches (not leaking), whereas in (B), the retinal vessels are not identifiable from the background due to dye extravasation (leaking). Both images were taken between 1.5 and 2.0 minutes from administration of the contrast agent.
Figure 1. 
 
Example of PCA channel 2 images before (A) and 4 days after intravitreal injection of 4 μg VEGF (B). In A, the retinal vessels have clearly identifiable contours and branches (not leaking), whereas in (B), the retinal vessels are not identifiable from the background due to dye extravasation (leaking). Both images were taken between 1.5 and 2.0 minutes from administration of the contrast agent.
Figure 2. 
 
PCA images of five rabbit retinas exemplifying the five different PCA scores. PCA 0: no leakage; PCA 1: leakage on channel 1 only; PCA 2: leakage on channels 1 and 2; PCA 3: leakage on channels 1, 2, and 3; PCA 4: leakage on all channels.
Figure 2. 
 
PCA images of five rabbit retinas exemplifying the five different PCA scores. PCA 0: no leakage; PCA 1: leakage on channel 1 only; PCA 2: leakage on channels 1 and 2; PCA 3: leakage on channels 1, 2, and 3; PCA 4: leakage on all channels.
Statistics.
Results are presented as means ± SD. Two-way ANOVA and Bonferroni posttests were used for statistical analyses. 
Results
VEGF Increases BREC Assay Permeability to PCM107
The newly developed PCM107 was tested in permeability assays in BRECs. BREC monolayers were grown to confluence on Transwell inserts and 2 days later were used for permeability measures. VEGF induced a 33% increase in permeability to the 40-kDa dextran PCM107 dye (Fig. 3), similar to permeability response of 70-kDa rhodamine B isothiocyanate dextran reported previously. 16  
Figure 3. 
 
Permeability of BREC to PCM107. (A) VEGF induced increase in permeability to PCM107 measured over a 4-hour time course. (B) Graph represents the mean ± SEM for control and VEGF-treated inserts.
Figure 3. 
 
Permeability of BREC to PCM107. (A) VEGF induced increase in permeability to PCM107 measured over a 4-hour time course. (B) Graph represents the mean ± SEM for control and VEGF-treated inserts.
Wst-1 Assay: No Evidence of Toxicity
After 4 hours of incubation with PCA components, both individually and in combination, BRECs showed no evidence of toxicity as assessed by WST-1 assay (Fig. 4). 
Figure 4. 
 
Cell viability of BRECs after incubation with the different PCA components and controls for 4 hours. Error bars indicate SD from triplicate samples. Data suggest that PCA components individually or combined are well tolerated by BRECs after 4 hours of treatment. FA, fluorescein sodium; PCA, the four PCA components combined.
Figure 4. 
 
Cell viability of BRECs after incubation with the different PCA components and controls for 4 hours. Error bars indicate SD from triplicate samples. Data suggest that PCA components individually or combined are well tolerated by BRECs after 4 hours of treatment. FA, fluorescein sodium; PCA, the four PCA components combined.
Polychromatic Angiography
A total of 154 PCAs were performed on 24 rabbits over 14 days, of which 146 PCAs were used for scoring. Some degree of leakage (i.e., PCA score ≥1) was seen in 67 PCAs. Of the 67 PCAs with some degree of leakage, 57 (85.1%) leaked PCM102, 19 (28.4%) leaked PCM107, and 7 (10.4%) of 67 leaked ICG (Fig. 5). 
Figure 5. 
 
Percentage of instances with leakage for each of the fluorophores in angiograms with PCA score ≥1.
Figure 5. 
 
Percentage of instances with leakage for each of the fluorophores in angiograms with PCA score ≥1.
Due to the pharmacokinetics of the different molecular weight components, the time needed for each channel to show evidence of dye extravasation was different: smaller molecules leaked earlier than larger ones. In case of the highest score leakages (PCA 4), all channels leaked approximately or before 2 minutes postinjection and leakage stabilized for an additional 1 to 5 minutes. Between 3 and 6 minutes postinjection, in most eyes with moderate to severe leakage (PCA 2 to PCA 4), the fluorophores from channels 1 and 2 accumulating in the vitreous affected the quality of the image. On channel 4, ICG fluorescence started to fade after 6 to 8 minutes postinjection. In all four channels, the quality of the images started to deteriorate 4 to 5 minutes after injection. Given these considerations, images obtained between 1.5 and 3.0 minutes from the time of intravenous injection of the contrast agent were used for PCA scoring purposes (Fig. 6). 
Figure 6. 
 
