Transplantation of Schwann Cell Line Clones Secreting GDNF or BDNF into the Retinas of Dystrophic Royal College of Surgeons Rats

Abstract

purpose. To assess the capacity of a retrovirus-engineered Schwann cell line (SCTM41), transfected with either a glial cell line–derived neurotrophic factor (GDNF) construct or a brain-derived neurotrophic factor (BDNF) construct, to sustain visual function in the dystrophic Royal College of Surgeons (RCS) rat.

methods. Cell suspensions were injected into the subretinal space of the right eye of 3-week-old dystrophic RCS rats through a transscleral approach. The left eye remained as an unoperated control. Sham-surgery animals received injections of carrier medium plus DNase to the right eye. All animals were placed on oral cyclosporine. At 8, 12, 16, and 20 weeks of age, animals were placed in a head-tracking apparatus and screened for their ability to track square-wave gratings at various spatial frequencies (0.125, 0.25, and 0.5 cyc/deg). At the end of the experiment, the animals were perfused and processed for histologic assessment of photoreceptor survival.

results. Animals with SCTM41-GDNF–secreting cells, on average, head tracked longer than animals with SCTM41-BDNF–secreting cells, and both performed better than those injected with the parent SCTM41 line. All tracked longer than sham-surgery or nonsurgical dystrophic eyes. Each cell type demonstrated preservation of photoreceptors up to at least 4 months of age, over and above the sham-surgery control.

conclusions. Engineered Schwann cells sustain retinal structure and function in the dystrophic RCS rat. Cells overexpressing GDNF or BDNF had a greater effect on photoreceptor survival than the parent line or sham surgery. This study demonstrates that ex vivo gene therapy and subsequent cell transplantation can be effective in preserving photoreceptors from the cell death that normally accompanies retinal degeneration.

Retinal degenerative diseases, such as age-related macular degeneration and retinitis pigmentosa, are the leading causes of blindness in the Western world. Currently, no therapy exists to cure these conditions, although several experimental approaches have been evaluated to promote photoreceptor survival. For instance, growth factor injections (e.g., bFGF, ciliary neurotrophic factor [CNTF], glial cell–derived neurotrophic factor [GDNF]) into the eyes of various rodent models, including light damage, ^{1 2 3} the RCS rat, ^{4 5} the retinal degeneration (rd) mouse, ^{6 7} the retinal degeneration slow (rds) mouse, ⁷ and transgenic rats ⁷ have been shown to slow the loss of photoreceptors. Usually, these factors are administered by single or multiple injections into the vitreous cavity. Because injected growth factors tend to have short-lived effects and multiple injections are undesirable, other methods of factor delivery have been explored for the long-term maintenance of photoreceptors. 

Viral-mediated gene therapy has been used to deliver growth factor transfects to the degenerating retina to prolong photoreceptor survival. As with the growth factor injection, investigators in some of the early adenovirus transfer studies in animals reported transient rescue, but researchers in more recent studies using recombinant adenoassociated virus (rAAV) carrying various growth factors have had greater success. For example, rAAV-GDNF has been injected subretinally into the TgN S334ter-4 transgenic rat (expressing a mutated rhodopsin gene), prolonging rod photoreceptor survival, and sustaining ERG a- and b-wave amplitudes until at least postnatal day (P)60. ⁸ Liang et al. ⁹ have described photoreceptor rescue up to 6 months after rAAV-mediated delivery of CNTF to P23H and S334ter rhodopsin transgenic rats and up to 8.5 to 9 months in Prph2 ^Rd2/Rd2 mice, although function was not similarly preserved. 

Delivery of more than one growth factor has so far not been possible using gene transfer because of the size of the constructs. Combinations of different growth factors, however, have been shown to be more efficacious than single factors alone. Thus, CNTF+brain-derived neurotrophic factor (BDNF) or CNTF+GDNF rescued rd mouse photoreceptors in organ culture, each factor presented singly, had little or no effect. ^{10 11} An alternative approach is to inject a donor cell population capable of producing a number of retinally active growth factors. 

This led us to explore the potential of Schwann cells (already shown to support repair after injury in the central nervous system), because they produce several factors including CNTF, BDNF, GDNF, and bFGF that are known to promote photoreceptor survival. Using the dystrophic RCS rat, which loses photoreceptors because of a failure of retinal pigment epithelial cells to phagocytose shed outer segments, we found ¹² that primary cultures of neonatal Schwann cells grafted into the subretinal space could delay photoreceptor cell loss for a significant period and maintain visual function. 

