Ординатура / Офтальмология / Английские материалы / Aging and Age Related Ocular Diseases_Lutjen-Drecoll_2000

The border between transplant and host retina could be identified more readily in the electron microscope and was seen in the vitreal part of the transplant IPL. Along this border, processes derived from grafted Müller cells arranged in a lamellar pattern were seen, but there was no sign of a limiting membrane with tight junctions. At regular intervals along the host side of the border, bundles of fimbriae of approximately 4±6 Ìm were also present. These fimbriae were interpreted as Müller cells by their content of glycogen granules. On the vitreal side of these fimbriae, aggregates of mitochondria were seen in the Müller cell cytoplasm. Also closely associated with the fimbriae, bundles of nerve fibers were visible on both sides of the border. They contained mature nerve cell processes, identified by their contents of microtubuli, as well as irregular larger structures filled with clear vesicles of different sizes. These structures lacked vacuoles and waste products and were interpreted as growth cones [43]. Bundles of nerve cell fibers were also present closer to the host IPL, passing the host inner nuclear cells, in close contact with Müller cell processes.

Graft-Host Connections

In short-term grafts, NF-160-kD-labeled fibers running from the transplant vertically towards the host were seen [35]. In 8 of the 14 long-term transplants, double labeling with PKC and parvalbumin revealed numerous small PKCand parvalbumin-labeled processes closely together in a thick intermediate plexiform layer located between host and graft [37]. These processes seemed to be derived from both host and graft rod bipolar and AII amacrine cells. In all these specimens, rod bipolar cells in the host extended sprouting dendrites towards the intermediate plexiform layer. Sprouting fibers from host AII amacrine cells were also common. Direct contacts between rod bipolar cells in the transplant and AII amacrine cells in the host were seen in 3 of the specimens. AII amacrine cells in the transplant were sometimes seen to extend fibers terminating on rod bipolar dendrites in the host.

Contacts between calbindin-labeled cone bipolar cells in the graft and AB5-labeled ganglion cells were not apparent. In one specimen, 126 days postoperatively, a few AB5-labeled fibers from the host came very close to calbindin-labeled cone bipolar axons from the graft, but no definite contact was seen. In the 10-month specimens, the distance between ganglion cell processes from the host and cone bipolar cell processes from the transplant was short, yet even then, no direct contacts were seen.

Morphology of Other Transplants

Adult Full-Thickness Grafts

One out of 4 adult full-thickness transplants could be identified in the light microscope [39]. This specimen was obtained 14 days after transplantation and the graft measured 0.6 mm in its vertical extent. It displayed laminated layers, but with a reduction in number of all cell types.

Embryonic Vibratome-Sectioned Grafts

Nine out of 14 embryonic vibratome-sectioned grafts survived the postoperative period of 11±27 days. They all displayed abnormal morphology with rosettes and sometimes also laminated retinas with reversed polarity [39]. The laminated segments often included many inner retinal cells, but these were always adjacent to the host RPE or Bruch's membrane.

Adult Vibratome-Sectioned Grafts

A total of 5 out of 20 adult vibratome-sectioned grafts could be identified with certainty [39]. These grafts displayed normal lamination and in all parts included the inner retina, suggesting that the vibratome sectioning in these cases had been incomplete. The photoreceptor outer segments faced the host RPE but were shorter than normal. No rosettes were seen. Four of the surviving transplants were from postoperative day 15±16 and measured from 0.4 to 1.3 mm in their vertical extent. One was from day 29 and measured 0.6 mm.

Embryonic Fragment Transplants

The fragment transplants were all in the form of rosettes, often with a certain degree of differentiation, and with photoreceptor outer segments facing in towards the lumen [36]. In the majority of specimens, large defects in the host RPE were seen. There was no apparent invasion of lymphocytes in any of the specimens, but in 2 out of 5 fragment transplants (49 days postoperatively), large clusters of cells which did not have the morphology of inflammatory cells were seen in the choroid and graft.

Immunogenicity ± MHC Expression

Normal Tissue and Controls

In normal adult and embryonic retina, no MHC-class- I- or -class-II-labeled structures were seen [36]. The choroid of adult specimens displayed single MHC-class-I- and -class-II-labeled cells. The MHC-class-I-expressing cells were round and had no processes. The MHC-class-

II-expressing cells had an elongated appearance and often displayed 1 or 2 dendritic processes. There was no MHC class I or class II labeling in the negative controls.

Full-Thickness Transplants

None of the 7 full-thickness grafts displayed any MHC class I and class II labeling, and there was no sign of increased MHC expression in the choroid when compared to normal specimens.

Fragment Transplants

Cells expressing MHC class I were found in 2 animals (31 and 49 days postoperatively). In the younger transplant, a few clusters of labeled cells were found in the rosettes. The older one displayed an array of numerous labeled cells in an area corresponding to the host choroid and scleral part of the graft. In this area, no RPE cells could be seen.

Cells expressing MHC class II were found in 4 of the 5 specimens. In 2 of these specimens (31 days postoperatively), clusters of a few labeled cells in the graft were present in areas with small RPE defects. In the other 2 (49 days postoperatively), large aggregates of MHC-class-II- expressing cells in the choroid and scleral part of the graft were seen in areas where the RPE was completely missing. The labeled cells in the choroid were mostly concentrated in the part adjacent to the transplant where they were considerably more frequent than in the normal rabbit. Their number decreased and returned to normal at some distance from the graft.

