Biocompatibility of a Synthetic Biopolymer for the Treatment of Rhegmatogenous Retinal Detachment

Copyright: © 2015 Sarfare S, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Biocompatibility of a Synthetic Biopolymer for the Treatment of Rhegmatogenous Retinal Detachment Shanta Sarfare1*, Yann Dacquay1, Syed Askari2, Steven Nusinowitz1 and Jean-Pierre Hubschman1 1Jules Stein Eye Institute, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, California 90095, USA 2Medicus Biosciences, 2528 Qume Drive, Unit 1, San José, California 95131, USA


Results:
The biopolymer was visualized in the subretinal space in vivo by SD-OCT and post-life by histology up to 1 week after the injection. There were no significant differences in ERG parameters between the two groups at 1 and 2 weeks post-injection. Minimal inflammatory response and loss of photoreceptor cells was only observed in the immediate proximity of the site of scleral perforation, which was similar in both groups. Overall integrity of the outer, inner retina and retinal pigment epithelial (RPE) layers was unaffected by the presence of the biopolymer in the subretinal space.
Conclusions: Functional and histological evaluation suggests that the synthetic biopolymer is non-inflammatory and non-toxic to the eye. It may represent a safe therapeutic agent in the future, for the treatment of rhegmatogenous retinal detachment.
which has the ability to effectively seal the retinal breaks in retinal detachment surgery.
Polyethylene glycol (PEG) derivatives with molecular weights 10 kilo Dalton (kDa) and higher are generally recognized as safe and nontoxic [20] and therefore have been extensively used in medical devices and as drug carriers in pharmaceuticals applications [21][22][23]. Our study aims to evaluate the potential ocular and in particular the retinal toxicity of a PEG-derived ophthalmic biopolymer through clinical, electrophysiological and histological evaluations.

Animals
Albino Balb/c mice 6-8 weeks of age purchased from Jackson Laboratories, ME maintained on 12/12 hours light/dark cycle were used in this study. All experiments were performed with IACUC-approved protocols and following the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

Functional evaluation
Electroretinography (ERG): Dark adapted rod-mediated ERGs were recorded to blue flashes (Wratten 47A, λ max =470 nm). Bilateral ERG recordings were performed on 16 mice, at baseline 1 week prior and 1 and 2-weeks post injection, as previously described [29]. ERGs were recorded from the cornea using a gold loop electrode, referenced to a similar gold wire in the mouth. The mouse eye was positioned in front of an opening in a large integrating sphere in which brief flashes of light at 1 Hz interval were presented. Responses were amplified 10,000 × (Grass P511 High Performance AC Amplifier), band-pass filtered (0.1-300 Hz), digitized using an I/O board (PCI-6221; National Instruments, Austin, TX) and averaged.
Histological evaluation: Light, electron microscopy and immunocytochemistry.
Paraffin embedded tissue (FFPE): Eyes from mice (group A, n=3; group B, n=13) 3 weeks post-injection were fixed in 10% neutral buffered formalin, paraffin embedded and sectioned. Consecutive 5-μm sections were stained with hematoxylin and eosin (H&E) for light microscopy and processed for TUNEL labeling or immunohistochemistry.
Plastic embedded tissue: Mouse eyes (group A, n=3; group B, n=6) were oriented with respect to site of injection using cautery iron or tissue marker pen, fixed with a mixture of 2% paraformaldehyde: 2.5% glutaraldehyde in 0.1M sodium phosphate buffer (pH 7.4) and processed as previously described [30]. Sections of 1µm thickness stained with Toluidine blue were imaged with a Zeiss Axiophot microscope and ultra-thin sections stained with uranium and lead salts were imaged using Zeiss EM910 electron microscope.