In the present work, we extended these studies to examine the potential of Schwann cell lines that have been genetically engineered to increase the production of specific growth factors. To this end, we used a retrovirus-engineered Schwann cell line (SCTM41), either without further manipulation or transfected with genes to express either GDNF or BDNF. The SCTM41-GDNF cell line has been shown to enhance survival of embryonic nigral neurons when transplanted into the rat striatum and substantia nigra. ¹³  

Methods

Generation of the SCTM41 Schwann Cell Line

The method has been reported in Wilby et al. ¹³ In brief, Schwann cells were derived from the sciatic nerves of neonatal CD rat pups (postnatal day 2) digested in a collagenase-trypsin mixture followed by treatment with cytosine arabinoside and complement-mediated lysis (using rabbit serum and IgM anti-Thy1-1) to reduce fibroblast contamination. 

The SCTM41 cell line is based on the constitutive expression of a synthetic ligand-gated proto-oncogene, tamoxifen-regulated human c-Myc. The retrovirus pBpuro MycMERG525R, which carries both a puromycin-resistance gene and the DNA encoding the c-myc/G525R (a mutant murine estrogen receptor) ^{14 15} fusion protein was used for the transformation process. Supernatant from producer cell lines (generous gift from Trevor Littlewood, Imperial Cancer Research Fund Laboratories, London, UK) was used as the source of retrovirus. Purified neonatal Schwann cells were cultured in medium containing the retrovirus and Polybrene (Sigma-Aldrich, Poole, UK). After 4 hours, this was replaced by DMEM (Life Technologies, Paisley, Scotland, UK) plus 10% fetal calf serum. After 3 days, puromycin and 4-hydroxytamoxifen (Semat Technical UK Ltd., St. Albans, UK) was added to the culture medium to select for virus-infected cells. Puromycin-resistant cells were replated in a 96-well plate and maintained in tamoxifen-containing medium. Colonies were observed, and the SCTM41 line was selected on the basis of morphology, expression of Schwann cell markers (S100, GAP-43, and laminin positive, only very weakly LNGFr positive, but negative for Thy1-1, a fibroblast marker) and promotion of axon growth. 

Transfection with the Rat GDNF and BDNF Constructs

The SCTM41 line expressing GDNF was produced as described in Wilby et al. ¹³ Briefly, the cDNA encoding rat GDNF was cut out of a vector (pBluescript; Invitrogen Corp., Paisley, Scotland, UK) using the restriction enzymes KpnI and XbaI and inserted into the equivalent site of the polylinker in a mammalian expression vector (pcDNA 3.1; Invitrogen Corp.) using KpnI and XbaI. This vector drives expression of the cloned DNA from the cytomegalovirus (CMV) promoter, which gives high-level expression in a wide variety of cell lines. Cells were transfected by incubating the plasmid mixed with transfection reagent (Lipofectamine; Invitrogen-Gibco, Paisley, Scotland, UK) and selected with G418 (800 μg/mL) 48 hours after transfection. Clones were analyzed for GDNF secretion with an ELISA assay (Promega-Biotek, Southampton, UK). 

Rat cDNA for BDNF was cloned into the NotI site of a modified form of the vector pRC CMV (Invitrogen), where neomycin-resistant selection had been replaced with that for hygromycin, using a vector selection cassette (Selectavector; R&D Systems, Abingdon, UK). The recombinant plasmids were identified by restriction digest and confirmed by sequencing. 

The DNA was introduced into the SCTM41 cells with transfection reagent (Lipofectamine; Invitrogen-Gibco). After 48 hours, the cells were split and selected with hygromycin at 400 μg/mL. Individual hygromycin-resistant clones were screened for expression of BDNF by Western blot using known amounts of recombinant BDNF as a standard. The BDNF antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA) was used at a concentration of 1:5000. Each sample was 20 μL after being concentrated 10 times. BDNF clones were additionally analyzed by ELISA (Promega-Biotek). 

Donor Cell Preparation

SCTM41 (the parent cell line), SCTM41-GDNF (clone 5), and SCTM41-BDNF (clone 6*) cells were grown on polylysine-coated flasks in DMEM supplemented with 10% fetal calf serum, glutamine, pyruvate, and tamoxifen (DMEMF). 