PCA images of rabbit retina at different times after contrast agent injection on day 11 post VEGF injection. Channels 1 and 2 images show the presence of leakage at 1.5 minutes. Images taken at 5 minutes on channels 1 and 2 became hazy due to the accumulation of fluorophore in the vitreous. Generally, image quality starts deteriorating after 4 to 5 minutes from the time of injection.
Figure 6. 
 
PCA images of rabbit retina at different times after contrast agent injection on day 11 post VEGF injection. Channels 1 and 2 images show the presence of leakage at 1.5 minutes. Images taken at 5 minutes on channels 1 and 2 became hazy due to the accumulation of fluorophore in the vitreous. Generally, image quality starts deteriorating after 4 to 5 minutes from the time of injection.
In Vivo VEGF Dose Response
Human recombinant VEGF165 induced retinal vascular leakage 2 days after intravitreal injection into the rabbit eye. Average PCA scores peaked on day 4 for group A (1.25 μg VEGF), whereas groups B (4 μg) and C (10 μg) peaked on day 7 (Table 3, Fig. 7). On day 7, there was a statistically significant difference between PCA scores of group A (0.6 ± 0.89, n = 5) and both groups B (2.67 ± 1.37, n = 6, P = 0.0154) and C (3.33 ± 0.52, n = 6, P = 0.00085). On day 11, there was also a statistically significant difference between group C (1.80 ± 1.1, n = 5) and both groups A (0, n = 5, P = 0.021) and B (0.33 ± 0.52, n = 6, P = 0.037). 
Figure 7. 
 
PCA scores before, and 2, 4, 7, 11, and 14 days after intravitreal VEGF injection in rabbits. Mean ± SD. *P ≤ 0.05, **P ≤ 0.001.
Figure 7. 
 
PCA scores before, and 2, 4, 7, 11, and 14 days after intravitreal VEGF injection in rabbits. Mean ± SD. *P ≤ 0.05, **P ≤ 0.001.
Table 3. 
 
Average PCA Scores and SDs Before and After VEGF Injection
Table 3. 
 
Average PCA Scores and SDs Before and After VEGF Injection
Pre VEGF Average (n)/SD Day 2 Average (n)/SD Day 4 Average (n)/SD Day 7 Average (n)/SD Day 11 Average (n)/SD Day 14 Average (n)/SD
Group A 0 1.6 2.4 0.6 0 0
1.25 μg VEGF (5)/0 (5)/1.67 (5)/0.89 (5)/0.89 (5)/0 (5)/0
Group B 0 1.33 2.20 2.67 0.33 0
4 μg VEGF (7)/0 (6)/1.03 (5)/0.83 (6)/1.37 (6)/0.52 (6)/0
Group C 0 1.57 2.86 3.33 1.80 0.33
10 μg VEGF (7)/0 (7)/0.79 (7)/1.21 (6)/0.52 (5)/1.1 (6)/0.82
In Vivo Effect of Bevacizumab
Intravitreal injection of 12.5 μg bevacizumab on day 2 to group E (4 μg VEGF) rabbits significantly lowered VEGF-induced leakage (PCA score ± SD 0.75 ± 0.96, n = 4) on day 7 when compared with the no antibody–treated group B (2.67 ± 1.37, n = 6, P = 0.032). There was a trend toward difference in PCA scores on day 7 between group D and group E (2 ± 0, n = 3, P = 0.08) (Table 4, Fig. 8). 
Figure 8. 
 
PCA scores before, and 2, 4, 7, 11, and 14 days after intravitreal VEGF ± bevacizumab (Bz) injection in rabbits. Mean ± SD. *P = 0.032, #P = 0.08.
Figure 8. 
 
PCA scores before, and 2, 4, 7, 11, and 14 days after intravitreal VEGF ± bevacizumab (Bz) injection in rabbits. Mean ± SD. *P = 0.032, #P = 0.08.
Table 4. 
 
Average PCA Scores and SDs After VEGF ± Bevacizumab on Day 2
Table 4. 
 
