Photoreceptor cell rescue in retinal degeneration (rd) mice by in vivo gene therapy.

Mutations in the beta subunit of the cGMP phosphodiesterase gene (beta PDE) can cause a recessively inherited retinal degeneration in several species, including mice, dogs and humans. We tested the possibility of altering the course of retinal degeneration in the rd mouse through subretinal injection of a recombinant replication-defective adenovirus that contains the murine cDNA for wild-type (beta PDE, Ad.CMV beta PDE. Subretinal injection of Ad.CMV beta PDE results in beta PDE transcripts and increased PDE activity and delays photoreceptor cell death by six weeks. The findings demonstrate cell rescue by in vivo gene transfer, thus supporting the feasibility of treating an inherited retinal degeneration by somatic gene therapy.

Retinal degeneration in mice lacking the gamma subunit of the rod cGMP phosphodiesterase.

The retinal cyclic guanosine 3',5'-monophosphate (cGMP) phosphodiesterase (PDE) is a key regulator of phototransduction in the vertebrate visual system. PDE consists of a catalytic core of alpha and beta subunits associated with two inhibitory gamma subunits. A gene-targeting approach was used to disrupt the mouse PDEgamma gene. This mutation resulted in a rapid retinal degeneration resembling human retinitis pigmentosa. In homozygous mutant mice, reduced rather than increased PDE activity was apparent; the PDEalphabeta dimer was formed but lacked hydrolytic activity. Thus, the inhibitory gamma subunit appears to be necessary for integrity of the photoreceptors and expression of PDE activity in vivo.

Transplantation of cultured rabbit retinal epithelium to rabbit retina using a closed-eye method.

We have developed a closed-eye technique for transplanting cultured rabbit retinal epithelial cells to Bruch's membrane of the rabbit. A glass micropipette containing a suspension of 3H-thymidine-labeled, cultured retinal pigment epithelial (RPE) cells is inserted through a pars plana incision and positioned adjacent to the neural retina. A jet stream from the pipette is used to make a small retinal hole and bleb detachment. Patches of host retinal epithelium lift off with the neural retina, creating areas of bare Bruch's membrane. The cell suspension is injected into the subretinal space, and labeled cells can be seen attached to Bruch's membrane as early as 1 hr later. The neural retina spontaneously reattaches within 24 to 48 hr, bringing photoreceptor outer segments in direct contact with the transplanted cells. Phagocytosis of outer segment material by transplanted cells can be seen as early as 24 hr after surgery. This closed-eye technique offers an advantage over the open-sky method used previously in that it allows for reattachment of the neural retina and at least a partial return of function in the transplanted retinal epithelium.

Retinal pigment epithelial transplants and retinal function in RCS rats.

PURPOSE: To determine if retinal pigment epithelium (RPE) transplantation maintains visual function in Royal College of Surgeons (RCS) strain of rats. METHODS: Twelve RCS rats received RPE transplants at 16 to 20 days after birth. The retinas were studied electrophysiologically and histologically from 3 to 10 months after transplantation and compared with 11 RCS controls and 11 normal rats of comparable ages. A microelectrode was guided to the transplant site visible by its pigmentation in the albinotic RCS retina to detect responses. RESULTS: Spontaneous ganglion cell activity was present in all retinas. Ganglion cell responses to light were detected in 9 of the 12 transplant eyes but not in any of the 11 controls. 96, 44, 140 units were encountered and 30%, 0%, 97% were driven by light respectively in transplant, control, and normal retinas. In transplants 36%, 29%, and 28% were driven at 3 to 4, 6 to 7, and 10 months after transplantation, respectively. Intraretinal ERGs with both a- and b-waves were recorded in 5 of the 8 transplants studied. None of the RCS controls studied had an IERG. The average IERG was 2.5 microV (SD = 1.9) in transplants and 59 microV (SD = 19) in normal retinas. The electrode track was traced to the transplant site in six of the seven retinas that were responsive to light and examined histologically. CONCLUSION: RPE transplants to RCS rats maintain retinal function in the transplant site for long periods of time.