Many of the cells labeled by MHC class I and class II had relatively large, elongated cell bodies with long dendritic processes, but the ones found inside rosettes were more round and lacked processes. There were no MHCexpressing cells in the vitreal part of the grafts or in the host retina.

In the fifth specimen (31 days postoperatively), no MHC class I or II labeling was found.

Discussion

Surgical Considerations

Development of the Transplantation Technique

The surgical procedure described for full-thickness transplantation evolved from a series of pilot experiments. The earliest technique involved lensectomy, vitrectomy and a 3-mm retinectomy. The transplant was then positioned in the defect produced by the retinectomy

and finally held in place by gas (20% sulfur hexafluoride) after a gas-fluid exchange. This procedure allowed for a more complete vitrectomy than the current one, and the graft was kept flat and in position. All of these cases, however, developed serious epiretinal membranes and eventually a traction detachment. To avoid epiretinal membranes, the procedure was simplified in the manner described above. Initially, only the first retinotomy was included, and in 50% of the operations, the graft ended up in the vitreous due to fluid reflux through the one existing retinotomy. After the introduction of the second retinotomy, this was no longer a problem. Other groups have succeeded in transplanting sheets through only one retinotomy [21], but we found the second retinotomy to be essential.

Vitrectomy

Vitrectomy was for several reasons used in the transplantation procedure. The established technique with an internal light source gives good access to the retina which enables the delicate microsurgery. The actual removal of vitreous also makes it easier for the retina to lift, allowing a localized bleb of retinal detachment to be created, and facilitates the positioning of the transplant subretinally. Although a complete removal of the vitreous is not possible with the lens in place, a central vitrectomy including the area of transplantation is a sine qua non for the success of the procedure. Other groups have tried to perform full-thickness transplantation without the three-port vitrectomy technique [44, 45]. The results, however, indicate a comparatively low frequency of successful grafts.

Vitrectomies have been safely performed in rabbit eyes for many years [46±48], but complications do occur and are well known. The size of the rabbit lens is considerably greater than its human counterpart, and navigating the vitrectomy instruments between this large lens and the retina can be a perilous task since neither tissue accepts maltreatment. The rabbit lens responds to operation trauma by cataract formation and inflammation leading to epiretinal membranes. Removing the lens altogether is not a good option, since lensectomy is in itself well known to precipitate inflammation leading to traction detachment [49]. Epiretinal membranes also develop in response to vitrectomy [50], but in our procedure this was effectively counteracted by using heparin in the infusion solution [51]. Retinal detachment is perhaps the most serious complication of vitrectomy but is usually caused by vitreous traction associated with lensectomy and did not occur in our series once we had omitted the lensectomy.

Placing the Donor Tissue

The most delicate part in the transplant procedure was positioning the graft flat and with correct polarity. The tendency of the donor tissue to arch enabled it to be drawn in with little trauma into the glass cannula. When placed in the subretinal space, it unfolded to a certain degree, but the margins of the graft were often elevated. Judged by the flat appearance in the light microscope, the combination of transplant growth and the diminishing bleb with the host retina pressing down on the transplant is evidently enough to flatten the bulk of the tissue.

Embedding retinal sheets in gelatin keeps them flat throughout the transplant procedure but has some disadvantages compared with the glass cannula used for fullthickness transplantation. Silverman and Hughes [20] worked with rats and mice and used a transcorneal approach to place vibratome-sectioned sheets subretinally. This procedure involved detachment of the retina from the ora serrata to the posterior pole. To avoid such trauma to the host retina we favored the transvitreal approach also used by Kaplan et al. [21] for our gelatin-embedded grafts [39]. This method requires an eye of approximately the same size as that of humans but minimizes detachment of the host retina. The instrument used for gelatin-embedded grafts is wider than the glass cannula and consequently demands a larger sclerotomy and retinotomy. This enhances the risk of perand postoperative hemorrhage due to scleral and choroidal damage. In addition, a large sclerotomy makes the intraocular pressure more difficult to maintain at a constant level. In our material, blood in the vitreous was evident in approximately 30% of the eyes receiving gelatin-embedded grafts. Most of these eyes also developed a total retinal detachment. Approximately 10% of the eyes receiving full-thickness grafts with the glass capillary displayed a vitreal hemorrhage, and retinal detachment was rare. A large retinotomy, apart from causing more damage, also makes the bleb of the host retina more unstable. Kaplan et al. [21] were forced to use hyaluronate sodium to maintain the height of the bleb, which in 1 out of 2 cases led to incomplete healing of the retinotomy. Another complication of the large retinotomy is that the transplant does not stay in place. In our full-thickness grafts approximately 5% of the grafts ended up in the vitreous, while 10% of the vibratome-sectioned ones did so.

To conclude, our results suggest that full-thickness grafts can be placed accurately with a standardized, safe operation technique. Gelatin-embedded grafts are difficult to transplant for physical reasons, and this procedure is, at least in our hands, associated with complications such as vitreous hemorrhage and retinal detachment.

Transplant Morphology

Organization and Survival

Good long-term photoreceptor survival has previously been reported in microaggregate (fragment) transplants [13]. However, the success rate was comparatively low and could not be influenced easily, as small pieces of donor tissue were placed at random subretinally. One important conclusion from the work on fragment transplants is that grafted photoreceptors will only survive if they are oriented correctly, i.e. facing the host RPE.

The majority of our embryonic full-thickness transplants displayed a laminated morphology and they extended over areas of more than one disk diameter. The photoreceptors showed well-developed outer segments facing the host RPE even after long survival times. The laminated flat morphology of the full-thickness grafts thus seems advantageous for photoreceptor cell survival.

