February 2011
Volume 52, Issue 2
Free
Physiology and Pharmacology  |   February 2011
Effects of Vitreous Liquefaction on the Intravitreal Distribution of Sodium Fluorescein, Fluorescein Dextran, and Fluorescent Microparticles
Author Affiliations & Notes
  • Lay Ean Tan
    From the Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde, Glasgow, United Kingdom; and
  • Werhner Orilla
    Allergan, Inc., Irvine, California.
  • Patrick M. Hughes
    Allergan, Inc., Irvine, California.
  • Susan Tsai
    Allergan, Inc., Irvine, California.
  • James A. Burke
    Allergan, Inc., Irvine, California.
  • Clive G. Wilson
    From the Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde, Glasgow, United Kingdom; and
  • Corresponding author: Clive G. Wilson, Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde, Glasgow G4 0NR, UK; c.g.wilson@strath.ac.uk
Investigative Ophthalmology & Visual Science February 2011, Vol.52, 1111-1118. doi:10.1167/iovs.10-5813
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Lay Ean Tan, Werhner Orilla, Patrick M. Hughes, Susan Tsai, James A. Burke, Clive G. Wilson; Effects of Vitreous Liquefaction on the Intravitreal Distribution of Sodium Fluorescein, Fluorescein Dextran, and Fluorescent Microparticles. Invest. Ophthalmol. Vis. Sci. 2011;52(2):1111-1118. doi: 10.1167/iovs.10-5813.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose.: The effects of vitreous liquefaction in the elderly on the distribution of drugs from intravitreal injections, depots, or devices remains unclear. The purpose of the present study was to develop a liquefied vitreous model that simulates the aged condition, to enable the study of clinically relevant drug distribution.

Methods.: Dutch-belted rabbits were used to develop a study model using hyaluronidase as a vitreolytic agent. The effects of experimental vitreous liquefaction were investigated on intravitreal sodium fluorescein, fluorescein isothiocyanate-dextran (MW 150 kDa), and a suspension of 1-μm fluorescent particles. The distribution of these model compounds was monitored by retinal angiography with a confocal laser scanning system and ocular fluorophotometer.

Results.: Hyaluronidase-treated vitreous humor (n = 6) was found to decrease the gel phase to 41% ± 9% (wt/wt; mean ± SD) compared with 81% ± 9% in the control eyes (n = 8; P < 0.05). The distribution of sodium fluorescein and fluorescein isothiocyanate dextran was greater in the liquefied vitreous than in the control. In comparison to the normal vitreous, fluorescent particles sedimented faster in the liquefied vitreous, and the distribution was more dispersed and scattered.

Conclusions.: A model of vitreous liquefaction in rabbits was successfully generated using intravitreal hyaluronidase. Small and large fluorescent molecules as well as particulates were distributed faster in liquefied vitreous than in the control. The results suggest enhanced convective flow and subsequent faster clearance in liquefied vitreous.