Average PCA Scores and SDs After VEGF ± Bevacizumab on Day 2
Pre VEGF Average (n)/SD Day 2 Average (n)/SD Day 4 Average (n)/SD Day 7 Average (n)/SD Day 11 Average (n)/SD Day 14 Average (n)/SD
Group B 0 1.33 2.20 2.67 0.33 0
4 μg VEGF (7)/0 (6)/1.03 (5)/0.83 (6)/1.37 (6)/0.52 (6)/0
Group D 0 1 2 2 0.33 0.33
4 μg VEGF +1.25 μg Bz (4)/0 (4)/1.15 (3)/0 (3)/0 (3)/0.58 (3)/0.58
Group E 0 2 1.25 0.75 0 0
4 μg VEGF +12.5 μg Bz (4)/0 (4)/0 (4)/0.96 (4)/0.96 (4)/0 (4)/0
Vascular Response and Resolution Patterns In Vivo
Proliferative Response.
In addition to inducing BRB leakage, VEGF induced retinal vascular proliferation. PCA allowed the visualization of three stages of vascular response (VR1 to 3, Fig. 9). In VR1, the retinal vessels became more tortuous and dilated compared with baseline. In VR2, in addition to the changes described in VR1, secondary and tertiary branches extended beyond the baseline vascular distribution zone by half optic disc diameter or more. While in VR3, the retinal vessels coalesced in a membranouslike structure composed of lobulated vessels. 
Figure 9. 
 
ICG images of three rabbit retinas exemplifying baseline and three levels of VR induced by VEGF injection.VR0: Baseline (before VEGF injection). VR1: Increased vascular tortuosity (*indicating area with increased tortuosity); VR2: VR1 + vascular extension beyond baseline greater than one-half disc diameter (#indicating extension of vascular branches); VR3: membranelike vascular formation (†).
Figure 9. 
 
ICG images of three rabbit retinas exemplifying baseline and three levels of VR induced by VEGF injection.VR0: Baseline (before VEGF injection). VR1: Increased vascular tortuosity (*indicating area with increased tortuosity); VR2: VR1 + vascular extension beyond baseline greater than one-half disc diameter (#indicating extension of vascular branches); VR3: membranelike vascular formation (†).
Effect of both VEGF and bevacizumab on vascular morphology as measured by the above-described VR scoring system (Fig. 10) was, generally, similar to their effect on BRB dysfunction as measured by PCA scores (Figs. 7, 8). 
Figure 10. 
 
VR scores after intravitreal injection of VEGF on day 0 (A) or VEGF on day 0 and bevacizumab on day 2 (B).
Figure 10. 
 
VR scores after intravitreal injection of VEGF on day 0 (A) or VEGF on day 0 and bevacizumab on day 2 (B).
Incomplete Resolution.
Most retinal vessels returned to baseline morphology by day 14. However, in a minority of retinas (3 and 1 retinas from groups A and B, respectively), the vessels at the edges of the vascular tree clump together in a fashion similar to that seen in patients with retinopathy of prematurity (Fig. 11). 
Figure 11. 
 
PCA images of rabbit retina 14 days after intravitreal VEGF injection. The vessels at the edges of the vascular tree clump together in a fashion similar to that seen in patients with retinopathy of prematurity
Figure 11. 
 
