PYK-1105: Preclinical Evaluation of a Novel Biodegradable Vitreous Substitute for Retinal Tamponade

Abstract

Purpose:

Current retinal tamponade strategies are limited by anatomic considerations (retinal break location), durability (short-term vs need for removal), and patient adherence (positioning, travel/altitude restrictions). Here we describe the preclinical safety and toxicology of a novel biodegradable hydrogel tamponade agent (PYK-1105) with the potential to improve both patient experience and outcomes after retina surgery.

Methods:

We studied in vitro performance to assess hydrogel gelation time, modulus, viscosity, degradation time, refractive index, and transmittance. In addition to studying in vitro and in vivo (mice and rabbits) biocompatibility, testing was performed to assess cytotoxicity, intraocular irritation, acute systemic toxicity, genotoxicity, and pyrogenicity. Furthermore, clinical safety was assessed using in vivo (rabbits and minipigs) response to vitrectomy with PYK-1105 insertion with the following measures: clinical examination, multimodal imaging, full-field electroretinography, and histopathology.

Results:

PYK-1105 met the predefined performance testing criteria for optimal tamponade and demonstrated excellent biocompatibility. Animal studies showed the PYK-1105 formulation to be well tolerated and nontoxic in mice, rabbits, and pigs.

Conclusions:

PYK-1105 holds promise as a new biodegradable tamponade agent that has the potential to improve both the patient experience and outcomes after retina surgery. Human pilot studies are warranted to further assess for safety and efficacy.

Introduction

Over the last 35 years there have been extraordinary advances in vitreoretinal surgical instrumentation and techniques, including the use of smaller-gauge instruments, faster cut rates, more stable fluidics, and improved visualization. The methods of providing retinal tamponade, despite their significant limitations, have changed little. The primary purpose of the retinal tamponade is to seal retinal breaks. The currently approved agents are hydrophobic and are effective because of their high surface tension. Additionally, gas and silicone oil both exert a buoyant force that helps to reapproximate the retina to the eye wall.

Physicians and patients alike are forced to make trade-offs between the tamponade choices that affect patient experience—face-down positioning, inability to fly (gas), altitude restrictions (gas), temporary poor vision (gas), refractive changes (oil), and need for a second surgery for tamponade removal (oil)—as well as outcomes—limited effectiveness in the presence of inferior retinal breaks, cataract progression, and development of macular edema.

A new method of retinal detachment (RD) repair without the use of gas or oil could lead to improved patient experience and outcomes. Thus, there is an unmet need for a safe, easily injectable, and biodegradable agent that can seal retinal breaks regardless of location and reduce the postoperative burdens imposed by gas and oil.

We present the preclinical assessment of PYK-1105, a novel hydrophilic, in situ cross-linking polymeric formulation for postoperative retinal tamponade. This medical device consists of a low-viscosity liquid that is injected into the vitreous cavity after the fluid-air exchange through standard small-gauge cannulas. In the intraocular environment, PYK-1105 rapidly transitions into a relatively stiff, cross-linked vitreous-like body that apposes retinal breaks. It is optically clear, neutrally refractive, and self-degrading after 2 weeks (Figure 1). Furthermore, the formulation can be manufactured at scale, is stable for long-term storage, and can be easily reconstituted in the operating room setting.

Figure 1.

Schematic representing different stages of retinal detachment (RD) repair using the hydrogel shows (A) RD prior to repair, (B) hydrogel insertion under air after RD repair and laser retinopexy, (C) hydrogel filled to the ora with rapid in situ cross-linking, and (D) the postoperative eye after hydrogel biodegradation and clearance.

In the following section, we present in vitro performance characteristics of the hydrogel, including gelation time, modulus, viscosity, degradation time, refractive index, and transmittance. In addition, in vitro and in vivo biocompatibility testing for toxicology was performed in mice and rabbits in compliance with International Standards Organization (ISO) 10993 (Biological Evaluation of Medicine Devices) guidelines. Furthermore, we present the safety data following vitrectomy and implantation in a series of rabbits (New Zealand White [NZW]) and minipigs (Yucatan) using a combination of clinical examination, full-field electroretinography, multimodal imaging (enhanced-depth imaging optical coherence tomography [OCT] and wide-field fluorescein angiography [FA]), and histopathology.

Methods

Device Description

PYK-1105 is a medical device formulated as a lyophilized product under Good Manufacturing Practice consisting of modified polyvinyl alcohol and modified polyethylene glycol as a cross-linker in a phosphate-buffered saline vehicle. Polyvinyl alcohol and polyethylene glycol are among the most well characterized and safest synthetic polymers known in the literature and are commonly used in cosmetics, as food additives, and in drug formulations, including topical ophthalmic preparations and recently as a vehicle for intravitreal drug delivery. ¹ PYK-1105 is a proprietary formulation of these modified polymers designed to meet the unique physiologic and biophysical requirements required for retinal tamponade.

