Triple Immune Suppression Increases Short-Term Survival of Porcine Fetal Retinal Pigment Epithelium Xenografts

Abstract

purpose. To determine the effect of triple drug immune suppression on RPE xenograft survival in the fetal pig after transplantation into the albino rabbit subretinal space.

methods. Primary RPE microaggregates (approximately 40,000 RPE cells) were injected into the subretinal space of 24 albino rabbits, with half the rabbits maintained on triple systemic immune suppression. RPE survival was estimated with a DNA probe (porcine DNA repeat element; PRE) against a porcine-specific repetitive chromosomal marker or a RAM-11 antibody against rabbit macrophages.

results. Numerous pigmented cells were visible in the subretinal space at all time points, but most pigment-containing cells 4 weeks or more after surgery were RAM-11 positive and PRE negative. The number of PRE-positive cells in the immune-suppressed group (4193 ± 2461, 1184 ± 1502, and 541 ± 324 at 4, 8, and 12 weeks, respectively) was greater than in immune-competent control animals (292 ± 506, 193 ± 173, and 111 ± 96), but the difference was only statistically significant at 4 weeks. The time-dependent decrease in PRE-positive cells was more pronounced in immune-suppressed animals. Image analysis performed on serial fundus photographs and fluorescein angiograms did not detect any difference in the appearance of the grafts in immune-suppressed versus immune-competent animals.

conclusions. Systemic immune suppression increased the 4-week survival of porcine RPE xenografts in the albino rabbit subretinal space, but there was poor survival in immune-suppressed and -competent animals 12 weeks after surgery. Many pigment-containing cells 4 or more weeks after surgery were PRE negative, indicating that they are of host origin.

In the normal human eye, the retinal pigment epithelium (RPE) is a hexagonally packed monolayer that lines the inner aspect of Bruch’s membrane between the neurosensory retina internally and the choriocapillaris externally. ¹ The RPE performs many important functions that are crucial to maintaining the integrity of the retina and choriocapillaris, including phagocytosis of the distal tips of photoreceptor outer segments, transport and isomerization of bleached visual pigments, and maintenance of the blood–outer retinal barrier. ¹ In addition to these functions, an intact RPE is necessary to maintain perfusion of the subjacent choriocapillaris. Drug-induced damage to the native RPE through intravitreal injection of ornithine, iodate, or iodacetic acid leads to secondary choriocapillaris atrophy. ^{2 3 4 5 6 7} Removal of RPE during submacular surgery in patients with subfoveal neovascularization may lead to progressive choriocapillaris atrophy, ^{8 9 10} and surgical RPE debridement with incomplete repopulation of the dissection bed in cats and pigs is associated with subjacent atrophy. ^{11 12 13 14}  

Approximately two decades ago an interest in RPE transplantation was fueled by the concept that primary RPE dysfunction is an early step in the pathogenesis of tapetoretinal degenerations, including retinitis pigmentosa. Subsequent genotyping studies of retinal degenerations demonstrate that the gene defect causing retinitis pigmentosa and similar retinal degenerations is nearly always caused by abnormalities in a gene expressed in photoreceptors, ^{15 16 17 18} with the exception of mutations in the RPE65 gene that cause some cases of Leber’s congenital amaurosis and the cellular retinaldehyde-binding protein (CRALBP) gene. ^{19 20} Thus, it is unlikely that RPE transplantation alone is sufficient to help most patients with tapetoretinal degenerations. However, a major need for successful RPE transplantation exists in the management of age-related macular degeneration (AMD). ^{21 22 23 24 25 26 27 28 29 30 31 32} Loss of the RPE precedes loss of choriocapillaris in patients with nonexudative AMD, and RPE transplantation may prevent or reverse choriocapillaris atrophy in these individuals. ^{27 28} The rationale for RPE transplantation in exudative AMD is deceptively simple. In this disease, the native RPE is excised with the choroidal neovascular complex during submacular surgery, ³³ and removal of RPE leads to progressive choriocapillaris atrophy and photoreceptor loss. ²⁸ RPE transplantation may prevent subsequent choriocapillaris atrophy and improve the visual prognosis. In initial studies, RPE transplantation in patients with exudative or nonexudative AMD has not resulted in significant improvement in vision, but graft survival may be hampered by the lack of immune suppression in most studies and damage to the host Bruch’s membrane with incomplete RPE repopulation of this structure. ^{21 22 23 24 25 26 27 28 29 30 31 32 34 35 36}  

Despite an extensive body of peer-reviewed literature on RPE transplantation, several fundamental questions remain unanswered about the behavior of transplanted RPE, due to difficulty identifying transplanted cells. In this study, we used in situ hybridization to detect a repeat portion of the porcine chromosome to identify donor porcine RPE cells after transplantation into the subretinal space of the albino rabbit. ³⁷ This probe allowed us to determine the short-term survival and behavior of fetal porcine RPE xenografts after subretinal transplantation, with and without systemic triple immune suppression. 

