Method for Quantitative Analysis of Vitreoretinal Adhesion in Ex Vivo Model

Abstract

Purpose: Posterior vitreous detachment (PVD) is implicated in numerous retinal pathologies. A necessary step in developing new therapies, an area of significant interest, is a quantifiable assessment of posterior vitreous adhesion (PVA) that is also clinically relevant.

Methods: A 23-gauge vitrector was used at varying levels of vacuum to attempt PVD induction in a porcine eye model injected with either balanced salt solution (BSS) (control) or plasmin (2, 3, or 5 U), which can pharmacologically induce PVD.

Results: The average minimum vacuum necessary to induce a PVD was 395 ± 28 mm Hg for BSS alone, 385 ± 58 mm Hg for 2 U of plasmin, 265 ± 53 mm Hg for 3 U of plasmin, and 145 ± 28 mm Hg for 5 U of plasmin. We demonstrated a dose-dependent response curve with increasing amounts of plasmin, leading to a statistically significantly lower minimum vacuum necessary to induce a PVD except between BSS and 2 U plasmin.

Conclusions: A dose-dependent relationship between plasmin concentration and PVD was demonstrated. We believe that this model offers significant benefits over prior work as it minimizes confounding manipulations and offers a quantitative assessment that is translatable to in vivo surgical models.

Translational Relevance: This is the first methodology to quantitatively assess the degree of vitreous adhesion in situ.

Introduction

The vitreous adheres to the internal limiting membrane of the retina with points of strong attachment at the vitreous base, peripheral blood vessels, and optic disc margins.¹ However, as a person ages, the vitreous liquifies and collapses due to the degeneration of the collagen meshwork. This process also weakens the adhesions between the vitreous and retina, ultimately leading to posterior vitreous detachment (PVD).^2,3 Aberrant adhesion and failure of complete PVD can lead to numerous retinal pathologies, including macular hole and vitreomacular traction.^4–6 There has been significant effort to develop nonsurgical interventions to modulate vitreous adhesion, including the clinical use of ocriplasmin (Jetrea; Thrombogenics, New York, NY, USA). Unfortunately, unexplained vision loss and electroretinography abnormalities in patients led to its discontinuation.^7–9 Nonetheless, given the tremendous therapeutic potential of PVD modulation, it remains an area of interest and active research. 

Several models have been developed to study PVD, most prominently, pharmacologic induction in postmortem porcine eyes. The earliest demonstration of plasmin for ocular in vitro use was by Sebag et al.¹⁰ in porcine eyes, which demonstrated that plasmin can be used to degrade links between components of the vitreoretinal interface, pharmacologically inducing a PVD. 

There is now a large body of literature focusing on pharmacologically inducing PVDs with plasmin and the effects this reagent has on the surrounding structures in the eye, including its dose- and time-dependent fashion. Currently, the primary method for evaluating PVD in animal models and postmortem human eyes relies on fixation and hemisectioning the eyes.^11,12 PVD status is assessed either grossly during dissection or with microscopy (light field or scanning electron).¹³ As these qualitative methods rely on dissection, it is not possible to isolate the effects of the plasmin from that of the mechanical manipulation or to quantify the degree of vitreous detachment. 

We developed a method that allows for testing of the strength of posterior vitreous attachment in a quantifiable way. Our method mimics surgical techniques to induce vitreous detachment in clinical practice, which could make it valuable in translation to clinical research. 

Methods

We assessed vitreous adhesion to the retina using varying levels of vacuum with a 23-gauge vitrector (Constellation Vision System; Alcon, Fort Worth, TX, USA) to engage the vitreous over the optic nerve and attempt to mechanically create a PVD. This was performed in an ex vivo porcine eye model. Control eyes with balanced salt solution (BSS) injection as well as those injected with increasing units of plasmin were assessed to determine at which level of vacuum a PVD could be induced. Ten eyes were used for each concentration (units) of plasmin injected. 

Preparation and Injection of Plasmin

Plasmin (human; Sigma-Aldrich, St. Louis, MO, USA) was obtained in a presuspended solution and stored in a –80°C freezer until used. The necessary plasmin was then thawed to room temperature per experiment. BSS (control) or 2 U, 3 U, or 5 U of room temperature plasmin was injected into porcine eyes 0.25 mm from the edge of the optic nerve posteriorly to focus the effects of the reagent in the posterior pole of the eye. Eyes were then placed in 37°C water bath for 1 hour to maximize plasmin activity.¹⁴ 