At the time of use, the 2 components are reconstituted and mixed together. The hydrogel begins to form several minutes after mixing, a process that accelerates rapidly after injection into the air-filled eye, initially transforming into a viscous liquid and then becoming a semisolid gel in approximately 3 to 5 minutes. It is recommended that all necessary surgical maneuvers, such as retinopexy, be performed under air prior to injection of the gel. Given that the mechanism of PYK-1105 is different from traditional hydrophobic tamponade agents, it is important to remove as much posterior subretinal fluid as possible.

The hydrogel biodegrades over a continuous 2-week period into molecular components that are smaller than 50 kDa, permitting transretinal clearance (see Figure 1). The redundancy of bonds within the gel structure allows the tamponade gel to maintain its shape and sealing effect during the residence period, but the gel begins to lose stability and noticeably softens by day 11; by day 14 the gel is completely dissolved. This slow, continual replacement of polymer components by water results in a steady maintenance of intraocular pressure (IOP) during the breakdown process.

Of note, the gel can be removed from the eye if needed. The process is similar to removal of a soft lens that occupies the entire posterior cavity. In our experience, removal is best achieved with aspiration using a soft-tip cannula with the silicone sleeve removed, or with a fragmatome if expeditious removal is desired. The small aperture of a 25-gauge cutter probe is too slow to facilitate rapid hydrogel removal.

In Vitro Performance Characteristics

Hydrogel gelation time, modulus, and viscosity were assessed by loading the newly formed hydrogel onto a TA Instruments ARES-G2 rheometer and analyzed using a 60-mm, 2° stainless steel cone at 0.1% strain with a frequency of 6.283 rad/s for 30 minutes both at 25 °C and 37 °C. Elastic or storage modulus (G′) and viscosity were recorded continually. Hydrogel degradation time was characterized by suspending the solid hydrogel in a cage within 37 °C phosphate-buffered saline followed by daily weight measurements until the hydrogel was gone. Measurements of refractive index and transmittance were obtained using an Abbe refractometer and BioTek PowerWave XS UV-VIS spectrophotometer, respectively.

Biocompatibility Testing

Cytotoxicity

Cytotoxicity testing (ISO 10993-5) was performed by incubating formed hydrogel for 72 hours in standard culture media, exposing L-929 cells to the extract or control extract media, and then examining the cells microscopically for abnormal morphology or degeneration.

Intraocular Irritation

To assess for intraocular irritation, hydrogel samples were extracted in saline and 0.2 mL of extract and saline control were injected intravitreally (right and left eyes, respectively) into 6 NZW rabbits. After 48 hours, the eyes were examined, animals were euthanized, and the vitreous was harvested and examined for white blood cells by hemocytometer (counts < 200 cells/mm³ were acceptable).

Acute Systemic Toxicity

To assess for acute systemic toxicity (ISO 10993-11), the hydrogel was extracted in saline and sesame oil (0.2 g hydrogel: 1 mL saline or sesame oil) and then injected intravenously and intraperitoneally, respectively, into 5 mice alongside controls. The animals were observed for signs of systemic toxicity after injection and at 4, 24, 48, and 72 hours post injection.

Genotoxicity

To assess for genotoxicity (ISO 10993-3), the mouse lymphoma forward gene mutation assay was performed in the standard fashion using hydrogel extractions and the mouse lymphoma L5178Y/TK^+/– cell line. Mean mutant frequency of the test extracts were compared with the appropriate positive and negative controls.

Pyrogenicity

Material-mediated pyrogenicity (United States Pharmacopeia 151 and in accordance with ISO 10993-11) was evaluated by injecting hydrogel extracts (4 hydrogel extracts: 20 mL in nonpyrogenic saline) intravenously into 3 NZW rabbits (10 mL/kg) and monitoring for changes in rectal temperatures over 3 hours.

Local Tissue Response

To assess for tissue response after implantation (ISO 10993-6), the hydrogel was implanted in the muscle tissue at 4 locations in 3 NZW rabbits for each time point (1 and 4 weeks) alongside appropriate controls. Animals were then euthanized accordingly, and the muscle sites were examined macroscopically and microscopically.

Endotoxin

The Food and Drug Administration’s guidance document, “Endotoxin Testing Recommendations for Single-Use Intraocular Ophthalmic Devices,” was followed using the limulus amebocyte lysate method according to United States Pharmacopeia 42.

