Preservation of Murine Whole Eyes With Supplemented UW Cold Storage Solution: Anatomical Considerations

Abstract

Purpose: Retinal ganglion cell (RGC) apoptosis and axon regeneration are the principal obstacles challenging the development of successful whole eye transplantation (WET). The purpose of this study was to create a neuroprotective cocktail that targets early events in the RGC intrinsic apoptotic program to stabilize RGCs in a potential donor eye.

Methods: University of Wisconsin (UW) solution was augmented with supplements known to protect RGCs. Supplements targeted tyrosine kinase signaling, histone deacetylase activity, K⁺ ion efflux, macroglial stasis, and provided energy support. Modified UW (mUW) solutions with individual supplements were injected into the vitreous of enucleated mouse eyes, which were then stored in cold UW solution for 24 hours. Histopathology, immunostaining of individual retinal cell types, and analysis of cell-specific messenger RNAs (mRNAs) were used to identify supplements that were combined to create optimal mUW solution.

Results: UW and mUW solutions reduced ocular edema and focal ischemia in globes stored in cold storage. Two major issues were noted after cold storage, including retinal detachment and reduction in glial fibrillary acidic protein staining in astrocytes. A combination of supplements resolved both these issues and performed better than the individual supplements alone. Cold storage resulted in a reduction in cell-specific mRNAs, even though it preserved the corresponding protein products.

Conclusions: Eyes treated with optimal mUW solution exhibited preservation of retinal and cellular architecture, but did display a decrease in mRNA levels, suggesting that cold storage induced cellular stasis.

Translational Relevance: Application of optimal mUW solution lowers an important barrier to the development of a successful whole eye transplantation procedure.

Introduction

Estimates conducted in 2015 indicated that approximately 36 million people globally were blind and nearly 252.6 million people had some form of visual impairment. These numbers are projected to increase to 115 million blind and 588 million with visual impairment.¹ The main disease causes of irreversible blindness, include glaucoma, age-related macular degeneration, and diabetic retinopathy. Other causes, such as trauma, also account for a high degree of visual impairment, with an estimated 1.7 million people becoming partially blind in the United States alone.² Injuries to the eye are also a common result of battlefield trauma, affecting between 15.8% and 24.6% of medical evacuations.³ The most common injuries are foreign object penetration and indirect damage to the optic nerve due to concussive force.³ Of these, 15% of cases result in enucleation or permanent blindness.⁴ 

Although many of the leading causes of blindness are manageable with treatment, there remains a major socioeconomic impact with irreversible blindness worldwide. This has led to the speculation that whole eye transplantation (WET) could be an option to restore vision.^5–7 Success of WET requires overcoming several major impediments, including adequate perfusion of the transplanted tissue, immunogenic rejection, the refractory nature of the retina including retinal ganglion cells (RGCs) and the regenerative capacity of RGC axons in the optic nerve. Some advances have addressed these impediments, such as the development of a partial facial exenteration that includes the globe and can be reconnected as a complete vascular bed,^8,9 combined with immunosuppressants. The critical issues of RGC survival and regeneration remain significant hurdles. 

RGCs are long projection neurons of the central nervous system. In mammals, damage to the RGC axons in the optic nerve initiates the intrinsic apoptotic program.¹⁰ Early molecular changes include signaling of the dual leucine zipper kinase (DLK)–Jun-N-terminal kinase (JNK) axis,^11,12 activation of histone deacetylases (HDACs), which contribute to the downregulation of RGC-specific gene expression,^13,14 and efflux of potassium ions which lead to cell atrophy.¹⁵ Additionally, optic nerve damage leads to changes in other retinal cell types, particularly microglia and macroglia, which may lose their homeostatic support functions and contribute to retinal pathology.^16,17 Studies have shown varying levels of therapeutic benefit by targeting each of these events individually in animal models of acute optic nerve damage. 

