Increased Spontaneous Retinal Ganglion Cell Activity in rd Mice after Neural Retinal Transplantation

Abstract

purpose. To study the functional success of neural retinal transplantation by means of retinal surface ganglion cell recordings.

methods. Eight-week-old C3H/HeJ (rd/rd) retinal degeneration mice received transplants (subretinal) in one eye only of neural retinal tissue isolated from newborn normal C57/BL6J mice. Four weeks after transplantation, ganglion cell responses were recorded directly from the retinal surface over the transplant, with a differential bipolar surface electrode. Measurements were performed, both with and without light stimulation. Similar recordings were performed in nontransplant areas of the transplant-recipient eyes, and in age-matched sham-treated and untreated control eyes. After the recordings, the eyes were processed for light and transmission electron microscopy.

results. Histologic examination showed that in some areas, transplanted cells were organized into small sheets and differentiated into photoreceptors with outer segments in intimate contact with the host RPE. No light-driven ganglion cell responses were recordable from the transplant-recipient or control eyes. However, the spontaneous ganglion cell activity was higher in the transplant areas (mean: 10.8 ±12.0 spikes/1.6 sec) compared with nontransplant areas of these recipient eyes (mean: 2.4 ± 5.7spikes/1.6 sec; P < 0.001), sham-treated eyes (mean: 2.5 ± 4.8 spikes/1.6 sec; P < 0.001), and the untreated control eyes (mean: 2.2 ± 4.4 spikes/1.6 sec; P < 0.001).

conclusions. Subretinal transplantation of neural retinal tissue results in a local increase of spontaneous ganglion cell activity. The increased activity may be due to the release of neurochemically active substances as a result of the presence of the transplant. Although light responses were not recordable, the technique of retinal surface ganglion cell recording may be useful for assessing the functional success of transplantation.

Neural retinal transplantation has been studied by a number of investigators in the hope that it may evolve into a means for visual restoration in patients with blindness caused by retinal degeneration. The eyes of patients and animals (rd mice) with retinal degeneration have been found to have a relative preservation of inner nuclear retinal cells, despite near total loss of photoreceptors. Thus, replacement of photoreceptors by transplantation has been suggested as a potential means of restoring vision to these eyes. Transplantation into the subretinal space of various animal models has shown that the grafted tissue can survive, differentiate into photoreceptors, and even form synaptic projections. ^{1 2 3 4 5 6 7 8 9} The functional capability of the transplanted tissue has also been well established. Seiler et al. ¹⁰ observed light-dependent distributional changes of proteins involved in phototransduction between the inner and outer segments of transplanted photoreceptors. In addition, Adolph et al. ¹¹ were able to measure a local electroretinogram (ERG) from the surface of the retinal graft. Despite these accomplishments, functional integration between the graft and the host retina has been more difficult to demonstrate. 

Del Cerro et al. ^{12 13} studied the suppressive effect of a warning-light flash on the startle reflex response to acoustic stimuli in light-blinded 344 Fisher rats to evaluate functional integration after transplantation. They observed partial restoration of this visually mediated behavior in some animals after retinal transplantation, ¹² but no recovery after sham injection, or injection of retinal cell homogenates or cerebellar cells. ¹³ Interpretation of behavioral changes in animals, however, can be very difficult and imprecise. In this same light-blinded rat strain, Silverman et al. ¹⁴ found some recovery of the pupillary reflex and visually evoked potentials (VEPs) after neural retinal transplantation. The pupillary reflex, however, is not a good predictor for retinal recovery, because there is no correlation between the size of the remaining photoreceptor population and the extent of pupillary constriction. ¹⁵ Although the VEP is a more objective measure, detection of a full-field VEP does not establish that the observed response is coming from the area of the transplant. 

Retinal ganglion cell (RGC) recordings are another well-established method of objectively measuring retinal function. Kuffler, ¹⁶ in 1953, was the first to record ganglion cell responses (action potentials) to light stimulation in the cat, by using a microelectrode detector system. Subsequently, a number of investigators have used ganglion cell recording techniques to enhance understanding of retinal circuitry. ^{17 18} In this study, in an effort to localize the effects of retinal transplantation more precisely, we studied the functional success of transplantation by recording RGC responses from the retinal surface of transplant-recipient retinal degeneration (rd) mice. 