Growth characteristics and ultrastructure of human retinal pigment epithelium in vitro.

The in vitro growth characteristics and morphology of human retinal pigment epithelium (RPE) cells from adult donor eyes (15 to 100 years of age) have been studied. Although RPE cells are viable in culture for several months, only a fraction of the cells actually divide. Rapid cell proliferation and confluency of a culture occur from 10 to 30 days after seeding. The time for confluency is both age- and media-dependent; the number of cells that are potential dividers and contribute to confluency decreases with increasing donor age. Since melanolipofuscin granules do not form in vitro and are diluted by cell division, stationary (nondividing) cells can be distinguished from the dividing cells by the presence of dense clusters of melanolipofuscin granules in the stationary cells. Confluent cultures contain a monolayer of relatively clear polygonal cells with densely pigmented stationary cells scattered throughout. Stationary cells can often represent as much as 95% of the original RPE cell population. Ultrastructurally all cells appear epithelioid, with apical-basal polarity, junctional complexes, and cytoplasmic organization characteristic of RPE cells in vivo. Stationary cells are extremely large, with melanolipofuscin granules clustered around the nucleus. Melanolipofuscin is a unique marker for nondividing cells and may facilitate studies of the age-dependent loss of replicability of human RPE.

Long-term photoreceptor transplants in dystrophic and normal mouse retina.

PURPOSE: To determine the long-term status of transgenic photoreceptors transplanted to the subretinal space of both rd mutant (receptorless) and normal mouse retina. METHODS: Microaggregates of neural retina from transgenic mice containing lacZ-labeled photoreceptors were transplanted to the subretinal space of adult rd mutant and normal mice. The transplant site was examined by light and electron microscopy at monthly intervals up to 9 months after transplantation surgery. RESULTS: Photoreceptors develop and survive well if transplanted with the proper orientation to the retinal pigment epithelium (RPE). The status of the photoreceptors, including outer segments and synaptic terminals, appear normal for at least 9 months after transplantation; they continue to express the lacZ reporter gene. Cones survive as well as rods. Transplants to the normal mouse develop normally, whereas the host photoreceptors displaced from the RPE degenerate. A barrier, formed by Müller cell processes, develops after photoreceptor degeneration in both normal and rd mouse retina and demarcates host from transplant tissue. Areas can be found in which neural processes have penetrated this barrier. There is no evidence of host-graft rejection. CONCLUSION: Transplanted progenitor photoreceptors develop and survive well for long periods of time in either the rd mutant or normal retina if they are properly positioned. In the former, they reconstitute a photoreceptor layer; in the latter, they replace the host photoreceptor layer, which degenerates after being displaced from the RPE. Areas of potential contact between donor and host neurons exist in these transplants.

Reconstruction of degenerate rd mouse retina by transplantation of transgenic photoreceptors.

Photoreceptors from neonatal transgenic mice with normally developing retinas were transplanted to the subretinal spaces of 2-3-month-old rd mutant mice that lack photoreceptors. The transgenic mouse photoreceptors express high levels of the lac Z reporter gene product, beta-galactosidase, which facilitated tracking the transplanted cells. Two sources were used for these cells: (1) dissection of retinal microaggregates containing photoreceptors and (2) papain-dissociated photoreceptors. Host retinas were examined after transplantation. Both methods led to survival of photoreceptors for at least 2 mo after transplantation. Relatively mature outer segments were found only in transplanted microaggregates; this occurred optimally when the cells were adjacent to the retinal pigment epithelium (RPE). beta-galactosidase-labeled outer segments associated closely with the apical processes of the host RPE, which, together with labeled phagosomes in the RPE cells, suggested functional interaction between the transplanted photoreceptors and the host RPE. This study is the first to the authors' knowledge to show electron microscopically that a morphologically normal-appearing photoreceptor layer can be reconstructed in an otherwise photoreceptorless retina.

Proteins from human retinal pigment epithelial cells: evidence that a major protein is actin.