Our strategy was to create a neuroprotective cocktail that stabilizes the retina of donor eyes, particularly the RGC population, and preserves it so that a secondary treatment could be applied to induce RGCs to begin regenerating their axons after transplantation. We chose to augment University of Wisconsin (UW) solution, which is often used to maintain the functional capacity of multiple organs and cells before transplantation.^18–20 For a full list of UW Solution components, see Table 1. Although UW solution has been used successfully to preserve peripheral nerves in allografts and in hand and limb transplantation,^21–24 it has not been tested on tissues of the central nervous system. The study described here shows the efficacy of UW solution alone or with supplements designed to neuroprotect RGCs, in the architectural preservation of retinas in enucleated eyes as a prelude for use in WET procedures. 

Table 1.

 

Composition of UW Solution

 

Materials and Methods

Animals

Mice were handled in accordance with the ARVO statement for the use of animals for research. Experimental protocols were approved by the Institutional Animal Care and Use Committee of the University of Wisconsin and the Animal Care and Use Review office of the Department of Defense. CB6/F1 mice, between the ages of 2 to 4 months, were used for experiments. For studies examining the activation response of nuclear factor κB (NFκB), we used the cis-NFκB^EGFP reporter line on a C57BL/6 background.²⁵ 

Intravitreal Injection of Modified UW (mUW) Solution and Storage of Enucleated Whole Eyes

Mice were euthanized by pentobarbital overdose and the head bathed in Betadine (Purdue Products, LLP, Stamford, CT). Eyes were enucleated and the surrounding tissues were cut away. For intravitreal injections, a Nanofil Syringe with a 35G Nanofil needle (World Precision Instruments, Sarasota, FL) was used to inject 6 µL of either UW solution (Bridge to Life Ltd., Northbrook IL), phosphate-buffered saline (PBS, 50 mM phosphate, 150 mM NaCl, pH 7.4) or mUW solution. Eyes were submerged fully in UW solution and stored at 4°C for 6 to 24 hours. Freshly enucleated eyes were used for baseline control samples. These were frozen or fixed for further analyses immediately following enucleation. Each series of experiments (histopathology, immunostaining, or messenger RNA [mRNA] analysis) used six eyes (randomly isolated from 3–4 mice) per treatment group. Each treatment group contained the same proportion of male and female mice. 

Supplements Tested in mUW Solution

Table 2 shows the supplements tested, including the source, rationale, and supporting literature. Supplements included kinase inhibitors sunitinib and SB203580, HDAC inhibitors trichostatin A (TSA) and valproic acid (VPA), K⁺ channel blockers tetraethylammonium (TEA) and barium chloride (BaCl₂), hydrocortisone, and L-lactate. Table 3 shows the stock and final concentrations of each supplement. Final concentrations were based on effective doses published for in vivo experiments. Osmolality of mUW solutions was measured in a VAPRO vapor pressure osmometer 5520 (Wescor, Logan, UT) according to the manufacturer. The pH of mUW solutions was measured using an UltraBASIC UB-10 pH meter (Denver Instrument Co., Arvada, CO). Because both the osmolality and pH of final test solutions fell into acceptable physiological parameters, neither of these variables were altered after the addition of the stock test reagent into UW solution. 

Table 2.

 

Description of Supplements Added to UW Solution With a Rationale for Using Them to Augment Protection of RGCs After Globe Enucleation and Cold Storage

 

Table 3.

 

Preparation of mUW Test Solutions

 

Quantitative Reverse Transcription Polymerase Chain Reaction (qPCR)

After cold storage, whole eyes were frozen in dry ice and stored at −80°C. RNA was isolated using the RNA/DNA/Protein Extraction Kit with RNase-Free DNase-1 Set (IBI Scientific, Dubuque, IA). RNA from each treatment group was pooled, and RNA concentration and quality was determined using a Thermo Nanodrop One-C spectrophotometer (Thermo Fisher Scientific, Waltham, MA). RNA was aliquoted and stored at −80°C. 