Before transplantation, cultures were rinsed with PBS and removed from the flasks using trypsin-EDTA. The reaction was stopped using DMEMF and the cells were centrifuged. After centrifugation, cells were resuspended in DMEM, counted, and diluted to the appropriate cell density (approximately 3.7 × 10⁶cells/100 μL). DNase 1, type IV (Sigma-Aldrich) was added to the final suspension to reduce cell aggregation. Sham injections comprised DMEM plus DNase. 

Transplantation Procedure

All animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, the Home Office (United Kingdom) Regulations for the Care and Use of Laboratory Animals, and the UK Animals (Scientific Procedures) Act (1986). 

Because donor cells were derived from CD rats (Charles River Laboratories, Margate, UK), host animals (including those that underwent sham surgery or were nonsurgical congenic and nondystrophic animals) were placed on ad libitum food and water containing cyclosporin A (CyA, 210 mg/L; Sandoz, Camberley, UK) 2 days before surgery and maintained on this for the duration of the experiment. Forty-two pigmented dystrophic RCS rats (postnatal days 23–25) received subretinal cell grafts in carrier medium plus DNase, or sham surgery (carrier medium plus DNase) on the right eye: GDNF-grafted animals (n = 11), BDNF-grafted animals (n = 9), SCTM41-grafted animals (n = 14), and sham-surgery animals (n = 8). The left eyes were used as the nonsurgical control. 

For the transplantation procedure, animals were anesthetized with intraperitoneal injections of ketamine (80 mg/kg) and xylazine (20 mg/kg), and the pupil was dilated using tropicamide (1% Mydriacyl; Alcon Laboratories, Hemel Hempstead, UK). Surgery was visualized by operating microscope (Leitz, Wetzlar, Germany), and cells or carrier medium (2 μL) were injected transsclerally into the subretinal space (dorsotemporal retina, right eye) by means of a fine glass capillary (inner diameter, 75–150 μm) attached by tubing to a 10-μL syringe (Hamilton, Reno, NV). Immediately after surgery, the size of the retinal detachment induced by the injection and the accuracy of graft placement was assessed visually by operating microscope. 

Behavioral Assessment

Transplant efficacy was tested using head-tracking behavior, a procedure based on an optokinetic test devised by Cowey and Franzini. ¹⁶ At 8, 12, 16, and 20 weeks of age, animals were placed individually into a stationary enclosed Perspex container surrounded by a motorized drum which revolved at a constant velocity (12 deg/s). Interchangeable panels (mean luminosity, 240 lux) of vertical black and white stripes with spatial frequencies of 0.125, 0.25, and 0.5 cyc/deg were placed on the wall of the outer drum. Each was presented, in random order, to the animal over three consecutive days. A test period consisted of four 1-minute sessions of rotation interspersed with 30-second rest intervals. The direction of rotation was alternated so that at the end of the test period they had experienced two 1-minute clockwise rotations and two 1-minute counterclockwise rotations. Presentation of the moving vertical stripes stimulated a turning reflex, so that a seeing animal involuntarily moved its head, tracking the lines. Animals were timed only for periods during which head-turning corresponded with the speed of rotation of the stripes. Habitual and other randomized movements were excluded. The Perspex container was cleaned between animals. At each time point, testing was performed by the same masked observer. To eliminate observer bias, some experiments were videotaped to permit independent assessment of head-tracking times. 

Tracking behavior was measured in animals grafted with SCTM41 cells, SCTM41-GDNF cells, or SCTM 41-BDNF cells and in sham-surgery animals. In addition, five control congenic, nonsurgical animals (receiving ad libitum oral CyA) were head tracked. 

All behavioral data at each time point (8, 12, 16, and 20 weeks of age) were subjected to a three-way analysis of variance (ANOVA) with one repeated measure (group × eye × grating frequency-repeated). A post hoc Student-Newman-Keuls analysis was applied to any significant factors to determine the source of the significance. Statistical analyses were performed on computer (Macintosh; Apple Computer, Cupertino, CA, running Statview; SAS Institute, Inc., Cary, NC). 