Two-dimensional gel electrophoresis coupled to fluorography was used to obtain the major 35S-methionine labeled protein patterns of cultured human retinal pigment epithelial cells (RPE) and fibroblasts. Comparisons of these profiles showed that: a) primary and subcultures of RPE were quite similar, each showing about 200 different proteins, b) the RPE protein patterns contained at least eight major acidic proteins (from 31 to 96 kilodaltons) not readily seen in fibroblasts; c) the fibroblast pattern showed three well-labeled, very acidic proteins, one of which (58 kilodaltons) appeared to be unique to fibroblasts; and d) a major complex of acidic protein of 43 kilodaltons with isoelectric points of 5.6 to 5.9 was a common protein in RPE and fibroblasts. These latter macromolecules were found to coelectrophorese with purified chicken muscle actin. Ultrastructural studies of cultured RPE supported the presence of actin microfilaments that were demonstrated specifically by labeling with fluorescent phallotoxin.

Patch transplants of human fetal retinal pigment epithelium in rabbit and monkey retina.

PURPOSE: To transplant human fetal retinal pigment epithelium (RPE) into the subretinal space of rabbits and monkeys as an organized monolayer without artificial support. METHODS: The method involves dissecting small patches of cultured RPE monolayers in sheets (1 to 5 mm2), sucking them into a glass pipette and injecting them into the subretinal space after producing a bleb detachment of the neural retina. RESULTS: These patches unfold and survive as a quasi-monolayer under the reattached neural retina intimately associated with the host photoreceptors and phagocytizing host outer segment material. Graft rejection is observed in most rabbits at 1 month but not in monkeys at 3 months after transplantation. CONCLUSIONS: Monolayer patches of cultured human fetal RPE can be transplanted to the subretinal space, where they survive in contiguity with healthy host outer segments. In primates, but not in rabbits, host-graft rejection does not occur for at least 2 to 3 months.

Transplantation of photoreceptors labeled with tritiated thymidine into RCS rats.

Tritiated thymidine was administered to newborn rats to label photoreceptors, about 50% of which are still dividing. These photoreceptors were enzymatically dissociated and separated from the remainder of the retina after the infant rat matured. These labeled photoreceptors were then transplanted into a foreign host retina in the region of the outer nuclear layer. The hosts were ocular, albinotic, Royal College of Surgeons (RCS) rats, congenic to the normal donors and at least 4 months old, a time when virtually all the photoreceptors have degenerated from their retinas. The transplant site was examined at various times after transplantation by light microscope autoradiography. Labeled photoreceptor cell bodies were found in clusters in the outer nuclear layer region for as long as 3 months after transplantation surgery.

Transplanted photoreceptors identified in dystrophic mouse retina by a transgenic reporter gene.

Dissociated photoreceptor cells from a transgenic strain of mice, containing a bovine promoter lac Z gene construct, were transplanted to the dystrophic retinas of a strain of mice (C3H, rd/rd) without obvious photoreceptors. The transgenic photoreceptor cells expressed beta-galactosidase and were distinguishable from the host retinal cells by light and electron microscopy after the 5-bromo-4-chloro-3-indolyl-beta-D-galactopyranoside histochemical reaction. These results showed that transplanted transgenic photoreceptor cells survived at least 1 month in the host retina and had a primitive outer segment and a well-developed synaptic terminal.

Transplanted retinal pigment epithelium modifies the retinal degeneration in the RCS rat.

Transplantation of dissociated retinal epithelial cells obtained from the retinas of normal, congenic pigmented strain of rats to Bruch's membrane and the subretinal space of dystrophic rats from the Royal College of Surgeon (RCS) strain can prevent photoreceptor cell degeneration in this retina for at least 4 months after transplantation. Host and transplant cells form close apposition with one another but can be distinguished by the presence of both phagosomes and melanin granules in the transplant and the absence of these inclusions in the host retinal epithelium. Transplanted cells show excessive amounts of phagosomal material within 48 hr after transplantation, implying that restoration of phagocytosis is responsible for the photoreceptor survival.

Tracking RPE transplants labeled by retroviral gene transfer with green fluorescent protein.