For qPCR, 2 µg of total RNA from each treatment group was combined with MMLV- reverse transcriptase, 10 mM dNTPs, and Oligo(dT) (Promega, Madison, WI) to synthesize cDNA. The cDNA was then diluted 10-fold, and 55 µL of cDNA from each group was mixed 1:1 with TaqMan Gene Expression Master Mix (Thermo Fisher Scientific) before being loaded into a TaqMan Custom qPCR Array Card (Thermo Fisher Scientific). Each sample was run in triplicate. Target genes on this card are listed in Table 4. Quantitative PCR analyses were then performed using the QuantStudio 7 Flex system (Thermo Fisher Scientific) and a qPCR procedure of 40 cycles of 15 seconds at 95°C and 60 seconds at 60°C. Changes in mRNA abundance of each gene of interest between the samples and baseline were quantified using the ΔΔCt method,^26,27 using the 18S rRNA abundance from each sample to normalize input. 

Table 4.

 

Gene mRNA Targets Interrogated Using qPCR TaqMan Array Cards

 

Histopathology and Immunofluorescence

Whole eye samples were fixed in 2% glutaraldehyde and 4% paraformaldehyde in PBS at 4°C overnight and then embedded in paraffin. Samples were then sectioned and stained with hematoxylin and eosin. The hematoxylin and eosin–stained samples were imaged with a Zeiss Axioimager Z2 upright microscope (Carl Zeiss Meditec, Oberkochen, Germany) and processed with Zen Blue image analysis software (v2.3, Carl Zeiss). Digital images were then evaluated using a four-point scoring system evaluating the integrity of multiple regions of the eye. Images were evaluated for: 

• (i) Protein leakage in vitreous;
• (ii) Protein leakage in the anterior chamber;
• (iii) Lens structural quality and integrity;
• (iv) Retinal edema; and
• (v) Retinal detachment.

Evaluation of (i–iv) was based on a scoring system of 1 to 4 with 1 being equivalent to baseline samples and 4 demonstrating the greatest level of pathology. Retinal detachment was scored as either 0, which represented a retina still in contact with the retinal pigmented epithelium (RPE), or 1, indicating detachment. Scoring was conducted by two masked observers who viewed two digital images each from a minimum of three eyes per treatment group. Differences in scores were discussed to yield a consensus score. Scores for each condition were added to yield a single score. 

Corneal thickness was measured with the ImageJ software (v1.53, National Institutes of Health, Bethesda, MD) measurement tool. Five separate measurements were acquired from one image that approximately bisected the entire globe each of three different eyes of the same treatment group and averaged. The location of the measurements were directly in the center, approximately 150 µm from the limbus on each side, and between the center and peripheral points of measurement. An unpaired t test was performed to determine significance between two groups and a one-way analysis of variance was performed to compare all the groups. A comparison of the variances between each test group and the baseline group (F test) showed no significant differences in variances, although a significant difference was found when comparing all variances using a Brown–Forsythe test (P = 0.01). 

For immunofluorescence, whole eye samples were fixed in 4% paraformaldehyde in PBS for 1 hour at room temperature, and then infiltrated with 30% sucrose in PBS overnight at 4°C, before embedding in Optimal Cutting Temperature media (Scigen, Paramount, CA) and frozen on dry ice. Eight-micrometer sections were taken through equator of the globe to visualize all ocular components. Sections were blocked with 2% goat serum in 0.3% Triton X-100/PBS overnight at 4°C in a humidified chamber. Slides were then treated with a primary antibody in blocking solution at a concentration of 1:1000. Primary antibodies were anti–RNA-binding protein with multiple splicing (cat# PA5-31231, RBPMS, Invitrogen Corporation, Rockford, IL), anti-glial fibrillary acidic protein (cat# PA1-10019, GFAP, Invitrogen Corporation), anti-rhodopsin (cat# MAB5356, RHO, Millipore Sigma, St. Louis, MO), or anti-protein kinase Cα (cat# sc-208, PKCα, Santa Cruz Biotechnology Inc., Dallas, TX). Slides were incubated either overnight (anti-RBPMS, anti-RHO, and anti-PKCα) or for 3 days (anti-GFAP) at 4°C in a humidified chamber. Slides were washed three times in PBS before applying the appropriate Alexa488 (Jackson Immuno-Research, West Grove, PA) conjugated goat anti-rabbit (anti-RBPMS, anti-GFAP) or goat anti-mouse (anti-PKCα, anti-RHO) IgG. Slides were then incubated overnight at 4°C in a humidified chamber. After washing in PBS, sections were mounted with Vectashield containing 4′6-diamidino-2-phenylindole (DAPI) for imaging with a Zeiss Axioimager Z2 upright microscope. 