PCA images of rabbit retina 14 days after intravitreal VEGF injection. The vessels at the edges of the vascular tree clump together in a fashion similar to that seen in patients with retinopathy of prematurity
Discussion
The current study uses PCA as a new semiquantitative method to assess BRB dysfunction in rabbits. Intravitreal injection of VEGF produced a differential effect on size-selective permeability of BRB during the 14-day observation period. Higher VEGF doses were associated with more prolonged periods of leakage and higher peak PCA scores. Blocking the effect of VEGF with bevacizumab had the reverse effect (i.e., shortening the period of leakage and lowered PCA scores). 
PCA assesses BRB permeability to materials of molecular weight between 376 Da and 70 KDa. The same dynamic range can be covered by using FA and ICG simultaneously. However, in this study the combined PCA2 (10 KDa leaking) and PCA3 (40 KDa leaking) scores constituted 75% of all angiographies with some degree of leakage (PCA scores ≥1), whereas only 15% and 10% had PCA1 (376 Da) and PCA4 (>70 KDa) scores, respectively. This indicates that changes in the BRB following VEGF intravitreous injection in rabbits causes permeability to molecules with molecular weights in the range of 10 KDa or greater and less than 70 KDa and delineates the importance of two intermediate molecular weights in PCA. 
In pilot studies, PCM03 (4 KDa) was used as a contrast agent for channel 2 but it was found to leak in all instances in which FA leaked (15 of 15 instances, data not shown). This led us to modify the molecular weight range by use of PCM102 (10 KDa). We also tested PCM103 (20 KDa) in channel 3, but its leakage was nearly identical to PCM102 (11 of 12 instances, data not shown). Accordingly, we increased the size of the channel 3 fluorophore to 40 KDa (PCM107). The dye load for both PCM102 and PCM107 was adjusted to prevent quenching. 
In addition to indicating the temporal relation of the start and disappearance of leakage, as reported by others, 1518 PCA allowed for more detailed analysis of BRB dysfunction by differentiating the effects of the different doses over time, which would not have been possible using FA and/or ICG alone. The PCA score pattern from groups C and E mirrored the changes in OCT measurements seen in rabbits given similar doses of VEGF or VEGF and bevacizumab, respectively. 15,16  
In addition to scoring of BRB dysfunction, PCA allowed for the visualization of vascular changes even in the presence of leakage in all but one channel. This permitted the development and use of the VR scoring system. Because PCA uses two long-wavelength channels (CH 3 and CH 4) that penetrate the RPE layer, it is possible to use images from these channels to study the choroid in nonalbino animals. Future studies could explore the effect of different VEGFs on retinal vasculature using PCA. 
The decision to use the rabbit eye for this experiment was based on two main reasons: the size of the eye is large enough to allow the use of a clinical fundus camera without changing the optics, and the relatively inexpensive cost of the experiments compared with using nonhuman primates. In addition, there were sufficient data on VEGF-induced leakage in rabbits to use for guidance and comparison. However, because the rabbit retina has significant morphologic differences compared with the human retina (merangiotic vascular distribution with no macula), it is important to validate the use of PCA in nonhuman primates (holoangiotic vascular distribution and a macula with a central avascular fovea). Animal models using VEGF-induced leakage are intended to mimic a state similar to diabetic retinopathy, one of the major retinal vascular diseases. It would be also of interest to validate the use of PCA in the assessment of pathologic processes modeling other retinal conditions, such as neovascular AMD and uveitis. 
One of the shortcomings of this study is the lack of histological evidence for the correlation between disease severity and PCA scores. However, in agreement with our results, Lightman et al. 13 have shown a correlation between the molecular weight of fluorescein-labeled dextran leaking from pathological vessels and electron microscopic evidence of retinal pathology. Another shortcoming of this study is the lack of information about the binding of PCM102 and PCM107 to plasma proteins or other plasma molecules. Nevertheless, because PCM102 and PCM107 leaked in instances where ICG did not leak, we may assume that both components are, at least partially, unbound to plasma proteins or other large molecules. This observation is confirmed by the renal excretion of the dyes (urine discoloration, data not included), which would have not been the case if PCM102 and PCM107 were completely bound to plasma proteins or other large plasma components. 