The volume of the human vitreous humor is approximately 4 mL and forms the largest (80%) compartment of the globe in the posterior segment of the eye. 1 Thus, the vitreous is a readily available drug reservoir for the treatment of retinal diseases. The main composition of the vitreous is water (>98%), but its three-dimensional microstructure is primarily constituted by the remaining 1% of structural elements—mainly, collagen and proteoglycans. 2 In juvenile vitreous humor, approximately 80% of the water is bound, trapped within the collagen-hyaluronan gel network, and the remainder (∼20%) is in the form of free water, constituting the liquid phase. 3  
With aging, the vitreous degenerates progressively and collagen fibrils, which are kept apart by proteoglycans in the juvenile eye, aggregate into coarse fibers. 4 6 The detached collagenous fragments, together with cells and pigment, are clinically perceived as floaters by most patients. In addition, the hyaluronan component segregates into a pool of free water, presumably as a result of depolymerization of its linear chain. 1 The vitreous of an elderly patient (>60 years old) will therefore have a higher content of free water with a gel–liquid ratio of approximately 1. 3 Most of the populations treated for many posterior segment diseases such as age-related macular degeneration are elderly by definition. Previous publications in the literature contributed by vitreoretinal surgeons have described the role of vitreous structure as a diffusional barrier 7,8 and have considered the implications of loss of vitreous gel structure in the movements of material around the eye. 9 13  
However, the preclinical designs of treatment regimens for most ocular problems are largely based on young animal models, rabbits, and rats, with a fully formed vitreous and a high gel content (∼80%). For this reason, the impact of vitreous liquefaction on the distribution of drugs and carriers has been largely ignored. The purpose of the present study was to quantitatively and reproducibly generate a partially (∼50%) liquefied vitreous model in laboratory rabbits that mimics the aged condition, to better assess drug disposition in the aged eye. 
Materials and Methods
Preparation of Hyaluronidase, Sodium Fluorescein, and Fluorescein Isothiocyanate-Dextran (FD) 150 kDa Solutions and 1 μM Fluorescent Particle Suspension
Lyophilized ovine testicular hyaluronidase with enzyme activity of ≥1000 IU/mg (MP Biomedicals, Irvine, CA) was reconstituted with 0.9% normal saline, and serial dilution was performed to produce hyaluronidase injection solutions of concentration 0.005 IU/20 μL. Fluorescein isothiocyanate-dextran, average molecular weight 150 kDa (Sigma, St. Louis, MO), was dissolved in 0.9% normal saline to produce a 0.01% (wt/vol) solution. A 10% (wt/vol) sodium fluorescein solution (Akorn, Inc., Buffalo Grove, IL) was diluted in 0.9% normal saline to 0.01% (wt/vol) before use. A 1-μm fluorescent polystyrene latex microsphere suspension (Fluoresbrite) was purchased from Polysciences, Inc. (Warrington, PA) and was used without further dilution. 
In Vivo Model of Vitreous Liquefaction
All animal-handling procedures were approved by the Allergan Animal Care and Use Committee and complied with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Fourteen Dutch-belted rabbits weighing 2 to 3 kg, with ages ranging from 3 months to 2 years, were involved in the developmental phase of the model. Before the start of the experiments, the eyes of all animals were examined externally to exclude abnormalities. Before intravitreal injections, the rabbits were anesthetized locally with topical 0.5% tetracaine HCl ophthalmic solution USP (Bausch & Lomb, Tampa, FL), and their pupils were dilated with topical 10% phenylephrine HCl (Akorn, Inc.) and 1% tropicamide USP (Alcon Laboratories, Fort Worth, TX). Twenty microliters of a hyaluronidase solution of 0.005 IU (n = 11) was intravitreally injected to induce vitreous liquefaction; another four animals received bilateral injections of 20 μL normal saline to act as the control. The animals were euthanatized 24 (n = 5) or 48 (n = 6) hours later with intravenous injection of pentobarbital (Eutha-6; Western Medical Supply Co., Arcadia, CA), and the eyes were enucleated and dissected to obtain the vitreous humor. The vitreous samples were then collected onto a Petri dish, and the gel–liquid ratio was determined by first weighing the entire vitreous and then weighing the liquid phase after removing the gel phase with a pair of forceps. The percentage of vitreous gel content was calculated as [(mass of gel phase/total mass of vitreous) × 100%]. The significance of differences in the percentage gel phase between normal and enzyme-treated eyes was evaluated by Student's t-test with the significance level defined at P < 0.05. During the study, evidence of ocular toxicity was monitored with a fundus camera. 
In Vivo Measurements of Fluorophores Distribution Using Retinal Angiography with a Confocal Laser Scanning System and Ocular Fluorophotometer
After the model was developed, nine animals were used to study the effects of vitreous liquefaction on distribution of material. Twenty microliters of 0.005 IU hyaluronidase solution was injected into the animal eyes unilaterally, leaving the contralateral eye untreated (normal vitreous) to act as a control. Forty-eight hours later, the animals received bilateral intravitreal injections of 10 μL 0.01% sodium fluorescein (n = 3) or 1% fluorescein isothiocyanate-dextran (FD; MW 150 kDa; n = 3) solution or a suspension of 1 μm fluorescent particles (n = 3). The distribution of the injected fluorescent molecules was monitored for a month with retinal angiography (HRA 2; Heidelberg Engineering, Vista, CA), with a wide-angle ocular Staurenghi scanning laser ophthalmoscope (SLO) retinal lens (Staurenghi 230; Ocular Instruments, Bellevue, WA) and an ocular fluorophotometer (Ocumetrics, Inc., San Francisco, CA). The retinal angiograph allows the imaging of fluorescent probes in both the x- and y-axes, while the ocular fluorophotometer detects the fluorescence intensity along the optical (z) axis. Ocular fluorophotometer data of the study groups were compared by using Student's t-test, with a confidence level of P < 0.05 considered statistically significant. 
Image Analysis
The experimental images obtained by the retinal angiograph were analyzed by using the image-processing feature in commercial software (MatLab; The MathWorks Inc, Natick, MA) to quantitatively compare the retention of the fluorescent mass of sodium fluorescein and FD 150 kDa in the enzyme-treated and contralateral normal vitreous. With this method, the degree of distribution and clearance of injected molecules was correlated to the fluorescence intensity (in pixels) detected within the vitreous chamber, with lower intensities indicating faster rate of transport. The relative distribution and clearance of both models can be compared by calculating the fluorescence intensity ratio of normal to liquefied vitreous. 
Results
Development of a Model of Vitreous Liquefaction
Figure 1 shows the degree of vitreous liquefaction induced by 0.005 IU of hyaluronidase at 24 or 48 hours after injection. The 24-hour postinjection period produced a lower degree of liquefaction (77.9% ± 6.1%; P > 0.05) compared with that of the 48-hour postinjection period. 
Figure 1.
 
Percentage vitreous gel phase in control and hyaluronidase-treated groups. The difference in percentage gel phase between the control and 0.005 IU (48 hours after injection) hyaluronidase-treated groups was statistically significant. *P < 0.05.
Figure 1.
 