Materials and Methods

Animals

Colonies of C3H/HeJ (rd/rd; Jackson Laboratories, Bar Harbor, ME) and C57/Bl6 (wild-type, Jackson Laboratories) mice were maintained at the Wilmer/Woods animal facility at the Johns Hopkins Hospital. Eight-week-old C3H/HeJ mice underwent transplantation in one eye only with neural retinal tissue isolated from the eyes of newborn (postnatal day 0) C57/Bl6 mice. Animals were anesthetized by intraperitoneal injection of a mixture of ketamine (60 mg/kg) and xylazine (8 mg/kg; both from Phoenix Scientific, Inc., St. Joseph, MO). Additional topical anesthesia was provided with 1% proparacaine hydrochloride drops (Alcon, Fort Worth, TX). Eyes were dilated with 1% tropicamide (Alcon). 

Transplantation Procedures

Donor tissue was prepared by gently dissecting the retina from the eye wall of eyes enucleated from newborn wild-type mice. The retinal tissue was placed in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Grand Island, NY) and cut with microscissors into fragments measuring 1 mm² or less in size. The retinal fragment suspension was then aspirated into a microsyringe. 

Transplantations of small retinal fragments were performed in one eye only at the equator (at the 12 o’clock position) according to the technique developed by Lazar and del Cerro. ¹⁹ Although rosettes may form in some areas of the transplant with this technique, some areas develop into small retinal sheets. Transplantation of larger retinal sheets is difficult because of the small size of the mouse eye. Briefly, a 27-gauge, partially sheathed butterfly needle connected to a microsyringe (containing the retinal fragment suspension) served as the delivery system. Under direct visualization through the pupil with a stereomicroscope, the needle was inserted through the sclera with the bevel facing the surgeon. The needle was rotated without changing its angle until the tip was in the subretinal space. The tip was advanced slightly to elevate the retina, and 2 to 4 μL of the retinal fragment suspension was injected into the subretinal space. Successful transplantation was confirmed by direct visualization, using a coverslip on the cornea and a surgical microscope. After successful transplantation, the whitish retinal transplants were readily identified posterior to the injection site, under the detached host retina. Before the (RGC recordings, fundoscopy was again performed (in the transplanted eyes only) to confirm the location of the transplants and to exclude any iatrogenic trauma to the retina caused by the procedure. For control animals, we used 10 age-matched, untreated animals, as well as 10 age-matched sham-injected animals in which the identical transplantation procedure was performed, but with subretinal injection of DMEM only (no retinal tissue). All animals were managed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and the experimental protocol was approved by the Johns Hopkins University School of Medicine Animal Care and Use Committee. 

Electrophysiologic Testing

RGC recordings were performed 4 weeks after retinal transplantation. At this age (12 weeks), in a non–transplant-recipient rd mouse, the photoreceptor degeneration was essentially complete. All the rods posterior to the equator had been lost, and only a few scattered cone nuclei (without outer segments) remained. ²⁰ Animals were dark adapted for at least 4 hours and then anesthetized with the same medications and dosages used for transplantation. A 30-gauge needle was used to make a small incision into the cornea just anterior to the limbus. Sodium hyaluronate (Healon GV; Pharmacia, Columbus, OH) was injected into the anterior chamber through the incision site to maintain the intraocular pressure. The cornea was excised and an intracapsular lensectomy was performed. Sodium hyaluronate was instilled to maintain the contour of the posterior pole and enable proper placement of the electrodes on the retinal surface. To remove the vitreous and expose the retinal surface for electrophysiologic testing, sodium hyaluronate was gently injected into the plane between the retina and the vitreous. The vitreous was allowed to slip away from the eye, and great care was taken to avoid traction on the retina. All these preparative procedures were performed under dim red-light illumination. Body temperature was maintained at 35°C with a feedback-controlled heating pad. 