PURPOSE: To determine whether human retinal pigment epithelium (RPE) can be modified by retroviral-mediated gene transfer and to monitor the human RPE cells in the subretinal space of living rabbits with scanning laser ophthalmoscopy (SLO). METHODS: Cultured human fetal retinal pigment epithelium (HFRPE) was exposed to green fluorescent protein (GFP)-transducing retroviral vectors, Moloney murine leukemia virus, and lentivirus. The cultured cells were followed by fluorescence microscopy. Suspensions of GFP-expressing HFRPE were transplanted into the subretinal space of pigmented rabbits, and the transplant sites were examined by SLO for fluorescence, including fluorescein and indocyanine green angiography. The rabbits were euthanatized at different times after transplantation, and the retinas were studied histologically. RESULTS: Retroviral gene transfer can introduce a foreign gene such as GFP into cultured HFRPE. Gene expression is maintained in cultured RPE for at least 3 months. The lentiviral vector transduced both nondividing and dividing cells; the Moloney vector only transduced the latter. GFP-expressing cells can be followed in the living retina. Their changes reflect the rejection response followed histologically. CONCLUSIONS: Cultured HFRPE could be transduced to express GFP for long periods of time by retroviral gene transfer. GFP allowed retinal transplants and gene expression to be monitored in vivo. These results provide a model for potential ex vivo gene therapy in the subretinal space.

Retarding photoreceptor degeneration in Pdegtm1/Pdegtml mice by an apoptosis suppressor gene.

PURPOSE: Mice (Pdegtm1/Pdegtm1) homozygous for a mutant allele of the gamma subunit of retinal cyclic guanosine monophosphate phosphodiesterase (PDE gamma) suffer a severe photoreceptor degeneration. To determine whether the antiapoptotic BCL2 gene is effective in delaying the cell death pathway in this new strain of mutant mice, a transgene encoding the BCL2 gene product was introduced by mating into the mutant background, and the resulting mice were examined for possible rescue of the retinal degeneration. METHODS: Electroretinograms (ERGs) of the Pdegtm1/Pdegtm1 mice carrying BCL2 were taken to monitor the responses to light. Light and electron microscopy of sections were used to examine degeneration at different times after birth. RESULTS: The ERGs of the mutants with the transgene were larger than those without the transgene at 2 and 3 weeks after birth. The maximum differences occurred at 2 weeks postpartum. At 4 weeks after birth, no ERG could be detected in either strain. Histologic analysis showed a greater preservation of photoreceptor nuclei in the Pdegtm1/Pdegtm1 mice containing the BCL2 transgene, which paralleled the electroretinography. CONCLUSIONS: The introduction of an antiapoptotic transgene BCL2 can delay temporarily and partially the degeneration of photoreceptors in a new autosomal-recessive murine model of retinal degeneration.

The effects of bFGF on RCS rat eyes.

Basic fibroblast growth factor (bFGF) has been implicated as a factor in retinal differentiation and disease. Recent studies have shown that subretinal or intravitreal injections of bFGF delay the retinal degeneration of the RCS rat but the global nature of this effect has been quantified for few test animals and the mechanism underlying this effect is not understood. In order to determine more accurately the global effects of intravitreal bFGF and to further elucidate the mechanism of bFGF promoted photoreceptor cell saving, we injected one of three bFGF doses into the vitreal cavities of young RCS rats. Using measurements from several eyes, we confirmed that a single intravitreal bFGF injection globally delays the RCS dystrophy. Test eyes contained fewer debris zone macrophages and more inner retinal macrophages than did control eyes at 1 month post injection. As bFGF's saving effect waned, the number of inner retinal macrophages decreased and the number of debris zone macrophages increased toward control levels. Dose-dependent cataract formation occurred in 100% of test eyes. Eyes that received the highest bFGF dose showed increased retinal vascularization at 1, 2 and 3 months post injection. The possible relationships between bFGF promoted photoreceptor survival and our findings are discussed.

Isolation of human fetal cones.