Percentage vitreous gel phase in control and hyaluronidase-treated groups. The difference in percentage gel phase between the control and 0.005 IU (48 hours after injection) hyaluronidase-treated groups was statistically significant. *P < 0.05.
In addition, a 0.005-IU dose with a 48-hour postinjection period generated the target degree of vitreous liquefaction in rabbits in a reproducible fashion. In comparison to the control, there was slight haziness in the vitreous chamber of the hyaluronidase-treated eye, but no gross ocular tissue changes were identified (Fig. 2). This model was then used to assess the effects of vitreous liquefaction on the disposition of three model compounds injected intravitreally. 
Figure 2.
 
Fundus assessments of the medullary rays and optic nerve head in control (left) and hyaluronidase-treated (right) eyes. (a) Medullary rays; (b) optic nerve head; (c) right medullary rays and optic nerve head.
Figure 2.
 
Fundus assessments of the medullary rays and optic nerve head in control (left) and hyaluronidase-treated (right) eyes. (a) Medullary rays; (b) optic nerve head; (c) right medullary rays and optic nerve head.
Effects of Vitreous Liquefaction on the Dispersion of Model Compounds
Sodium Fluorescein.
Figure 3 shows a set of representative retinal angiography images for comparison of the intravitreal distribution of sodium fluorescein in both normal (control) and liquefied (hyaluronidase-treated) vitreous at 2 and 5 hours after injection. The fluorescent depot in the liquefied vitreous spread faster from the injection pocket and perhaps from the vitreous cavity. An analysis constructed with the image-analysis program used in our laboratories (MatLab; The MathWorks) indicated that sodium fluorescein clearance, as measured by fluorescence intensity, was 1.4 and 1.5 times faster in the liquefied vitreous at 2 and 5 hours, respectively, than that measured in the control. Figure 4 shows the distribution profiles along the z-axis of the eye at 2 and 5 hours for both models. At the same injected dose, sodium fluorescein concentration was found to be lower in the liquefied vitreous at all times (P < 0.05). However, the fluorescent mass distributed with a similar concentration gradient in both models. The plateau in the control eye data at 2 hours between 17 and 21 mm from the origin probably indicates penetration into the lens. The probe did not continue a forward diffusion, since the vectoral fluid flow in the lens opposed diffusion, unloading absorbed fluorescein back into the vitreous humor at 5 hours after administration. 14 The behavior was less obvious in the liquefied vitreous, presumably because the concentration gradient cannot be sustained. Table 1 is a summary of the retinal angiography and ocular fluorophotometry results for the study models. 
Figure 3.
 
Representative retinal angiograph images obtained from one animal 2 hours and 5 hours after receiving a 10 μL intravitreal injection of sodium fluorescein solution. The fluorescent mass was well retained in the intact vitreous after 5 hours, but was more diffuse in the liquefied vitreous.
Figure 3.
 
Representative retinal angiograph images obtained from one animal 2 hours and 5 hours after receiving a 10 μL intravitreal injection of sodium fluorescein solution. The fluorescent mass was well retained in the intact vitreous after 5 hours, but was more diffuse in the liquefied vitreous.
Figure 4.
 
Ocular fluorophotometry of intravitreally injected sodium fluorescein depot at 2 (left) and 5 (right) hours after injection (n = 3). On these graphs, 0 mm represents the vitreoretinal interface; 15 mm represents the back of the lens.
Figure 4.
 
Ocular fluorophotometry of intravitreally injected sodium fluorescein depot at 2 (left) and 5 (right) hours after injection (n = 3). On these graphs, 0 mm represents the vitreoretinal interface; 15 mm represents the back of the lens.
Table 1.
 
Summary Data for Sodium Fluorescein and FD 150 kDa in the Study Models
Table 1.
 