The grafts contained many of the normal retinal components, and their organization in most cases mirrored the one found in the normal rabbit retina. Specifically labeled ganglion cells, however, were not common, even though a GCL was often seen. In short-time transplants we found cells in the GCL labeled with AB5, a specific ganglion cell marker [52], and also NF-160-kD-labeled fibers in the NFL, indicating that ganglion cells were present. Many cells in the GCL, however, were not labeled either by AB5 or choline acetyltransferase, suggesting that a number of the cells were not standard ganglion or amacrine cells.

Rosette Formation

Fragment transplants are well known to form rosettes with an initial degree of layering [8]. A small fraction of our full-thickness transplants developed into rosettes which indicates that other factors than the donor tissue integrity may be important in rosette formation of neuroretinal grafts. Why rosettes appeared in these transplants is not quite clear. All transplants were mechanically separated from the RPE in the donor eye, and it is possible that excessive force, enough to cause an abnormal development, was used in these cases. Another cause of rosette development of full-thickness grafts may be reversed polarity, either peroperatively or during the first postoperative days when then host retinal bleb was still present.

Two of our long-term grafts developed into rosettes and displayed a degenerated morphology which is in accordance with the fate of other rosetted transplants [8]. The fact that transplants rich in rosettes as well as disorganized parts of otherwise laminated transplants tend to

degenerate, again confirms that a normal organization of the transplanted retina is important for its survival.

Embryonic versus Adult Donor Tissue

The good survival of immature retinal transplants is well documented [13, 16, 20], while studies on adult transplants are not as common. Our results indicate that embryonic transplants survive better than their adult counterparts. In our relatively small group of adult fullthickness grafts, only one survived, and this graft showed severe signs of degeneration [39]. Adult retina is very sensitive to ischemia, which is evident by the rapid degenerative changes occurring post mortem [53]. Differences in tolerance to drastic environmental changes, such as dissection and transplantation may in part be responsible for the poor survival of adult grafts. In our 5 (out of 20) surviving adult vibratome-sectioned grafts, large parts of the inner retina remained which of course indicates that vibratome sectioning in these grafts was not complete and that the surviving transplants were in effect almost full thickness. The implication of this observation is interesting however ± adult transplants can survive, at least for short periods, if they are gelatin embedded and the inner retina remains. The finding is consistent with those of Silverman et al. [54] who in their vibratome-sectioned transplants observed a better preservation of the photoreceptors when portions of the inner retina were included in the transplant.

In conclusion the embryonic full-thickness transplants develop well and survive for at least 10 months. From photoreceptor outer segments to IPL their morphology is comparable to the normal retina.

Graft-Host Integration and Connections

Degeneration of the Host Retina

A prerequisite for good graft host integration of fullthickness transplants is degeneration of the host outer retina. The host retina, even in our short-time grafts, very much resembled a retina affected by degenerative disease with an almost complete loss of the outer nuclear layer. The retina of the rabbit is merangiotic, i.e. vessels are confined to a horizontal band at the myelinated nerve fibers, and this makes it vulnerable when separated from the RPE. Degeneration of the outer nuclear layer of the host has also been reported in fragment transplants to holangiotic (completely vascularized) retina [17], indicating that other mechanisms than separation from the RPE may be involved in this phenomenon.

Integration

Two factors prevented graft-host fusion in the shortterm specimens: (1) the well-developed inner retinal layers in the graft and (2) the presence of disorganized cells without proper lamination between the laminated layers of graft and host. Longer surviving times led to a loss of the innermost layers of the transplant and to a decrease in the number of intervening nonlaminated cells. In long-term specimens, complementary elements from host and graft together almost reconstructed the normal retinal appearance with all layers present.

Connections

In the 2 specimens examined in the electron microscope, we were able to identify and characterize the border between graft and host. Fiber bundles containing mature nerve fibers as well as growth cones on both sides of this border were found. The close relationship of these neurites with Müller cell fimbriae and mitochondria at regular intervals along the border suggests that neuron sprouting might be guided by these cells. This is comparable to normal retinal neuron development in which Müller cells are known to play an essential role [29, 32, 55].

Growth cones are usually associated with developing neurites [42, 56] but are also present in adult retinal regenerating neurons, especially when stimulated by an embryonic environment [57] or retinal detachment [58]. Using immunohistochemistry, the growth cones of our specimens were found to be both host and graft derived, originating predominantly from rod bipolar and AII amacrine cells. The sprouting activity seemed directed towards the formation of an intermediate plexiform layer which most likely represented a vitreally displaced transplant IPL fused with inner layers of the host. The abundance of neurites from both host and graft in this layer suggests that it might function as a switchboard between neurons of graft and host. The intermediate plexiform layer was not organized in sublaminae, and the order of neurons was not always clear. The direct connections found between grafted rod bipolar cells and host AII amacrine cells however indicate that connections between interneurons normally found in the rod pathway [59] can be established.

The AB5 and calbindin double-label experiments showed that both cone bipolar and ganglion cells survived for at least 10 months in the host, and that they often retained their normal contacts in the host IPL. This finding is of great importance since they constitute the last two elements in the rod pathway and thus are critical for any graft-host transmission of visual information. In severe

retinitis pigmentosa, 70% of the ganglion cells perish, possibly because of transsynaptic degeneration [60]. This can be compared with our specimens, where the outer retina of the host rapidly degenerates but where most of the ganglion cells seem to survive long postoperative times. Whether host ganglion cells survive because of input from the transplant remains to be investigated.