Testing in the Surgical Model

Four NZW rabbits and 5 Yucatan minipigs underwent general anesthesia followed by a 25-g pars plana vitrectomy performed in the standard surgical fashion using the Accurus vitrectomy machine (Alcon). After removal of the vitreous, a complete fluid-air exchange was performed, and the hydrogel was injected to the level of the ora serrata. A small anterior air bubble was allowed to remain in the Berger’s space. Animals then underwent clinical examinations daily and ophthalmic examinations at postoperative days 1, 7, 14, 21, and 28, which included measurements of IOP. Animals were treated with polymixin and prednisolone acetate drops topically 4 times a day for 1 week to the operative eye. Statistical analysis of IOP was performed using the mixed-effect model repeated measures test (Stata, StataCorp).

Multimodal imaging was performed at postoperative weeks 2 and 4, which included spectral domain (SD)-OCT (Spectralis), widefield FA (Spectralis), and color fundus photography (iNview, Volk). Full-field electroretinograms (ERGs) were performed at postoperative week 4 on all animals in dark-adapted and light-adapted conditions with a 6-step, modified International Society for Clinical Electrophysiology of Vision’s protocol using the RETeval Device (LKC Technologies). ERG reports for functional assessment and statistical analysis consisted of bright flash b-wave amplitudes in dark-adapted animals, bright flash b-wave amplitudes in light-adapted animals, and 30-Hz implicit times and amplitudes in light-adapted animals. Operative eyes were compared with a normative control data set created by compiling nonoperative eye data (the contralateral eyes) from all animals assessed that examination day. Statistical analysis of amplitudes and implicit times was performed using the Wilcoxon-Mann-Whitney test.

After 4 weeks, the eyes were harvested, fixed in Davidson’s solution for 24 hours followed by formalin (4%), embedded in paraffin, sectioned, and then stained with hematoxylin-eosin. The pathological interpretation was performed by a masked veterinary ophthalmic pathologist (StageBio), and representative images were acquired using a light microscope.

Results

PYK-1105 was designed to be minimally viscous during injection into the eye through small-gauge ports, cross-link rapidly in the eye, and biodegrade approximately 2 weeks after insertion. Furthermore, the polymer is optically clear and has a neutral refractive index (Table 1).

Table 1.

PYK-1105 Performance Assessment.

Parameter	Result
Time to injection (duration of minimal viscosity after mixing at 25 °C), min	10
Time to gel formation after injection (duration until viscous gel formation at 37 °C), min	4
Time until degradation, d	11-14
Refractive index	1.3385
Transparency	>90% across visual spectrum

Abbreviation: PYK-1105, a novel biodegradable hydrogel tamponade agent. Rheology curves for the storage modulus (G′) and complex viscosity over time at 25 °C and 37 °C (Supplementary Figure 1) were used to calculate the time to injection and time to gel formation. Degradation time was measured in vitro by weighing the hydrogel mass over time (Supplementary Figure 2). Transparency of the gel was assessed over the visible spectrum (Supplementary Figure 3).

To assess in vivo safety and biocompatibility, a series of tests were conducted in accordance with ISO 10993 and the Food and Drug Administration’s guidance for establishing the safety of an intraocular medical device. The details of each test are described in the Methods section and the results are summarized in Table 2. Cytotoxicity testing performed on mouse cell lines was unremarkable. Intraocular irritation was assessed by injecting hydrogel extracts into rabbit eyes, and no significant white blood cell response was detected using a hemocytometer. In mice, no acute systemic toxicity was detected following systemic administration. The mouse lymphoma assay was negative for mutagenic potential. Implantation studies on the backs of rabbits showed no significant change macroscopically as compared with the negative control article and minimal to no reactions as compared with the negative control article microscopically. Finally, endotoxin levels were below strict levels required for intraocular devices.

Table 2.

Results of Biocompatibility Testing Demonstrating Excellent Biocompatibility in Standard ISO-10993 Assays.

Test	Result
Cytotoxicity	Noncytotoxic
Intravitreal injection	Nonirritant
Acute systemic toxicity	Nontoxic
Material-mediated pyrogenicity	Nonpyrogenic
Endotoxin	<0.2 EU/device
Implantation	No macroscopic reaction; minimal to no microscopic reaction
Genotoxicity	Nonmutagenic

Abbreviations: EU, Endotoxin units; ISO, International Organization for Standardization.

