Graefe's Arch Clin Exp Ophthalmol (1990)228:46a 466 Graefe's Archive lot Clinical and Experimental

Ophthalmology © Springer-Verlag 1990

Panretinal photocoagulation aggravates experimental proliferative vitreoretinopathy* Peep V. Algvere 1, Kerstin Halln/is 1, Eva Dafgfird 2, and Anders H66g g

1 Department of Ophthalmology and z Department of Pathology, Karolinska Institute and Hospital, Box 60500, S-10401 Stockholm, Sweden

Received November 15, 1989 / Accepted February 20, 1990

Abstract. The effects o f panret inal pho tocoagu la t ion (PRP) on intravitreal prol i ferat ion were evaluated in an experimental model for proliferative v i t reore t inopathy (PVR). Thir ty- three p igmented rabbits underwent a rgon laser P R P in one eye. Cul tured fibroblasts (2 x 105 or 4 × 105 cells) were implanted into the intact vitreous o f bo th eyes o f each animal either 3 days after P R P (when acute laser lesions were present) or 4 weeks after P R P (scarred laser lesions). P V R was assessed by indirect o p h t h a l m o s c o p y for 4 -8 weeks. Histological examina- t ion included staining with monoc lona l antibodies against glial fibrillary acidic prote in (GFAP) . In bo th groups significantly more severe stages o f P V R devel- oped after P R P than in the controls. Total retinal detach- ments ensued in 13 pho tocoagu la t ed eyes versus five controls. P R P induced invasion o f macrophages , prolif- erat ion o f retinal p igment epithelium and a conspicuous Mfiller cell response, enhancing in t raocular inf lamma- tion, which st imulated intravitreal prol i ferat ion and ag- gravated PVR.

Introduction

Pho tocoagu la t i on is widely used in the t rea tment o f reti- nal breaks and detachments . Focal adhesions between the neural ret ina and retinal p igment epithelium (RPE) are thereby produced. In vascular occlusive diseases, such as retinal b ranch vein occlusion and diabetic retino- pathy, pho tocoagu la t i on m a y inhibit the prol iferat ion o f f ibrovascular tissue. The pathogenet ic mechanisms for the mode o f pho tocoagu la t i on act ion m a y be mani- fold and are no t yet fully unders tood.

However , our clinical experience indicates that in ret- inal de tachment (RD) with proliferative vi treoret ino-

* Presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology, Sarasota, Florida, 2 May 1989 Offprint requests to." P.V. Algvere, St. Erik's Eye Hospital, Flem- inggatan 22, S-112 82 Stockholm, Sweden

pa thy (PVR) pho tocoagu la t i on may in fact stimulate the g rowth o f preretinal tissue. In order to evaluate these effects o f pho tocoagula t ion , experimental R D and P V R were p roduced by intravitreal implanta t ion o f cultured fibroblasts [1]. In this paper the results with respect to the development and progression o f R D are reported and the histological features, which show a conspicuous glial cell response, are described.

Materials and methods

Thirty-three pigmented rabbits (initial weight 1.8-2.8 kg) under- went panretinal photocoagulation (PRP) in one eye, using a blue- green argon laser and general anesthesia (ketamine 35 mg/kg and xylazine 5 mg/kg body weight). Laser burns were delivered through a corneal three-mirror contact lens, the eyes receiving 810 to 1070 lesions (0.15 to 0.25 W, duration 0.1 s, size 0.5 ram, Fig. 1).

Heterologous fibroblasts, originating from human embryonic lung fibroblasts (WI 38, Flow Laboratories) were used. The fibro- blasts were cultured in modifed Eagle's medium (MEM) supple- mented with 10% (v/v) newborn calf serum and antibiotics (penicil- lin 0.2 mg/ml, streptomycin 0.2 mg/ml). Stocks were maintained under humidified 5% CO2/95% air at 37 ° C. The cells were har- vested with trypsin/EDTA at 4 ° C and passaged at a split ratio of 1 : 2. For experimental purposes cells in passage 20-30 were har- vested when the cells were confluent. The cell solution was thereaf- ter diluted in MEM to a concentration of 1 x 106 or 2 x 106 cells/ml. A volume of 0.2 ml was injected intravitreally.