Very few of the cone bipolar cells found in the transplant made direct contact with ganglion cell processes from the host. The cone pathway is difficult to analyze since only a fraction of the cone bipolar cells can be labeled with commercially available antibodies. In the normal rabbit retina, however, contacts between calbin- din-labeled cone bipolar and AB5-labeled ganglion cell processes were readily seen, and these results indicate that graft-host connections in the cone pathway of our specimens are not common.

Immunology of Neuroretinal Transplants

The results of our immunological study indicate a difference in immune response depending on donor tissue integrity. The lack of any detectable expression of MHC in the full-thickness transplants suggests a low degree of immunogenicity in these grafts which is encouraging and in accordance with their good long-time survival. The different state of the fragmented donor tissue does not readily explain why these grafts display an enhanced MHC expression, and a further analysis of this phenomenon is therefore necessary. The literature on host immune responses to neuroretinal grafts is limited, but more knowledge based on other types of transplants is available.

A classic graft rejection is triggered by allogeneic MHC molecules [61], and when cells expressing these antigens are grafted to an incompatible host, an immune response leading to rejection ensues [62, 63]. The response is characterized by MHC upregulation in graft-derived cells followed by upregulation in host cells and subsequent T cell infiltration [64, 65]. The immune response seen in our fragment transplants differs from the one seen in classic rejection in two ways: first, as will be further discussed, most cells that expressed MHC appeared to be host derived; secondly, the reaction was limited to an increase in MHC-expressing cells without any obvious infiltration of lymphocytes, indicating that a full-blown graft rejection was not at hand.

The Origin of MHC-Expressing Cells

The tight junctions of the RPE constitute the outer blood-retina barrier which prevents the passage of large molecules and cells to and from the retina. This barrier was broken in both full-thickness and fragment transplantation procedures by the injection of fluid into the subretinal space which swept away many of the RPE cells. In the fragment transplants, the RPE defects appeared to have healed only partially and were often associated with an increase in MHC-labeled cells, many of which had a dendritic morphology. The concentration of labeled cells in the choroid and scleral part of the graft and the presence of cells with similar morphology in the normal choroid support the assumption that they were host derived. The breakdown of the blood-retina barrier in fragment transplants thus seems to permit dendritic cells from the host choroid to invade the graft. The delayed healing of RPE defects associated with these grafts as compared with normal RPE healing [66] indicates that the fragmented donor tissue affects the host RPE, either directly or through a response from the host immune system. The RPE defects in full-thickness grafts on the other hand healed well, and the remaining small breaks were never associated with any increase in MHC-expressing cells. This suggests that full-thickness grafts do not affect the host RPE in the same manner as their fragmented counterparts and that a disturbance in the blood-retina barrier alone is not enough to provoke a reaction from the host immune system.

The labeled cells found in the lumen of rosettes in the graft were presumed to be macrophages as previously described [67]. Whether these cells were host or graft derived is difficult to ascertain, but they probably represent transformed RPE cells [68] which have not been removed during the dissection of the donor neuroretina.

The Host Immune Response

We found no MHC expression in normal embryonic retinas suggesting that triggers of the immune response in fragment transplants involved predominantly minor instead of major histocompatibility antigens [69]. The apparent absence of lymphocytes and the abundance of dendritic cells in the grafts suggest that the rejection process may be different from the one seen in classic graft rejection. Dendritic cells are known for their constitutive expression of MHC class II antigens [70, 71] and have been found in a variety of tissues including the choroid where they can increase in number in response to experimental uveitis [72]. To our knowledge, they have not been described in retinal transplants previously, but host-

derived dendritic cells have been implicated in experimental brain transplants [73] and have also been found in human renal grafts [74]. Their exact role in the rejection process is uncertain, but one possible function is to extract antigens from the graft and present them to host T cells [75]. Fragmenting the graft may in this setting increase antigen exposure to dendritic cells, which would help to explain the difference in the immune response between fragmented and full-thickness transplants.

Future Considerations of Neuroretinal Transplants

The important question regarding the ability of embryonic full-thickness grafts to transfer visual information to the host has not yet been answered. One disadvantage with our model is that the recipient rabbits have normal retinas, which makes electrophysiological and behavioral testing difficult. To get more information on graft-host transmission of useful visual information, animals suitable for vitrectomy also displaying retinal degeneration will probably have to be used in the future.

Another important issue is how grafts are to be obtained in a future clinical setting. Our experiments confirm earlier findings on the superiority of embryonic grafts over adult ones, which makes access to immature tissue desirable. Human retinal cells from aborted fetuses have already been used in clinical trials involving cell suspension transplants [19], and fetal brain transplants have been used for a number of years in surgical treatment of parkinsonian syndromes [6]. Examining other sources of donor tissue is important, however, since the amount of human fetal tissue is limited. Possible sources include tissue from cell cultures and xenogeneic grafts [26], in which the low immunogenicity of full-thickness grafts may prove favorable.

Conclusions

Full-thickness embryonic retinas can be transplanted into adult rabbits with a minimum of complications.

The transplants develop into laminated retinas with layers parellel to the host RPE and display most of the normal retinal morphology.

The survival time of the transplants is at least 10 months, after which they still retain their laminated appearance and show no signs of photoreceptor degeneration.