Materials and Methods

Preparation of RPE for Transplantation

The harvesting procedure was performed at Diacrin (Charlestown, MA) 24 hours before the tissue was shipped to our laboratory at the Harkness Eye Institute (New York, NY). Porcine skin gelatin powder (300 blooms rigidity; Sigma-Aldrich, St. Louis, MO) was sterilized by autoclaving, and 300 mM sucrose (Sigma-Aldrich) was added before the powder was dissolved in Dulbecco’s phosphate-buffered saline (DPBS; Invitrogen-Gibco, Grand Island, NY). The solid gelatin solution was stored at room temperature. Immediately before use, the gelatin was melted in a 37°C water bath and 50 μL of 25% gelatin and 300 mM sucrose was spread onto a glass slide and solidified at 4°C for at least 6 minutes. 

RPE were harvested by using a modification of techniques described previously. ^{38 39} Freshly enucleated fetal porcine eyes were cleaned of extraocular tissue. Scissors were introduced at the optic nerve into the vitreous cavity and the incision was extended through the sclera, choroid, and retina toward the iris. Three additional radial incisions were made, and these cuts were completed with a number 10 scalpel and the tissue cut into 3-mm squares. The retina was removed carefully and the tissue was incubated with 6.25 U/mL dispase (Invitrogen-Gibco) for 8 hours at 4°C. The loosened RPE sheets were separated from the remainder of the ocular tissue and placed on a drop of 25% gelatin with the apical surface of the RPE facing upward. Contamination with choroidal cells was avoided by visualizing the RPE sheets under a dissecting microscope during harvesting. The gelatin film containing the RPE sheet was then incubated in a humidified atmosphere of 5% CO₂ and 95% air at 37°C for 2 minutes to allow the gelatin to melt and encase the RPE sheet. The specimen was kept at 4°C for 4 minutes to solidify the liquid gelatin, covered with 20 μL of 10% gelatin and 300 mM sucrose in DPBS at 37°C, returned to 4°C for 6 minutes, and shipped to our laboratory overnight in CO₂-free medium (CFM; Gibco-Invitrogen) at 4°C. On receipt in the laboratory, the RPE-containing sheet was triturated into small microaggregates in DMEM and diluted to a concentration of approximately 40,000 cells/10 μL for injection into the subretinal space. Cytokeratin staining before transplantation indicated that all the cells were of epithelial origin. ³⁸  

Cell Viability Analysis

Triturated RPE cells were transferred to a Petri dish containing MEM and incubated at 37°C for 5 minutes to melt the gelatin. Viability was assessed with a cytotoxicity kit (Live/Dead Viability/Cytotoxicity; Molecular Probes, Eugene, OR) containing calcein and ethidium homodimer, as described previously. ⁴⁰ At least three different areas containing approximately 250 cells each were counted under 100× magnification. The viability of the triturated RPE cells was expressed as the average of the ratio of live cells to the total number of cells in at least three different areas. ⁴⁰ Grafts with viabilities lower than 90% were not used for transplantation. 

Delivery of Fetal Porcine RPE into Rabbit Subretinal Space

All surgical procedures were performed in one eye of albino (2–4 kg) rabbits cared for in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Twenty-four animals underwent transplant surgery. Animals were sedated with intramuscular ketamine and xylazine, and a sterile field was established around one eye. A glass pipette was introduced through the pars plana and vitreous cavity to enter the subretinal space under direct visualization. A bleb neurosensory retinal detachment was created by injecting approximately 10 μL of DMEM followed by 10 μL of cell suspension containing approximately 40,000 RPE cells in microaggregates into the subretinal space. The instrument was withdrawn, and the single sclerotomy was closed with a 7-0 coated Vicryl suture (Ethicon, Somerville, NJ). Two milligrams dexamethasone (Anpro Pharmaceuticals, Arcadia, CA) and 20 mg gentamicin sulfate (Elkins-Sinn, Cherry Hill, NJ) were injected subconjunctivally. Subcutaneous buprenorphine 0.005 mg/kg (Reckitt and Colman Pharmaceuticals, Inc., Richmond, VA) was administered for postoperative pain control. 

Immune Suppression

Half the rabbits received daily systemic immune suppression with prednisone, cyclosporine, and azathioprine from the day of surgery until the time of death. The drug dosages used were adapted from the regimen used after renal transplantation. Animals received azathioprine (5 mg/kg per day intramuscularly for 3 days, 2.5 mg/kg per day thereafter), prednisone (1 mg/kg per day for 2 weeks, 0.5 mg/kg per day for 2 weeks, 0.4 mg/kg per day for 2 weeks, and 0.3 mg/kg per day thereafter), and cyclosporine (5 mg/kg per day orally for 1 month, 4 mg/kg per day for the next month, and 3 mg/kg per day thereafter). Serum cyclosporine levels were determined in some of the animals, and the cyclosporine dose was adjusted to maintain the desired trough blood level (350–400 ng/mL for 1 month, 300–350 ng/mL for the next month, and 250–300 ng/mL thereafter). All animals receiving immune suppression were monitored for drug toxicity with blood tests every 2 weeks, including a complete blood count, platelets, and serum electrolytes. The cyclosporine dose was reduced if blood urea nitrogen or creatinine increased up to 50% and discontinued if the blood urea nitrogen or creatinine increased to more than 50% above baseline levels. Azathioprine was discontinued if there was a 50% decrease in white blood cell count. None of the animals required lowering or discontinuation of azathioprine or cyclosporine due to drug toxicity. 

