Hepatocyte Growth Factor and its Role in the Pathogenesis of Retinal Detachment

Abstract

purpose. Hepatocyte growth factor (HGF) regulates barrier function of retinal pigment epithelial (RPE) cells. The purpose of this study was to determine whether overexpression of HGF in the RPE induces retinal detachment (RD).

methods. E1/E3-deleted adenoviral vectors encoding HGF (Ad CMV.HGF), green fluorescent protein (Ad CMV.GFP), or connective tissue growth factor (AdCMV.CTGF) were injected subretinally in adult pigmented rabbits (5 × 10⁴ plaque-forming units [pfu]/eye). Animals were observed for up to 28 days with fundus photography. HGF expression in the retina and vitreous was determined using immunohistochemistry, and ELISA. Histopathologic examinations were performed with light and electron microscopy.

results. Control eyes injected with AdCMV.GFP showed GFP expression almost exclusively in the RPE monolayer. Eyes injected with AdCMV.HGF showed strong HGF immunopositivity in RPE cells at the injection site. Elevated HGF levels were found in the vitreous peaking at postinjection day 7, diminishing to baseline by postinjection day 28. Eyes injected with AdCMV.HGF developed chronic RD and chronic inflammation in the choroid within the time frame of HGF expression. Groups of proliferating RPE cells were seen in the subretinal space in the region of the RD, and in some cases multilayered cellular membranes developed. No RD and minimal morphologic changes were seen in the eyes injected with AdCMV.GFP or AdCMV.CTGF.

conclusions. Overexpression of HGF in RPE induces chronic, serous RD with subretinal proliferation of RPE. This work provides insight into the pathogenesis of RD and suggests that HGF should be further investigated as a target for therapeutic intervention in RD.

Retinal detachment (RD) is characterized by separation of the neural retina from the retinal pigment epithelium (RPE) with collection of fluid within the intervening subretinal (interphotoreceptor) space. ¹ In the normal adult eye, RPE cells form a nonproliferating monolayer of polarized cells that are essential for the maintenance and survival of photoreceptors. ² Detachment of the retina from the RPE surface initiates a complex series of molecular and cellular changes in both the retina and RPE. ^{1 3 4} Based on the forces that induce the detachment, four major types of RD have been described: rhegmatogenous (RRD), traction, combined traction-rhegmatogenous, and exudative or serous (SRD). ¹ In RRD, a full-thickness break in the retina allows liquefied vitreous fluid to leak under the retina and separate the retina from the RPE. Traction RD occurs when the retina is pulled off the RPE by tractional forces without a retinal tear. In contrast, SRD is characterized by an accumulation of fluid under the retina without retinal tears or traction and is a consequence of a wide variety of vascular, inflammatory, and neoplastic diseases of the retina. ¹  

Human tissues from pure RD are difficult to obtain, because most surgical specimens are complicated by treatments or secondary disorders such as proliferative vitreoretinopathy (PVR). Animal models of RD are therefore of great value in determining pathogenetic mechanisms and to test novel therapeutic approaches to the disease. Although animal models of RRD and traction RD have been developed using ocular trauma, surgical tearing of the retina, intravitreal injection of fibroblasts or RPE, or transgenic overexpression of growth factors, ^{1 5 6} it has been more problematic to model SRD. Occlusion of the choroidal circulation, injury of the RPE and choriocapillaris without occlusion, and toxin-induced damage to the choroidal endothelium all induce SRD ^{7 8 9 10 11 12} ; however, these models are associated with severe damage to the RPE and/or choriocapillaris. A less destructive model of RD has been developed by infusing fluid under the retina in cats in which the lens and vitreous have been removed. ¹³ The artificial detachment allows detailed analysis of the secondary effects on the retina and RPE. 