In long-term specimens, a graft-host adaptation occurs where neurons from both entities coalesce to form an intermediate plexiform layer. Graft-host connections exist, and the participating neuronal types are the ones seen in the normal rod pathway of the retina.

Fragmented neuroretinal transplants trigger a host immune response characterized by an increase in MHCexpressing cells which is not present in full-thickness grafts, indicating that the integrity of the donor tissue is important for favorable host integration.

Acknowledgements

This work was supported by a BioMed2 EU grant, the Swedish Medical Research Council (project 2321), the Faculty of Medicine at the University of Lund, the Segerfalk Foundation, the Crafoord Foundation, the H. and L. Nilsson Foundation, the Royal Physiographic Society in Lund, the Clas Groschinsky's Memorial Foundation, the Swedish Society for Medical Research, the Edwin Jordan Foundation for Ophthalmic Research and Crown Princess Margareta's Foundation for Blind Children.

References

1Berson EL, Rosner B, Sandberg MA, Hayes KC, Nicholson BW, Weigel-DiFranco C, Willett W: A randomized trial of vitamin A and vitamin E supplementation for retinitis pigmentosa. Arch Ophthalmol 1993;111:761± 772.

2Weleber RG, Kennaway NG: Clinical trial of vitamin B6 for gyrate atrophy of the choroid and retina. Ophthalmology 1981;88:316±324.

3RetNet. World Wide Web URL: http:/ /www.sph.uth.tmc.edu/Retnet.

4Murata T, Kimura H, Sakamoto T, Osusky R, Spee C, Stout TJ, Hinton DR, Ryan SJ: Ocular gene therapy: Experimental studies and clinical possibilities. Ophthalmic Res 1997;29:242± 251.

5Björklund A, Stenevi U: Growth of central catecholamine neurones into smooth muscle grafts in the rat mesencephalon. Brain Res 1971;31:1±20.

6Widner H, Rehncrona S: Transplantation and surgical treatment of parkinsonian syndromes. Curr Opin Neurol Neurosurg 1993;6:344±349.

7Turner JE, Blair JR: Newborn rat retinal cells transplanted into a retinal lesion site in adult host eyes. Brain Res 1986;391:91±104.

8Bergström A, Ehinger B, Wilke K, Zucker CL, Adolph AR, Szel A: Development of cell markers in subretinal rabbit retinal transplants. Exp Eye Res 1994;58:301±313.

9Juliusson B, Mieziewska K, Bergström A, Wilke K, Van-Veen T, Ehinger B: Interphotoreceptor matrix components in retinal cell transplants. Exp Eye Res 1994;58:615±621.

10Aramant R, Seiler M, Ehinger B, Bergström A, Adolph AR, Turner JE: Neuronal markers in rat retinal grafts. Dev Brain Res 1990;53:47± 61.

11Aramant RB, Seiler MJ: Fiber and synaptic connections between embryonic retinal transplants and host retina. Exp Neurol 1995;133: 244±255.

12Adolph AR, Zucker CL, Ehinger B, Bergström

A:Function and structure in retinal transplants. J Neural Transplant Plasticity 1994;5: 147±161.

13Gouras P, Du J, Kjeldbye H, Yamamoto S, Zack DJ: Long-term photoreceptor transplants in dystrophic and normal mouse retina. Invest Ophthalmol Vis Sci 1994;35:3145±3153.

14Bergström A, Ehinger B, Wilke K, Zucker CL, Adolph AR, Aramant R, Seiler M: Transplantation of embryonic retina to the subretinal space in rabbits. Exp Eye Res 1992;55:29±37.

15Sharma RK: Retinal Transplants ± Growth, Differentiation, Integration and Survival; thesis, Lund University, 1999.

16del Cerro M, Notter MF, del Cerro C, Wiegand SJ, Grover DA, Lazar E: Intraretinal transplantation for rod-cell replacement in light-dam- aged retinas. J Neural Transplant 1989;1:1± 10.

17Juliusson B, Bergström A, van Veen T, Ehinger

B:Cellular organization in retinal transplants using cell suspensions or fragments of embryonic retinal tissue. Cell Transplant 1993;2: 411±418.

18del Cerro M, Ison JR, Bowen GP, Lazar E, del Cerro C: Intraretinal grafting restores visual function in light-blinded rats. Neuroreport 1991;2:529±532.

19del Cerro M, Das TP, Lazar E, Jalali S, DiLoreto D, Little C, Sreedharan A, del Cerro C, Rao GN: Fetal neural retinal grafts into human retinitis pigmentosa (RP). Soc Neurosci 1996; 22:319.

20Silverman MS, Hughes SE: Transplantation of photoreceptors to light-damaged retina. Invest Ophthalmol Vis Sci 1989;30:1684±1690.

21Kaplan HJ, Tezel TH, Berger AS, Wolf ML, Del-Priore LV: Human photoreceptor transplantation in retinitis pigmentosa: A safety study. Arch Ophthalmol 1997;115:1168±1172.

22Medawar PB: Immunity to homologous grafted skin. III. The fate of skin homografts transplanted to the brain, to the subcutaneous tissue, and to the anterior chamber of the eye. Br J Exp Pathol 1948;29:58.

23Kaplan HJ, Streilein JW, Stevens TR: Transplantation immunology of the anterior chamber of the eye. II. Immune response to allogeneic cells. J Immunol 1975;115:805±810.

24Zhang X, Bok D: Transplantation of retinal pigment epithelial cells and immune response in the subretinal space. Invest Ophthalmol Vis Sci 1998;39:1021±1027.