Both eyes of each rabbit were injected intravitreally with fibro- blasts through the superior sclera I 2 mm from the limbus, using aseptic conditions and microscopic control, under general anesthe- sia. Prior to the injection a paracentesis was performed in order to reduce the intraocular pressure. Eight rabbits received 4 x 105 cells (included in group II, see next paragraph), and the rest of the animals 2 x 105 fbroblasts.

In ten rabbits the cells were implanted intravitreally 3 days after PRP when the laser-treated eye showed acute laser lesions (group I), and in 23 animals the cells were implanted after 4 weeks when the treated eye displayed scarred laser lesions (group II).

The rabbits were examined once a week using a binocular indirect ophthalmoscope for 4-8 weeks, depending on whether bi- lateral total RD had developed. Experimental PVR was graded into five stages, according to Fastenberg et al. [5]. The animals were killed by overdoses of pentobarbital, their eyes subsequently enucleated and fixed in a buffered 4% solution of formaldehyde.

462

Fig. 1. Fundus photograph showing medullary wing and photocoagulated retina with laser lesions in the living rabbit ] h after treatment

The globes were dissected under an operating microscope, sec- tioned, and processed for histological examination.

Selected sections were stained with glial fibrillary acidic protein (GFAP) antibodies. Endogen peroxidase activity was eliminated from the sections by a 0.6% H202 solution in methanol/distilled water (80/20 by volume) for 30 min. Nonspecific staining was coun- teracted by incubation of the sections with 1.5% horse serum in TBS-buffer for 20 min. Monoclonal antibodies raised against hu- man GFAP (M761 Dakopatts Laboratories) were used in a 1:20 dilution and incubated overnight at 4°-8 ° C. To identify the anti- gen-antibody binding, the avidin-biotin-peroxidase complex meth- od was used (Vectastain, Vector Laboratories). Diaminobenzidine was used as chromogen.

The sections were counterstained with Mayer's hematoxylin. Positive controls were the rabbit cerebellar tissue and optic nerve head. Negative controls were normal untreated rabbit retina and sections of treated rabbit eyes in which the antibody was replaced by serum from nonimmunized mice.

R e s u l t s

The fibroblasts implanted intravitreally were seen to grow in bo th eyes o f all o f the rabbits. The eye o f each animal which was no t pho tocoagu la ted (the control eye), showed considerable changes o f P V R in a large number o f cases in response to the fibroblasts (Table 1).

In ten animals which had monocu la r acute laser le- sions when the fibroblasts were implanted (3 days after PRP), total R D developed in five pho tocoagu la ted eyes versus one control eye. The experiments showed that the laser-treated eyes gradual ly advanced to more severe changes than the controls (Student 's t-test P = 0.04).

Similar results were obta ined in eyes with scarred laser lesions (cells implanted 4 weeks after PRP) ; more advanced sequelae o f intravitreal prol iferat ion devel- oped following P R P than in the controls (23 rabbits).

Table 1. Experimental proliferative vitreoretinopathy (PVR) in eyes implanted with fibroblasts 3 days (group I) and 4 weeks (group II) after photocoagulation and in controls

Photo- Stage of experimental PVR a Total coagulation no. of eyes

1 2 3 4 5

Group I No 0 3 1 5 1 10 Yes 0 0 3 2 5 l0 Group H No 5 4 3 7 4 23 Yes 1 2 3 9 8 23

a Stage 1, intravitreal membrane; stage 2, focal traction, blood ves- sel elevation; stage 3, localized detachment of medullary ray; stage 4, retinal detachment in posterior pole; stage 5, total retinal detachment

This difference was statistically significant (Student 's t- test P - - 0 . 0 1 ; Wi lcoxon ' s test P = 0 . 0 2 ) . In this series, eight pho tocoagu la t ed eyes developed total R D versus four o f the controls. Accordingly, slight vitreous changes, such as vitreous membranes only (stage 1), were more c o m m o n in the controls, and were present in five control eyes versus one laser-treated eye (Table 1).