Growth factors and cytokines are thought to play important roles in the maintenance of normal retinal function and in the retinal dysfunction that accompanies retinal detachment. Recently, we reported that hepatocyte growth factor (HGF), a glycoprotein that mediates epithelial–mesenchymal interactions in many tissues, ^{14 15 16} has autocrine–paracrine activity in RPE ¹⁷ and is a major regulator of RPE barrier function. ¹⁸ These studies led to our hypothesis that HGF may have the ability to induce retinal detachment in vivo in the absence of a retinal break. Although we and others have implicated HGF in the pathogenesis of RRD, and RRD complicated by PVR, ^{19 20 21 22} there was no direct evidence that HGF induces RD. 

Herein, we report that adenovirus-mediated overexpression of HGF in RPE induces chronic retinal detachment within the time frame of HGF overexpression. The detachment is associated with separation, migration, and proliferation of RPE cells, and inflammatory responses in the choroid. This represents an important new model of RD that allows detailed study of the role of HGF in RD, permits new insights into the secondary changes in the retina, and facilitates preclinical testing of novel therapeutic interventions for patients with SRD. 

Materials and Methods

Recombinant Adenovirus

Recombinant adenoviruses based on adenoserotype-5, lacking the E1/E3 region and expressing human HGF or GFP (kind gifts of Katherine P. Ponder, Washington University, St. Louis, MO) ²³ or connective tissue growth factor (CTGF, the kind gift of FibroGen Inc, South San Francisco, CA), and driven by a cytomegalovirus (CMV) promoter were used. Virus was generated to titers of 5 × 10¹⁰ plaque-forming units (pfu)/mL. For subretinal injections, viral stocks were diluted with sterile saline to achieve the desired titers. 

Subretinal Injection

All experiments were performed in accordance with the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic and Vision Research. One eye from each adult pigmented rabbit was used in these experiments. The subretinal injection was performed as described previously. ^{24 25} Briefly, the animals were anesthetized by a mixture of ketamine (40 mg/mL) and xylazine (6 mg/mL) delivered intramuscularly. The eye was topically anesthetized with 4% amethocaine drops, and the pupils were dilated with 1% tropicamide and 2.5% phenylephrine hydrochloride drops. The conjunctiva was cut close to the limbus to expose the sclera. A 32-gauge needle was used to make a shelving puncture of the sclera. A thin glass micropipette attached to a 250-μL glass syringe was then passed through this hole in a tangential direction under an operating microscope. The micropipette was gently touched to the retina at a distance of 2 to 3 disc diameters from the optic disc. A total volume of 50 μL of AdCMV.HGF, AdCMV.GFP, AdCMV.CTGF, or recombinant human HGF (R&D Systems, Minneapolis, MN) was injected, raising a circular bleb. The success of each subretinal injection was further confirmed by indirect ophthalmoscopy. Photographs of the fundus were taken at various intervals after injection using a handheld camera (fx-50 R; Kowa Company Ltd., Tokyo, Japan). 

Measurement of HGF in Vitreous

HGF vitreous concentration was determined using a human HGF enzyme-linked immunosorbent assay (ELISA) kit (Quantikine; R&D System Inc., Minneapolis, MN) according to the manufacturer’s protocol. The lower limit of sensitivity for this assay in our laboratory is 200 pg/mL. Posterior vitreous (50 μL) was collected from enucleated eyes at postinjection days 3, 7, 14, and 28 before processing of the eyes for histology. Vitreous samples were centrifuged at 5000g at 4°C to remove cells, and the supernatant frozen at −70°C. 

Light and Electron Microscopy

For light microscopy and immunohistochemistry, dissected posterior segments of eyes were embedded in optimal cutting temperature (OCT) compound, and 8-μm serial sections of the injection area were prepared. Immunoperoxidase staining of retinal sections was performed with monoclonal anti-HGF antibody, anti-cytokeratin antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), anti-CD4, anti-CD8, or anti- macrophage antibody (Spring Valley, Inc., Woodbine, MD). Sections were stained for apoptotic cells with a peroxidase in situ apoptosis detection kit (ApopTag; Intergen Company, Purchase, NY). For electron microscopy, eyes were fixed in half-strength Karnovsky fixative (2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M phosphate buffer, pH 7.2–7.4) at 4°C for 48 hours, postfixed in 1% osmium tetroxide for 2 hours, dehydrated in a series of graded alcohols, and embedded in mounting medium (Poly/Bed 812; Polyscience, Warrington, PA). One-micrometer-thick sections were cut, stained with toluidine blue, and examined by light microscopy. Ultrathin sections were stained with uranyl acetate and lead citrate and examined with a transmission electron microscope (model EM10B; Zeiss, Thornwood, NY). 

