A New Model of Retinal Pigment Epithelium Transplantation With Microspheres FREE

SUCCESSFUL retinal pigment epithelium (RPE) transplantation requires that the donor cells retain their polarity and function, avoid formation of clumps or multilayers, and maintain their viability. Cell attachment to a substrate prevents RPE apoptosis and dedifferentiation after transplantation.^{1 - 2} In previous studies,^{3 - 6} subretinal provision of RPE cells was carried out in the form of a cell suspension, RPE patches, or RPE cells grown on artificial substrates. Cell suspension provision has the limitations of reflux from the iatrogenic retinotomy site and irregular distribution of the donor cells in the subretinal space.⁷ Retinal pigment epithelium patch grafts, although probably the most physiologic, have not been shown to proliferate in vivo.^{8 - 9}

Three-dimensional cell culture systems with various attachment substrates offer new possibilities for long-term viability and donor cell functions.^{10 - 18} Successful use of 3-dimensional microcarriers for transplantation to the liver¹⁰ and brain¹¹ has been reported by several investigators. In 3-dimensional microcarriers, the cultured cells are distributed at the outer surfaces and within the body of the particles.¹² In a 3-dimensional carrier, more cell contacts are generated compared with a monolayer state, thereby facilitating cell proliferation and spreading.^{13 - 14} The chemistry of the extracellular matrix itself can also modulate various aspects of cell behavior, including adhesion, proliferation, and migration.¹⁵

In this study, we developed a 3-dimensional culture system for human fetal retinal pigment epithelial (HFRPE) cells in the form of microspheres, and evaluated their growth in the subretinal space after transplantation. Fibrinogen, which is a powerful stimulator of cell attachment and proliferation,¹⁹ was used in the matrix. We used a multicellular spheroid, which is 1 form of a 3-dimensional cell culture system.²⁰

Cross-linked fibrinogen films were prepared under sterile conditions by mixing fibrinogen, 90 mg, and flavin mononucleotide, 1.3 mg (Sigma-Aldrich Corp, St Louis, Mo), in 5 mL of deionized water.²¹ Four drops (≈80 µL) of the mixture were spread evenly on the bottom of a 30-cm Petri dish. The mixture was left under UV light for 12 hours. This allowed the formation of 20- to 50-g thick, yellowish, transparent, slightly sticky films that could easily be separated from the bottom of the dish with fine forceps. The film was cut into smaller 1×1-mm pieces that were used for HFRPE monolayer implantation.

Human fetal eyes at 17 to 22 weeks of gestation were used in this study. The eyes were obtained from the Anatomic Gift Foundation, Laurel, Md, and from the The University of Chicago Hospitals, Chicago, Ill, after therapeutic abortions. The eyes were enucleated and processed under aseptic conditions. They were subsequently dissected circumferentially posterior to the ora serrata. After gentle removal of the vitreous and retina, RPE cells were separated from the choroid by forceps in large monolayer sheets. No digestive enzymes were used during the separation. After separation, the large HFRPE pieces were cut with microscissors into approximately 1×1-mm pieces. The HFRPE monolayer pieces were placed on the surface of the matrix and incubated in high-glucose Dulfecco modified eagle medium supplemented with 15% fetal bovine serum, levoglutamide, and a combination of penicillin G sodium and streptomycin sulfate. For the first 24 hours, the matrix films bearing the cells were kept attached to the bottom of the culture dish. The films were then easily detached from the dish with forceps and kept in a floating state until microspheres formed. The films bearing the cells became rounded and formed oval or round conglomerates, ie, microspheres, covered with HFRPE. Microsphere formation took place 7 to 10 days after attachment of HFRPE pieces to the matrix. Three-week-old microspheres were used for transplantation. The growth pattern and morphologic characteristics of HFRPE microspheres attached to the floor of 8-well chamber slides (NUNC, Naperville, Ill) were studied as in vitro controls.