Summary Data for Sodium Fluorescein and FD 150 kDa in the Study Models
Sodium Fluorescein FD 150 kDa
1 h 2 h 5 h 1 h 2 h 5 h 2 d 3 d 6 d 9 d 15 d 21 d 30 d
HRA intensity ratio 1.1 1.4 1.5 0.87 0.84 0.97 1.04 1.15 2.6 3.8 >10 >10 2.17
OF Student's t-test P <0.05* P >0.05 P <0.05*
FD 150 kDa.
Figure 5 is a set of representative retinal angiography wide-angle images obtained from one animal for comparison of the distribution of FD 150 kDa in the vitreous cavity over a period of 1 month. The figure also shows that the FD 150 kDa solution formed a defined depot that lasted for at least 5 hours after injection and gradually spread across the vitreous chamber thereafter, as indicated by the increasing diameter of the fluorescent mass. There was no significant difference in the distribution of FD 150 kDa between the models until day 3. However, from day 6 onward, the amount of FD 150 kDa that remained in the vitreous was lower in the liquefied model than in the control. There was no residual FD 150 kDa detected optically in the liquefied vitreous model on day 30 but a low degree of fluorescence intensity remained visible in the normal vitreous, suggesting FD 150 kDa had a shorter intravitreal half-life in the liquefied vitreous. Figure 6 shows the graph generated with the image-analysis program to calculate the fluorescence intensity of FD 150 kDa remaining in the vitreous humor over time. The y-axis on the plot represents fluorescence intensity (in pixels) remaining within the vitreous cavity and the scale is in arbitrary units. FD 150 kDa was cleared faster in the liquefied vitreous compared with the normal vitreous model. Figure 7 shows the plots obtained from the ocular fluorophotometry study. At 1 hour after injection, most of the injected FD 150 kDa remained at the site of injection in both models. However, the 2- and 5-hour data suggest that FD 150 kDa moved toward the anterior chamber in the liquefied vitreous, but remained in place in the normal vitreous. On day 2, FD 150 kDa started to spread across the optical axis, as shown by the more even distribution of fluorescence. In addition, the fluorescence gradient was steeper in the liquefied vitreous than in the normal vitreous on days 2 and 3. On day 6, the gradient decreased, indicating that the initial phase of diffusive equilibration was completed. The data were in agreement with the retinal angiograph's wide-angle image, where the fluorescence intensity was evenly distributed across the chamber in the liquefied vitreous. In the normal vitreous, the general fluorescence gradient did not change considerably over similar time periods. As the FD 150 kDa was cleared from the vitreous cavity of the study models, the fluorescence intensity reduced in both x- and y-axes as detected by retinal angiography and ocular fluorophotometry. Statistical analyses showed that the fluorescence intensity of FD 150 kDa was significantly higher (P < 0.05) in normal vitreous on days 21 and 30, which agreed with the angiography observations. 
Figure 5.
 
Retinal angiograph wide-angle images of intravitreally injected FD 150 kDa in the normal and liquefied vitreous models from 1 hour to 30 days after injection.
Figure 5.
 
Retinal angiograph wide-angle images of intravitreally injected FD 150 kDa in the normal and liquefied vitreous models from 1 hour to 30 days after injection.
Figure 6.
 
Plot generated (MatLab; The MathWorks) based on retinal angiograph images comparing the fluorescence intensity of FD 150 kDa from 2 hours to 30 days after injection in both liquefied and normal vitreous models. Data are the mean (n = 3) ± SD. The curves were consistent with the angiography findings.
Figure 6.
 
Plot generated (MatLab; The MathWorks) based on retinal angiograph images comparing the fluorescence intensity of FD 150 kDa from 2 hours to 30 days after injection in both liquefied and normal vitreous models. Data are the mean (n = 3) ± SD. The curves were consistent with the angiography findings.
Figure 7.
 
Ocular fluorophotometry data obtained from the study animals showing the fluorescence intensity along the optical axis from the back of the lens (15 mm) through to surface of the retina (0 mm).
Figure 7.
 
Ocular fluorophotometry data obtained from the study animals showing the fluorescence intensity along the optical axis from the back of the lens (15 mm) through to surface of the retina (0 mm).
One-Micrometer Particle Suspension.
The intravitreal distribution of 1 μm fluorescent particles was assessed with retinal angiography. The images obtained are shown in Figure 8. Gravitational forces resulted in sedimentation of microparticles in the liquefied vitreous model on day 9, whereas microparticles in the normal vitreous settled later. In the normal vitreous, microparticles sedimented without much lateral diffusion as opposed to the liquefied vitreous, where the particles appeared more dispersed and scattered. 
Figure 8.
 
A representative set of angiograph images obtained from one of the three animals involved in the study. The images were taken from 1 hour until 30 days after injection. The upper part of the image represents the superior area of the rabbit vitreous. Microparticles remained relatively in place at the injection site (superior-temporal) in both models for at least 5 hours after injection.
Figure 8.
 