DNA Fragmentation Assay Using Terminal Deoxynucleotidyl Transferase dUTP Nick-End Labeling (TUNEL)

TUNEL staining, to identify dying cells, was performed using the DeadEnd Fluorometric TUNEL System (Promega). Fixed eyes were embedded, and frozen sections were obtained as described above. As a control, some sections were treated with 10 units/ml RQ1 DNase I (Promega) for 20 minutes at room temperature. Slides were then stained according to the manufacturer's protocol. After washing, slides were mounted with Vectashield and imaged as described elsewhere in this article. 

Results

Initial Studies With UW Cold Storage Solution

Our initial experiments compared the effects of cold storage of mouse eyes emersed in either PBS or unmodified UW solution. Straight emersion of globes into either solution resulted in varying levels of retinal edema, within 6 hours for PBS and 24 hours for UW solution. This effect was mitigated for both solutions if they were injected directly into the vitreous cavity immediately after enucleation (Supplemental Fig. S1), although edematous regions of the retina, optic nerve, and cornea persisted in eyes injected with PBS and incubated for 24 hours (Supplemental Figs. S2, S3, and S4). TUNEL staining of sections revealed no evidence of nuclear DNA degradation in either PBS- or UW solution-treated retinas at any time point up to 24 hours (Supplemental Fig. S5). Finally, we used eyes from cisNFκB^eGFP transgenic reporter mice to assess regions exhibiting cellular ischemic responses.²⁸ As a control, we injected eyes with 100 ng tumor necrosis factor α 24 hours before harvesting. Only eyes treated with PBS exhibited any evidence of NFκB activation, and this activity was restricted to the sclera of the limbus near Schlemm's canal (Supplemental Fig. S6). 

Characteristics of mUW Solutions

After the addition of all test reagents, the osmolality and pH of the mUW solution was measured. Table 5 shows that all test mUW solutions moderately increased the osmolality of the starting solution with the most dramatic effect observed for 15 µM TSA. All solutions retained a pH of 7.4. The addition of BaCl₂ to UW solution caused a marked increase in turbidity, possibly as a consequence of the chelating characteristic of lactobionate, which is known to chelate both iron and calcium ions.²⁰ Nevertheless, this solution, and the combination of VPA and BaCl₂, was also tested in our experiments. 

Table 5.

 

Osmolality of mUW Test Solutions at 20°C

 

Table 6.

 

Pathology Scores of Hematoxylin and Eosin–Stained Sections of Globes Incubated in Cold Storage for 24 Hours

 

Histopathology of Eyes After Cold Storage

Hematoxylin and eosin staining was performed for histopathological analyses of all treatment groups. Longitudinal sections of whole eyes (Fig. 1) were used to evaluate eyes from each treatment group in the categories as described in the Materials and Methods. The top five performing solutions were hydrocortisone, VPA, TEA, L-lactate, and BaCl₂ (Table 6). All five outperformed UW solution alone. Eyes injected with hydrocortisone, TEA, or L-lactate did not exhibit protein leakage into the vitreous chamber. Hydrocortisone, TEA, VPA, and BaCl₂ showed mild leakage into the anterior chamber, and L-lactate had moderate protein leakage into this chamber. Retinal edema was moderate in eyes injected with TEA and mild in eyes injected with VPA, L-lactate, and hydrocortisone. Eyes treated with BaCl₂ had levels of retinal edema similar to that of the baseline control. The remaining groups—VPA/BaCl₂, sunitinib, TSA, and SB203580—performed worse than basic UW solution and exhibited severe pathology in at least one of the five areas analyzed. These groups were not considered as candidates for the optimal mUW solution. 