PURPOSE: To describe a method for isolating a monolayer of human fetal cone photoreceptors and to compare their structure and ultrastructure before and after preparation. METHODS: Eyes from human fetuses (fetal week 20 to 24) were dissected and the neural retina of the developing fovea identified, cut out and placed on 10% gelatin. A VISX Star excimer laser was used to remove the inner retinal layers. The isolated cone monolayers were cultured for 18 hours and compared with untreated retinas by light microscopy and transmission electron microscopy. RESULTS: Excimer laser ablation removed the inner nuclear and ganglion cell layers leaving a monolayer of fetal cones. These cones survive in culture for at least 18 hours. The laser ablation disorganized the ultrastructure of the synaptic pedicles of these cones, left their plasma membranes intact. CONCLUSIONS: The developing central retina of human fetal eyes provides a source of fetal cones, which can be isolated from inner retinal cells using the excimer laser. Such a monolayer of human fetal cones may be useful for transplantation or biochemical studies.

Reporter gene expression in cones in transgenic mice carrying bovine rhodopsin promoter/lacZ transgenes.

Rhodopsin gene expression has been used as a model system to study the mechanisms regulating photoreceptor gene expression. Previous transgenic experiments using rhodopsin promoter/lacZ fusion constructs identified some of the cis-acting DNA elements responsible for photoreceptor cell-specific expression. However, the issue of rod specificity vs. photoreceptor (rod and cone) specificity of the elements was not resolved. To address this issue, the specificity of reporter gene expression in the retinas of transgenic mice carrying bovine rhodopsin promoter/lacZ (beta-galactosidase) fusion genes was assessed using X-gal staining and electron microscopy. Two independent transgenic lines, one carrying a rhodopsin promoter fragment extending from -2174 to +70 base pairs (bp) relative to the messenger RNA start site and another line carrying a fragment from -222 to +70 bp, both showed reporter gene expression in cones as well as rods, although the level of staining appeared to be less in the cones than in the rods. These results demonstrate that the -2174 to +70 bp and -222 to +70 bp bovine rhodopsin promoter fragments are not rod-specific in transgenic mice and indicate that the existence of rod promoter mediated-expression in cones must be considered when interpreting results from transgenic experiments utilizing the rhodopsin promoter.

The ultrastructure of transplanted rabbit retinal epithelium.

Retinal epithelial cells from pigmented rabbits have been cultured and radiolabelled during division with 3H-thymidine and transplanted to the subretinal space of albino rabbits using a pars plana transvitreal approach. The host rabbits were maintained on cyclosporine immunosuppression after transplantation surgery. The transplant cells survive and maintain a morphologically normal appearance for at least 5 months. The pigmented epithelial cells retain their pigmentation and become integrated with the host retinal epithelium forming junctional complexes with both the host and other transplant cells. The transplant cells contact host photoreceptors with their apical processes and phagocytize outer segments. There appears to be no significant cell division among the transplant cells.

Muller cells of chicken retina synthesize 11-cis-retinol.

The amounts of endogenous retinyl palmitate, retinol and retinaldehyde were measured in the neural retina and retinal pigment epithelium (RPE) of predominantly cone (chicken), rod (rat) and more mixed (cat, human) retinae. The ratio of 11-cis to all-trans isomers of retinyl palmitate and retinol in the neural retina and the RPE increases progressively with the increase in diurnality of the species from rat to chicken. The membrane fractions of both chicken and bovine RPE enzymically isomerize all-trans retinol to 11-cis-retinol. Chicken neural retina membranes enzymically form 11-cis-retinol and all-trans-retinyl palmitate from all-trans-retinol. Light and electron microscopy revealed no contamination of chicken neural retina by RPE. Muller cells from chicken retina were isolated, cultured and characterized by immunocytochemical localization of cellular retinaldehyde-binding protein. Cultured chicken Muller cells form all-trans-retinyl palmitate, 11-cis-retinol and 11-cis-retinyl palmitate from all-trans-retinol and release most of the 11-cis-retinol into the medium. The results indicate that chicken neural retina and Muller cells in particular synthesize 11-cis-retinoids from all-trans-retinol.