Thus, P R P did no t reduce the incidence o f R D in any group studied and irrespective o f whether 4 x 105 or 2 x 105 fibroblasts were implanted. In the total case material (33 rabbits), P R P was followed by significantly more severe complicat ions o f intravitreal prol iferat ion (Student 's t-test P = 0 . 0 0 1 ; Wi lcoxon ' s test P = 0 . 0 0 4 ) , with total R D in 13 eyes versus 5 o f the controls.

Histological examinat ion disclosed modera te to se- vere cellular prol i ferat ion and fo rmat ion o f f ibrovascu-

463

Fig. 2. Low-power micrograph of detached retina (R) which is ad- herent to photocoagulation scar at the pigment epithelial level (PE). A firm, cellular structure forms the retinal adhesion (arrow). Five weeks after implantation of fibroblasts which was performed 4 weeks following photocoagulation. (H + E, x 100)

lar tissue in front of the optic nerve head both in eyes with and without PRP. There was a conspicuous glial cell response in the vitreous anterior to the optic nerve head, as demonstrated by the staining with G F A P anti- body (Fig. 4). No attempt was made to quantitate the cellular response in detail, but there was no evidence of inhibition of the cellular growth and migration in the vitreous in response to PRP.

The laser lesion formed a stronghold for adhesion between the neural retina and RPE and was, to a consid- erable extent, resistant to vitreous traction (Fig. 2). However, the laser lesions were often incapable of coun- teracting continued vitreous traction but were broken (Fig. 3).

In the laser lesions there was a considerable prolifera- tion of RPE and, notably, of glial cells which expressed G F A P immunoreactivity. In these instances, glial cells were seen to make contact with the RPE, forming sub-

Fig. 3. Micrograph of attenuated retinal attachment (open arrow) to subretinal membrane (arrowheads') which is adherent to photo- coagulation scar at the level of the pigment epithelium. R, detached retina. (H + E, x 200)

retinal membranes (Fig. 5). Clusters of inflammatory cells accumulated at the internal limiting membrane (ILM) over the laser lesion. In addition, glial cells were noted to penetrate the ILM and there was an extensive formation of epiretinal membranes which expressed GFAP immunoreactivity (Fig. 6). Areas around the pho- tocoagutation lesions revealed large numbers of elongat- ed cell bodies which stained with G F A P antibody and could be traced from the ILM to the outer limiting mem- brane and, occasionally, to the subretinal space; these cells were identified as Miillerian glia (Fig. 5). The glial elements constituted the predominant cell type in the detached and atrophic retina (Fig. 7).

No affinity for GFAP antibody was found in the retina of the normal (nontreated) control eye, in agree- ment with a previous report [4]. As expected, there was a conspicuous G F A P immunoreactivity of the glial ele- ments (astrocytes) of the optic nerve head (Fig. 4). How- ever, the attached, peripheral retina outside the laser- treated region displayed G F A P immunoreactivity of the Mtiller cells (Fig. 8).

Discussion

The experiments show that PRP does not reduce the incidence of tractional RD in response to intravitreal proliferation induced by fibroblasts. Actually, total RD

464

465

Figs. 4-8. GFAP immunoreactivity in rabbit retina

Fig. 4. Micrograph of optic nerve head (ON) with prepapillary cellular proliferation (arrow) in front of papillary vessels (V). GFAP immunoreactivity (brownish staining) is present in optic nerve (astrocytes) and in prepapillary, proliferating cells (arrow), indicating glial origin. ( x 300)