Anatomy

For anatomic assessment, two animals from each group were processed for anatomy at 8 weeks of age to provide comparative baseline data, but most were processed at 16 to 20 weeks. Animals were anesthetized with pentobarbitone sodium (Euthatal; Rhone Merieux, Harlow, UK; 200 mg/mL), perfused intracardially with PBS followed by 2.5% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M phosphate buffer plus 0.01% picric acid. A small volume of the same fixative was injected into the front of the eye with a 30-gauge needle. An additional puncture with the same needle provided a drainage hole to relieve any resultant increase in pressure. A suture was placed ventrally to permit orientation of the eye. The eyes were excised and left in the same fixative overnight. After washing in 0.1 M sodium cacodylate buffer, the lenses were removed, and the retina was postfixed in 1% osmium tetroxide in buffer, dehydrated through graded alcohols and epoxy propane to agar resin (Agar Scientific, Stansted, UK). 

Tissue Analysis

Semithin sections were cut through the region of the transplant and stained with toluidine blue (Sigma-Aldrich). A temporal-to-nasal orientation was maintained throughout by placing a suture in the ventral corneoscleral junction as a marker. Photoreceptors were counted in representative sections (adjacent to or within 200 μm of the injection site) from five retinas in each group in approximately 100-μm bins across the retina. Mean cell counts for each bin location in each experimental group across the retinas (temporal to nasal) were generated and compared. 

In addition, some ultrathin sections were taken from each group, stained with alcoholic uranyl acetate and lead citrate, and viewed with an electron microscope (model 1010; JEOL, Tokyo, Japan). 

Results

Cell Types In Vitro

Schwann cells in primary culture typically have a spindle-shaped morphology and divide very slowly, if at all. The SCTM41 parent cell line and the SCTM41-GDNF and SCTM41-BDNF cells had more extensive cytoplasm and they divided rapidly in vitro. However, they continued to express Schwann cell markers. ¹³  

GDNF and BDNF Secretion by SCTM41 and Transfected Clones

Medium was collected from 25-cm² flasks while the cells were actively dividing. Four milliliters of medium was added to each flask for 48 hours and then removed. GDNF secretion was measured with an ELISA (Promega-Biotek), courtesy of Philippe Horellou (see Ref. ¹³ ) and was found to be approximately 92 ng/10⁶ cells per day. SCTM41 did not secrete detectable levels of GDNF. 

For BDNF, clone 6* was analyzed by Western blots (Fig. 1) and ELISA. The production rate of BDNF was calculated to be approximately 336 pg/10⁶ cells per day. The parent cell line produced only trace amounts of BDNF. 

Behavioral Assay for Visual Function

Head-tracking times for all groups over the total period examined are presented in Figure 2 . 

Because head-tracking to a moving stimulus is initiated by an image moving from the temporal to nasal retina, tracking in the clockwise direction measured visual function in the untreated left eye, and tracking in the counterclockwise direction measured function in the treated/transplant-recipient right eye. Hence, in Figure 2 the nonsurgical eye results derive from the times the animal spent following the clockwise rotation. 

At 8 weeks of age, the sighted, nondystrophic group was able to track all the gratings (Fig. 2) . A higher level of tracking was observed in the right eye than in the left, but the difference was nonsignificant. Performance in the sham-surgery group was similar to that in the nondystrophic group at the coarse (0.125 cyc/deg) and medium (0.25 cyc/deg) spatial frequencies, but a decline in head-tracking behavior was seen at the finer grating of 0.5 cyc/deg (Fig. 2) . Furthermore, there were no significant differences in performance between sham-surgery and untreated eyes. An almost identical pattern of results was observed in the BDNF and SCTM41 groups compared with the sham group. Both the BDNF and SCTM41 groups were able to track the coarse and medium gratings, but again a reduction in tracking was observed for the finer 0.5-cyc/deg grating (Fig. 2) . A similar profile was observed in the untreated eyes in the GDNF group. However, a significant difference was detected between the transplant-recipient and untreated eyes in the GDNF group that was largely due to a substantial amount of tracking in the transplant-recipient versus untreated eye at the finer 0.5-cyc/deg grating (F = 9.01; df = 1.18; P < 0.01). 

At 12 weeks of age (Fig. 2) , the nondystrophic rats were tracking at similar levels in both right and left eyes. In the sham group, a further decline in tracking was observed for all gratings in both the sham-surgery and untreated eyes. No tracking was detected in the sham group with the 0.5-cyc/deg grating. In the experimental groups (BDNF, GDNF, and SCTM41), a greater amount of head-tracking was observed in the transplant-recipient eye compared with the untreated eye (eye: F = 36.43, df = 3.74, P < 0.0001; group × eye: F = 8.51, df = 3.74, P < 0.001). At the most discriminating frequency (0.5 cyc/deg), only the eyes with the GDNF and the BDNF cell transplants showed tracking responses. The GDNF group performed significantly better than all the other groups, with better tracking times at all grating widths (group: F = 10.07, df = 3.74, P < 0.0001; post hoc Student-Newman-Keuls test). In addition, both the GDNF and BDNF transplant-recipient eyes tracked significantly more than the sham-surgery eyes. 