Figure 1.

 

Images of whole eyes stored in mUW solutions for 24 hours at 4°C, stained with hematoxylin and eosin. These images are representative of the images that were used for evaluation in the scoring system listed in Table 6 (see Materials and Methods for a complete description of the scoring criteria). The optimal mUW solution had the best overall performance based on the subjective scoring protocol, and is the only solution, aside from the baseline, which did not demonstrate retinal detachment after cold storage. Scale bar, 1 mm.

The reagents comprising the optimal mUW solution (2 mM VPA/5 mM TEA/10 mM L-lactate/0.36 µg/mL hydrocortisone) were chosen based on their ranking relative to the baseline. TEA, a potent K⁺ channel inhibitor, was substituted for BaCl₂ to eliminate the increase in turbidity associated with BaCl₂. The combination of reagents in the optimal solution yielded histopathology scores that were similar to the baseline in all areas, with the exception of mild protein leakage into the anterior chamber. Remarkably, the retinas of all tested eyes treated with the optimal mUW solution remained attached to the RPE layer. This characteristic seemed to be a function of L-lactate, because eyes treated with optimal mUW solution lacking L-lactate again exhibited retinal detachment (Supplemental Fig. S7). Overall, the combination of test reagents in the optimal solution performed better than any of the individual components alone. 

Histopathological examination also revealed a marked increase in cornea thickness (Fig. 2) (P < 0.0001 for all test reagents relative to baseline). Importantly, the increase in thickness was not attributable to the swelling of epithelial, stromal, or endothelial cells that were detected commonly in PBS-treated eyes. Instead, swelling seemed to be restricted to extracellular regions of the stroma. 

Figure 2.

 

Corneal thickness was measured using ImageJ software and compared against the baseline. (A–L) Representative images of the corneas from all mUW solutions tested. Scale bar, 300 µm. (M) A graph showing the average corneal thickness from three samples each of all solutions. Data from eyes stored in PBS are also included (gray bar). All groups exhibited retinal swelling in the extracellular regions of the stroma (***P < 0.0001 by analysis of variance). Optimal mUW solution demonstrated the least swelling of all groups tested but was still significantly greater than baseline corneas (t test, ***P < 0.0001).

Immunostaining for Retinal Cell Type-Specific Markers in Eyes After Cold Storage

Frozen retinal sections were stained for RPBMS (RGCs), PKCα (rod bipolar cells), GFAP (astrocytes and Müller cells), and RHO (rod photoreceptors) to assess the effect of cold storage solutions on the presence of markers of different retinal cell types. RBPMS (Fig. 3) and RHO (Fig. 4) staining was relatively uniform across all the different test solutions and not markedly changed from the baseline. Similarly, staining for PKCα was present in all test reagents, although cell morphology was considered suboptimal in the sunitinib, SB203580, and TEA supplemental reagents, which were marked by fewer stained cells and disorganization of their dendritic arbor regions (Fig. 5). This effect was not carried over in the optimal mUW solution, which contained TEA. The most prominent difference in staining was for GFAP, which was generally reduced or absent across most of the test reagents with the exception of VPA+BaCl₂ and L-lactate (Fig. 6). Importantly, there was sporadic evidence of GFAP upregulation in Müller cells in the L-lactate condition, suggesting that a minority of these cells were undergoing reactive gliotic changes in response to cold storage. This effect was not observed in optimal mUW solution, which exhibited GFAP staining in astrocytes similar to the baseline condition. 

Figure 3.

 

Immunostaining of retinas for RBPMS (an RGC-specific marker) in freshly isolated eyes (A) or eyes kept in cold storage for 24 hours after injection of UW solution (B) or mUW solutions containing the supplement(s) indicated (see Tables 2 and 3) (C–L). Optimal mUW solution contained VPA, TEA, hydrocortisone, and L-lactate. The ganglion cell layer (GCL) and robust RBPMS staining appeared relatively normal in all conditions tested. Scale bar, 50 µm. 4′6-Diamidino-2-phenylindole (DAPI) counterstain; INL, inner nuclear layer; ONL, outer nuclear layer.