25Larsson J, Juliusson B, Holmdahl R, Ehinger B: MHC expression in syngeneic and allogeneic retinal cell transplants in the rat. Graefe's Arch Clin Exp Ophthalmol 1999;237:82±85.

26Ehinger B, Bergström A, Seiler M, Aramant RB, Zucker CL, Gustavii B, Adolph AR: Ultrastructure of human retinal cell transplants with long survival times in rats. Exp Eye Res 1991; 53:447±460.

27Ghosh F, Arnér K, Ehinger B: Transplant of full-thickness embryonic rabbit retina using pars plana vitrectomy. Retina 1998;18:136± 142.

28Kelly SJ: Studies of the developmental potential of 4- and 8-cell stage mouse blastomeres. J Exp Zool 1977;200:365±376.

29Germer A, Kuhnel K, Grosche J, Friedrich A, Wolburg H, Price J, Reichenbach A, Mack AF: Development of the neonatal rabbit retina in organ culture. 1. Comparison with histogenesis in vivo, and the effect of a gliotoxin (alpha-ami- noadipic acid). Anat Embryol 1997;196:67± 79.

30Stone J, Holländer H, Dreher Z: `Sunbursts' in the inner plexiform layer: A spectacular feature of Müller cells in the retina of the cat. J Comp Neurol 1991;303:400±411.

31Reichenbach A, Stolzenburg JU, Eberhardt W, Chao TI, Dettmer D, Hertz L: What do retinal Müller (glial) cells do for their neuronal `small siblings'? J Chem Neuroanat 1993;6:201±213.

32Willbold E, Brummendorf T, Rathjen FG, Schwarz H, Weiss B: The neural cell recognition molecule F11 is expressed on Müller cells and Schwann cells in vitro. J Brain Res 1997; 38:71±80.

33Campenot RB: Local control of neurite development by nerve growth factor. Proc Natl Acad Sci USA 1977;74:4516.

34Lund RD, Hankin MH: Pathfinding by retinal ganglion cell axons: Transplantation studies in genetically and surgically blind mice. J Comp Neurol 1995;356:481±489.

35Ghosh F, Bruun A, Ehinger B: Immunohistochemical markers in well laminated retinal transplants. Ophthalmic Res 1999;31:5±15.

36Ghosh F, Larsson J, Wilke K: MHC expression in fragment and full-thickness allogeneic embryonic retinal transplants. Submitted.

37Ghosh F, Johansson K, Ehinger B: Long-term full-thickness embryonic rabbit retinal transplants. Invest Ophthalmol Vis Sci 1999;40: 133±142.

38Ghosh F, Bruun A, Ehinger B: Graft-host connections in long-term full thickness embryonic rabbit retinal transplants. Invest Ophthalmol Vis Sci 1999;40:126±132.

39Ghosh F, Juliusson B, Arnér K, Ehinger B: Partial and full-thickness neuroretinal transplants. Exp Eye Res 1999;68:67±74.

40Ames A III, Nesbett FB: In vitro retina as an experimental model of the central nervous system. J Neurochem 1981;37:867±877.

41Juliusson B, Mieziewska K, Bergström A, Wilke K, Van-Veen T, Ehinger B: Interphotoreceptor matrix components in retinal cell transplants. Exp Eye Res 1994;58:615±621.

42Casini G, Grassi A, Trasarti L, Bagnoli P: Developmental expression of protein kinase C immunoreactivity in rod bipolar cells of the rabbit retina. Vis Neurosci 1996;13:817±831.

43del Cerro MP, Snider RS: Studies on the developing cerebellum: Ultrastructure of the growth cones. J Comp Neurol 1968;133:341±362.

44Seiler MJ, Aramant RB: Intact sheets of fetal retina transplanted to restore damaged rat retinas. Invest Ophthalmol Vis Sci 1998;39:2121± 2131.

45Schuschereba ST, Silverman MS: Retinal cell and photoreceptor transplantation between adult New Zealand red rabbit retinas. Exp Neurol 1992;115:95±99.

46Gross JG, Schneiderman TE, Wiley CA, Berge- ron-Lynn G, Faber DW, Freeman WR: Experimental endoretinal biopsy. Am J Ophthalmol 1990;110:619±629.

47Hutchins RK, D'Amico DJ, Casey V-NJ, Morin B: Experimental retinectomy in the rabbit. Retina 1990;10:72±77.

48Miller B, Ryan SJ: Experimental vitrectomy: A lens-sparing technique in the rabbit eye. Arch Ophthalmol 1985;103:1382±1384.

49Brinkman CJ, Otto AJ, Breebaart AC: Ocular inflammatory activity following different techniques of lens extraction and vitrectomy in rabbits. Curr Eye Res 1990;9:1137±1140.

50Stern WH, Fisher SK, Anderson DH, O'Donnell JJ, Erickson PA, Lewis GP, Guerin CJ, Borgula GA, McDermott MA: Epiretinal membrane formation after vitrectomy. Am J Ophthalmol 1982;93:757±772.

51Johnson RN, Balyeat E, Stern WH: Heparin prophylaxis for intraocular fibrin. Ophthalmology 1987;94:597±601.

52Fry KR, Tavella D, Su YY, Peng YW, Watt CB, Lam DM: A monoclonal antibody specific for retinal ganglion cells of mammals. Brain Res 1985;338:360±365.

53Johnson NF, Grierson I: Post-mortem changes in the rabbit retina. Acta Ophthalmol 1976;54: 529±541.