Inducing Posterior Vitreous Detachment

The 23-gauge cannulas were placed 4 mm away from the corneal limbus. BSS infusion was attached inferotemporally, and infusion pressure was set to 30 mm Hg. Widefield viewing was used to visualize the vitreous and retina. Core and retrolental vitrectomy were performed to allow for movement in the mid-vitreous. A mixture of dilute triamcinolone and blue dye was instilled in the posterior vitreous around the optic nerve to enhance visualization of the posterior vitreous. The vitrector was used to aspirate the peripapillary vitreous (cutter was turned off) with increasing levels of vacuum (starting at 50 mm Hg and increasing in 50-mm Hg increments) to test the tensile strength of the vitreous attachment. Once PVD was induced, it was confirmed by repeat staining of the vitreous. If a PVD was induced with minimum vacuum (50 mm Hg), the eye was deemed to have a preexisting PVD. 

Results

A total of 66 porcine eyes were used in this experiment. However, 3 eyes had preexisting retinal detachment, 17 eyes had preexisting globe rupture, and 6 eyes had preexisting PVD (3 in eyes with BSS control and 3 in eyes with 3 U of plasmin). Consequently, 40 eyes were included in the final data set. 

Increased Plasmin Units Injected Was Associated With Lower Vacuum Needed to Induce a PVD

Minimum vacuum needed to perform PVD near the optic disc was averaged and 95% confidence intervals (CIs) and P values were calculated per unit of plasmin (Fig.). Statistical significance was set as P values less than or equal to 0.05. Data demonstrated that increased plasmin led to a decrease in the minimum vacuum necessary to induce a PVD. There was no statistically significant difference between 0 U and 2 U, with the average minimum vacuum needed to perform PVD being 395 ± 28 mm Hg and 385 ± 58 mm Hg, respectively. There was a statistically significant difference (P ≤ 0.05) between the minimum vacuum necessary to induce a PVD with 2 U, 3 U, and 5 U of plasmin, with averages of 385 ± 58 mm Hg, 265 ± 53 mm Hg, and 145 ± 28 mm Hg, respectively. These data are consistent with prior work showing increasing concentrations of plasmin increase posterior vitreous detachment, supporting the external validity of our methodology.¹³ 

Figure.

 

Effect of plasmin concentration injected on the average minimum vacuum needed to induce a PVD.

Discussion

Prior experimental methods for evaluating vitreous adhesion ex vivo relied upon dissection, which necessarily affected the vitreous and its relationship to ocular structures. This is a known limitation, and others have used the number of attempts required to achieve successful intraoperative separation of posterior hyaloid as more representative of the vitreous attachment strength in situ.¹⁵ However, this method is subjective and not easily translatable between different investigators preforming the assessment. We developed a method to assess vitreous adhesion, more specifically the vacuum required to induce PVD in situ, using quantitative, clinically relevant techniques. 

Others have established that varying concentrations of plasmin affect the development of PVD.⁷ Our data demonstrate an expected dose-dependent response curve of intravitreal plasmin on the average minimum vacuum needed to induce a PVD, strengthening the validity of our model. 

The data demonstrate statistically significant findings utilizing a novel methodology, but the techniques used within the experiment have their limitations. Porcine eyes had varying quality, which was minimized by throwing out eyes with preexisting globe ruptures, retinal detachments, and existing PVDs. We could only assess prior PVDs after prepping the eyes for experimentation to minimize disruption of the vitreous. Based on our preliminary and experimental data, eyes with the high-dose plasmin (5 U) required >100 mm Hg of vacuum to induce a PVD, so we assumed any eyes where a PVD was induced with 50 mm Hg of vacuum (the lowest setting) was most likely a preexisting PVD, and therefore, the eye was not used for data collection. Another limitation is that the 50-mm Hg increments were used to test the minimum vacuum necessary to induce a PVD to demonstrate the sensitivity and efficacy of the proposed method. However, using 50 mm Hg may reduce the resolution of the measurements as the minimum vacuum may be between the intervals tested. 

We had a total of 26 eyes excluded due to globe rupture, retinal detachment, or preexisting PVD. None of the eyes with 5 U of plasmin had “preexisting PVD,” making it less likely that we were excluding eyes that simply had a complete PVD induced by the plasmin. 

Given the tremendous impact vitreous status has on medical and surgical retinal disease, it is important to develop models to assess the impact of various interventions on it. We believe the current model offers significant benefits over prior work as it minimizes confounding manipulation and offers a quantitative assessment of posterior vitreous adhesion. 

Acknowledgments

Supported by the UCSF Vision Core shared resource of the NIH/NEI P30 EY002162. 

Disclosure: T. Suresh, None; L. Ong, None; C.B. Marotta, None; W. Keller, None; D. Schwartz, None; F. Brodie, None