Figure 4.

 

Immunostaining of retinas for RHO (a rod photoreceptor cell specific marker) in freshly isolated eyes (A) or eyes kept in cold storage for 24 hours after injection of UW solution (B) or mUW solutions containing the supplement(s) indicated (see Tables 2 and 3) (C–L). Optimal mUW solution contained VPA, TEA, hydrocortisone, and L-lactate. RHO staining is most prominent in the outer segments (OS) of the rod photoreceptors, a pattern that was relatively preserved in each of the test conditions, although reduced intensity was noted in both hydrocortisone and L-lactate conditions. The intensity was robust, however, in optimal mUW solution, which contained both supplements. Scale bar, 50 µm. 4′6-Diamidino-2-phenylindole (DAPI) counterstain. INL, inner nuclear layer; IS, inner segments; ONL, outer nuclear layer.

Figure 5.

 

Immunostaining of retinas for PKCα (a rod bipolar cell specific marker) in freshly isolated eyes (A) or eyes kept in cold storage for 24 hours after injection of UW solution (B) or mUW solutions containing the supplement(s) indicated (see Tables 2 and 3) (C–L). Optimal mUW solution contained VPA, TEA, hydrocortisone, and L-lactate. Rod bipolar cell nuclei reside in the outermost region of the inner nuclear layer (INL) and send axons to the dendritic arbors of the RGCs in the ganglion cell layer (GCL). Compared with baseline retinas, these cells seemed to be disorganized and abnormal in mUW solutions containing sunitinib (C), SB203580 (D), and TEA (G). Retinas from eyes stored in optimal mUW solution, which contains TEA, exhibited cells virtually identical to baseline retinas (L). Scale bar, 25 µm. 4′6-Diamidino-2-phenylindole (DAPI) counterstain, which is shown as a separate channel on the left of each image so that bipolar cell morphology can be clearly shown. ONL, outer nuclear layer.

Figure 6.

 

Immunostaining of retinas for GFAP (a cell-specific marker for astrocytes and Müller cells, the latter being in a reactive state) in freshly isolated eyes (A) or eyes kept in cold storage for 24 hours after injection of UW solution (B) or mUW solutions containing the supplement(s) indicated (see Tables 2 and 3) (C–L). Optimal mUW solution contained VPA, TEA, hydrocortisone, and L-lactate. Astrocytes reside on the surface of the retina near the ganglion cell layer (GCL). Overall, cold storage in UW solution and most of the supplemented solutions resulted in a decrease or absence of GFAP staining except for (J) hydrocortisone, (K) L-lactate, and (L) optimal mUW solution. The combination of VPA and BaCl₂ (I) seemed to induce increased in expression in cells restricted to the astrocyte layer, and L-lactate also contained sporadic examples of Müller cell end feet processes containing GFAP (arrows in K). Scale bar, 50 µm. INL, inner nuclear layer; ONL, outer nuclear layer. 4′6-Diamidino-2-phenylindole (DAPI) counterstain.

Assessment of Cell Type-Specific mRNA Abundance in Retinas After Cold Storage

In addition to immunostaining, we also monitored the change in mRNA abundance for several retina cell-types using qPCR. In most test reagents, there was a decrease in mRNA levels for all target genes, with the exception of UW solution that was supplemented with only VPA (Fig. 7). This effect of VPA was not carried over to either UW solution modified with both VPA and BaCl₂, or to the final optimal mUW solution. Notably, the decrease in transcript abundance was not appreciably reflected in a loss of immunostaining for many of the same gene products (the exception being GFAP), suggesting that most immunostaining was of latent protein that had not been turned over and that mRNA levels may be a more sensitive quantitative measurement of cells that had entered a period of stasis while in cold storage. 

Figure 7.