Color fundus imaging of subretinal bleb
Due to co-injection of fluorescein with the biopolymer, a hyper fluorescent spot with a distinct outline marking the detachment can be seen in subretinal biopolymer injected eyes ( Figure 1). The retinal vessels are clearly seen in front of the hyper fluorescent area, indicating a true subretinal injection without leakage in the vitreous space. In fellow eyes with no injections, there is no sign of retinal change. The fluorescein cleared rapidly and the subretinal blebs start regressing in

Generation of the ophthalmic biopolymer
The ophthalmic pre-formulation developed by Medicus Biosciences (San José, California) comprises of three compounds: one containing 8 nucleophilic groups (8-arm-PEG 20K amine), the second also containing 8 nucleophilic groups and 8 degradable acetate groups (8-arm-PEG 20K amine acetate) and the third containing 4 electrophilic groups (4-arm-PEG 20K amide ester) in a stoichiometric ratio; an aqueous buffer and a viscosity enhancer [24][25][26][27]. The polymer was designed to be degraded in 2-weeks by selecting degradable acetate groups, numbers of crosslinking sites and the polymer concentration in the solution. Retinal adherence and the viscosity of the polymer were enhanced with the use of hydroxyl methyl propyl methyl cellulose. The polymer demonstrated in previous study similar adhesive properties as the retina when tested in our laboratory [28]. All reactive PEG materials were purchased from JenKem Corp, Plano, Texas. The firmness, adhesion, and elastic modulus of the polymer were characterized by a Texture Analyzer (TA.XT.plus) with Exponent software (v6.0) using various probes. The components in the injection kit were mixed, filtered and injected into the subretinal space as a liquid biopolymer which solidifies at pH 7.4 at a pre-set time of 90-300 sec and adheres to the tissue.

Subretinal injections
Eyes were dilated with 2.5% phenylephrine (HUB Pharmaceuticals, Rancho Cucamonga, CA) and 0.5% tropicamide (Akorn Pharmaceuticals, Lake Forest, IL). Mice were anesthetized with ketamine (100 mg/kg; Phoenix Pharmaceuticals) and xylazine (8 mg/ kg; Lloyd Laboratories). Goniovisc (2.5%) (HUB Pharmaceuticals, Rancho Cucamonga, CA) was applied to the eye to maintain corneal clarity and optical interface. Of the 31 mice in the study, 9 mice were assigned to Group A (subretinal BSS) and 22 mice to Group B (subretinal biopolymer). 0.2 µl biopolymer or BSS was injected in one eye for each mouse of the two groups and the non-injected fellow eye was used as control.
Fluorescein to a final concentration of 0.01% was co-injected to facilitate visualization during injection. A multi-purpose rigid telescoping endoscope (11 cm length × 2.7 mm diameter) with a 30° beveled tip (Karl Storz, Tuttlingen, Germany) was used for direct visualization while performing the subretinal injections. The endoscope tip was illuminated via a fiber optic cable connected to a xenon light source (Xenon Nova 175, Karl Storz, Tuttlingen, Germany) and connected to a high definition camera and monitor (IMAGE 1 HUB, Karl Storz, Tuttlingen, Germany). The endoscope tip was apposed to the corneal surface, while the injection was performed with a trans-scleral/choroidal approach using a 33 g beveled needle attached to a 10 µl Nanofil syringe (World Precision Instruments, Sarasota, FL).

Imaging
Fundus imaging: Mice were anesthetized with 5% isoflurane (Abbott Laboratories, Abbott Park, IL) and visualization was performed with the endoscope camera as described above.

Spectral domain optical coherence tomography (SD-OCT):
Ultra high resolution SD-OCT imaging was performed on both eyes from all groups, at baseline 1 week prior and 1 and 2-weeks post-injection using Bioptigen SD-OCT system (Research Triangle Park, NC). A series of 100 b-scans were collected, stacked and aligned spatially to form a registered three-dimensional rendering of retinal volume.

SD-OCT imaging of retinal structure
Bilateral SD-OCT imaging encompassing the detached and neighboring attached retina was performed for all mice 1 and 2 weeks post-injection. Local disruption of the retinal layers was seen at the site of the needle puncture wound for all injected eyes ( Figure 2). In biopolymer injected eyes, a small detachment persisted 1 week postinjection ( Figure 2C) whereas in the BSS injected eyes, the retina was completely re-attached ( Figure 2B). The photoreceptor and RPE layers were disrupted directly under the detachment and hyporeflective material corresponding to the biopolymer and hyperreflective cellular debris could be seen under the retina. There was no evidence of any abnormalities in the lens, vitreous and anterior segment structures in any of the injected groups.