The nondystrophic group at 16 weeks of age (Fig. 2) maintained a similar pattern of tracking across the spatial frequencies. However, there was a continuing and equal decline in head-tracking in both treated and untreated eyes in the sham-surgery group. A similar decline in performance was noted in untreated eyes in the experimental groups. In both the BDNF group and the GDNF group, the transplant-recipient eyes performed significantly better than the untreated control eyes (eye: F = 46.15, df = 3.74, P < 0.001; group × eye: F = 12.11, df = 3.74, P < 0.0001). While a difference was evident graphically between transplant-recipient and untreated eyes in the SCTM41 group, it was not significant. Again, at 16 weeks, only the BDNF- and GDNF-treated eyes tracked at 0.5 cyc/deg. Both the BDNF- and GDNF-treated eyes performed significantly better than the sham-surgery eyes. 

Lastly, at 20 weeks of age (Fig. 2) , head-tracking behavior in the nondystrophic group was similar to that at previous time points. Head-tracking performance in the sham-surgery group, however, had declined to zero for all spatial frequencies in both sham-surgery and untreated eyes. Furthermore, in the experimental groups (BDNF, GDNF, and SCTM41), no head-tracking was detected in the untreated eyes. Once again, however, transplant-recipient eyes in the BDNF and GDNF groups were able to track coarse (0.125 cyc/deg) and medium (0.25 cyc/deg) spatial frequencies (group × eye: F = 7.34, df = 3.74, P < 0.001). By 20 weeks of age in the SCTM41 group, because of the continued decline in the performance in the untreated eyes and the sustained performance in the transplant-recipient eyes, the transplant-recipient eyes tracked significantly better than the untreated control eyes (eye: F = 30.11, df = 3.74, P < 0.001; group × eye: F = 7.34, df = 3.74, P < 0.001). Head-tracking at 0.5 cyc/deg had deteriorated to near zero in all transplant-recipient eyes, although BDNF-, GDNF-, and SCTM41-grafted eyes still performed significantly better than sham-surgery eyes at the coarse (0.125 cyc/deg) and medium (0.25 cyc/deg) spatial frequencies (group × eye: F = 7.34, df = 3.74, P < 0.001; post hoc Student-Newman-Keuls test). 

It should be noted that although head-tracking times were measured by a single observer to eliminate observer differences, observer bias can be discounted because ANOVA showed that there was no significant difference between head-tracking times scored by two independent assessors from videotaped experiments. 

Anatomic Assessment

All the photomicrographs (Figs. 3 4 and 5) were taken from the dorsal retina. 

At 8 weeks of age the nonsurgical dystrophic eye had a somewhat disorganized outer nuclear layer (ONL) between two and six cells deep, and there was a considerable zone of amorphous photoreceptor outer segment debris between the RPE and the ONL (Fig. 3A) . Inner and outer segments were absent. Pyknotic cells were evident in the ONL. A sham-surgery retina at 8 weeks is shown in Figure 3B . The ONL was up to six cells deep in the best-preserved area of retina, and disorganized photoreceptor debris was present between the RPE and the photoreceptors. Occasional pyknotic cells were evident. Transplant-recipient retinas showed greater preservation of the ONL, approximately four to eight cells deep (Figs. 3C 3D) , but inner and outer segments were seen only in those retinas grafted with GDNF- and BDNF-secreting cells (e.g., Fig. 3D ). The GDNF-grafted eye illustrated in Figure 3D is from the same animal whose nonsurgical eye is shown in Figure 3A . Retinas grafted with the parent Schwann cell line had an intermediate appearance, with good preservation of the ONL and inner segments but lacking outer segments (Fig. 3C) . They also had a substantial debris zone. 