All procedures conformed to the Association for Research Guidelines in Vision and Ophthalmology on the Use of Animals in Ophthalmic and Vision Research as well as The University of Chicago guidelines for animal experimentation. After a standard 3-port vitrectomy, a localized iatrogenic retinal bleb was created with balanced salt solution^{22 - 23} in 1 eye of 9 albino and 9 pigmented rabbits. A microsphere containing the HFRPE cells was introduced into the vitreous cavity with a blunt micropipette and transferred to the subretinal space. Nine albino and 9 pigmented rabbits were used as controls. The controls underwent transplantation with a bare matrix of comparable size. Eyes that showed any signs of bleeding during the procedure were excluded from the study. The eyes were evaluated at 7, 14, and 30 days after surgery by indirect ophthalmoscopy and fundus photography. The rabbits underwent euthanasia, and the eyes were enucleated at 7, 14, and 30 days after the surgery for histological and immunohistochemical studies. Three albino and 3 pigmented rabbits were euthanatized at each point after the transplantation. Controls were followed up and euthanatized similarly. Paraffin-embedded and cryostat sections were used for these studies.

To identify the donor cells in the pigmented eyes that underwent transplantation, the sections were stained with anti–human monoclonal HLA-ABC antibody specific for human tissue. The antibody showed no cross-reaction with rabbit tissues. Monoclonal anti–human pancytokeratin was used as an epithelial marker. CD5 monoclonal antibody (Sigma-Aldrich Corp), specific for rabbit panlymphocytes, was used to assess the gross immune response.

For HLA-ABC (catalog number M 736; DAKO, Carpinteria, Calif) and CD5 immunostaining, the cryosections were fixed in cold acetone for 10 minutes. Primary antibodies were used in a 1:10 dilution, and the slides were incubated in a moist chamber at room temperature for 1 hour. For monoclonal antipancytokeratin (catalog number C-2562; Sigma-Aldrich Corp) immunostaining, the sections were stained in 10% buffered formaldehyde solution for 10 minutes. The primary antibody was used at a dilution of 1:120 for 2 hours in a moist chamber at 37°C. The secondary antibodies used were sheep anti–mouse immunoglobulin–rhodamine B, 1:10, or immunoglobulin-fluorescein, 1:30 (Rosh Molecular Biochemical, Indianapolis, Ind), for 1 hour at 37°C. After a final washing in distilled water, the specimens were covered with mounting medium and examined under a microscope (model BH-2; Olympus, Osaka, Japan). Photographs were taken with a camera (model 35AD-4; Olympus, Japan). Film (Ektachrome 320T; Kodak, Rochester, NY) was used for all fluorescence pictures. Monoclonal human anti-CD4 (Sigma-Aldrich Corp) was used as a negative control substrate. The in vitro control specimens were stained similarly for anti–human monoclonal HLA-ABC antibody and for monoclonal antipancytokeratin.

Three albino and 3 pigmented rabbits (6 eyes) were studied in each group that underwent transplantation and in each control group at each period. In total, 36 eyes were studied.

The transplanted microsphere (Figure 1, A) appeared as a pigmented subretinal lesion (Figure 1, B and C). On day 7, subretinal hyperpigmentation was noted in 2 rabbit eyes in proximity to the transplanted HFRPE microspheres. None of the eyes showed intraocular inflammation.

Four of 6 eyes showed hyperpigmentation around the transplanted microsphere (Figure 1, D). In those eyes in which hyperpigmentation was present from day 7, an increase in its size with the formation of pseudopodia was noted around the donor tissue source. No intraocular inflammation was seen.

Five of 6 eyes showed hyperpigmentation around the microspheres. No ophthalmoscopic evidence of inflammation or infection was noted at the 30-day follow-up.

In summary, the extent of hyperpigmentation around the microspheres varied among the eyes, ranging from no hyperpigmentation (7 eyes: 3 albino and 4 pigmented eyes) to prominent hyperpigmentation, with some extending as far as 3 to 4 disc diameters away from the initial donor site (Figure 2, A-C). A total of 11 of 18 rabbits that underwent transplantation showed subretinal hyperpigmentation adjacent to the microspheres.

In the control eyes, no subretinal hyperpigmentation was noted. Starting from day 7, prominent whitening due to chorioretinal atrophy, with no change in size with time, was seen in all control pigmented rabbit eyes around the transplanted matrix (Figure 2, D). Similar chorioretinal atrophy was noted in the control albino eyes at the site of the transplanted matrix.