For differential RGC recordings, the tips (separated by 200 μm) of two 125-μm-diameter parylene-coated tungsten electrodes (A-M Systems, Everett, WA) were positioned on the retinal surface with the aid of a microscope and a micromanipulator (MX-100; Newport, Irvine, CA). The ground electrode was placed in the animal’s mouth. A halogen light source was used as the light stimulus, and the eyes were directly illuminated (full-field illumination) with a fiber optic positioned at a distance of 1.0 to 1.5 cm from the retina. The onset of the light stimulus was at 0.1 second into the recording cycle. One-second and 200-ms light stimuli were used for the recordings. Ten eyes each were used for RGC recordings in the four experimental groups: (1) the transplant areas of transplant-recipient eyes posterior from the visible injection site; (2) the non-transplant areas of these eyes on the side opposite the transplant (i.e., 6 o’clock position); the corresponding areas (equatorial retina at 12 and 6 o’clock) of (3) control and (4) sham-injected eyes. For each group, three separate locations within a given experimental host retinal area (2 × 2 mm) were sampled. Ten cycles (1 cycle = 1.6 seconds) of RGC responses were recorded at each location with the light turned either on or off (shutter sound only). Animals were excluded if a retinal detachment developed during the preparative procedures for the recordings or during the recordings themselves. Light-driven RGC responses were recorded from a normal B57/Bl6 mouse before each experimental session, to confirm that all recording equipment was functioning properly. The output of the recording electrodes was fed into a differential AC preamplifier (band-pass: 300-2000 Hz; gain: 10-fold; model EX-1; Dagan, Minneapolis, MN) which was connected to another AC amplifier (gain: 2000-fold; model AM502; Tektronix, Beaverton, OR). Total gain was 20,000 fold. A speaker and an oscilloscope were connected to the system to allow both audio and video monitoring of the recording conditions and the RGC responses. The output of the amplifier was fed into in a computer. Data were acquired synchronously to the shutter action with a custom-written software program (Labview; National Instruments, Austin, TX) and analyzed with statistical software (MatLab; The MathWorks, Inc., Natick, MA). Thresholds were then set manually for each file. For the construction of peristimulus time histograms (PSTHs), spikes were grouped into 0.1-second intervals and counted only if the amplitude was at least twice the amplitude of baseline. Total spike counts during the recording interval of 1.6 seconds were compared between the transplant areas, and the various control eyes (nontransplant areas of recipient eyes, sham-injected eyes, and untreated eyes) using a two-tailed unpaired t-test. 

Histology Methods

For histologic investigations, animals were killed with an intraperitoneal injection of pentobarbital (200 mg/kg). The eyes were then enucleated and immersion fixed with 3% glutaraldehyde in 0.1% 1 M phosphate buffer for 24 hours. For light and electron microscopy, specimens were postfixed in 1% veronal acetate-buffered osmium tetroxide, dehydrated in a series of graded ethanols, and embedded in plastic. For transmission electron microscopy, ultrathin sections were stained with uranyl acetate and lead citrate. 

Results

Histology

Light and electron microscopic examination revealed the presence of grafted tissue in the subretinal space in all transplant recipients in which recordings were obtained. The grafts varied in size, with the dimensions ranging from 1 to 3.5 mm in greatest linear dimension. In all cases, in the transplant-recipient retina, rosettes formed in some areas, and in other areas and small retinal sheets with photoreceptor differentiation in other areas (Figs. 1 2) . Transplanted photoreceptor outer segments were intimately associated with the host retinal pigment epithelium (RPE). At the inner aspects of the photoreceptors, many triad ribbon-type and conventional synapses with numerous synaptic vesicles were found. However, it was not possible to determine whether the synapses were between transplanted cells or between transplanted cells and host neurons. Nontransplant areas in the recipient eyes and the control untreated eyes uniformly showed only a few remaining cones, but no evidence of outer segments. In both transplant-recipient and control eyes, the inner nuclear and ganglion cell layers were well preserved in all animals. There was no host outer nuclear layer and only rare outer nuclear layer cell bodies (presumably cones), outside the grafted area. No photoreceptor outer segments were present, except within the graft area. 