Histology of mouse eyes following subretinal injection
Light microscopy: We examined perfusion-fixed, plastic embedded and Toluidine-blue stained sections of both injected and non-injected eyes at 1 day post-injection. Compared to the non-injected eyes ( Figures  3A and 3D), the BSS-injected eyes showed some disorganization of the photoreceptor outer segments and vacuoles in the RPE layer ( Figures  3B and 3E). In the subretinal biopolymer injected eyes, the biopolymer could be seen as Toluidine blue-stained amorphous material in the subretinal space 1 day post-injection ( Figures 3C and 3F). At this time point, the photoreceptor outer segments overlying the detachment appeared compressed and partially disorganized but the inner segments and rest of the retinal morphology appeared normal in both BSS and biopolymer injected eyes. Occasional rounded macrophagelike cells could be seen infiltrating into the biopolymer layers from the RPE-choroidal layer at this stage ( Figure 3G). Retinal morphology was completely normal immediately adjacent to the detachment in the subretinal biopolymer injected eyes ( Figure 3H).
We examined eyes from H&E stained FFPE sections from all injected groups 3 weeks post-injection. The retina of the BSS injected eyes appeared normal ( Figures 4A and 4C). As expected, the biopolymer could not be detected in the subretinal or vitreous space at this time point (Figures 4B and 4D). A sharp demarcation was seen between the   Electron microscopy: In order to examine in greater detail the effect of biopolymer on RPE and its recovery after reattachment, we used transmission electron microscopy to observe the ultrastructure of the RPE and subretinal space. The retina appears reattached 1 day postinjection of BSS ( Figure 5B). In control non-injected ( Figure 5A) and subretinal BSS-injected eyes ( Figure 5B), the rod outer segments are close to the RPE monolayer with the tips of the outer segments enclosed by the apical processes of the RPE. 1 day after injection of biopolymer in the subretinal space, amorphous "blobs" (p) of biopolymer are seen overlying the RPE ( Figure 5C). The RPE apical microvilli (v) appear shorter and compressed consistent with temporary remodeling due to retinal detachment [32]. When assessed at 3 weeks post-injection of biopolymer, the RPE apical microvilli once again appear normal, the biopolymer has been absorbed and the photoreceptor outer segments (os) are apposed to the RPE ( Figure 5D).

Apoptotic or necrotic cell death determination by TUNEL staining
We assessed ongoing apoptotic cell death at three weeks postinjection by TUNEL staining. Extremely sparse TUNEL-positive nuclei were observed exclusively in the outer nuclear layer (ONL) of both noninjected and BSS-injected eyes ( Figure 6). In the biopolymer injected eyes, apoptotic nuclei were observed in the photoreceptor layer only at the needle puncture site with none to extremely sparse TUNEL labeling in the retina distal to the injection site.

Photoreceptor recovery and inflammatory markers
We examined sections containing the injected quadrant of the eye for markers of photoreceptor recovery and inflammation 1 week post-injection. Proper localization of rhodopsin to the photoreceptor outer segments was observed in the biopolymer injected eyes ( Figures  7A and 7B). Equitable distribution of retinal glial cells (microglia; macroglia) which produce inflammatory cytokine as visualized with the marker Ionized calcium binding adaptor molecule1 (Iba1) was observed in both the control BSS-injected and biopolymer-injected eyes ( Figures 7C and 7D). No evidence of these phagocytic microglia was observed in the outer or inner nuclear layer. Immunoreactivity for pan-macrophage marker F4/80 is seen in subpopulations of phagocytic microglia in several retinal degenerative diseases [33], ischemiainduced retinopathy [34] and laser-induced retinal injury models [35]. However F4/80-reactive macrophages were not observed in the retina or the subretinal space of either BSS-injected or biopolymer-injected eyes ( Figures 7E and 7G). F4/80-positive microglia/macrophages were normally observed in the ciliary body from the same sections ( Figures  7F and 7H), as has been previously reported in wild-type mouse retina [36,37].
To determine gliosis, we used the marker for glial fibrillary acidic    protein (GFAP) which labels the Müller cell "endfeet" and astrocytes in the inner limiting membrane and nerve fiber layer (NFL). GFAP labeling was observed in the NFL corresponding to the distribution of retinal astrocytes and Müller cell "endfeet" in control non-injected and BSS-injected eyes (Figures 7I and 7J). In the biopolymer-injected eyes, minimally increased GFAP reactivity was seen at the site of the needle stick ( Figure 7K), but GFAP labeling remained normal at all sites immediately distal from the site of needle stick ( Figure 7L).