Overall, our findings augment the reported correlation between the permeability of BRB to increasing molecular weights and disease progression and regression. 12,13 These data also suggest the potential utility of PCA as a tool to assess and stratify the severity of BRB dysfunction in animal models and, supplemented by previous FITC-dextran reports, 1214 provide a rationale for furthering the development of PCA for the assessment of BRB dysfunction in humans. 
Acknowledgments
Supported in part by Grant 17-2011-518 from JDRF. 
Disclosure: S.R. Tari, PCAsso Diagnostics LLC (I, E), P; M. Youssif, PCAsso Diagnostics LLC (E); C.M. Samson, PCAsso Diagnostics LLC (I, C); R.L. Harris, None; C.-M. Lin, None; U.B. Kompella, PCAsso Diagnostics LLC (F), P; D.A. Antonetti, None; G.R. Barile, PCAsso Diagnostics LLC (I, C), P 
References
Smith RT Lee CM Charles HC Farber M Cunha-Vaz JG. Quantification of diabetic macular edema. Arch Ophthalmol . 1987; 105: 218–222. [CrossRef] [PubMed]
Lobo CL Bernardes RC Santos FJ Cunha-Vaz JG. Mapping retinal fluorescein leakage with confocal scanning laser fluorometry of the human vitreous. Arch Ophthalmol . 1999; 117: 631–637. [CrossRef] [PubMed]
Rosenfeld PJ Brown DM Heier JS Ranibizumab for neovascular age-related macular degeneration. N Engl J Med . 2006; 355: 1419–1431. [CrossRef] [PubMed]
Gragoudas ES Adamis AP Cunningham ET Jr Feinsod M Guyer DR; VEGF Inhibition Study in Ocular Neovascularization Clinical Trial Group. Pegaptanib for neovascular age-related macular degeneration. N Engl J Med . 2004; 351: 2805–2816. [CrossRef] [PubMed]
Phillips RP Ross PG Tyska M Sharp PF Forrester JV. Detection and quantification of hyperfluorescent leakage by computer analysis of fundus fluorescein angiograms. Graefes Arch Clin Exp Ophthalmol . 1991; 229: 329–335. [CrossRef] [PubMed]
Hee MR Puliafito CA Duker JS Topography of diabetic macular edema with optical coherence tomography. Ophthalmology . 1998; 105: 360–370. [CrossRef] [PubMed]
Salminen L Chioralia G Schreiber S. Distribution of intravenously injected dextrans in the posterior part of the rat eye. Bibl Anat . 1977; 16: 69–72. [PubMed]
Chioralia G Salminen L Baurmann H Kremer F. Fluorescein-labelled dextrans as tracer substance for experimental angiograms. Acta Ophthalmol (Copenh) . 1976; 54: 665–667. [CrossRef] [PubMed]
Rabkin MD Bellhorn MB Bellhorn RW. Selected molecular weight dextrans for in vivo permeability studies of rat retinal vascular disease. Exp Eye Res . 1977; 24: 607–612. [CrossRef] [PubMed]
Bellhorn MB Bellhorn RW Poll DS. Permeability of fluorescein-labelled dextrans in fundus fluorescein angiography of rats and birds. Exp Eye Res . 1977; 24: 595–605. [CrossRef] [PubMed]
Salminen L Chioralia G Schreiber S. Effect of acute ocular hypotony on the blood-ocular barrier. Trans Ophthalmol Soc U K . 1977; 97: 621–622. [PubMed]
McNaught EI Foulds WS Johnson NF. The permeability of the posterior blood ocular barrier after xenon photocoagulation: a study using fluorescein labeled dextrans. Br J Ophthalmol . 1981; 65: 473–477. [CrossRef] [PubMed]
Lightman SL Caspers-Velu LE Hirose S Angiography with fluorescein-labeled dextrans in a primate model of uveitis. Arch Ophthalmol . 1987; 105: 844–848. [CrossRef] [PubMed]
Atkinson EG Jones S Ellis BA Molecular size of retinal vascular leakage determined in patients with posterior uveitis. Eye . 1991; 5: 440–446. [CrossRef] [PubMed]
Arana LA Pinto AT Chader GJ Fluorescein angiography, optical coherence tomography, and histopathologic findings in a VEGF(165) animal model of retinal angiogenesis. Graefes Arch Clin Exp Ophthalmol . 2012; 250: 1421–1428. [CrossRef] [PubMed]
Phillips BE Cancel L Tarbell JM Antonetti DA. Occludin independently regulates permeability under hydrostatic pressure and cell division in retinal pigment epithelial cells. Invest Ophthalmol Vis Sci . 2008; 49: 2568–2576. [CrossRef] [PubMed]
Saati S Agrawal RN Louie S Chader GJ Humayun MS. Effect of multiple injections of small divided doses vs single injection of intravitreal bevacizumab on retinal neovascular model in rabbits. Graefes Arch Clin Exp Ophthalmol . 2010; 248: 457–466. [CrossRef] [PubMed]
Ameri H Chader GJ Kim JG Sadda SR Rao NA Humayun MS. The effects of intravitreous bevacizumab on retinal neovascular membrane and normal capillaries in rabbits. Invest Ophthalmol Vis Sci . 2007; 48: 5708–5715. [CrossRef] [PubMed]
Figure 1. 
 