 

Assessment of cell-type specific mRNA abundance in retinas, 24 hours after cold storage. Heat map showing the relative change in transcript abundance compared to freshly isolated eyes. The heat scale bar on the right reflects the ratio of mean mRNA levels of experimental eyes divided by the same transcript level measured in the baseline eyes. Genes included in the mini-array are detailed in Table 4. qPCR results demonstrate a general decrease in mRNA abundance of all marker genes in all mUW groups, with the exception of VPA. The effect of VPA, when in combination with other supplements (including optimal mUW solution) was not retained, however. PRs, photoreceptors.

Discussion

A successful WET has been reported recently,²⁹ which apparently has resolved the barrier of vascular perfusion by a procedure using hemifacial flap transplantation that includes the eye. The use of antirejection medications also mitigates the possibility of immune rejection of the donor tissue. Establishing a fully functional eye, however, remains elusive because the two most serious obstacles of preventing neurodegeneration, principally of the RGCs, and regeneration of severed axons to appropriate targets in the brain, have not been overcome. In this study, we tested modifications of UW cold storage solution that were both designed to mitigate apoptotic signaling within the RGCs so that they survive and to preserve the associated cells in the retina so that they could function normally once the transplant was complete. Preserving RGCs does not guarantee that they will regenerate, and we envision that additional treatments will need to be applied to transplanted donor eyes to induce the preserved RGCs to begin axonal regeneration. 

Several of the components were tested specifically based on their known protective effects on RGCs. HDAC inhibitors were selected because HDAC activity is selectively upregulated in damaged neurons,^30–33 including several retinal cell types such as RGCs.^34–38 Importantly, the inhibition of HDAC activity is strongly protective in a diverse range of neurodegenerative conditions,^39–43 including RGCs in models of acute optic nerve damage.^37,44–49 Of the two pan-HDAC inhibitors tested (VPA and TSA), we selected VPA because it yielded minimal pathological effects, such as the generation of extracellular fluid when assessed by histology, helped to preserve GFAP expression in astrocytes, and performed extremely well in maintaining mRNA abundance levels in retinas of cold stored eyes. 

We also tested K⁺ channel blockers based on the rationale that one of the earliest pathological events associated with neuronal death is an efflux of cellular K⁺ ions^15,50–52 that leads to a phenomenon known as the apoptotic volume decrease.^53–55 Studies using agents that block the K⁺ efflux show that they provide resiliency to neurons in the face of an apoptotic challenge.^53–55 In rodent models of acute and chronic optic nerve damage, RGCs exhibit nuclear atrophy,^56,57 which can be attenuated if the eyes are injected with BaCl₂ to block the K⁺ efflux.⁵⁸ The addition of BaCl₂ to UW solution yielded generally favorable results histologically, even though it created a precipitate that made the mUW solution turbid. As a consequence, BaCl₂ was abandoned in favor of TEA, which has also been used extensively in the literature as a K⁺ channel blocker. 

The third major category that was targeted with supplements was protein kinase activity using the pan-kinase inhibitor sunitinib or the selective p38MAPK inhibitor SB203580. Sunitinib was chosen because it had been shown previously to affect the axis between DLKs and JNKs,⁵⁹ which is the principal mediator of initiating the activation of the intrinsic apoptotic program in RGCs in response to axonal damage.^11,12,60,61 Consistent with this finding, sunitinib delivery to the eye in rat optic nerve damage models reveals a promising neuroprotective effect on the RGC population.^62,63 The activation of p38MAPK is also associated with the DLK–JNK axis⁶⁴ and is implicated in activating the intrinsic apoptotic pathway by regulating the activity of the proapoptotic BAX protein,^65–67 as well as mediating K⁺ channel opening during the apoptotic volume decrease.⁶⁸ The inhibition of p38MAPK has been shown to have a protective effect on RGCs in animal models of optic nerve damage.^69–71 Surprisingly, the addition of either of these kinase inhibitors to UW solution yielded poor overall outcomes to retinal architecture after cold storage, with both leading to excessive protein accumulation in the anterior and posterior chambers and regions of focal retinal edema. 