54Silverman MS, Hughes SE, Valentino TL, Liu

Y:Photoreceptor transplantation: Anatomic, electrophysiologic and behavioral evidence for the functional reconstruction of retinas lacking photoreceptors. Exp Neurol 1992;115:87±94.

55Reichenbach A, Ziegert M, Schnitzer J, PritzHohmeier S, Schaaf P, Schober W, Schneider

H:Development of the rabbit retina. V. The question of `columnar units'. Dev Brain Res 1994;79:72±84.

56Greiner JV, Weidman TA: Embryogenesis of the rabbit retina. Exp Eye Res 1982;34:749± 765.

57Vanselow J, Muller B, Thanos S: Regenerating axons from adult chick retinal ganglion cells recognize topographic cues from embryonic central targets. Vis Neurosci 1991;6:569±576.

58Lewis GP, Linberg KA, Fisher SK: Neurite outgrowth from bipolar and horizontal cells after experimental retinal detachment. Invest Ophthalmol Vis Sci 1998;39:424±434.

59Vaney DI, Young HM, Gynther IC: The rod circuit in the rabbit retina. Vis Neurosci 1991; 7:141±154.

60Santos A, Humayun MS, de Juan E Jr, Greenburg RJ, Marsh MJ, Klock IB, Milam AH: Preservation of the inner retina in retinitis pigmentosa: A morphometric analysis. Arch Ophthalmol 1997;115:511±515.

61Rao K, Lund RD, Kunz HW, Gill TJ III: The role of MHC and non-MHC antigens in the rejection of intracerebral allogeneic neural grafts. Transplantation 1989;48:1018±1021.

62Hutchinson IV: Cellular mechanisms of allograft rejection. Curr Opin Immunol 1991;3: 722±728.

63Lund RD, Banerjee R, Rao K: Microglia, MHC expression and graft rejection. Neuropathol Appl Neurobiol 1994;20:202±203.

64Litchfield TM, Whiteley SJ, Yee KT, Tyers P, Usherwood EJ, Nash AA, Lund RD: Characterisation of the immune response in a neural xenograft rejection paradigm. J Neuroimmunol 1997;73:135±144.

65 Mannon RB, Coffman TM: Immunologic

mechanisms of transplant rejection. Curr Opin

Nephrol Hypertens 1992;1:230±235.

66Oganesian A, Bueno E, Yan Q, Spee C, Black J, Rao NA, Lopez PF: Scanning and transmission electron microscopic findings during RPE wound healing in vivo. Int Ophthalmol 1997; 21:165±175.

67Aramant RB, Seiler MJ: Human embryonic retinal cell transplants in athymic immunodeficient rat hosts. Cell Transplant 1994;3:461± 474.

68Mueller-Jensen K, Machemer R, Azarnia R: Autotransplantation of retinal pigment epithelium in intravitreal diffusion chamber. Am J Ophthalmol 1975;80:530±537.

69Nicholas MK, Stefansson K, Antel JP, Arnason BG: An in vivo and in vitro analysis of systemic immune function in mice with histologic evidence of neural transplant rejection. J Neurosci 1987;18:245±257.

70Choudhury A, Pakalnis VA, Bowers WE: Characterization and functional activity of dendritic cells from rat choroid. Exp Eye Res 1994;59: 297±304.

71McMenamin PG: The distribution of immune cells in the uveal tract of the normal eye. Eye 1997;11:183±193.

72Yang P, de Vos AF, Kijlstra A: Macrophages and MHC class II positive cells in the choroid during endotoxin induced uveitis. Br J Ophthalmol 1997;81:396±401.

73Lawrence JM, Morris RJ, Wilson DJ, Raisman G: Mechanisms of allograft rejection in the rat brain. Neuroscience 1990;37:431±462.

74Wakabayashi T, Onoda H: Interdigitating reticulum cells in human renal grafts. Virchows Arch A Pathol Anat Histopathol 1991;418: 105±110.

75Janeway CA, Travers P: Immunobiology: The Immune System in Health and Disease, ed 3. London, Current Biology Ltd/Garland Publishing Inc, 1997, vol 12, p 29.

Key Words

Rabbit W Choroid W Retinal cell transplants W Ultrastructure

Abstract

Purpose: Allogenic rabbit-to-rabbit retinal cell transplants survive in the choroid, which is not as expected because it has not been shown that this is an immuneprivileged site. We have therefore examined the ultrastructure of such transplants, looking for features that might explain the phenomenon. Methods: Rabbit retinal tissue fragment transplants were produced with previously described methods. The donor age was 15 days and the transplants were examined by standard electron microscopy when the transplants were 1±2 months (3 transplants) or 3±4 months old, of postconception age (3 transplants). Results: The transplants survived and developed as expected from previous observations. Rosettes were seen, but they were not as common as in transplants produced with the same technique in the subretinal space of rabbits. Photoreceptor outer segments were not seen in the transplants. At 1 month, there was an incomplete sheath of Müller cells around the transplants, and a complete one at 3±4 months. There was also a well-developed basement membrane around the transplant at 3±4 months, but less so at 1 month. Blood vessels did not enter the transplant. The fenestrations in the choriocapillaris were not affected as long as the pigment epithelium was normal. Conclusions: The

enclosure of the transplants by Müller cells might help to insulate them from the immune system of the host, but it is a late phenomenon and it is not likely to have much effect for the first few weeks after the transplantation. We suspect that either the rabbit choroid is an immune-privi- leged site, even though there is no previous direct evidence for this, or that the retinal tissue itself is responsible for the prolonged survival at this site.