Sequential Image Analysis

We obtained color photographs, red-free photographs, and fluorescein angiograms at 2, 4, 8, and 12 weeks after surgery. A camera (Canon USA, Lake Success, NY) with filters matched to the excitation and emission wavelengths of fluorescein was used for angiography. Red-free and color photographs were obtained using the same optical system. Intravenous access was established, and 2 mL of 10% sodium fluorescein was injected intravenously. Fundus photographs were taken with black-and-white film (T-max 400; Kodak, Rochester, NY) immediately before and in rapid sequence after fluorescein injection. The transplant area was photographed during the transit phase of the fluorescein angiogram. Late-phase photographs were taken up to 15 minutes after fluorescein injection. We then performed digital image analysis of red-free photographs to determine whether there was any quantifiable difference in the appearance of the transplant in immune-suppressed versus immune-competent animals at different time points after surgery. The average brightness and contrast of sequential photographs were standardized using Image J software (http://rsb.info.nih.gov/ij; developed by Wayne Rasband and provided in the public domain by the National Institutes of Health, Bethesda, MD) before measuring the size and average gray-scale brightness of the pigmented patch. We then used commercial image-analysis software (Photoshop; Adobe Systems, Mountain View, CA) to prepare digitally subtracted images for immune-suppressed and -competent animals. 

Tissue Processing

The eye was enucleated rapidly at different time points after death, immersed in 4% paraformaldehyde in PBS, fixed for at least 48 hours, and stored in PBS at pH 7.4. Intact globes were opened with a full-thickness circumferential incision posterior to the ora serrata. The posterior eyecup was inspected, and the transplant area was excised from the eyecup. The excised specimens were processed for light microscopy by washing in PBS, dehydrating in graded alcohol, clearing in xylene, embedding in paraffin, and cutting to 5-μm thickness. 

For in situ hybridization, paraffin sections mounted onto glass slides were deparaffinized, hydrated, and treated at 50°C for 4 minutes with pepsin (Research Genetics, Huntsville, AL). After rinsing with sodium citrate-saline (SSC), the sections were postfixed at room temperature for 10 minutes in 4% paraformaldehyde, rinsed again, dehydrated through a graded ethanol series, and air dried. A digoxigenin-labeled PCR probe containing the porcine DNA repeat element (PRE) was added to the slides at a concentration of 100 ng/mL in 65% formamide and 2× SSC. ⁴¹ The DNA was denatured at 95°C for 5 minutes, and hybridization was performed at 37°C overnight. The slides were washed twice with PBS after a series of posthybridization washes with the hybridization buffer for 10 minutes each at room temperature (i.e., 37°C and 42°C). The slides were incubated for 2 hours with an alkaline phosphatase-conjugated monoclonal antibody against digoxigenin (Roche Diagnostics, Nutley, NJ), then developed with 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium (BCIP/NBT; Zymed, San Francisco, CA). Bleaching with 10% hydrogen peroxide reduced the natural pigmentation of the cells. The tissue was counterstained with nuclear fast red (Polyscientific, Bay Shore, NY). The number of PRE-positive RPE cells was estimated by counting the number of PRE-positive cells in every 20th section through the transplant bed and then correcting for the distance between counted sections (100 μm). We performed control studies by using this staining technique to detect PRE-positive RPE in sections of obtained from healthy porcine eyes embedded in the same manner as described earlier. 

Rabbit macrophages were detected with the monoclonal antibody RAM 11 (Dako Corp., Carpinteria, CA), which binds an uncharacterized cytoplasmic antigen in rabbit macrophages and has low cross-species reactivity. ⁴² Paraffin-embedded tissue sections mounted onto glass slides were deparaffinized, hydrated, and washed for 5 minutes in PBS. Sections were blocked with 1.5% normal horse serum for 20 minutes and then incubated with RAM 11 antibody diluted 1:100 overnight at 4°C. Sections were washed with PBS three times for 10 minutes each, and the antibody was detected with a streptavidin biotin alkaline phosphatase kit (Vector Laboratories, Burlingame, CA). The biotinylated secondary antibody was incubated for 30 minutes at room temperature followed by a 5-minute wash with PBS, a 30-minute incubation with the biotinylated alkaline phosphatase/streptavidin complex reagent, and another 5-minute PBS wash. Red stain (Vector Red Alkaline Phosphatase Substrate Kit I; Vector Laboratories) was used as the enzyme substrate. Sections were washed in PBS and rinsed with water, and the natural pigmentation was bleached with 10% hydrogen peroxide. An irrelevant isotype-matched antibody (anti-TNP; BD-PharMingen, San Diego, CA) was used as a negative control and showed no background staining of RPE transplants. RAM 11 did not stain normal rabbit RPE cells or porcine RPE control cells. 