Electrophysiology

Light-driven ganglion cell responses were recorded in the normal mice (Fig. 3) , but the eyes of transplant-recipient, noninjected, and sham-treated rd mice showed no increase in ganglion cell activity with the light stimulus. The level of spontaneous ganglion cell activity, however, was significantly increased in the area of the transplant (mean: 10.8 ± 12.0 spikes/1.6 sec; Figs. 4 5 ) when compared with the nontransplant areas (mean: 2.4 ± 5.7spikes/1.6 sec; P < 0.001), sham-injected eyes (mean: 2.5 ± 4.8 spikes/1.6 sec; P < 0.001), and the control (mean: 2.2 ± 4.4 spikes/1.6 sec; P < 0.001; Fig. 6 ). There was no significant difference between the nontransplant areas and the control eyes. 

Discussion

This study describes the use of retinal surface ganglion cell recordings to evaluate functional integration after retinal transplantation. Although previous investigators have measured local electroretinograms (ERGs) from the graft surface, ¹¹ the presence of a recordable ERG does not necessarily signify the development of functional connections. The a-wave of the ERG is derived from the photoreceptors, and the b-wave is derived from cells in the inner nuclear (bipolar cells). Because the transplanted cells may differentiate into bipolar cells and photoreceptors, the transplant alone may produce a recordable ERG. If light-driven responses were measured from the host ganglion cells, however, this could provide stronger evidence that functional connections have formed between the transplant and host retinal cells. Connectivity can also be studied by measuring responses from higher order centers, such as the superior colliculus ²¹ or the visual cortex ²² itself, as described by Lund and coworkers. An advantage of recordings from the visual cortex (such as VEPs) is that they provide information that is more representative of actual visual perception. Full-field flash VEPs, however, do not have the spatial selectivity of RGC responses to allow direct assessment of function of the transplant itself. 

RGCs and some amacrine cells are the only cells in the retina known to generate spike responses (action potentials). Other neuronal cells (photoreceptor, bipolar, and horizontal cells) show only graded responses. With our recording technique, the differential bipolar electrodes are gradually advanced with a micromanipulator, until the electrodes just contact the retinal surface. Both visual (oscilloscope) and audio cues are used to monitor the electrode approach. With this method, unintentional penetration of the retinal surface is nearly impossible. Without retinal penetration, we have not been able to record responses from deeper retinal layers. Thus, it is unlikely that the spike responses are from the deeper amacrine cells; they are more likely a manifestation of RGC activity. This technique does not allow, however, the differentiation of host and transplant ganglion cells if transplanted ganglion cells were to migrate into the host ganglion cell layer. We have not observed such a migration, however, after transplantation of green fluorescent protein–labeled donor tissue into the subretinal space of rd mice (Sadda SR, unpublished data, 2002). 

In this study we did not observe the restoration of light-driven RGC responses after retinal transplantation, despite the survival of the graft and differentiation of the transplanted cells into photoreceptors with outer segments. One possible explanation for the absence of light responsiveness is that the photoreceptors were not functioning normally. Although the transplanted photoreceptor outer segments were certainly in contact with the host RPE in many areas, in other areas the photoreceptors were organized into rosettes, with outer segments crowded in the lumen of the rosette. A rosette configuration is likely to be less compatible with normal function than a layered arrangement. It is important to note, however, that Adolph et al. ¹¹ were able to record an ERG, even from grafts that were organized into rosettes. Moreover, we sampled multiple locations over the graft site, and thus we were also likely to have recorded over an area with a retinal sheet. Another possible explanation for the failure to record a light response may be the failure of connections to form between the transplant and the host retina. Although synapses were observed at the graft–host interface, we could not determine whether the connections were between transplanted neurons and host neurons. Furthermore, even if connections formed between the host and the transplant, they may not have been functional. We have previously suggested that the possible absence of graft–host connections may be due to an inability of the adult host retina to form new synaptic contacts. Transplantation into younger host animals (2-week-old rd mice) has demonstrated the presence of light-driven RGC responses over the graft in a small percentage of animals. ²³ In addition, Woch et al. ²⁴ demonstrated light-driven superior colliculus responses in Royal College of Surgeons (RCS) rats (aged 37–69 days), after cotransplantation of intact sheets of fetal retina and RPE. Although a light response was observed in these somewhat older animals, the underlying mechanism (e.g., rescue effect) of the response in this animal model of retinal degeneration is uncertain. Regardless, if transplantation is to be of significant therapeutic value in patients with retinal degeneration, it is important that restoration of light responses be conclusively demonstrated in the setting of an adult degenerated retina. 