Example of PCA channel 2 images before (A) and 4 days after intravitreal injection of 4 μg VEGF (B). In A, the retinal vessels have clearly identifiable contours and branches (not leaking), whereas in (B), the retinal vessels are not identifiable from the background due to dye extravasation (leaking). Both images were taken between 1.5 and 2.0 minutes from administration of the contrast agent.
Figure 1. 
 
Example of PCA channel 2 images before (A) and 4 days after intravitreal injection of 4 μg VEGF (B). In A, the retinal vessels have clearly identifiable contours and branches (not leaking), whereas in (B), the retinal vessels are not identifiable from the background due to dye extravasation (leaking). Both images were taken between 1.5 and 2.0 minutes from administration of the contrast agent.
Figure 2. 
 
PCA images of five rabbit retinas exemplifying the five different PCA scores. PCA 0: no leakage; PCA 1: leakage on channel 1 only; PCA 2: leakage on channels 1 and 2; PCA 3: leakage on channels 1, 2, and 3; PCA 4: leakage on all channels.
Figure 2. 
 
PCA images of five rabbit retinas exemplifying the five different PCA scores. PCA 0: no leakage; PCA 1: leakage on channel 1 only; PCA 2: leakage on channels 1 and 2; PCA 3: leakage on channels 1, 2, and 3; PCA 4: leakage on all channels.
Figure 3. 
 
Permeability of BREC to PCM107. (A) VEGF induced increase in permeability to PCM107 measured over a 4-hour time course. (B) Graph represents the mean ± SEM for control and VEGF-treated inserts.
Figure 3. 
 
Permeability of BREC to PCM107. (A) VEGF induced increase in permeability to PCM107 measured over a 4-hour time course. (B) Graph represents the mean ± SEM for control and VEGF-treated inserts.
Figure 4. 
 
Cell viability of BRECs after incubation with the different PCA components and controls for 4 hours. Error bars indicate SD from triplicate samples. Data suggest that PCA components individually or combined are well tolerated by BRECs after 4 hours of treatment. FA, fluorescein sodium; PCA, the four PCA components combined.
Figure 4. 
 
Cell viability of BRECs after incubation with the different PCA components and controls for 4 hours. Error bars indicate SD from triplicate samples. Data suggest that PCA components individually or combined are well tolerated by BRECs after 4 hours of treatment. FA, fluorescein sodium; PCA, the four PCA components combined.
Figure 5. 
 
Percentage of instances with leakage for each of the fluorophores in angiograms with PCA score ≥1.
Figure 5. 
 
Percentage of instances with leakage for each of the fluorophores in angiograms with PCA score ≥1.
Figure 6. 
 
PCA images of rabbit retina at different times after contrast agent injection on day 11 post VEGF injection. Channels 1 and 2 images show the presence of leakage at 1.5 minutes. Images taken at 5 minutes on channels 1 and 2 became hazy due to the accumulation of fluorophore in the vitreous. Generally, image quality starts deteriorating after 4 to 5 minutes from the time of injection.
Figure 6. 
 
PCA images of rabbit retina at different times after contrast agent injection on day 11 post VEGF injection. Channels 1 and 2 images show the presence of leakage at 1.5 minutes. Images taken at 5 minutes on channels 1 and 2 became hazy due to the accumulation of fluorophore in the vitreous. Generally, image quality starts deteriorating after 4 to 5 minutes from the time of injection.
Figure 7. 
 
PCA scores before, and 2, 4, 7, 11, and 14 days after intravitreal VEGF injection in rabbits. Mean ± SD. *P ≤ 0.05, **P ≤ 0.001.
Figure 7. 
 
PCA scores before, and 2, 4, 7, 11, and 14 days after intravitreal VEGF injection in rabbits. Mean ± SD. *P ≤ 0.05, **P ≤ 0.001.
Figure 8. 
 
PCA scores before, and 2, 4, 7, 11, and 14 days after intravitreal VEGF ± bevacizumab (Bz) injection in rabbits. Mean ± SD. *P = 0.032, #P = 0.08.
Figure 8. 
 
PCA scores before, and 2, 4, 7, 11, and 14 days after intravitreal VEGF ± bevacizumab (Bz) injection in rabbits. Mean ± SD. *P = 0.032, #P = 0.08.
Figure 9. 
 
ICG images of three rabbit retinas exemplifying baseline and three levels of VR induced by VEGF injection.VR0: Baseline (before VEGF injection). VR1: Increased vascular tortuosity (*indicating area with increased tortuosity); VR2: VR1 + vascular extension beyond baseline greater than one-half disc diameter (#indicating extension of vascular branches); VR3: membranelike vascular formation (†).
Figure 9. 
 