Results

Optimization of Transgene Delivery into Retinal Cells

The dose–response, longevity, and location of transgene expression using AdCMV.GFP were first established. Eight rabbits were subjected to subretinal injection with AdCMV.GFP. Two rabbits of each group were injected with 5 × 10⁴, 5 × 10⁶, 5 × 10⁷, or 5 × 10⁸ pfu/eye, and killed at postinjection day 7 or 28. Funduscopic examination revealed that all retinal blebs resolved within 48 hours as subretinal fluid was absorbed (Figs. 1A 1B) . Eyes injected with the highest viral concentration (5 × 10⁸ pfu/eye) showed outer retinal degeneration and acute choroidal inflammation. In contrast, eyes injected with an AdCMV.GFP concentration of 5 × 10⁶ pfu/eye showed only minor morphologic changes, including slight folding of neural retina within the area of reattachment and mild RPE vacuolization (data not shown). Retinal blebs injected with 5 × 10⁴ pfu/eye were morphologically indistinguishable from those injected with buffered saline alone and were not associated with choroidal inflammation (Fig. 1C) . These results are consistent with previous studies, showing that short-term retinal blebs induce minimal morphologic change in retina-RPE. ²⁶  

As a result of these studies, adenoviral injections were used at a concentration of 5 × 10⁴ pfu/eye. For each vector, at least three rabbits were studied at each time point (days 3, 7, 14, and 28). GFP was strongly expressed in retinal sections of AdCMV.GFP (5 × 10⁴ pfu/eye) injected eyes at postinjection day 7 (Fig. 2) . Consistent with previous reports, ^{27 28} GFP expression was localized to the RPE monolayer with only rare GFP-positive cells in the outer retina. GFP-positive RPE cells were limited to the region of the original bleb area. The sections from eyes at postinjection day 28 showed no detectable GFP in the RPE cells (data not shown), indicating the transient nature of the adenovirus-mediated transgene expression. 

Expression of HGF in RPE Cells

Eyes injected with AdCMV.HGF (5 × 10⁴ pfu/eye) were strongly immunoreactive for HGF at postinjection day 7 within the area of retinal detachment (Fig. 3B) , whereas control eyes without subretinal adenoviral injection or those injected with an equivalent dose of AdCMV.GFP showed no RPE or retinal HGF staining (Fig. 3A) . Vitreous HGF levels were measured at postinjection days 3, 7, 14, and 28 (Fig. 4) . The peak vitreous HGF concentration after AdCMV.HGF injection was reached at day 7 and thereafter declined to baseline levels at day 28. HGF levels were below the limit of detection in vitreous samples from AdCMV.GFP-injected eyes. 

Effect of HGF Expression on the Morphology of the RPE–Retina Interface

The funduscopic, histologic, and ultrastructural appearances of AdCMV.GFP and AdCMV.HGF injected eyes at postinjection day 7 are shown in Fig. 5 . Funduscopic examination of eyes injected with AdCMV.GFP showed normal reattached retina (Fig. 5A) , whereas the AdCMV.HGF injected eyes demonstrated retinal detachment, and areas of depigmentation and pigment aggregation in the detached area. A demarcation line was seen at the boundary of the attached and detached retina (Fig. 5B) . Hyper- and hypopigmentation of the fundus has been shown to be associated with RPE migration and proliferation in the subretinal space. ²⁹  