A representative set of angiograph images obtained from one of the three animals involved in the study. The images were taken from 1 hour until 30 days after injection. The upper part of the image represents the superior area of the rabbit vitreous. Microparticles remained relatively in place at the injection site (superior-temporal) in both models for at least 5 hours after injection.
Discussion
In this study, a partially liquefied vitreous model was successfully developed in Dutch-belted rabbits for ocular drug delivery research. To the best of our knowledge, no study to date has quantitatively established a partially liquefied vitreous model with hyaluronidase. Several studies have described intervention for vitreoretinal problems, with hyaluronidase used to induce posterior vitreous detachment (PVD) with varying success. 15 17 Tanaka and Qui, 18 Gottlieb et al., 19 and others 20,21 have demonstrated that vitreous liquefaction is feasible after hyaluronidase injection, but without quantifying the degree of liquefaction. 
In addition, our data show that in an unmodified rabbit vitreous, only ∼20% of its content was composed of a liquid phase, which was similar to that in young human adults. 3 Hyaluronidase liquefied the vitreous by cleaving the glycosidic bonds of the vitreous hyaluronan and changed its structural conformation. 22 The resultant depolymerization altered its hydration property, leading to the release of bound water, as indicated by the increase in vitreous liquid volume and decrease in the gel wet weight observed in this study. We concluded that 0.005 IU/20 μL with a 48-hour exposure achieved the target degree of liquefaction representative of elderly vitreous. Bracketing this concentration with single-point administrations at higher and lower amounts of enzyme suggested that the concentration and exposure time are optimal. In further in-house studies with this model, we have seen no further changes in the liquid–gel ratio between 3 and 30 days after injection of the enzyme into similar animals at the same doses. This observation was further supported by the literature, which has reported that hyaluronidase-induced disaggregation of vitreous hyaluronic acid remains for at least 1 month before being replaced by the newly formed hyaluronic acid. 21 Based on the old observations by Pirie, 21 it was possible that partial vitreous liquefaction persisted for the entire duration of the experiment. Nevertheless, we were unable to confirm in this study whether the liquefaction generated was in the form of bulk fluid or several liquid pockets embedded in the gel phase. This process must be further investigated. 
The interanimal variability (∼9%) was similar in both the normal and the hyaluronidase-treated groups, emphasizing the robustness of the liquefied model. In addition, it has been reported that hyaluronidase is nontoxic to rabbit retina at a dose of 1 U. 19 The safety of this enzyme was not a prime objective for investigation in this study, but no gross ocular tissue changes were observed with fundus examination. 
Three model compounds were selected to assess the impact of partial vitreous liquefaction on the intravitreal movements of molecules with different molecular weights. The molecular size of sodium fluorescein can be a representative compound to many intravitreal antimicrobial agents and steroids used in the treatments of ocular infection and inflammation. FD 150 kDa is a large molecule, and thus the molecular size might model for antibodies such as bevacizumab. While 1 μm fluorescent particles are representative of the sizes used in microparticulate delivery systems, they are also ideal for examining intravitreal convective flow, since diffusive movement is limited for this object size. Drug movements within the vitreous chamber and their intravitreal lifetime depend largely on molecular size. Our data show that FD 150 kDa spread slower and had a longer intravitreal half-life of approximately 30 days compared with sodium fluorescein, which was cleared within 48 hours. Microparticles, which were the largest among others, remained in the vitreous for longer than 1 month. These common findings with respect to larger molecules being retained better in the vitreous than smaller molecules have been documented. 23 The likely explanation for this phenomenon is that the diffusion process of larger molecules is limited by the barrier functions of the collagen–hyaluronan network. 24 Unlike small molecules, which can diffuse in all directions, macromolecules have to move between the porous gel meshwork and thus, the transport process is largely depending on convective flow. 23,25,26  
Ocular fluorophotometry showed the presence of concentration gradient across the vitreous chamber with the highest concentration near the anterior chamber and the lowest at the vitreoretinal surface, suggesting diffusion was leading toward the retina. The clearance pathways of sodium fluorescein and FD 150 kDa were not investigated in this study. However, Araie and Maurice, 27 and Cunha-Vaz and Maurice, 28 based on their work examining the vitreous sections of the frozen eye after intravitreal injection of fluorescent molecules, have related this type of distribution pattern to retinal clearance. Fluorescein is a small hydrophile, and its clearance from the vitreous appears to be aided by a small nucleotide transporter, P2Y2. 29 However, FD 150 kDa is a relatively large, hydrophilic molecule, that is not transported by this route, and permeation through the retinal layer remains in question. Dias and Mitra 30 attributed this observation to the possibility that FD 150 kDa was cleared via the paracellular or endocytic mechanism present at the retina layers. In addition, the movements of microparticles were more restrictive and slower than sodium fluorescein and FD 150 kDa, perhaps due to its hydrophobic interactions with collagen fibers. 7  