ICG images of three rabbit retinas exemplifying baseline and three levels of VR induced by VEGF injection.VR0: Baseline (before VEGF injection). VR1: Increased vascular tortuosity (*indicating area with increased tortuosity); VR2: VR1 + vascular extension beyond baseline greater than one-half disc diameter (#indicating extension of vascular branches); VR3: membranelike vascular formation (†).
Figure 10. 
 
VR scores after intravitreal injection of VEGF on day 0 (A) or VEGF on day 0 and bevacizumab on day 2 (B).
Figure 10. 
 
VR scores after intravitreal injection of VEGF on day 0 (A) or VEGF on day 0 and bevacizumab on day 2 (B).
Figure 11. 
 
PCA images of rabbit retina 14 days after intravitreal VEGF injection. The vessels at the edges of the vascular tree clump together in a fashion similar to that seen in patients with retinopathy of prematurity
Figure 11. 
 
PCA images of rabbit retina 14 days after intravitreal VEGF injection. The vessels at the edges of the vascular tree clump together in a fashion similar to that seen in patients with retinopathy of prematurity
Table 1. 
 
PCA Filters and Contrast Agent Components
Table 1. 
 
PCA Filters and Contrast Agent Components
Imaging Channel Filters Contrast Agent MW Dose, mg/kg
Excitation Emission Name Conjugation
Ch 1 480/20 525/20 Na fluorescein Partially free 376 Da 5
Ch 2 561/10 600/40 PCM 102 10 KDa dextran ≈10 KDa 15
Ch 3 645/35 690/50 PCM 107 40 KDa dextran ≈40 KDa 10
Ch 4 775/50 845/55 Indocyanine green Plasma components >70 KDa* 1
Table 2
 
Study Design
Table 2
 
Study Design
n Day 0 Day 2 Day 4 Day 7 Day 11 Day 14
PCA VEGF165 PCA Bevacizumab PCA PCA PCA PCA
Group A
 5 + 1.25 μg + + + + +
Group B
 6 + 4 μg + + + + +
Group C
 7 + 10 μg + + + + +
Group D
 4 + 4 μg + 1.25 μg + + + +
Group E
 4 + 4 μg + 12.5 μg + + + +
Table 3. 
 
Average PCA Scores and SDs Before and After VEGF Injection
Table 3. 
 
Average PCA Scores and SDs Before and After VEGF Injection
Pre VEGF Average (n)/SD Day 2 Average (n)/SD Day 4 Average (n)/SD Day 7 Average (n)/SD Day 11 Average (n)/SD Day 14 Average (n)/SD
Group A 0 1.6 2.4 0.6 0 0
1.25 μg VEGF (5)/0 (5)/1.67 (5)/0.89 (5)/0.89 (5)/0 (5)/0
Group B 0 1.33 2.20 2.67 0.33 0
4 μg VEGF (7)/0 (6)/1.03 (5)/0.83 (6)/1.37 (6)/0.52 (6)/0
Group C 0 1.57 2.86 3.33 1.80 0.33
10 μg VEGF (7)/0 (7)/0.79 (7)/1.21 (6)/0.52 (5)/1.1 (6)/0.82
Table 4. 
 
Average PCA Scores and SDs After VEGF ± Bevacizumab on Day 2
Table 4. 
 
Average PCA Scores and SDs After VEGF ± Bevacizumab on Day 2
Pre VEGF Average (n)/SD Day 2 Average (n)/SD Day 4 Average (n)/SD Day 7 Average (n)/SD Day 11 Average (n)/SD Day 14 Average (n)/SD
Group B 0 1.33 2.20 2.67 0.33 0
4 μg VEGF (7)/0 (6)/1.03 (5)/0.83 (6)/1.37 (6)/0.52 (6)/0
Group D 0 1 2 2 0.33 0.33
4 μg VEGF +1.25 μg Bz (4)/0 (4)/1.15 (3)/0 (3)/0 (3)/0.58 (3)/0.58
Group E 0 2 1.25 0.75 0 0
4 μg VEGF +12.5 μg Bz (4)/0 (4)/0 (4)/0.96 (4)/0.96 (4)/0 (4)/0
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×