Although light-driven responses over the graft were not observed in this study, an apparent increase in spontaneous ganglion cell activity was detected in the transplant areas. This effect appeared to be fairly localized, because no increase in activity was observed in the nontransplant areas of these same eyes. Sham injection of DMEM into the subretinal space did not result in an increase in activity, suggesting that the surgical procedure itself was not a causative factor. One could speculate, however, that the creation of a retinal detachment may be sufficient to trigger increased spontaneous activity and that the detachment may have persisted for a longer duration in transplant-recipient eyes. We did not observe such a discrepancy, however; the retina appeared to flatten within a week’s time, whether transplanted tissue was injected or not (because no break was created in the retina, the retina flattened quickly). Thus, the transplanted tissue itself appears to be the most likely agent responsible for the local increase in ganglion cell activity. A possible mechanism for this transplantation effect may be the elaboration of neurochemically active mediators by the transplanted neurons or the remaining host retinal cells. Regardless of the mechanism of action, neural retinal transplantation into the subretinal space did not restore light responsiveness to the eyes of 12-week-old rd mice as assessed by retinal surface ganglion cell activity. Nonetheless, RGC recordings may still be a useful method for measuring the functional success of various future transplantation approaches. 

 Supported by the Foundation Fighting Blindness, Hunt Valley, Maryland; the Grousbeck Family Foundation, Menlo Park, California; the Rabensteiner Funds, Vienna, Austria; and the Funds for the Promotion of Scientific Knowledge, Vienna, Austria.

 Submitted for publication September 19, 2001; revised March 5, 2002; accepted May 29, 2002.

 Commercial relationships policy: N.

 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.

 

Light micrographs of a transplant in the eye of a 12-week-old rd mouse 4 weeks after transplantation. (A) The transplant (tr) was located between the host inner nuclear layer (inl) and RPE and showed variable organization, with cells forming small sheets in some areas (at left) and rosettes (at right) in others. (B) Higher magnification of the area of the retinal sheet. Transplanted cells were in contact with cells in the inner nuclear layer (inl) and elaborated photoreceptor outer segments (arrows). (C) Nontransplanted area of same mouse eye. Photoreceptors are absent and the host inner nuclear layer is in direct contact with the RPE (arrows). Magnification: (A) ×100; (B, C) ×400.

Figure 2.

 

Transmission electron micrographs of transplant-recipient eye of the same 12-week-old rd mouse shown in Figure 1 , in the area of the retinal sheet. (A) Transplanted photoreceptor outer segments demonstrate the characteristic stacked disc arrangement. (B) Ribbon-type synapses (arrows) were present within the transplant. Magnification: (A) ×40,000; (B) ×10,000.

Figure 3.

 

RGC response to a 1-second light stimulus in a normal C57/Bl6 mouse. (A) Raw data showed a burst of increased RGC activity after the light was turned on (shutter open). (B) PSTHs (grouping spikes into 0.1-second intervals) of 10 cycles recorded with the light turned on shows evidence of increased RGC activity both after the light was turned on (ON-response) and after it was turned off (OFF-response).

Figure 4.

 

RGC response to a 1-second light stimulus in the transplant area in a 12-week-old rd mouse eye. (A) Raw data show a few spikes with no apparent increase in activity with light stimulation. Typical ganglion cell action potentials were observed (B).

Figure 5.

 

RGC activity in the transplant areas in 12-week-old rd mouse eyes. Averaged PSTHs with spikes grouped into 0.1-second intervals are shown. No difference in spike counts was observed with light stimulation (A) compared with no light (B).

Figure 6.

 

Comparison of RGC activity among control groups. Averaged PSTHs are shown in response to light stimulation in (A) nontransplant areas of transplant-recipient eyes (B) untreated eyes, and (C) sham-injected eyes. The level of activity was similar in all three groups and showed no increase with light stimulation.