Effect of biopolymer on retinal function
Dark-adapted ERGs were recorded at baseline, and at 1-and 2weeks post-injection. Rod b-wave amplitudes as a function of retinal illuminance were analyzed by finding the parameters of the best fit using the Naka-Rushton function as previously described [38]. We assessed potential retinal toxicity of the biopolymer by comparing the interocular difference in V max , the maximum saturated b-wave response before and after injections, to the interocular difference in V max at baseline for each mouse. Vmax at baseline, expressed as a percentage difference, ∆V max %, was not statistically significant between eyes (∆V max =0.67%, t=0.32, df=17, P=0.75).
For group A (subretinal BSS), there was a slight weakening of function in the injected eye at 1-week post injection, although the change in V max was not statistically significant (df=23, ∆V max =-3.3%, t=0.76, P>0.48). (In this and in subsequent analyses, a "-"∆Vmax indicates that the injected eye performed worse than the non-injected control eye.) Although the number of eyes tested at 2-weeks post injection was small (n=3) there is a suggestion of continued weakening of the response (∆Vmax=-8.70%, no statistical test was performed due to the small number of mice tested).

Discussion
Retinal breaks or holes combined with vitreoretinal tractional forces are the causes of rhegmatogenous retinal detachments [39]. The annual incidence of retinal detachment is 1 in 10,000 or 1 in 300 over the lifetime [1,40,41], often requiring therapeutic surgery to prevent permanent loss of vision. These retinal openings allow vitreous fluid to enter the subretinal space and enlarge the retinal detachment. Over time, loss of contact of the photoreceptor layer with the supporting RPE starts a cascade of processes leading to Müller glia proliferation and ultimately photoreceptor cell death and vision loss [42,43]. Additionally, RPE cells can migrate into the vitreous through the break in the retina where they can release inflammatory cytokines and cause proliferation of epiretinal and subretinal membranes described as PVR [44,45]. PVR is a major complication of retinal detachment surgery, exerting traction on the retina and causing recurrent retinal detachment [46,47]. Currently laser photocoagulation is used during PPV surgery to create chorioretinal adhesion around the retinal break, and silicone oil or gas tamponade agents are used to fill the vitreous cavity and keep the neuroretina attached to the RPE until the retinopexy becomes effective [5][6][7][8][48][49][50]. These tamponade agents cause considerable patient discomfort and complications [51]. Additionally, it takes up to 2 weeks for the laser scar to develop, during which time the unsealed tear may allow the release of RPE, inflammatory cells and serum components from the subretinal space into the vitreous [52,53].
A retinal adhesive is designed to be used during surgery to form a temporary seal of the retinal tear, which eliminates the fluid flow from the vitreous cavity into the subretinal space and also prevent RPE migration into the vitreous while the laser retinopexy takes effect [54]. This would eliminate the need for long-term tamponade agents. A successful retinal patch must stick to the retinal surface, seal the tear and remain in place until the retinopexy is effective. It would also be advantageous it this polymer can be applied to the retinal site through a 23-27 gauge cannula as a liquid and polymerize in vivo. Previously, cyanoacrylate or fibrin-based glues, mussel protein adhesive and other hydrogels have been tested in animal models or clinical cases for the treatment of retinal detachment [14][15][16][17][18][19]. These agents have limitations such as severe retinotoxicity, inflammation, difficulty of delivery or limited efficacy.