Results

Four weeks after transplantation, there was heavy pigmentation at the transplant site and fluorescein angiography showed a blocking defect with no leakage or staining in both immune-suppressed and -competent animals (Figs. 1 2 , respectively). In immune-suppressed animals, there was a layer of heavily pigmented cells under the retina at the transplant site, with some cystic spaces suggesting acinar formation (Fig. 1C) . Some pigment was present on the inner aspect of the retina and the retinal thickness and lamellae appeared normal. PRE-positive cells were visible within the transplant bed (Fig. 1D) , although many of the pigment-containing cells were PRE negative and positive for the macrophage marker RAM 11 (Fig. 1E) . In one eye, 4 weeks after surgery, there were numerous PRE-positive cells distributed around a lumen, consistent with the known ability of epithelial cells to form acini (Fig. 1F) . Pigmented cells lining the lumen were predominantly PRE positive, whereas pigmented cells not lining the lumen were predominantly RAM 11 positive (Figs. 1G 1H) . There were some pigmented cells in the subretinal space in the immune competent group 4 weeks after surgery, and pigment was visible within the choroid as well (Fig. 2C) . The inner and outer retina appeared intact. There were occasional PRE-positive cells (Fig. 2D) and numerous RAM 11–positive cells in the subretinal space (Fig. 2E) . In one animal, a choroidal neovascular membrane developed, possibly from perforation of Bruch’s membrane during surgery (not shown). We performed companion studies, using this staining technique to determine the sensitivity and specificity of the PRE staining technique in sections of obtained from healthy porcine eyes and nontransplant regions of rabbit eyes embedded in the same manner as described earlier. One control was run each time the PRE staining was performed. In a total of eight runs, the percentage of porcine RPE cells that stained PRE positive was 84.0% ± 7.7%. There was no significant staining of rabbit RPE sections with the PRE marker. ⁴²  

Twelve weeks after surgery, the transplant site could still be identified by heavy pigment in the subretinal space with blockage by pigment, but no leakage or staining on fluorescein angiography in immune-suppressed and -competent animals (Figs. 3 4 , respectively). In immune-suppressed animals, viable pigmented cells were visible in the subretinal space (Figs. 3C 3F) . Some of the pigmented cells were PRE positive (Figs. 3D 3G) . Many of the pigmented subretinal cells were RAM 11 positive (Figs. 3E 3H) . In immune-competent animals, there were numerous pigmented cells in the subretinal space (Figs. 4C 4F) , with some PRE-positive cells (Figs. 4D 4G) and RAM 11–positive cells (Figs. 4E 4H) . 

The number of PRE-positive cells was higher in the immune-suppressed group compared with the immune-competent group 4 weeks after surgery (4193 ± 2461 vs. 292 ± 506, respectively; P < 0.05; Fig. 5 ). By 8 weeks after surgery, there were more PRE-positive cells in immune-suppressed animals than in immune-competent animals (1184 ± 1502 vs. 193 ± 173, respectively) but this difference was not statistically significant because of the small number of eyes in each group. By 12 weeks after surgery, there was no statistically significant difference in the number of PRE-positive cells in the immune-suppressed versus immune-competent group (541 ± 324 vs. 111 ± 96, respectively). 

We also determined the average cyclosporine level in immune-suppressed animals up to 8 weeks after surgery (Fig. 6) . Cyclosporine levels were below target levels at 2 weeks after surgery and were above target levels at later time points, despite measurement of levels every 2 weeks with adjustment of cyclosporine dose thereafter. None of the animals showed an increase in blood urea nitrogen or creatinine from cyclosporine toxicity and no animal had a 50% decrease in white count that was attributable to azathioprine toxicity (data not shown). There was no statistical correlation between cyclosporine level and the number of surviving cells, although the number of animals was small. 

We performed digital image analysis of red-free photographs to determine whether there was any quantifiable difference in the appearance of the transplant in immune-suppressed versus immune-competent animals at different time points after surgery. Images were obtained at 2, 4, 8, and 12 weeks after surgery. There was no significant change in either size of the pigmented patch or gray-scale value of the transplant site between 2 and 12 weeks after surgery. The ratio of the size of the pigmented patch (12:2 weeks) in the immune-suppressed group versus the immune-competent group was 95.06% ± 8.97% and 87.46% ± 7.97%, respectively, but this difference was not statistically significant (P = 0.252, t-test). The change in gray scale value of the patch (12:2 weeks) in the immune-suppressed and -competent groups was 102.72% ± 0.90% and 120.91% ± 7.50%, respectively (P = 0.053, t-test). Fluorescein angiography demonstrated a blocking defect in the area of the pigmentation, with no leakage or staining of the transplant site in either immune-suppressed or -competent animals at all time points. 