Macroglia include astrocytes, which are situated in close proximity to the nerve fibers of the RGCs in the inner retina, and Müller glia, which have cell bodies in the inner nuclear layer and extend processes to both the nerve fiber layer and the base of the photoreceptor outer segments. These cells play critical roles to neuronal maintenance, including providing energy substrates to the neurons in the form of L-lactate,^72,73 homeostasis of ion and small molecule environments,^74,75 and neurotrophic support.^76–78 To assess the effect of cold storage solutions on the macroglia, we interrogated both the pattern and level of expression of GFAP. This served two purposes. The first was to determine if Gfap gene expression remained restricted to astrocytes, and therefore normal, and the second was to evaluate the level of glial activation, which is characterized by an increase in Gfap expression in both astrocytes and Müller cells.^79,80 Initial studies with mUW solutions showed that, rather than increases in GFAP, there was a dramatic decrease in expression, suggesting that astrocytes were susceptible to the cold storage process. Hydrocortisone, however, applied to retinal explants, was shown to stabilize GFAP expression, and actually increase expression which was associated with an increase in the regenerative potential of the RGCs.^81,82 The addition of hydrocortisone to UW solution was able to help preserve GFAP staining in the astrocyte layer. 

Given that macroglial activity may be suppressed partially during cold storage, and therefore unable to provide critical energy support to adjacent neurons, we also tested the addition of L-lactate in UW solution. The rationale for this addition was that the transplanted cells would have immediate access to this energy substrate and provide a buffer period to allow macroglia to recover. L-lactate had several effects on the efficacy of mUW solution. By itself, it was able to help preserve GFAP immunostaining in the astrocyte layer, but it also induced moderate upregulation of GFAP protein accumulation in Müller cells. The reasoning for this is not clear, but it is known that L-lactate is also a critical energy substrate for these cells, and that they are actively involved in L-lactate homeostasis in healthy retinas.⁸³ Unexpectedly, L-lactate, only in combination with other supplements in optimal mUW solution, prevented retinal detachment. L-lactate is also an energy substrate for RPE cells.⁸⁴ These cells surround the distal ends of photoreceptor outer segments and provide structural and metabolic support,⁸⁵ so it is likely that the L-lactate helps to maintain this important role for the RPE by keeping them active. It is not clear, however, why the effect of L-lactate is dependent on the other supplements in the optimal mUW solution and remains an area needing further investigation. 

In combination, the supplements that comprise the optimal mUW solution provided exceptional preservation of retinal and cellular architecture after 24 hours in cold storage. Cells, although seeming to be normal, did exhibit a decrease in mRNA abundance across the retina, suggesting that cold storage was pushing the cells into a period of stasis that was characterized by reduced levels of transcription but not mRNA turnover. Additional testing is required to determine if the period of stasis is reversible. We anticipate extending these studies to rat eyes in a rodent model of hemifacial transplantation.^8,9 

Acknowledgments

The authors thank Christian Jobin for providing the cis-NFκB^EGFP reporter mice, Joel Dietz for managing the mouse colony and assisting in the quantitative PCR experiments, and Santoshi Kinoshita at the Translational Research Initiative in Pathology (TRIP) laboratory at the University of Wisconsin-Madison for cutting sections analyzed in this study. This work was supported by Department of Defense grant W81XWH-16-0775 (to K.W.), core grant P30 EY016665 (Department of Ophthalmology and Visual Sciences at the University of Wisconsin-Madison), a grant from the Lion's Eyebank of Wisconsin (to R.W.N.), and unrestricted funding from Research to Prevent Blindness, Inc (Department of Ophthalmology and Visual Sciences at the University of Wisconsin-Madison). 

The components and use of the modified UW solution as described in this report are the subject of U.S. provisional patent (#63/566646) filed by the Wisconsin Alumni Research Foundation, Madison, Wisconsin. The authors of this manuscript have no further conflicts of interest to disclose. 

Disclosure: N.A. Muench, None; H.M. Schmitt, None; C.L. Schlamp, None; A.-J.A. Su, None; K. Washington, None; R.W. Nickells, None