Moldow et al. 31 showed that the diffusion profile of fluorescein in the elderly liquefied vitreous was more homogeneous as opposed to a healthy vitreous. In addition, Foulds et al. 9 reported that hyaluronidase increases the clearance of tritiated water. These studies demonstrated that small molecules are distributed and cleared faster in the liquefied vitreous system, an observation that was also illustrated in our study on sodium fluorescein. Since diffusion is the main drive for the transport of small molecules, it is possible that diffusivity in liquefied vitreous is enhanced. 
On the other hand, the movement of large molecules depends on the convective flow process. 23 In partially liquefied vitreous, FD 150 kDa was cleared faster, which may indicate greater convective movement in the liquefied vitreous circulation. A lower concentration of FD 150 kDa was detected in the liquefied vitreous, suggesting a faster rate of distribution and that the capability of the vitreous to sustain injected doses had decreased significantly. The plot derived from ocular fluorophotometry suggests that there was a temporary forward flux in the partially liquefied vitreous model during the first few hours after injection. The same observation was not noted in normal vitreous. This result indicates that the flow processes in the liquefied vitreous generated this movement. Interestingly, forward flux was not evident in the case of sodium fluorescein. We can attribute this finding to the fact that sodium fluorescein diffused faster than FD 150 kDa, and thus if present, it may take place at the first few minutes after injection, which was not captured in this study. The temporary forward flux of FD 150 kDa in partially liquefied vitreous resulted in a steeper vitreous concentration gradient. However, the gradient pattern remained the same with the level of FD 150 kDa reducing from anterior to posterior at later time points. Hence, we deduced that partial vitreous liquefaction increased the rate of material clearance, but materials still moved along the same concentration gradient. More importantly, this observation could explain the need for more frequent injections of bevacizumab among the elderly population (mean age, 68.5 years; 3.75 injections) compared with younger individuals (mean age, 39.5 years; 1.75 injections) seen clinically by Spielberg and Leys. 32 Therefore, for larger molecules such as antibodies and antibody fragments injected into the vitreous, the faster rate of clearance in partially liquefied vitreous has to be taken into consideration as part of current research investigating optimal dose regimens in elderly patients. 
In the case of 1 μm fluorescent particles, retinal angiography (HRA2; Heidelberg Engineering) was used to investigate the difference between the normal and liquefied vitreous models. The fluorescence signals from the microparticles measured with the ocular fluorophotometry had a high level of noise leading to unreliable data. Angiography data show that clusters of fluorescent microparticles were more scattered and sedimented faster under gravitational pull in the partially liquefied vitreous system versus the control eyes. Therefore, tissue exposure on the floor of the inferior quadrant of the eye may be greater in the more liquefied vitreous due to faster sedimentation of injected microparticles. In contrast, the local injection site experiences a higher concentration in a more gel-like vitreous, particularly if there is reflux back to the point of injection. Maurice 33 proposed that particle sedimentation may increase the risk of dangerously high drug concentration achieved at the retina, especially if the patient is sleeping in a supine position. Localized multinuclear giant-cell reaction has also been reported around retina tissues that were in contact with the polymeric microspheres. 34 This information will be beneficial in the design of microparticulate formulations in optimizing the particle size, drug loading, duration and rate of drug release and in targeting the desired site of action in the elderly population. In summary, we have quantitatively established a reproducible partially liquefied vitreous model in laboratory rabbits that represents the degree of vitreous liquefaction that occurs in the general population ∼60 years of age. Therefore, it can serve as a useful investigative tool in the future of ocular treatments for the elderly population. We have determined the percentage gel phase of unmodified rabbit vitreous and the result showed a close similarity to young human adults. Therefore, the use of juvenile animal models may underestimate the influence of convective forces present in the elderly patients, thereby overestimating drug efficacy. The joint operation of ocular fluorophotometry and retinal angiography in the present study provided a better understanding of intravitreal events. We managed to correlate data from both techniques in a quantitative manner. Unlike previous studies, we were able to relate the degree of vitreous liquefaction to material transport. Therefore, data provided in this study are more relevant to actual clinical situations. More studies are currently being undertaken in our laboratory to further understand the influence of vitreous liquefaction on drugs released from controlled release devices. However, the limitation of our current model is that it has neglected the influence of collagen fibers formed on aging, which may be important for drug molecules prone to protein binding. Future studies should take this factor into consideration. 
Footnotes
 Presented in part at the ARVO 2009 Summer Eye Research Conference, National Institutes of Health, Bethesda, Maryland, July 31 to August 1, 2009.
Footnotes
 Supported by an open educational scholarship from Allergan, Inc. (LET).
Footnotes