Discussion

Short-Term RPE Survival after Transplantation

In the present study, donor RPE cells were identified at all time points after surgery, but the number detected decreased with time. Four weeks after injection of 40,000 RPE cells into the subretinal space of the albino rabbit, we detected approximately 4200 (10.5% survival) and 1200 (3% survival) PRE-positive cells in immune-suppressed and -competent animals, respectively. By 12 weeks after surgery, only 100 to 200 PRE-positive cells (0.25%–0.50% of the injected cells) were detected. These numbers are comparable to the short-term survival rates after transplantation of dopamine-producing neurons into the brain of rats with chemically induced Parkinson’s disease. Approximately 5% to 6% of human dopamine-producing neurons harvested from 6.5- to 8-week-old fetuses survived after grafting into cyclosporin A–immunosuppressed rats, with no detectable survival in immune-competent rats and a lower survival rate of cells from 11-week-old donors. ⁴³ The survival rates of dopamine-producing neurons in patients receiving grafts of human nigral tissue has ranged from 1% to 4% (see Ref. ⁴³ for review). 

There was no discernible difference in the ophthalmoscopic or fluorescein angiographic appearance of the transplant area between immune-suppressed and -competent animals. It is tempting to draw inferences about the behavior of the transplant on the basis of sequential fundus photography or fluorescein angiography, but the present study suggests that sequential fundus photography cannot detect differences in graft behavior between the immune-competent and -suppressed animals. We suspect this is because dying cells release subretinal pigment that is ingested by host RPE cells or macrophages and sometimes is brought into the choroid. This could lead to no change in the topographical distribution of pigment viewed by ophthalmoscopy or sequential fundus photography, despite loss of transplanted cells. An increase in size of the pigmented patch may represent growth or spreading of the transplant, but may also represent migration of pigment-laden macrophages away from the transplant bed. Fluorescein angiography did not detect differences in behavior of the transplanted cells between immune-suppressed and -competent animals in the present study. None of the animals demonstrated leakage from the choroidal circulation, despite the loss of some cells by mechanisms inhibited partially by systemic immune suppression. It is not known whether sequential fluorescein angiography performed in a species with an intact retinal circulation, such as the nonhuman primate or pig, would lead to similar conclusions. 

Immunity in the Subretinal Space and Role of Immune Suppression in Graft Survival

In this study, the number of PRE-positive cells in immune-suppressed animals was greater than in immune-competent animals at 4 and 8 weeks after surgery, although the latter was not statistically significant because of the small number of animals. This implies that systemic immune suppression promotes short-term survival of pig RPE xenografts, but does not guarantee longer survival. We did not observe prominent lymphocyte infiltration or other signs of acute graft rejection in these animals, yet our results suggest that immune mechanisms were responsible partly for loss of the graft, despite the relative immune privilege of the subretinal space. Hyperacute rejection, an antibody-mediated event leading to acute vascular occlusion and loss of organ xenografts, was not observed in the present study. Despite the absence of overt graft rejection, our short-term 4-week data show that the immune system plays a role in regulating xenograft survival. 

Our data are consistent with findings in prior studies that show slow deterioration of grafted cells within a few months of transplant surgery. Porcine RPE cells placed into the subretinal space of non–immune-suppressed RCS rats induced a dramatic rescue effect compared with sham-injected control animals, with a delay in photoreceptor cell degeneration. ⁴⁴ These xenografts underwent a slow functional deterioration reflected by a decline in their capability to rescue adjacent photoreceptors. ⁴⁴ The number of 5-bromo-2-deoxyuridine (BrdU)-labeled human fetal RPE injected subretinally into rabbit eyes declines starting from day 14 after transplantation, suggesting possible rejection. ⁴⁵ Grafted RPE can preserve visual function after subretinal transplantation in animal models, such as the RCS rat. ^{46 47}  

Other investigators have demonstrated the importance of the immune system and the relative immune privilege of the subretinal space in determining graft survival. Several important facts have been established. First, the subretinal space is an immune-privileged site, but the immune privilege of this space is relative rather than absolute. RPE allografts, from C57BL/6 newborn mice, when implanted into the subretinal space and anterior chamber of adult BALB/c mice, survive longer than subconjunctival implants with suppression of donor-specific delayed hypersensitivity. ⁴⁸ RPE cells transplanted from two rat strains with incompatible major histocompatibility complex (MHC) haplotypes were used as healthy RPE donor cells and transplanted into the subretinal space of the RCS rat. ⁴⁹ Grafts with disparity in MHC classes I and II lost the ability to rescue photoreceptor cells more readily than grafts with disparity at MHC class II alone. ⁴⁹ There was increased photoreceptor loss in RCS rats challenged with spleen cells from the donor 2 weeks after transplantation, ⁴⁹ and rejection of previously healthy intraocular RPE grafts within 2 weeks after donor-specific challenge. ⁴⁸ Second, neonatal murine RPE is an immune-privileged tissue, owing its immune privilege to the capacity to prevent immune rejection rather than to inhibit sensitization. ^{50 51} It is not certain whether this relative immune privilege extends to fetal or neonatal RPE from other species, including humans. Third, immune suppression can increase short-term graft survival in certain animal models but may not be sufficient to guarantee it. Immune suppression with intraocular cyclosporin A increases short-term survival of human fetal RPE xenografts in the rabbit subretinal space, ⁵² although daily cyclosporine (20 mg daily intramuscularly) does not completely prevent loss of RPE grafted from brown rabbits into albino rabbits. ^{53 54}  