Fig. 5. Detached retina, locally adhering to photocoagulation scar, and subretinal membrane (open arrows). Elongated Mfiller cell bodies through the whole retina and in subretinal membrane show GFAP immunoreactivity. Photoreceptor outer and inner segments are lost (black arrow). One month after implantation of fibroblasts which was performed 3 days following photocoagulation. ( x 200)

Fig. 6. Area around photocoagulation scar, showing epiretinal membranes and Mfiller cell extensions penetrating internal limiting mem- brane (open arrows). Mfiller cells make contacts to proliferating retinal pigment epithelium (RPE; black arrow). ( x 300)

Fig. 7. Detached and degenerated retina with epiretinal membrane (arrow) and two laser-induced adhesions to RPE (asterisks). Mfillerian glial cells constitute the framework of the degenerated retina, mediate vitreoretinal traction (arrow), and make adhesions to RPE. Eight weeks after implantation of fibroblasts which were injected 4 weeks following photocoagulation. ( x 200)

Fig. 8. Attached retina outside the photocoagulated region, expressing GFAP immunoreactivity of Mfiller cells. Eight weeks after implanta- tion of fibroblasts into eye with acute laser lesions. ( × 300)

466

occurred more often in laser-treated eyes than in the controls.

These results were quite unexpected, since photocoa- gulation produces focal adhesions between the neural retina and RPE (as depicted in Fig. 2). However, the experiments showed that the laser lesions generally were not strong enough to resist vitreoretinal traction but were broken. It is thus reasonable to assume that the contracting forces of the type caused by the vitreo-retinal traction were increased after PRP, since tractional R D was more common in these eyes. It is also conceivable that photocoagulat ion may induce focal retinal attenua- tion which facilitates the format ion of retinal tears and detachment f rom vitreous traction.

The laser lesion in its acute stage displayed a consid- erable inf lammatory activity, including the invasion of macrophages into the retina and vitreous. It is well known that the acute lesion is associated with a break- down of the blood-retinal barrier [10, 11].

Four weeks after PRP, pigment-laden macrophages were still active, and proliferating RPE and glial cells participated in the wound repair. Thus, the inf lammato- ry activity in the scarred laser lesion had not subsided at the time of the intravitreal implantat ion of fibroblasts. There was a conspicuous glial cell response both on the retinal surface, noted as epiretinal membranes expressing GFAP, and in the subretinal space, seen as proliferating tissue sheets. Previous work on postphotocoagulat ion retinal repair in the rabbit has shown macrophages [6] and migrating RPE cells 5 weeks after t reatment [17]. In the rabbit retina, astroglia is found only at the area of the medullary rays [14]. The photocoagulated region outside that area is thus lacking astrocytes and the GFAP-immunoreact ive cells there are almost exclusively Mfillerian glia. It is concluded that Mfiller cells are able to express G F A P in response to intravitreal cellular pro- liferation. The present study confirms the conspicuous activity of Mfillerian glia in the repair of the photocoa- gulation lesion [8].

In large laser lesions, there is a long-lasting, and sometimes permanent, damage to the RPE barrier func- tion as indicated by the penetration of horse-radish per- oxidase [11, 16]. Following breakdown of the blood-ocu- lar barrier, various serum components gain access to the retina and vitreous. These components include fibro- nectin and platelet-derived growth factor, both of which stimulate RPE migration [3, 18].

Inf lammatory cells, such as macrophages [13], and proliferating RPE are likely to stimulate other cells by the secretion of various growth factors. For example, activated macrophages stimulate flbroblast growth [9], collagen synthesis [7], vascular proliferation [12], and RPE migration [2].

A recent report on a cell-injection model of PVR suggests that localized retinal cryopexy (one or five freezes) results in a higher frequency of R D than in con- trols [15]. The present study indicates that laser photo- coagulation induces breakdown of the blood-retinal bar- rier, retinal invasion of macrophages, and proliferation

of RPE and glial (Miiller) cells enhancing intraocular inflammation. The inf lammatory response to PRP, thus, significantly stimulates intravitreal cellular proliferation.