Copyright © 2000 S. Karger AG, Basel

Introduction

Fetal retinal cells can be transplanted to the interior of the eye [for reviews, see 1±4] where they continue to develop, differentiating into mature cells. It is not necessary for the transplant to have the same cell recognition antigens as the host, and it is generally enough that the transplant comes from the same species as the host. The interior of the eye thus appears to be immunologically relatively inert.

The reasons for the special immune-privileged status of the interior of the eye are not fully known and are likely to be complex. One of them is that there is an active downregulation of the immune responses in the interior of the eye (the ACAID system [5]), making it immunologically privileged. The system appears to be present in mammals only [6]. It is regarded as a dynamic state in which the immune response to ocular antigens is modi-

ABC

Fax + 41 61 306 12 34 E-Mail karger@karger.ch www.karger.com

fied by the eye itself. The immune deviation involves the appearance of unique immunoregulatory factors in the aqueous humor. Antigen-presenting cells in the eye pick up some antigen and leave the eye through Schlemm's canal to enter the blood stream. In the spleen, they activate a unique spectrum of antigen-specific T and B cells. When returning to the eye these cells induce a specific ocular immune tolerance to the antigen. There is no systemic delayed hypersensitivity nor do any complementfixing antibodies appear in the circulation. This immune privilege may be the eye's way of protecting its vital functions from immunopathogenic injury.

The cells of the pigment epithelium of the eye are connected by tight junctions, which prevent the free movement of water and solutes across the epithelium. The endothelial cells of the retinal blood vessels also constitute a barrier for most substances. This is often referred to as the blood-retina barrier or the blood-aqueous barrier. The choroid is thus walled off from the retina by the pigment epithelium, and since it is highly vascularized and since its capillaries are fenestrated, blood components have rapid access to it [7]. Transplants with cell and tissue recognition antigens different from those of the host might therefore be expected to survive poorly. Preliminary experimental results in the monkey by Eichhorn et al. [8] suggest that when the uveoscleral outflow is increased with prostaglandin F2· isopropyl ester, the ACAID reaction is no longer inducible. This may be a direct effect of the prostaglandin, or the uvea may not be an immuneprivileged site in this species. In rabbits, on the other hand, Bergström [9] noted that transplants develop and survive also in the choroid, suggesting that the environment is for some reason more protective in this species than in, for instance, rats in which transplants have not been seen to survive [Ehinger and Larsson, unpubl. data]. We have therefore studied the ultrastructure of retinal cell transplants to the rabbit choroid, examining possible anatomical features that might protect them. There is no previous such study available. We have found certain anatomical barriers that appear to play a role, but it seems unlikely that they are the only factors involved.

Materials and Methods

Transplants were produced in outbred pigmented rabbits as previously described by Bergström et al. [10, 11]. The procedure is a development of the ones described by Lopez et al. [12], Gouras et al. [13] and by Turner and Blair [14]. All animals were purpose-bred for experimental laboratory use and were obtained from certified local breeders. They were of a mixed rural brown strain. Six transplants have been examined in the electron microscope, all produced with

tissue obtained from fetuses taken 15 days after conception. Four fetal retinas were used for each transplantation. Three transplants were obtained 3±4 weeks after the transplantation (28, 28 and 21 days) and three after 3±4 months (109, 116 and 123 days).

The host rabbits were sacrificed with an overdose of barbiturates. The transplanted eye was enucleated and bisected a little anteriorly of the equator. Small pieces of the entire posterior segment containing sclera, choroid and retina were fixed according to Ito and Karnowsky [15]. The tissue pieces were embedded in Epon, sectioned and stained with uranyl acetate according to standard electronmicroscopic procedures. Sections taken from different parts of the transplants and surrounding tissues were examined in a Zeiss 902 EM microscope.

The animals were treated according to the ARVO convention for the use of animals in ophthalmic and vision research and according to pertinent Swedish laws and regulations. Appropriate permits for the work were given by the Swedish government committee on animal experimentation ethics.

Results

Transplants were readily identified in the choroid, having a morphology that is distinctly different from that of the choroid. They were composed of tightly packed cells with immature, usually oval nuclei, as has been previously described by e.g. Bergström et al. [16]. The three 1-month transplants were in places thicker than the regular choroid, leaving only narrow strips of it on either side. In 2 cases the transplants partly extended into the subretinal space. The 4-month transplants were thinner. In 1 case, the transplant was seen to reach into the vitreous through the retinotomy made at the time of surgery. In 2 cases, parts of the transplant extended from the choroid into the subretinal space.

The overall morphology of the transplants was the same irrespective of their location in the choroid. Rosettes were present in the choroidal transplants, but they were less prominent than what e.g. Sharma et al. [3] and others have usually seen in standard 2- or 4-week subretinal transplants. The cells tended to form whorls and small sheets, but no regular lamination was apparent (fig. 1).

It was not the purpose of this study to examine the different cell types that might be present in the transplant. Nevertheless, it was clear from the morphological observations that photoreceptors, amacrine cells and bipolar cells had developed, because the synaptic structures and other markers for such cells were easily and repeatedly observed (fig. 2±4). In both 1-month and 4-month transplants, neuropil regions were regularly seen with nerve cell processes of a morphology similar to that seen in the regular inner plexiform layer (fig. 4). Photoreceptor outer segments were rudimentary and not as common as in