Identification of Transplanted Cells

In contrast to previous studies, we identified transplanted cells by performing in situ hybridization with a DNA probe (PRE) against a repetitive segment of the porcine chromosome and simultaneously used the RAM 11 antibody to identify rabbit macrophages after surgery. ^{41 42} Use of the PRE probe allows us to draw inferences about the behavior of transplanted RPE cells that would not be possible without it. Others have used pigment as a marker to identify transplanted cells, but released pigment can be ingested by host RPE cells or macrophages. ^{55 56 57 58 59 60 61 62 63 64 65} Numerous pigmented cells were visible in the subretinal space at all time points in the present study. Some of these pigment-laden cells were PRE or RAM 11 positive, but many of the pigment-containing cells were both PRE and RAM negative. Cells that contained pigment and were negative for both PRE and RAM 11 staining probably represent host RPE cells, sometimes dislodged from Bruch’s membrane, that have ingested pigment released by dying transplanted cells. Thus, the present study calls into question the observations and conclusions of previous investigations that relied on pigment alone to identify transplanted RPE cells. Other markers have also been used to identify transplanted cells with mixed results, including pretransplantation labeling of donor cells with di-I, nuclear yellow, or BrdU. ^{55 56 57 58 59 60 61 62 63 64 65} Recent advances have included the use of a Y-chromosome probe to identify male feline RPE cells transplanted into the female cat’s subretinal space ⁶⁶ and Barr body techniques to identify female porcine RPE cells transplanted into the male pig’s subretinal space (Jones Z, et al. IOVS 1997;38:ARVO Abstract 1572). Both techniques produce a high specificity rate for donor tissue, and their sensitivity can be as high as 50% to 70% (Jones Z, et al. IOVS 1997;38:ARVO Abstract). ⁶⁶ Green fluorescent protein has a high sensitivity and high specificity, but this technique is limited to identification of passaged cells genetically modified to express the protein. ^{67 68 69} Development of a transgenic animal expressing green fluorescent protein should allow this marker to be used to track primary RPE harvested from mice, with eventual expansion to other species. ⁷⁰  

Barriers to Successful RPE Transplantation

The current investigation was designed to determine the effect of immune suppression on survival of RPE microaggregates after subretinal transplantation and is one of several studies that should be considered in understanding graft survival in patients with AMD. Under ideal circumstances, RPE transplantation would be performed in an animal model of AMD, but no complete animal model for AMD exists. Thus, our understanding of RPE graft behavior after subretinal transplantation in AMD must be spliced together from a variety of incomplete models, including in vivo transplantation studies in animal models with intact basement membrane, tissue culture studies of RPE attachment to normal and damaged Bruch’s membrane, and understanding of cell transplant survival in other diseases such as Parkinson’s disease. Despite these limitations, several patterns emerge. We have shown that RPE harvested for transplantation must reattach to a substrate to avoid programmed cell death, ⁷¹ and cell surface contact increases graft survival in vitro. ^{34 35 36} In the present study, the cells were harvested as sheets and then dissociated into microaggregates, and it is interesting to note that the largest number of surviving RPE at 4 weeks was in animals with RPE cells arranged in acini with extensive cell–cell contact (see Fig. 1F 1G 1H ). The surface onto which the RPE cells are seeded has a significant impact on RPE survival, because RPE attachment and proliferation are highest on young versus old basal lamina of Bruch’s membrane. ^{34 35 36} RPE cells seeded onto basal lamina have a higher survival and proliferation rate and lower apoptosis rate than cells seeded onto the inner collagen layer or elastin layer of Bruch’s membrane. ^{34 35 36} In addition, it is not known whether there is a mismatch between porcine RPE membrane receptors and extracellular matrix ligands within the Bruch’s membrane in the rabbit. This possible mismatch exists in all xenograft studies involving epithelial cell transplantation. Other factors that may influence graft survival include details of harvesting techniques including the presence or absence of cell–cell contacts, and disruption of cell–substrate interactions are also important in graft survival. 