Acknowledgement. This study was supported by the Crown Princess Margareta Foundation and by the O.E. and Edla Johansson Scien- tific Foundation, Stockholm, Sweden.

References

1. Algvere P, Kock E (1976) Experimental fibroplasia in the rabbit vitreous. Retinal detachment induced by autologous fibro- blasts. Graefe's Arch Clin Exp Ophthalmol 199:215-222

2. Burke JM, Twining SS (1987) Vitreous macrophage elicitation: generation of stimulants for pigment epithelium in vitro. Invest Ophthalmol Vis Sci 28 : 1100-1107

3. Campochiaro PA, Glaser BM (1985) Platelet-derived growth factor is chemotactic for human retinal pigment epithelial cells. Arch Ophthalmol 103:576-579

4. Ekstr6m P, Sanyal S, Narfstr6m K, Chader GJ, van Veen T (1988) Accumulation of glial fibrillary acidic protein in Mfiller radial glia during retinal degeneration. Invest Ophthalmol Vis Sci 29:1363-1371

5. Fastenberg DM, Diddle KR, Dorey K, Ryan SJ (1982) The role of cellular proliferation in an experimental model of mas- sive periretinal proliferation. Am J Ophthalmol 93 : 565-572

6. Gloor BP (1974) On the question of the origin of macrophages in the retina and the vitreous following photocoagulation (auto- radiographic investigations by means of 3H-thymidine). Graefe's Arch Clin Exp Ophthalmol 190:183 194

7. Hunt TK, Knighton DR, Thakral KK, Goodson WH, Andrews WS (1984) Studies on inflammation and wound healing: anglo- genesis and collagen synthesis stimulated in vivo by resident and activated macrophages. Surgery 96:48-54

8. Ishigooka H, Hirata A, Kitaoka T, Ueno S (1989) Cytochemi- cal studies on pathological Mfiller cells after argon laser photo- coagulation. Invest Ophthalmol Vis Sci 30: 509-520

9. Martin BM, Gimbrone MA, Unanue ER, Cotran RS (1981) Stimulation of non-lymphoid mesenchymal proliferation by a macrophage-derived growth factor. J Immunol 126:1510-1515

10. Noth JM, Vygantas CH, Cunha-Vaz J (1978) Vitreous fluoro- photometry evaluation of xenon photocoagulation. Invest Oph- thalmol 17:1206-1209

11. Peyman GA, Bok D (1972) Peroxidase diffusion in the normal and laser coagulated primate retina. Invest Ophthalmol 11:35- 45

12. Polverini PJ, Leibovich SJ (1984) Induction of neovasculariza- tion in vivo and endothelial proliferation in vitro by tumor- associated macrophages. Lab Invest 51:635-642

13. Rappolee DA, Mark D, Banda MJ, Werb Z (1988) Wound macrophages express TGF-alfa and other growth factors in vivo: analysis by mRNA phenotyping. Science 241:708-712

14. Schnitzer J (1985) Distribution and immunoreactivity of glia in the retina of the rabbit. J Comp Neurol 240:128-142

15. Sen HA, Robertson TJ, Conway BP, Campochiaro PA (1988) The role of breakdown of the blood-retinal barrier in cell-injec- tion models of proliferative vitreoretinopathy. Arch Ophthal- mol 106:1291-1294

16. Wallow IH (1984) Repair of the pigment epithelial barrier fol- lowing photocoagulation. Arch Ophthalmol 102 : 126-135

17. Wallow IH, Tso MOM, Fine BS (1973) Retinal repair after experimental xenon arc photocoagutation. A comparison be- tween rhesus monkey and rabbit. Am J Ophthalmol 75 : 32-52

18. Yeo JH, Sadeghi J, Campochiaro PA, Green WR, Glaser BM (1986) Intravitreous fibronectin and platelet-derived growth factor. New model for traction retinal detachment. Arch Oph- thalmol 104: 417-421