 Disclosure: L.E. Tan, Allergan, Inc. (F); W. Orilla, Allergan, Inc. (E); P.M. Hughes, Allergan, Inc. (E); S. Tsai, Allergan, Inc. (E) J.A. Burke, Allergan, Inc. (E); C.G. Wilson, Allergan, Inc. (C)
The authors thank Stephen Marshall and Paul Murray (Electronic and Electrical Engineering Department, University of Strathclyde, Glasgow) for help in composing the computer program for retinal angiography (HRA; Heidelberg Engineering) wide-angle image analysis. The details of the program will be described in another publication. 
References
Sebag J . Macromolecular structure of the corpus vitreous. Prog Polym Sci. 1998;23:415–446. [CrossRef]
Bishop PN . Structural macromolecules and supramolecular organisation of the vitreous gel. Prog Retin Eye Res. 2000;19(3):323–344. [CrossRef] [PubMed]
Balazs EA Denlinger JL . Aging changes in the vitreous. Aging and Human Visual Function. New York: Alan R. Liss; 1982;45–47.
Sebag J. Balazs EA . Morphology and ultrastructure of human vitreous fibers. Invest Ophthalmol Vis Sci. 1989;30(8):1867–1871. [PubMed]
Sebag J Balazs EA . Human vitreous fibres and vitreoretinal disease. Trans Ophthalmol Soc UK. 1985;104(Pt 2):123–128. [PubMed]
Sebag J . Age-related changes in human vitreous structure. Graefes Arch Clin Exp Ophthalmol. 1987;225:89–93. [CrossRef] [PubMed]
Peeters L Sanders NN Braeckmans K . Vitreous: a barrier to nonviral ocular gene therapy. Invest Ophthalmol Vis Sci. 2005;46:3553–3561. [CrossRef] [PubMed]
Pitkänen L Ruponen M Nieminen J . Vitreous is a barrier in nonviral gene transfer by cationic lipids and polymers. Pharm Res. 2003;20(4):576–583. [CrossRef] [PubMed]
Foulds WS Allan D Moseley H . Effect of intravitreal hyaluronidase on the clearance of tritiated water from the vitreous to the choroid. Br J Ophthalmol. 1985;69:529–532. [CrossRef] [PubMed]
Sebag J Ansari RR Suh KI . Pharmacologic vitreolysis with microplasmin increases vitreous diffusion coefficients. Graefes Arch Clin Exp Ophthalmol. 2007;245:576–580. [CrossRef] [PubMed]
Lee SS Ghosn C Yu Z . Vitreous VEGF clearance is increased following vitrectomy. Invest Ophthalmol Vis Sci. 2010;51:2135–2138. [CrossRef] [PubMed]
Chin HS Park TS Moon YS . Difference in clearance of intravitreal triamcinolone acetonide between vitrectomized and nonvitrectomized eyes. Retina. 2005;25(5):556–560. [CrossRef] [PubMed]
Ficker L Meredith TA Gardner S . Cefazolin levels after intravitreal injection. Invest Ophthalmol Vis Sci. 1990;31(3):502–505. [PubMed]
Stepanova LV Marchenko IY Sychev GM . Direction of fluid transport in the lens. Bull Exp Biol Med. 2005;139(1):50–51. [CrossRef] [PubMed]
Harooni M McMillan T Refojo M . Efficacy and safety of enzymatic posterior vitreous detachment by intravitreal injection of hyaluronidase. Retina. 1998;18:16–22. [CrossRef] [PubMed]
Hikichi T Kado M Yoshida A . Intravitreal injection of hyaluronidase cannot induce posterior vitreous detachment in the rabbit. Retina. 2000;20(2):195–198. [CrossRef] [PubMed]
Wang ZL Zhang X Xu X . PVD following plasmin but not hyaluronidase: implications for combination pharmacologic vitreolysis therapy. Retina. 2005;25:38–43. [CrossRef] [PubMed]
Tanaka M Qui H . Pharmacological vitrectomy. Semin Ophthalmol. 2000;15:51–61. [CrossRef] [PubMed]
Gottlieb JL Antoszyk AN Hatchell DL . The safety of intravitreal hyaluronidase, a clinical and histologic study. Invest Ophthalmol Vis Sci. 1990;31:2345–2352. [PubMed]
Winkler BS Cohn EM . Hyaluronidase and retinal function. Arch Ophthalmol. 1985;103:1743–1746. [CrossRef] [PubMed]
Pirie A . The effect of hyaluronidase injection on the vitreous humor of the rabbit. Br J Ophthalmol. 1949;33:678–684. [CrossRef] [PubMed]
Narayanan R Kuppermann BD . Hyaluronidase for pharmacologic vitreolysis. Dev Ophthalmol. 2009;44:20–25. [PubMed]
Park J Bungay PM Lutz RJ . Evaluation of coupled convective-diffusive transport of drugs administered by intravitreal injection and controlled release implant. J Control Release. 2005;105(3):279–295. [CrossRef] [PubMed]
Worst JGF Los LI . Functional anatomy of the vitreous. Cisternal Anatomy of the Vitreous. Amsterdam: Kugler Publications; 1995;33–48.
Stay MS Xu J Randolph TW Barocas VH . Computer simulation of convective and diffusive transport of controlled-release drugs in the vitreous humor. Pharm Res. 2003;20(1):96–102. [CrossRef] [PubMed]
Xu J Heys JJ Barocas VH . Permeability and diffusion in vitreous humor: Implications for drug delivery. Pharm Res. 2000;17(6):664–669. [CrossRef] [PubMed]
Araie M Maurice DM . The loss of fluorescein, fluorescein glucuronide and fluorescein isothiocyanate dextran from the vitreous by the anterior and retinal pathways. Exp Eye Res. 1991;52:27–39. [CrossRef] [PubMed]
Cunha-Vaz JG Maurice DM . The active transport of fluorescein by the retinal vessels and the retina. J Physiol (Lond.) 1967;191:467–486. [CrossRef] [PubMed]
Takahashi J Hikichi T Mori F . Effect of nucleotide P2Y2 receptor agonists on outward active transport of fluorescein across normal blood-retina barrier in rabbit. Exp Eye Res. 2004;78(1):103–108. [CrossRef] [PubMed]
Dias CS Mitra AK . Vitreal elimination kinetics of large molecular weight FITC-labeled dextrans in albino rabbits using a novel microsampling technique. J Pharm Sci. 2000;89:572–578. [CrossRef] [PubMed]
Moldow B Sander B Larsen M . The effect of acetazolamide on passive and active transport of fluorescein across the blood-retina barrier in retinitis pigmentosa complicated by macular oedema. Graefes Arch Clin Exp Ophthalmol. 1998;236:881–889. [CrossRef] [PubMed]
Spielberg L Leys A . Intravitreal bevacizumab for hyopic choroidal neovascularization: short-term and 1-year results. Bull Soc Belge Ophtalmol. 2009;312:17–27. [PubMed]
Maurice D . Practical issues in intravitreal drug delivery (review). J Ocul Pharmacol Ther. 2001;17(4):393–401. [CrossRef] [PubMed]
Giordano GG Refojo MF Arroyo MH . Sustained delivery of retinoic acid from microspheres of biodegradable polymer in PVR. Invest Ophthalmol Vis Sci. 1993;34:2743–2751. [PubMed]
Figure 1.
 