There were no notable differences noted in the inflammatory response at various times after transplantation. For better assessment of the sections, we defined 3 tissue regions: a region over the microsphere, which included only the microsphere with overlying retina and underlying choroid; a region close to the microsphere, which included the area where the microsphere was always seen with adjacent migrated cells; and a farther region, which included sections where only donor cell monolayer was seen.

Light microscopy showed that the areas corresponding to the transplanted microspheres were composed of a circumscribed region of highly pigmented HFRPE cells. The cells were residing in the subretinal space as thick multilayers (Figure 3, A). Loss of photoreceptors was typically noted immediately above and in close proximity to the transplanted tissue. Migration of transplanted HFRPE cells into the overlying neurosensory retina was noted at the site of microsphere implantation. Fragments of the matrix appeared as eosinophilic material between the HFRPE cells. Human fetal retinal pigment epithelial cells were seen in 2 eyes in close proximity to the initial donor tissue at 7 days after the transplantation. At 14 days after the transplantation, the HFRPE cells were identified at farther regions from the microsphere. In both albino and pigmented eyes, the donor cells formed a monolayer in the subretinal space. This corresponded to the hyperpigmentation site that was seen ophthalmoscopically around the microsphere (Figure 3, B-F, and Figure 4, A and C). The cells forming the transplanted microspheres as well as pigmented cells seen at the subretinal space showed positive immunostaining for HLA-ABC monoclonal antibodies (Figure 4, C, and Figure 5, A-D). Neurosensory retina was preserved above the pigmented cell monolayers located at the distant site from the microsphere (Figure 3, C).

The area with the transplanted microsphere was also studied by scanning electron microscopy at 30 days after the transplantation (Figure 6, A and B) in 2 albino eyes. The retina was locally "glued" to the underlying microsphere, and the microsphere appeared flattened. Two layers of RPE could be identified. The cells in the top layer appeared more rounded, with filamentous cell-cell junctions, possibly representing donor tissue, while the bottom layer showed flatter, more polygonal cells, probably the host RPE. The HFRPE cells that grew out from the microsphere appeared rounded and were of different sizes compared with the host RPE cells. They showed monolayer formation around the microsphere, and they formed long filamentous cell-cell junctions. In some areas, the monolayers were not continuous; and in some areas, only 2 or 3 donor cells were seen.

The cellular response to the transplanted tissue was strictly local and was present around the microsphere, mainly in the underlying choroid. Compared with the controls, there was minimal choroidal thickening with mononuclear cell infiltration beneath the microsphere itself. No inflammatory or lymphocytic responses were seen in the areas where the HFRPE cells were distributed as monolayers.

A striking difference was noted in the control eyes in which only bare matrix microspheres were transplanted (Figure 7, A and B). A markedly thickened, infiltrated choroid was evident under the transplanted matrix, with loss of photoreceptors in the overlying retina. Lymphocytes and other mononuclear cells invaded the entire area of the subretinal space and the retina around the matrix. This could represent a mixed nonspecific inflammatory and immune response. This cellular response was local and was confined to the area of the transplanted bare matrix. The reaction was more severe than in the eyes that underwent HFRPE transplantation. Multiple cells with engulfed eosinophilic material, possibly the matrix, were identified in the area. At 30 days after the surgery, some of the control eyes showed chorioretinal atrophy, with no extracellular matrix present. Immunostaining with CD5 monoclonal antibody, which recognizes rabbit panlymphocytes, showed more intense lymphocytic infiltration in control eyes than in the eyes that underwent HFRPE transplantation (Figure 8, A-C).

Subretinal RPE transplantation has shown promise in the rescue of overlying receptors in some experimental degenerative retinal diseases.²⁴ This may be important in the management of various diseases affecting the RPE. Our 3-dimensional culture system offers a new approach for the provision of donor tissue into the subretinal space.