One-micrometer plastic sections stained with toluidine blue showed essentially normal morphology of RPE and retina in the AdCMV.GFP-injected eye (Fig. 5C) . AdCMV.HGF-injected eyes showed retinal detachment with subretinal fluid and inflammatory cells, indicating loss of RPE barrier function (Fig. 5D) . RPE cells in the region of detachment were separated from one another and mounded with cytoplasm protruding into the subretinal space. Electron microscopy confirmed the normal structure of the RPE-photoreceptor interface in AdCMV.GFP-injected eyes (Fig. 5E) . The RPE cells in the retinal detachment zone of the AdCMV.HGF-injected eyes were mounded in appearance, and their pigment granules were localized in the apical cytoplasm, microvilli, and adjacent subretinal space, suggesting pigment granule extrusion (Fig. 5F) . 

Frozen sections of AdCMV.HGF-injected eyes were examined 1, 2, and 4 weeks after injection. At 1 week after injection (Fig. 6B) , a zone of retinal detachment was seen corresponding to the area of HGF overexpression in the RPE. Consistent with the plastic sections, RPE cells were mounded in appearance, and an inflammatory response was present in vascular choroid. The retina was intact with preservation of the outer nuclear layer. At 2 weeks after injection (Fig. 6C) , the RPE in the region of detachment contained clumps of subretinal pigmented and nonpigmented cells, and chronic choroidal inflammation was still prominent. The outer nuclear layer of the retina unexpectedly remained intact with no apparent loss of photoreceptor cells. The presence of apoptotic cells was studied using TUNEL stain. Only rare, isolated TUNEL⁺ photoreceptor nuclei were seen at days 3, 7, and 14 after RD, consistent with the histologic preservation of the outer nuclear layer (results not shown). Immunoperoxidase staining of the choroid revealed a mixed infiltrate of inflammatory cells. CD4⁺ T cells were predominant (>50%), with the remainder divided among CD8⁺ T cells, macrophages, and rare neutrophils. In the subretinal space, most of the inflammatory cells were macrophages (results not shown). In several of the eyes examined 2 weeks after AdCMV.HGF injection, subretinal multilayered cellular membranes were identified. The cells within these membranes stained strongly for cytokeratin, suggesting that they originated from the RPE (Fig. 7) ; however, their lack of pigmentation and polarity indicate loss of a mature RPE phenotype. By 4 weeks after injection, the retinal detachment had resolved, consistent with loss of HGF expression in retina and vitreous. Sections confirmed retinal reattachment (Fig. 6D) , but the retina showed prominent loss of photoreceptor cell nuclei, degeneration of residual RPE cells, and resolution of the choroidal inflammation. 

To determine whether any of the changes observed in the AdCMV.HGF-injected eyes were due to nonspecific effects of protein secretion into the subretinal space, we examined the effect of a single subretinal injection of recombinant HGF (100 and 500 ng in 50 μL PBS). Subretinal blebs resolved rapidly, and no significant morphologic or funduscopic changes were identified (data not shown). Because HGF has dramatic effects on RPE explants treated in vitro, ¹⁸ these experiments suggest the clearance of HGF from the subretinal bleb. Multiple subretinal injections of recombinant HGF were not attempted due to inflammatory complications associated with repeat surgeries. As a further control, parallel experiments were performed with subretinal injection of AdCMV.CTGF. CTGF is a secreted cysteine-rich heparin-binding polypeptide growth factor that participates in fundamental biological processes, including wound healing and fibrosis. ^{19 30} CTGF has strong biological effects on RPE and choroidal endothelial cells in vitro ³¹ and is expressed in RPE in the context of retinal wound healing responses. ^{19 31} In eyes in which AdCMV.CTGF was delivered subretinally at the same concentration as AdCMV.HGF, prominent secretion of CTGF in the subretinal space was found; however, this was not associated with RD or an inflammatory response (3 days, 1 week, or 2 weeks; data not shown), suggesting that adenovirus-mediated protein secretion alone, at the level found in this model, does not induce RD. Although extensive evidence supports the role of CTGF in fibrotic diseases, including ocular disorders affecting the retina, ^{19 30 31} CTGF alone may have little effect on the normal resting RPE monolayer in nondisease conditions. 