Percentage vitreous gel phase in control and hyaluronidase-treated groups. The difference in percentage gel phase between the control and 0.005 IU (48 hours after injection) hyaluronidase-treated groups was statistically significant. *P < 0.05.
Figure 1.
 
Percentage vitreous gel phase in control and hyaluronidase-treated groups. The difference in percentage gel phase between the control and 0.005 IU (48 hours after injection) hyaluronidase-treated groups was statistically significant. *P < 0.05.
Figure 2.
 
Fundus assessments of the medullary rays and optic nerve head in control (left) and hyaluronidase-treated (right) eyes. (a) Medullary rays; (b) optic nerve head; (c) right medullary rays and optic nerve head.
Figure 2.
 
Fundus assessments of the medullary rays and optic nerve head in control (left) and hyaluronidase-treated (right) eyes. (a) Medullary rays; (b) optic nerve head; (c) right medullary rays and optic nerve head.
Figure 3.
 
Representative retinal angiograph images obtained from one animal 2 hours and 5 hours after receiving a 10 μL intravitreal injection of sodium fluorescein solution. The fluorescent mass was well retained in the intact vitreous after 5 hours, but was more diffuse in the liquefied vitreous.
Figure 3.
 
Representative retinal angiograph images obtained from one animal 2 hours and 5 hours after receiving a 10 μL intravitreal injection of sodium fluorescein solution. The fluorescent mass was well retained in the intact vitreous after 5 hours, but was more diffuse in the liquefied vitreous.
Figure 4.
 
Ocular fluorophotometry of intravitreally injected sodium fluorescein depot at 2 (left) and 5 (right) hours after injection (n = 3). On these graphs, 0 mm represents the vitreoretinal interface; 15 mm represents the back of the lens.
Figure 4.
 
Ocular fluorophotometry of intravitreally injected sodium fluorescein depot at 2 (left) and 5 (right) hours after injection (n = 3). On these graphs, 0 mm represents the vitreoretinal interface; 15 mm represents the back of the lens.
Figure 5.
 
Retinal angiograph wide-angle images of intravitreally injected FD 150 kDa in the normal and liquefied vitreous models from 1 hour to 30 days after injection.
Figure 5.
 
Retinal angiograph wide-angle images of intravitreally injected FD 150 kDa in the normal and liquefied vitreous models from 1 hour to 30 days after injection.
Figure 6.
 
Plot generated (MatLab; The MathWorks) based on retinal angiograph images comparing the fluorescence intensity of FD 150 kDa from 2 hours to 30 days after injection in both liquefied and normal vitreous models. Data are the mean (n = 3) ± SD. The curves were consistent with the angiography findings.
Figure 6.
 
Plot generated (MatLab; The MathWorks) based on retinal angiograph images comparing the fluorescence intensity of FD 150 kDa from 2 hours to 30 days after injection in both liquefied and normal vitreous models. Data are the mean (n = 3) ± SD. The curves were consistent with the angiography findings.
Figure 7.
 
Ocular fluorophotometry data obtained from the study animals showing the fluorescence intensity along the optical axis from the back of the lens (15 mm) through to surface of the retina (0 mm).
Figure 7.
 
Ocular fluorophotometry data obtained from the study animals showing the fluorescence intensity along the optical axis from the back of the lens (15 mm) through to surface of the retina (0 mm).
Figure 8.
 
A representative set of angiograph images obtained from one of the three animals involved in the study. The images were taken from 1 hour until 30 days after injection. The upper part of the image represents the superior area of the rabbit vitreous. Microparticles remained relatively in place at the injection site (superior-temporal) in both models for at least 5 hours after injection.
Figure 8.
 
A representative set of angiograph images obtained from one of the three animals involved in the study. The images were taken from 1 hour until 30 days after injection. The upper part of the image represents the superior area of the rabbit vitreous. Microparticles remained relatively in place at the injection site (superior-temporal) in both models for at least 5 hours after injection.
Table 1.
 
Summary Data for Sodium Fluorescein and FD 150 kDa in the Study Models
Table 1.
 
Summary Data for Sodium Fluorescein and FD 150 kDa in the Study Models
Sodium Fluorescein FD 150 kDa
1 h 2 h 5 h 1 h 2 h 5 h 2 d 3 d 6 d 9 d 15 d 21 d 30 d
HRA intensity ratio 1.1 1.4 1.5 0.87 0.84 0.97 1.04 1.15 2.6 3.8 >10 >10 2.17
OF Student's t-test P <0.05* P >0.05 P <0.05*
×
×

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.

×