By 16 weeks of age, only isolated photoreceptors remained in sham-surgery (Fig. 4A) and nonsurgical retinas, in accordance with our previous studies (e.g., Ref. ¹⁷ ). However, retinas grafted with BDNF- or GDNF-secreting cells still retained several layers (two to five) of preserved photoreceptors (Figs. 4C 4D) . In any one section of the dorsal retina, up to one half of the ONL showed some degree of preservation. Inner and outer segments were also preserved but only in limited regions (Figs. 5B 5C) . Retinas with SCTM41 grafts showed discontinuous regions of preservation, approximately two to three cells deep, with inner but no outer segments (Fig. 4B) . The retinas illustrated in Figures 4 and 5 were from animals showing the best head-tracking performance in each category at this age. 

Around the region of the graft, there was some blood vessel formation in the outer retina adjacent to the RPE (Fig. 4C) . This could have been induced by the graft but is more likely to be a manifestation of the usual vascular changes that occur in the dystrophic RCS retina with time. ¹⁸ There was no evidence of excessive donor cell division (in accord with Wilby et al., ¹³ ) nor was there any indication of Müller cell gliosis. 

Microglia were visible in the debris zone that accumulated between the retinal pigment epithelium and photoreceptors. At the light microscope level they were identified as uniformly small cells, but ultrastructurally they had the typical appearance of microglia (Fig. 5A) . The number of microglia appeared to be similar to those observed at earlier time points in nonsurgical dystrophic RCS rats. In a few transplant-recipient retinas, there was evidence of macrophages around the graft. 

Mean photoreceptor cell counts for each group at 20 weeks are shown in Figures 6A 6B 6C . Nonsurgical and sham-surgery animals had very few remaining photoreceptors, and most of those were pyknotic. More photoreceptors were rescued in the three groups of animals that received cell grafts than in the nonsurgical and sham-surgery retinas. All had a significant region of increased rescue in the temporal retina compared with sham-surgery or nonsurgical eyes (F = 5.37, df = 4.20, P < 0.01), associated with the site of transplantation. A significantly wider area of rescue was observed in the GDNF-transplanted eyes, approximately 5.1 mm, compared with approximately 1.7 to 1.9 mm in BDNF- or SCTM41-grafted eyes, approximately 1.7 to 1.9 mm (F = 66.40, df = 5.448, P < 0.001; post-hoc Student-Newman-Keuls test). 

Discussion

The present results show that subretinal grafts of Schwann cell lines both preserve photoreceptors from degeneration and slow the deterioration of visual function in the RCS rat. The SCTM41-GDNF Schwann cell line produced more sustained preservation of photoreceptors and function than SCTM41-BDNF grafts. It is not known why BDNF and GDNF produced different degrees of rescue. It may simply be that the BDNF level achieved by the present number of cells was less than optimal. Alternatively, the two growth factors may activate different pathways. It is clear, however, that the cell lines engineered to secrete GDNF and BDNF showed better functional and anatomic rescue than the parent cell line, adding support to the concept that growth factor delivery is an important component. The parent cell line, by being homologous, provided an appropriate control not often available in experiments of this kind. 

Both direct injection of GDNF and adenovirus delivery of GDNF to the eye have been shown to slow photoreceptor loss in various rodent models of retinal degeneration (e.g., Refs. ^{6 8 19} ), and GDNF receptors have been detected on photoreceptors ^{20 21} and on Müller glia. ²¹ In vitro studies have confirmed that GDNF can promote photoreceptor survival, ^{20 22} reduce apoptosis, and augment opsin expression, ²² suggesting a role as survival and differentiation factors. However, low concentrations of GDNF appear not to exert a rescue effect, ²³ and endogenous levels of GDNF fail to prevent photoreceptor loss in the rd mouse. ⁶ In accord with this, we found in pilot experiments preliminary to the present sequence of studies, that SCTM41-GDNF-secreting cells injected at the lower level of 2.2 × 10⁵cells/100 μL failed to give better head-tracking than the parent SCTM41 line injected at the same concentration, suggesting a dose effect (Lawrence JM, unpublished results, 2001). 

BDNF injected into the eye has also been shown to reduce photoreceptor loss after injection in vivo ^{2 7} ; however, in isolated cell culture ²³ BDNF had no effect on rod outer segment survival. Because TrkB receptors (for BDNF) have not been identified on photoreceptors (except on red-green cones ²⁴ ) but have been identified on RPE ^{25 26} and Müller cells, ²⁵ the protective effect of most survival factors is likely to be mediated indirectly through these cells. In support of this, Wahlin et al., ²⁷ showed that BDNF, CNTF, and bFGF did not activate intracellular signaling pathways in photoreceptors but did so in Müller cells. ²⁷ Because of the pigmentation in the RPE cells, they were unable to identify reactivity in RPE cells. 