Discussion

Adenoviral vectors have been useful for the transduction of retinal tissues. Intravitreal delivery of adenovirus targets ganglion and Müller cells, ^{32 33} whereas subretinal delivery preferentially targets RPE cells. ^{24 25} The targeting of adenovirus to the apical surface of the RPE is thought to occur by ανβ5 integrin–mediated binding and internalization. ^{34 35} Adenoviral protein expression has been reported as early as 48 hours and extending as long as 60 days after exposure to vector in some animal models. ^{36 37} In the experiments presented herein, the adenoviral-induced expression of HGF was first detected in the RPE and vitreous at day 7, remained elevated during weeks 2 and 3, and returned to undetectable levels by day 28. The active secretion of HGF by the RPE and its accumulation in the vitreous is consistent with our previous work showing that HGF is secreted by cultured RPE cells and accumulates in medium supernatant. ¹⁷  

HGF is a pleiotropic growth factor that has mitogenic, morphogenic, and motogenic activities when cultured with epithelial cells. ¹⁵ We have shown in an earlier study that HGF induces disassembly of intercellular junctional complexes and increased paracellular permeability in RPE monolayers. ¹⁸ In this study, overexpression of HGF in RPE layer induced chronic RD within the time frame of HGF expression. Ordinarily, active ion transport by the RPE maintains retinal attachment by dehydrating the subretinal space. ^{38 39} HGF-induced loss of RPE polarity and intercellular junction integrity may impair barrier function and removal of fluid from the subretinal space, leading to retinal detachment. 

Overexpression of HGF in the RPE is associated with migration and aggregation of RPE and formation of multilayered subretinal cellular membranes. Although the vitreous levels of HGF are elevated in these eyes, alterations are limited to the area of detachment, suggesting that apposition of RPE and neural retina is important in maintaining mitogenic and motogenic quiescence of the RPE. Previous studies have shown that experimental retinal detachment can induce a proliferative response in RPE cells within 24 hours that continues as long as 8 weeks if the detachment is maintained. ^{40 41} Molecular mediators of RPE proliferative and migratory responses in vivo are largely unknown. HGF has been shown to regulate RPE cell proliferation and migration in an IL-1β/retinal hole–induced rabbit model of PVR ³⁸ and to induce migration and proliferation in cultured RPE cells. ^{17 42 43} When proliferating RPE cells aggregate and form a subretinal membrane, they may impair reattachment and interfere with photoreceptor function and survival. 

RPE cells infected with AdCMV.HGF secrete a biologically active form of HGF. ²³ In vivo, the morphologic changes induced in the RPE at the site of infection are similar to those seen in vitro after stimulation with recombinant HGF, suggesting that HGF is the mediator of the RD. We also considered the possibility that accumulation of protein in the subretinal space may induce RD in a nonspecific manner; however, previously published work and our current work with adenoviral expression of CTGF suggest that this is very unlikely. Adenoviral ⁴⁴ or adenoassociated viral (AAV) ⁴⁵ expression of a secreted form of vascular endothelial growth factor (VEGF) ^{44 45} in rat RPE in vivo does not result in serous retinal detachments. Although some animals show development of small subretinal hemorrhagic exudates, these effects are thought to be due to VEGF acting on the vasculature. Similarly, subretinal injection of adenovirus expressing pigment epithelium–derived factor (PEDF) stimulates chronic pigmentary change but no retinal detachment. ⁴⁶ The latter study is particularly relevant to this discussion, because PEDF is secreted from the apical domain of the RPE cell to the subretinal space. ⁴⁷ Furthermore, subretinal transplantation of iris pigment epithelial cells that have been transduced ex vivo with an adenoviral vector expressing human PEDF, show prominent long-term expression and secretion of PEDF from the transplanted cells, and transplanted eyes show no evidence of retinal detachment. ⁴⁸ In the present study, a parallel experiment in which AdCMV.CTGF was injected into the subretinal space of the rabbit did not inhibit bleb resolution or induce RD, even though CTGF protein was secreted from the adenovirally infected RPE into the subretinal space. 