Additional lessons can be gleaned from the neurobiology of transplantation of dopamine-producing cells in Parkinson’s disease (reviewed by Brundin and Hagell ⁷² ). Prior cryopreservation improves graft survival and function after transplantation of a porcine ventral mesencephalon single-cell suspension into the striatum of 6-hydroxydopamine-lesioned rats. ⁷³ In vitro preincubation of human fetal tissue strands with IGF-I and bFGF improves cell survival and the behavioral outcome of dopamine-producing neurons transplanted into the same animal model. ⁷⁴ Cell survival after transplantation in Parkinson’s disease also depends on technical aspects of graft preparation including harvesting technique and the design of the delivery cannula, the age of the donor, the age of the recipient, tissue storage methods, details of the surgical technique, and the extent of postoperative immune suppression. ⁷²  

At the current time, the ideal population for transplantation to reverse or prevent RPE-mediated disease is not known. Several sources of transplanted cells have been tested to date, including adult and fetal human RPE cells transplanted into a small number of patients with AMD, ^{21 22 23 24 25 26 27 28} autologous RPE harvested from other regions of the same eye, ^{31 32} and autologous iris pigment epithelium harvested as patches or sheets from the periphery. ^{29 30} In addition, workers have proposed the use of stem cells ⁷⁵ and RPE xenografts. ⁷⁶ The use of xenografts offers some distinct advantages over the use of human tissue, including greater availability of donor tissue, the possibility of using inbred animals or established cells lines engineered with desired immune and functional characteristics, and the use of fetal tissue without the legal and ethical barriers that accompany use of fetal human tissue. ⁷⁶ However, xenografts may be more likely to be rejected than same species grafts and have the theoretical potential to introduce pathogenic agents across species lines. Despite these concerns, xenografts have already been implanted into the central nervous system of patients with Parkinson’s disease ^{76 77 78 79 80 81 82} and avascular xenografts of porcine tissue have been used for cardiac valve transplantation. ^{83 84 85 86} The role of xenografts in the management of RPE-mediated disease or dysfunction remains to be determined. 

Although in the present study xenografts did not survive in the subretinal space in the long term, even with immune suppression, these data do not mean that xenografts will not work in this setting. Modification of several variables, such as harvesting technique, tissue handling before transplantation, delivery of donor cells as sheets rather than suspensions, and the simultaneous delivery of cell survival factors, have increased graft survival in Parkinson’s disease, and similar efforts may be necessary to guarantee long-term RPE survival. Successful long-term survival of xenografts would usher in a new era of management of human ocular disease, because porcine tissue may offer a solution to the shortage of human RPE available for transplantation. ^{87 88 89} In theory, porcine xenografts introduce the probability of cross-species transmission of viruses, including porcine endogenous retrovirus (PERV), which is a C-type retrovirus permanently integrated in the porcine genome. Two PERVs, type A and B, productively infect human cells and are therefore considered to constitute a potential risk in pig-to-human xenotransplantation. ⁸⁹ The concern about transspecies transmission of PERV has received much attention. To date, several hundred patients have been exposed to living porcine cells and tissues, including pancreatic islet cells, skin, and whole livers and spleens for extracorporeal blood perfusion. ^{90 91 92} Reverse transcription-polymerase chain reaction (RT-PCR) and protein immunoblot analyses performed on serum from 160 patients who had been treated with various living porcine tissues up to 12 years earlier detected no viremia in any patient. Peripheral blood mononuclear cells from 159 of the patients were analyzed by PCR using PERV-specific primers. No PERV infection was detected in any of the patients from whom sufficient DNA was extracted to allow complete PCR analysis (97% of the patients). Dinsmore et al. ⁹³ found no evidence of PERV provirus integration into host DNA after transplantation of fetal porcine neuronal cells to the central nervous system of 24 patients with intractable neurologic disorders such as Parkinson’s disease, Huntington’s disease, or epilepsy, and there was no transfer of PERV from porcine fetal neuronal cells to human cells in vitro. Although these results are encouraging, they do not exclude the possibility of pig-to-human transmission of PERV or other viruses. ⁸⁹  

The present study has several limitations. First, there was often some reflux of cells into the vitreous cavity after subretinal injection of RPE microaggregates. This introduces the possibility of animal-to-animal variation with underestimation of survival rates at different time points after surgery. Second, average cyclosporin A levels were lower than target levels during the first 2 weeks after surgery and thereafter were higher than target levels. This raises the possibility that graft rejection may have occurred due to inadequate cyclosporine levels. However, immune-suppressed animals were also maintained on azathioprine and prednisone, and the ideal cyclosporine concentration for suppression of subretinal RPE rejection is not known, as target levels were derived from data on solid organ transplantation. It is not known whether modification of the immune-suppression protocol, including higher doses of these three drugs or the addition of newer immune suppressive agents, would have yielded better results. Third, we do not know whether specific issues associated with our harvesting technique and subsequent surgery, such as the use of dispase and gelatin during the harvesting, or details of our surgical technique may improve cell survival after transplantation surgery. Fourth, the results we obtained are applicable strictly to microaggregate xenografts in an animal model with healthy retina, native RPE and choriocapillaris, and intact basement membrane. These results cannot be extrapolated directly to the behavior of RPE cells transplanted as sheets, RPE seeded onto older or surgically damaged Bruch’s membrane in vivo or in vitro, or into a subretinal space with an abnormal basement membrane and an abnormal milieu of soluble factors. 