In transplantation studies, preservation of function is a necessary measure of success. Improved electroretinogram recordings have been observed after GDNF administration to the rd mouse ⁶ and the TgN S344 ter-4 rat. ⁸ In this study, we used head-tracking behavior which, in previous studies, has been shown to be a good predictor of successful grafts. ²⁸ The best-performing graft-recipient eyes tracked for far longer periods than those with sham surgery and GDNF-treated eyes showed better preservation of function than eyes treated in any other way. The pattern of head-tracking in the engineered rat Schwann cells is similar to that described for engineered human retinal pigment epithelial cells, ¹⁷ with a good response early on that deteriorates with time. Thus, at 9 to 10 weeks after surgery, dystrophic animals with engineered cell grafts (both RPE and Schwann) were able to track all gratings and performed significantly better than sham-surgery or nonsurgical dystrophic animals. By 17 to 20 weeks after surgery, sham-surgery animals in both experiments were unable to track any of the gratings. Animals with engineered cell (RPE and Schwann) grafts, however, performed significantly better at both the 0.25 and 0.125 gratings but showed little or no tracking with the 0.5 grating. Although dystrophic rats with GDNF- or BDNF-engineered cell grafts show a deterioration of head-tracking behavior with time, it not necessarily mean that all visual function has been lost, because photoreceptors remain (Fig. 6) , and electrophysiological studies recording from the superior colliculus and visual cortex of animals whose eyes had received engineered RPE cells ¹⁷ or immortalized human RPE cell grafts ²⁸ showed continued function for at least 6 to 8 months. It is not known why head-tracking behavior deteriorates over time when there is evidence of vision using other tests, but these studies have shown that good head-tracking behavior at earlier time points correlates with good performance in other visual tests. ²⁸ It should be noted, however, that the donor cells are allografts and there may be a slow loss of cells because of some form of rejection. ²⁹  

Although levels of CyA were not determined in this study, Lund et al., ¹⁷ found a mean blood level of 321.6 ± 21.9 μg/L in RCS rats where CyA from the same source was administered in an identical dose and manner. However, CyA is poorly absorbed, ³⁰ and a sub-group of T cells have been shown to become resistant to immunosuppressive drugs (such as CyA, see Ref. ³¹ ). These cells may be important contributors to chronic graft rejection. Although there was no evidence of cell infiltration or vessel-cuffing in the initial weeks after grafting, it may be that as the vessels in the RCS rat become leaky (a feature of the vascular changes in this animal with time, e.g., Ref. ¹⁸ ) and the CyA treatment becomes less effective, some nonclassic form of rejection occurs, possibly reducing the area of the graft. This is the subject of further studies. Even when syngeneic donor cells are used in the RCS rat, there is a reduction in the number of rescued photoreceptors with time. ¹² Possibly the extensive remodeling of the retina that occurs in the dystrophic RCS rat ¹⁸ affects graft survival. 

This study shows that Schwann cell lines engineered to overexpress growth factors can prolong photoreceptor survival and function, supporting our original findings with syngeneic grafts of primary Schwann cells. ¹² In accord with the findings of Wilby et al., ¹³ the engineered cells performed better than the parent cell line, suggesting that it is the additional growth factor production that is affecting photoreceptor survival. It remains to be seen whether rescue cam be further improved by grafting a mixture of the two engineered cell lines to a single eye or adding yet other cells engineered to make additional factors. 

The results add to recent work ³² that has indicated that ex vivo gene therapy coupled with transplantation may be a useful approach for preserving photoreceptors in retinal degenerative conditions. Transfected cells can be carefully monitored before introduction into the eye for factor production levels, for stability of factor production, and for any indication that the introduced factors might change the behavior of the cell in a detrimental way. 

 Supported by the Medical Research Council Cooperative Component Grant (RDL, JML); the Wynn Foundation, Foundation Fighting Blindness, and National Eye Institute Grant EY 14038 (RDL); Wellcome Trust Award (DJK); and the International Spinal Research Trust (EMM). RDL is the Calvin and JeNeal Hatch Professor of Ophthalmology, University of Utah. The Wellcome Trust provided funds for the purchase of the JEOL 1010 transmission electron microscope and Leica ultramicrotome used in this study.