The transfer of the HFRPE cell–containing microspheres into the subretinal space is simple and reproducible. The adjustable size and spherical shape of the donor tissue makes it easy to insert into the subretinal space. Because the HFRPE cells in our model form compact tissue conglomerates, there is less chance for cell reflux.⁷ To the best of our knowledge, in previous studies²⁵ of RPE transplantation, there was no evidence of donor cell proliferation or migration in the subretinal space. Some studies^{26 - 28} suggest that the subretinal space is an immune-privileged environment where cell proliferation is kept under strict control. Transplanted cells may need some kind of initial stimulation to migrate or proliferate actively in this environment. Fibrinogen and the 3-dimensional state of HFRPE tissue could contribute to the ability of the cells in our model to grow out from the initial source. Recent studies done in our laboratory provide indirect evidence of the importance of the modulatory effect of the matrix on cell behavior in the subretinal space. Human fetal retinal pigment epithelial cells grown as microspheres on a synthetic polymer matrix showed notably less potential for subretinal spread compared with cells grown on a fibrinogen matrix.²⁹ Some cell types reexpress their original in vivo characteristics in a 3-dimensional culture and maintain cell-specific functions that are lost in monolayer cultures. The cells in 3-dimensional cultures express increased DNA synthesis and proliferation.^{30 - 31} Donor cells provided to the recipient as 3-dimensional culture systems show prolonged survival and the ability to migrate from the initial source and establish themselves among the host tissues.^{10 - 11} In addition, fibrinogen is a known potent stimulator for cell proliferation and migration.³² Related to it, fibronectin is an important constituent of the Bruch membrane and the surrounding RPE.³³ Human fetal retinal pigment epithelial cells grown on cross-linked fibrinogen matrix in a 3-dimensional state with tight cell-cell contacts may become activated^{34 - 35} and possess the potential for migration and proliferation after being brought into the subretinal space. Although there was notable damage to the overlying retina, it appeared to be only local and restricted to the site of the microsphere. The damage can be comparable in size with a large laser burn. The retina appears preserved, however, above the HFRPE monolayers at more distant sites from the microsphere. The growth of HFRPE cells outside the maternal source of donor tissue, shown in our studies, may provide an opportunity to transplant microspheres in an extrafoveal region with secondary spreading to the subfoveal space.

The ophthalmoscopic observation of subretinal hyperpigmentation around the transplanted microspheres corresponded histologically to a monolayer of pigmented cells in albino rabbits. Immunohistochemically, the transplanted cells were identified by staining for HLA-ABC antibody, and similarly showed migrating HFRPE cells from the initial source with monolayer formation close to the transplanted microsphere.

All control eyes transplanted with bare matrix showed a notably higher inflammatory response and increased lymphocytic infiltration compared with the HFRPE transplanted eyes. Retinal pigment epithelium can modulate the functions and behavior of other cells, such as lymphocytes, vascular endothelial cells, and macrophages.^{36 - 38} Recent studies³⁹ have shown that some tumor cells grown as spheroids show increased resistance to lymphocyte lysis and inhibition of lymphocyte penetration compared with the cells grown as monolayers. Human fetal retinal pigment epithelial cells transplanted as multicellular spheroids, ie, microspheres, may similarly possess lymphocyte inhibitory qualities. Retinal pigment epithelium cells have been shown to release transforming growth factor β family proteins⁴⁰ that have immunosuppressive functions and that can inhibit neovascularization.^{41 - 42} In addition, cell types grown in a 3-dimensional culture system show increased levels of intracellular cytokines, including transforming growth factors.⁴³ Human fetal retinal pigment epithelial cells cover the matrix, possibly preventing its direct contact with the subretinal tissues, resulting in less intense inflammation compared with controls. Some studies²⁴ explain the rescuing effect of RPE transplantation to cytokine release by healthy donor cells. Three-dimensional cultures of HFRPE cells may be a better source of the cumulative release of different trophic cytokines,⁴³ compared with monolayers, due to the high accumulation of healthy cells.

In conclusion, the provision of donor cells into the subretinal space as microspheres is reproducible and tech nically easy, and it may decrease the chances for iatrogenic damage to the retina. The donor cells can spread and survive in the subretinal space for at least 1 month. Subretinal transplantation of HFRPE cells as a 3-dimensional culture system may become a useful model for studying donor and host cell interactions.

Accepted for publication June 8, 1999.

Supported in part by Research to Prevent Blindness Inc, New York, NY, and the Bergman Foundation, Chicago, Ill.

Presented in part at the Chicago Ophthalmological Society Meeting, Chicago, Ill, May 20, 1998, and was awarded the Beem-Fisher prize.

Reprints: Arutun Oganesian, MD, Department of Ophthalmology and Visual Sciences Center, University of Chicago, 939 E 57th St, Chicago, IL 60637 (e-mail: aoganesi@midway.uchicago.edu).