Despite these limitations, this study sheds some insight on the role of immune suppression in controlling RPE cell survival after subretinal transplantation. Successful repopulation of denuded and/or damaged Bruch’s membrane with donor RPE cells necessitates a complete understanding of the multiple factors that influence graft survival after transplantation and of the basic biology of the RPE and its extracellular environment in health and disease. 

 Supported by funds from Diacrin Inc.

 Submitted for publication November 18, 2002; revised March 13, 2003; accepted March 16, 2003.

 Disclosure: L.V. Del Priore, Diacrin Inc. (F); O. Ishida, Diacrin Inc. (F); E.W. Johnson, Diacrin Inc. (E); Y. Sheng, Diacrin Inc. (F); D.B. Jacoby, Diacrin Inc. (E); L. Geng, Diacrin Inc. (F); T.H. Tezel, Diacrin Inc. (F); H.J. Kaplan, Diacrin Inc. (F)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Four weeks after surgery with immune suppression. First animal: (A) The transplant site was identified by heavy pigment in the subretinal space. (B) Fluorescein angiography revealed blockage of choroidal fluorescence by pigment, with no late staining or leakage from rabbit choroid. (C) There was heavy intracellular pigment at the transplant site (center) with cystic spaces (arrow) at the margin of the transplant bed. The outer and inner retinal lamellae appeared intact over the transplant site. Occasional pigmented cells (arrowhead) were present internal to the internal limiting membrane. (D) After the pigment was bleached, occasional blue PRE-positive cells (inset, arrows) were present within the transplant bed. (E) Many pigmented cells were RAM 11 positive (red patch, arrows) indicating they were macrophages. Second animal: (F) There was a large lumen (⋆) lined with pigmented cells. Cystic spaces (arrows) were present at one end of the transplant bed. (G) Many PRE-positive cells (inset, dark blue cells, arrows) lined the lumen. (H) Pigmented cells that did not surround the lumen were RAM 11 positive (red cells, arrows), indicating they were rabbit macrophages. Note the paucity of RAM 11–positive cells around the lumen.

Figure 2.

 

Four weeks after surgery without immune suppression. (A) Transplant site was identified by pigment in the subretinal space. (B) Fluorescein angiography revealed blockage of choroidal fluorescence by pigment with no late staining or leakage. (C–E) The retina was detached during processing; images reconstructed by joining separated images. (C) There was some subretinal pigment (arrows) and choroidal pigment (arrowheads) at the transplant site. The outer and inner retinal lamellae appeared intact. (D) Occasional PRE-positive cells (dark blue, arrows) were present among the pigmented PRE-negative cells. (E) RAM 11–positive cells were present in the subretinal space (red, arrows).

Figure 3.

 

Twelve weeks after surgery with immune suppression. First animal: (A) Scattered pigment was present at the transplant site. (B) Fluorescein angiography demonstrated blockage from pigment without leakage or staining. (C) A heavily pigmented monolayer was visible, with occasional intraretinal pigment (arrow). (D) Some PRE-positive cells (dark blue nuclei, arrows) were present within the monolayer. (E) A patch of RAM 11–positive cells (red, arrowheads) and PRE-positive cells (dark blue nuclei, arrows) were visible in same section with dual labeling. Second animal: (F) Scattered subretinal intracellular pigment was visible. (G) Solitary PRE-positive cell (dark blue nucleus, arrow) visible. (H) Many subretinal cells were RAM 11 positive (red, arrows).

Figure 4.

 

Twelve weeks after surgery without immune suppression. First animal: (A) The transplant site was identified by heavy pigment in the subretinal space. (B) There was blockage by pigment but no leakage or staining on fluorescein angiography. (C) There were numerous pigmented cells arranged mainly in a monolayer along Bruch’s membrane. The retinotomy (arrow) was visible. (D) Some PRE-positive cells were visible (dark blue nuclei, arrows). (E) RAM 11–positive cells (red, arrows) were visible. Note that some pigmented cells were both PRE and RAM 11 negative (compare C, D, and E). Second animal: (F) A mound of pigmented cells was visible at the transplant site (arrow). (G) Some cells were PRE positive (dark blue nuclei, arrows). (H) There are many RAM 11–positive cells (red, arrows).

Figure 5.

 

Number of PRE-positive cells estimated from serial sections as a function of time after transplantation surgery. (▪) Immune suppression; (□), no immune suppression. Error bars, SD.

Figure 6.

 

Average serum cyclosporine levels at different time points after surgery. Open brackets adjacent to bars represent target cyclosporine levels at different time points after surgery. Numbers refer to number of rabbits in which the cyclosporine level was measured at each time point. Error bars, SD.