Medicus Biosciences has developed a novel PEG-derived adhesive biopolymer having the mechanical properties required to seal the retinal tears and avoiding the complications of current treatments. The reactants can be in vivo polymerized in PBS at physiological pH. Reacting liquid monomers injected at the site of the retinal break would adhere to the retinal surface and seal the lesion when gelled at a predetermined time. Medicus has manipulated the viscoelastic properties of the biopolymer such that it behaves identically to the healthy human retina [55]. The biopolymer can be delivered as a liquid through conventional minimally invasive instrumentation found in ophthalmological operating rooms, allowing sufficient working time for application to multiple, large retinal breaks. The transparency of the adhesive should allow maintenance of visual function of the patient in case the biopolymer needs to be used in the macular area. The biopolymer dissolves in about 2 weeks after application leaving no permanent residue. The kinetics of the polymerization and disaggregation can be easily modified and controlled to be adapted to the needs. This allows for maintaining the neuro-retina attached to RPE and re-establishment of normal photoreceptor outer segment-RPE connections.
Since the adhesive is designed to be delivered to the retinal surface during vitreo-retinal surgery, we decided to examine the safety and toxicity of the biopolymer when in direct contact with the mouse retina. Although the best way to avoid possible positive or negative effects of melanin on oculotoxicity is to test both pigmented and non-pigmented animals [56], we used albino mice in this study as lack of pigment in the RPE and choroid allowed for easier visualization of the needle tip during the trans-scleral injection approach for delivery in the subretinal space.
To test the retinal toxicity of the polymer, we decided to inject the biopolymer in the subretinal space using a trans-scleral approach. We understand that, in the normal life, this polymer will be delivered mainly on the retinal surface through the vitreous cavity, with only a limited amount accessing the subretinal space through the retinal opening. We decided to consider this approach given the very large relative size of the mouse lens and the fact that the vitreous in mice is firmly attached to the retina. In this model, an intravitreal injection of the polymer would have been impossible and would not allow direct contact of the polymer with the retinal surface. We observed slightly weaker ERG responses in all injected eyes including BSS controls, which we attribute to traumatic damage to the small mouse eye from the injection itself. There was no significant reduction in ERG parameters due to the biopolymer and our functional assays were not consistent with a toxic response to the biopolymer.
We show that the biopolymer is safe, well-tolerated and causes no inflammatory response or retinal abnormalities. Clumps of dislodged RPE cells and photoreceptor outer segment debris that can be seen under the detachment are attributed to mechanical damage caused by injection pressure [57]. Proliferation of Müller glial cells recognized by GFAP upregulation is observed in retinal inflammation due to various stress stimuli [32,[58][59][60]. In the biopolymer-injected eyes we observed no signs of glial proliferation outside the immediate site of needle entry wound. Loss of photoreceptors was observed in both experimental as well as BSS control eyes in front of the site of sclerochoroidal perforation, indicating that it is from the physical trauma of the injection and not related to the biopolymer itself. Photoreceptor cell death by apoptosis peaks at 3 days after retinal detachment and gradually diminishes [61]. Three weeks after biopolymer injection in the subretinal space, we did not see an increase in apoptotic or necrotic cell death due to the biopolymer, suggesting that the biopolymer did not produce extensive and global retinal toxicity. Our observation that the retina had completely reattached in the injected quadrant of the eye 2-3 weeks post-injection indicates that the biopolymer is non-persistent and is likely cleared by the natural processes of phagocytosis and fluid absorption of the RPE and circulating retinal macrophages. Since 8-10 rows of photoreceptor nuclei were normally present overlying the detachment created by the polymer, we expect complete structural and functional recovery at later time points.
In summary, our study shows that subretinal administration of the adhesive biopolymer has excellent biocompatibility. We found no evidence of retinal toxicity or inflammation from the biopolymer injections using anatomical, histological and functional evaluations. Further studies are needed to confirm the safety and efficacy of this biopolymer for the treatment of retinal detachment in humans.