Eyes injected with a low-titer AdCMV.GFP generated strong transgene expression without an inflammatory cell response, consistent with the relative immune privilege of the subretinal space. ⁴⁹ In contrast, low-titer AdCMV.HGF-injected eyes showed prominent chronic choroidal inflammation within the time frame of HGF expression, suggesting that the inflammation was due to HGF expression rather that the adenoviral vector. Consistent with these results, transgenic mice overexpressing HGF in the digestive tract experience development of rectal inflammatory bowel disease, and monocytes cultured with HGF show increased invasiveness and chemokine expression. ^{50 51} Although RD itself may induce reactive inflammation, ^{1 2} the extent of the response in this study is much greater than that expected by RD alone. The possibility that HGF overexpression could induce autoimmune uveitis, as a result of an immune response to antigen exposed in degenerating RPE cells, must also be considered. CD4⁺ T-cell predominance is characteristic of many forms of autoimmune uveitis ⁵² ; however, the clear association of the inflammation with elevated HGF levels and the prompt resolution of the inflammation with decline of HGF in the presence of increased RPE degeneration, suggests that this is unlikely. 

Photoreceptor cell death is a typical response to RD and is the major cause of visual loss in these patients. Apoptotic death of photoreceptors in RD can be associated with activation of caspase-3, -7, and -9, or relocalization of apoptosis-inducing factor in the detached retinas. ^{29 53 54 55 56 57 58} This study unexpectedly showed that the outer nuclear layer of eyes with RD due to HGF overexpression remained intact, even after 2 weeks of RD. Although rabbit eyes with experimental RD typically show numerous apoptotic photoreceptor nuclei as early as 3 days after detachment, we were unable to show any more than rare apoptotic photoreceptors over a 2-week period when RD was induced by overexpression of HGF. HGF has been shown to be antiapoptotic in several cell types, ^{15 16 59 60} and this effect has been thought to be a result of caspase inhibition, ⁵⁸ or sequestration of Fas, ⁶¹ or modulation of the nuclear relocalization of apoptosis-inducing factor. ⁶² It has also been shown to have a strong neuroprotective effect in animal models of transient focal cerebral ischemia and retinal ischemia–reperfusion injury. ^{63 64} It is important to establish the mechanism of photoreceptor protection in this model of RD. 

In the current study in vivo overexpression of HGF in the RPE induced profound morphologic changes in retina and choroid. Prolonged RD was associated with migration and aggregation of RPE cells, cellular subretinal membranes, choroidal inflammation, and preservation of photoreceptors. Subsequent loss of HGF expression was associated with reattachment of the retina and loss of RPE cells and photoreceptors. These data suggest that HGF may play a critical role in the pathogenesis of SRD and may be neuroprotective for photoreceptor cells. 

 Supported in part by Grant EY02061 and Core Grant EY03040 from the National Institutes of Health, Bethesda, Maryland; by grants from the Arnold and Mabel Beckman Foundation and FibroGen, Inc.; and by a grant to the Department of Ophthalmology from Research to Prevent Blindness, Inc.

 Submitted for publication April 7, 2003; revised August 10, 2003; accepted September 13, 2003.

 Disclosure: M. Jin, None; Y. Chen, None; S. He, (P); S.J. Ryan, None; D.R. Hinton, FibroGen, Inc. (F, C, P, R)

 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.

 

Funduscopic (A, B) and histologic (C) photographs of the retina after subretinal injection with AdCMV.GFP (5 × 10⁴ pfu/eye). A circular bleb (arrows) was formed immediately after injection (A) and was resolved at postinjection day 3 (B). Histologic examination at the center of the bleb at postinjection day 7 showed minimal morphologic alterations in the RPE and retina and no choroidal inflammation (C). Magnification: (A, B) ×50; (C) ×100.

Figure 2.