Because PYK-1105 demonstrated ideal performance characteristics and biocompatibility, the safety of the polymer was further assessed with a series of preclinical vitreoretinal animal studies in both rabbits and pigs. The animals underwent vitrectomy, fluid-air exchange, injection of the hydrogel, and close clinical follow-up as described earlier. In rabbits and pigs, no significant ophthalmic clinical findings were observed (Supplementary Figure 4). Mild surface irritation was noted in all animals on postoperative days 1 and 7, consistent with typical postoperative changes. IOP differences were not significantly different between surgical eyes and the contralateral nonoperative control eyes at any time point in rabbits (Figure 2A) and pigs (Figure 2B). One pig was eliminated from the analysis because there was significant intraocular bleeding observed on the postoperative day 1 examination with development of vitreous haze. From both our prior experience and the experience of others in the literature, ^{2 -4} vitreous hemorrhage that occurs postoperatively in the porcine eye is a potent nidus for a fibrinoid, proliferative vitreoretinopathy (PVR)–like reaction and confounds safety assessment.

Figure 2.

Intraocular pressure (IOP) measurements after surgery. (A) Rabbits and (B) pigs both had no significant difference in postoperative IOP between control eyes (nonoperative alone) and hydrogel eyes (vitrectomy with hydrogel injection). Post-op indicates postoperative.

Experimental animals were further assessed using multimodal imaging 4 weeks after surgery, or approximately 2 weeks following hydrogel degradation. In rabbits, the fundus photography (Figure 3A) and retinal SD-OCT (Figure 3B) did not demonstrate any detectable abnormality. Similarly, in pigs, findings from fundus photography (Figure 3C) and retinal SD-OCT (Figure 3D) were normal. Because pigs have a more human-like vascularized retina, FA was performed to assess for any signs of subtle vasculitis, retinitis, or other changes. At early, middle, and late frames there were no detectable abnormalities on angiography (Figure 3, E-G). Anterior-segment OCTs were also obtained and demonstrated a normal anterior chamber configuration in all animals (Supplementary Figure 5).

Figure 3.

Multimodal imaging in animals 1 month after hydrogel insertion. Representative observations from rabbit fundus (A) photograph and retinal (B) optical coherence tomography were normal. Representative findings from pig fundus (C) photograph and (D) retinal optical coherence tomography were also normal. Pig fluorescein angiography (E) early, (F) middle, and (G) late frames did not demonstrate any vascular leakage, retinal ischemia, retinitis or other detectable pathology.

To assess retinal function, full-field ERGs were performed at postoperative week 4 in rabbits and pigs as described earlier. Rod function, as measured by the b-wave amplitude in response to a dark-adapted bright flash, was unchanged in the operative eye compared with the contralateral eye normative data set in both rabbits (Figure 4A) and pigs (Figure 4B). Additionally, cone responses were assessed by comparing light-adapted bright-flash amplitudes and 30-Hz flicker amplitude and implicit times, which were also found to be normal (Figure 4, C and D).

Figure 4.

Electrophysiology in rabbits’ and pigs’ retinas tested 1 month after hydrogel implantation. Representative tracings for (A) rabbit and (B) pigs under specified testing conditions. The physiologic parameters between nonoperative control eyes and experimental eyes were averaged and compared in (C) rabbits and (D) pigs. No significant differences were detected.

After 4 weeks, the operative eyes were harvested, and a histopathologic analysis was performed by a masked veterinary ophthalmic pathologist. Typical postsurgical changes were noted in the operative eyes, and there were no significant histopathologic abnormalities observed (Figure 5).

Figure 5.

Histopathology of animals after hydrogel implantation. Representative photomicrographs of normal retinal histopathology 1 month after hydrogel implantation in (A and B) rabbits and (C and D) pigs.

Conclusions

PYK-1105 demonstrates local and systemic biocompatibility without toxicity or evidence of anatomic or functional changes in 3 animal models and standardized in vitro testing models. This lack of any safety signal highlights that PYK-1105 is a promising agent for human studies of retinal tamponade, for which its potential advantages of biodegradability, lack of patient restrictions (travel, altitude, and positioning), and complete retina tamponade (inferior, superior, nasal, temporal, and posterior) may be evaluated for efficacy with respect to anatomic and functional outcomes.

Numerous materials have been investigated to replace gas and oil including inflatable balloon-like vitreous replacements, noncross-linked vitreous substitutes, and cross-linked vitreous substitutes. Beginning in the 1980s, uncross-linked polymer solutions, including viscoelastic substances, were tried as retinal tamponade agents, but these polymers were either ineffective or associated with in vivo complications such as opacification and inflammation. ^{5 -7} Cross-linked polymers were found to be more effective tamponade agents but required large sclerotomies to facilitate intraocular insertion because the material would shear when injected through the small-gauge cannulas.