 Submitted for publication; January 29, 2003 revised July 9, 2003; accepted August 3, 2003.

 Disclosure: J.M. Lawrence, None; D.J. Keegan, None; E.M. Muir, None; P.J. Coffey, None; J.H. Rogers, None; M.J. Wilby, None; J.W. Fawcett, None; R.D. Lund, None

 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.

 

Clone of SCTM41 cells transfected with pRc/CMV/hyg/BDNF: Western blot for BDNF. Lanes 1 and 2: conditioned medium (CM) from different batches of the SCTM41 parent cell line. Lane 3: CM from SCTM41/BDNF clone 6*. Each sample is 20 μL and concentrated 10×. Lanes 4, 5, and 6: recombinant BDNF at 10, 100, and 200 pg, respectively.

Figure 2.

 

Comparison of head-tracking performances with different grating periodicities through time (8–20 weeks of age). Graphs show head-tracking times (in seconds) of congenic RCS rats (left and right eyes), and dystrophic rats with right eyes grafted with BDNF- or GDNF-secreting cells, or with the parent cell line (SCTM), or with sham surgery alone compared with nonsurgical dystrophic left eyes. Each grating frequency is plotted separately (0.5, 0.25, or 0.125 cyc/deg).

Figure 3.

 

Semithin sections of nonsurgical, sham-surgery and transplant-recipient RCS rat retinas at 8 weeks of age. (A) A nonsurgical dystrophic retina. The disorganized, narrower outer nuclear layer (ONL) is seen with an extensive amorphous debris zone (DZ) lying between the ONL and the retinal pigment epithelium (RPE). Pyknotic nuclei (arrowheads). (B) A sham-surgery retina. The ONL is still relatively well-organized although some pyknotic nuclei can be seen (arrowhead). No inner (IS) and outer segments (OS) can be identified, and the DZ is filled with OS debris. (C) A retina grafted with the parent cell line (SCTM41). The ONL is more organized and wider. IS are present, and the DZ is narrower. (D) A retina grafted with GDNF-secreting SCTM41 cells. The ONL is well-organized and wide. Photoreceptor IS and OS are present. There is no DZ in this region of the retina. Scale bar, 50 μm.

Figure 4.

 

Semithin sections of sham-surgery and transplant-recipient RCS rat retinas at 16 weeks of age. (A) A sham-surgery retina. Very few photoreceptors remain in a discontinuous ONL. The DZ is much narrower. (B) A retina grafted with the parent cell line. The ONL and the debris zone (DZ) are more extensive than in the sham-surgery animal. (C) A retina grafted with BDNF-secreting SCTM41 cells. The ONL remains relatively well-organized in this region of the retina. Inner segments (IS) and a few outer segments (arrow) remain. The DZ is extensive, and microglia are present. Blood vessels (bv) close to the RPE may be induced by the graft but are more likely to be the result of the ongoing vascular changes that occur in the dystrophic RCS rat. (D) A retina grafted with GDNF-secreting SCTM41 cells. The ONL remains, four cells deep in places. A DZ with microglia is present. It is not possible to discern IS and OS in this particular section, but see Figure 5C . Scale bar, 50 μm.

Figure 5.

 

Electron micrographs of transplant-recipient RCS rat retinas at 16 weeks of age. (A) A retina grafted with the parent cell line. There is a reasonable preservation of ONL and inner segments (IS) are present. Only outer segment (OS) debris is present in the debris zone (DZ), which also contains a microglial cell (Mg). OLM, outer limiting membrane. (B) A retina grafted with BDNF-secreting SCTM41 cells. This region of retina is very well preserved, although it covers only a relatively small area of retina. The ONL is well-organized, and IS are present, as are stacks of OS. (C) Retina grafted with GDNF-secreting SCTM41 cells. There is good preservation of the ONL and IS. Arrow: stacks of OS. Scale bars: (A, B) 2 μm; (C) 7 μm.

Figure 6.

 

Mean photoreceptor numbers at 20 weeks. The graphs represent mean photoreceptor counts across the retina (temporal to nasal) for nonsurgical, sham-surgery, SCTM41-BDNF, and SCTM41-GDNF, and SCTM41 groups. Mean photoreceptor counts for retinas grafted with (A) SCTM41-BDNF–treated, (B) SCTM41-GDNF–treated, and (C) untreated SCTM41 cells plotted against nonsurgical and sham-surgery retinas.

The authors thank Robin Howes for valuable assistance.