 

Frozen sections of retina 7 days after subretinal injection with AdCMV.GFP (5 × 10⁴ pfu/eye). Confocal microscopy reveals GFP expression (green fluorescence) in the RPE layer (A, B). Propidium iodide staining (red) labels nuclei in the outer nuclear layer (ONL) and choriocapillaris (CC) (A). OS identified the outer segments of the photoreceptors. Magnification, ×400.

Figure 3.

 

HGF protein expression was determined by immunohistochemistry. Seven days after subretinal injection with Ad.CMV.HGF, the retina showed HGF immunopositivity in RPE cells (B, arrows), whereas the control eye injected with an equivalent amount of Ad.CMV.GFP showed no HGF staining at a corresponding location (A). Note the retinal detachment and marked inflammatory response in the choroid of the AdCMV.HGF-injected eye (B). Magnification, ×200.

Figure 4.

 

Vitreous HGF levels were measured 3, 7, 14, and 28 days after subretinal adenoviral vector injection using a human HGF ELISA. The peak level of HGF expression was reached at postinjection day 7 and fell to baseline by day 28. HGF was not detected in the vitreous of AdCMV.GFP-injected eyes (lower threshold of detection is 200 pg/mL).

Figure 5.

 

Pathologic changes 7 days after subretinal injection of AdCMV.GFP (A, C, E) or AdCMV.HGF (B, D, F). Fundus photographs of eyes injected with AdCMV.GFP showed reattachment of the retina without significant pigmentary change (A). In contrast, eyes injected with AdCMV.HGF (B) showed chronic retinal detachment, areas of depigmentation and pigment aggregation, and a demarcation line between attached and detached retina (arrows). One-micrometer plastic sections from AdCMV.GFP-injected eyes showed intact photoreceptor and RPE layers (C). One-micrometer sections from the AdCMV.HGF-injected eyes showed retinal detachment and choroidal inflammation (D). The RPE cells were separated from one another and showed mounded morphology (D, arrows). Electron microscopic view of an AdCMV.GFP-injected eye showing that the outer segments of the photoreceptors reestablished appropriate relationships with the RPE (E). The AdCMV.HGF-injected eye showed retinal detachment and mounded RPE cells with pigment granules confined to microvilli and apical cytoplasm (F). OS, outer segment; RPE, retinal pigment epithelium; BM, Bruch’s membrane; CC, choriocapillaris. Magnification: (A, B) ×50; (C, D) ×200; (E, F) ×2500.

Figure 6.

 

Light micrographs of frozen sections from an AdCMV.GFP-injected eye (A), and AdCMV.HGF-injected eyes at 1 (B), 2 (C), and 4 (D) weeks. Control eye 7 days after subretinal injection of AdCMV.GFP showed retinal reattachment with intact RPE and photoreceptors (A). Seven days after AdCMV.HGF injection, the retina showed retinal detachment, RPE cells had a mounded morphology and some had lost their pigmentation, and an inflammatory response was visible in the choroid (B). Fourteen days after AdCMV.HGF injection, the chronic retinal detachment persisted, choroidal inflammation was still present, and groups of nonpigmented cells extended from the RPE layer to the subretinal space (C, arrow). Note that the outer nuclear layer was well preserved at both postinjection days 7 and 14 (A, B). Twenty-eight days after AdCMV.HGF injection, the retina had reattached and the choroidal inflammation had resolved (D). Note the prominent loss of photoreceptor and RPE cells (D). Hematoxylin and eosin stain. Magnification, ×200.

Figure 7.

 

Immunoperoxidase stain for cytokeratin expression in retinal sections from day 14 control (AdCMVGFP) and AdCMV.HGF-injected eyes. The eye injected with AdCMV.GFP (A) showed an intact cytokeratin-positive RPE cell monolayer (cytokeratin is identified by red chromogen). The eye injected with AdCMV.HGF (B) contained a cellular subretinal membrane composed of depigmented, cytokeratin-positive cells, indicating their origin from the RPE. Magnification, ×200.