One way of achieving the functionality of a cross-linked polymer but allowing injection through small-gauge cannula-based systems is to develop a polymer that cross-links in situ, either by using an external activator or by self–cross-linking polymeric systems. Several groups have developed such systems; however, many of these systems are permanent hydrogel implants that do not degrade. It is unclear what hazards may occur with long-term residence of hydrogel materials, as methyl acrylate and 2-hydroxyethyl acrylate hydrogel (MIRAgel, MIRA Inc) scleral buckles have previously demonstrated. ⁸

One in situ cross-linking polymer that has shown promise is ABV-1701 (Vitargus), a product under development by American BriVision. In a recent pilot study, ABV-1701 was injected into 11 eyes of patients undergoing vitrectomy for RD or nonclearing vitreous hemorrhage. ⁹ They reported severely elevated IOP in 3 of 11 patients, with 1 patient requiring glaucoma surgery. Importantly, ABV-1701 is a large molecule (a functionalized hyaluronic acid of 320 kDa), ¹⁰ which may lead to elevated IOP when the molecule is cleared from the eye. The degradation components are comparable in molecular size and concentration to many hyaluronic acid viscoelastic devices, which are known to cause elevated IOP when left in the eye after surgery. ¹¹ Limiting the size of the breakdown products is a critical consideration when designing a biodegradable hydrogel product. The PYK-1105 breakdown species are primarily 30 to 50 kDa with a maximal species size of 100 kDa. No significant elevation in IOP was detected in either rabbits or pigs at any time point.

One important limitation of performing preclinical experiments for retinal tamponade is the inability to evaluate efficacy. When RDs are created in animal models, the majority spontaneously reattach with or without the presence of a tamponade. ¹² In our own experience we have observed similar results, with most control animals having spontaneous retinal reattachment without tamponade (balanced salt solution tamponade), and a small subset developing severe PVR and remaining detached. The presence or absence of a tamponade agent did not seem to affect the outcome (data not shown). Therefore, the scope of our animal surgical studies has been to demonstrate safety of the biomaterial and to use in vitro model systems to demonstrate physical characteristics consistent with the ability to tamponade a retinal break. Early human experience with other hydrogel formulations, such as ABV-1701, further supports the proof of concept of using hydrophilic in situ cross-linking polymeric systems for retinal tamponade.

Finally, a general limitation we observed in preclinical models of ocular surgery studies is the propensity of animal eyes to develop inflammation and a fibrinous proliferative reaction that can mimic PVR. In the literature, in the pig model following uncomplicated vitrectomy, rates of PVR close to 30% have been demonstrated, with a higher risk when performing a retinotomy, such as in attempts to create an RD. ⁴ Other groups have reported comparable results (38%) of fibrinous PVR-like exudation in minipigs, specifically the Göttingen breed, stating “the postoperative fibrin exudations and formation of synechiae reflect to a certain extent a species-specific predisposition to develop inflammatory reaction following any intraocular intervention.” ² Although these exudative responses can clear spontaneously, when severe they can lead to the formation of fibrinous membranes that can result in tractional RDs. Pigs seem to be particularly prone to intraoperative or postoperative bleeding, which significantly increases the risk of this type of PVR reaction occurring after retina surgery. Consistently, in our experience, a subset of pigs have this surgical bleeding complication, which seems to occur in both control and experimental animals at the same rate, and it occurs at different rates depending on the breed of pig.

Notably, eyes undergoing this response do not have scleral inflammation, nor do they have anterior cell or flare, suggesting that this fibrinous reaction is a response to surgical injury and is not a true uveitis. As noted previously, 1 pig in this cohort of animals developed this type of reaction after having postoperative bleeding noted at the sclerotomy sites on postoperative day 1. On histopathology, fibrinous exudate was appreciated but there was a notable lack of neutrophils, lymphocytes, or evidence of an infectious process. As discussed earlier, that animal was not included in the analysis. Similar fibrin formation can be seen in rabbits, especially juveniles, but was not observed in our experience, possibly related to our use of older, larger rabbits. ³

PYK-1105, a Good Manufacturing Practice–manufactured biodegradable hydrogel, is a novel retinal tamponade and vitreous substitute that may ameliorate many of the disadvantages of current methods of tamponade. In this series of toxicologic and preclinical safety evaluations, there is no apparent evidence of clinical, electrophysiologic, angiographic, or histopathologic toxicity to the animal retina. Additionally, the biophysical properties of the material strongly suggest it would serve as a beneficial tamponade for the treatment of RDs. Pilot human studies are warranted to further assess the safety and efficacy of this promising new therapy.