Ultraviolet-B Enhances Corneal Stromal Response to 193-nm Excimer Laser Treatment

Ultraviolet . .B Enhances Corneal Stromal Response to 193 . . nm Excimer Laser Treatment Zoluin Zsolt Nagy, MD, 1 ' 2 Paul Hiscott, PhD, 1•3 Berthold Seitz, MD, 1 Ursula Shlotzer#Schrehardt, PhD/ MiklOs Simon, ]r, PhD, 4 Ildik6 Suveges, MD, 2 Gottfried 0. H. Naumann, MD 1 Purpose: The purpose of the study was to evaluate the biomicroscopic, light microscopic, and electron microscopic effects of ultraviolet-B (UV-B) exposure on the outcome of photorefractive keratectomy (PRK). Methods: A total of 24 pigmented rabbits were used in the study. One eye of 16 rabbits received a 193-nm, 45-JJm deep (-5.0 diopter) excimer laser PRK. Twenty-one days after PRK, eight of the laser-treated eyes were exposed to 100 mJ/cm2 UV-B (280315 nm) UV radiation by placing the rabbits in a standard clinically used dermatologic chamber for 7 minutes. Eight PRK-treated rabbits received no further treatment. The remaining eight non-PRK-treated rabbits received 100 mJ/cm2 UV-B only to one eye. Subepithelial haze was assessed before and after UV irradiation. Corneal morphology was assessed 4, 8, 12, and 16 weeks after UV-B exposure, using light microscopic and transmission electron microscopic (TEM) techniques. Results: Untreated eyes exposed to 100 mJ/cm2 UV-B only exhibited photokeratitis for 2 days, but showed no haze and were normal histologically at all intervals. The PRKtreated UV-B irradiated eyes exhibited a significant increase of stromal haze compared to eyes receiving PRK alone. Histologically, the main difference between the UV-B irradiated and nonirradiated post-PRK eyes was the presence of anterior stromal extracellular vacuolization in the UV-B-exposed eyes. The vacuolated foci were confined to the PRK treatment area and showed increased keratocyte density and disorganization of normal collagen lamellae. TEM showed activated keratocytes containing abundant rough endoplasmic reticulum, prominent Golgi zones, and extracellular vacuoles filled with amorphous material. The haze and morphologic changes showed a tendency to incomplete resolution over the period of 16 weeks. Conclusions: The UV-B exposure during post-PRK stromal healing exacerbates and prolongs the stromal healing response and is manifest biomicroscopically by augmentation of subepithelial haze. The findings suggest that excessive ocular UV-B exposure should be avoided during the period of post-PRK stromal repair and that UV-B may modulate the response of tissues to 193-nm excimer laser and perhaps other laser energy in general. Ophthalmology 1997; 104:375-380

Originally received: March 6, 1996. Revision accepted: October 2, 1996. ' Department of Ophthalmology, Friedrich Alexander Universitat, Erlangen-Niimberg, Germany. 2 1st Department of Ophthalmology, Semmelweis Medical University, Budapest, Hungary. 3 Department of Ophthalmology, Liverpool, United Kingdom. 4 Department of Dermatology, Friedrich Alexander Universitat, Erlangen-Niimberg, Germany. Supported by grants from the Deutscher Akademischer Austauschdienst

Deep corneal photoablations with the 193-nm excimer laser are associated with stromal haze and scarring as, for example, after photorefractive keratectomy (PRK) for high degrees of myopia. 1 However, even patients receiv(A/94/16005, Z. Z. Nagy) and the Alexander von Humboldt Foundation (#15233, P. Hiscott). Reprint requests to Zoltan Zsolt Nagy, MD, 1st Department of Ophthalmology, Semmelweis Medical University, H-1081 Budapest, Torno u. 25-29, Hungary.

375

Volume 104, Number 3, March 1997

Ophthalmology

Table 1. Summary of Experimental Design and Results 4 wks

Treated Group 1. Control eyes (n = 8) 2. Only UV-B exposed eyes (n = 8) 3. PRK treated (n = 8)

4. PRK + UVB treated (n = 8)

Haze*

8 wks

Morphology of Anterior Stroma

Haze

12 wks

Morphology of Anterior Stroma

Haze

16 wks

Morphology of Anterior Stroma

Haze

Morphology of Anterior Stroma

Nil

N

Nil

N

Nil

N

Nil

N

Nil

N

Nil

N

Nil

N

Nil

N

0.5 :!:: 0.5

Vacuolation detectable with TEM only Vacuolation detectable with light microscopy and TEM

0.25 :!:: 0.1

N

0.01 :!:: 0.03

N

Nil

N

Wavy, interwoven collagen fibers

0.75 :!:: 0.25

Wavy, interwoven collagen fibers

2.5 :!:: 1.76

1.5 :!:: 0.84

Vacuolation detectable with TEM only

1.0 :!:: 0.56

PRK = photorefractive keratectomy; N = normal; TEM = transmission electron microscopy; N = normal.

* Haze

assessed using Hanna's scale.

ing PRK for mild or moderate degrees of refractive error may have haze and scarring develop. 1 Recently, the authors observed a substantial increase of stromal haze and regression of achieved refractive effect in patients treated for moderate degrees of myopia and subsequently exposed to increased environmental ultraviolet-B (UV-B) radiation. 2 Because UVB (wavelength, 280-320 run) has a biotoxic effect, 2- 8 we postulated that UV-B augments the response of the corneal stroma to PRK. The purpose of the current study was to investigate this hypothesis in an animal model.

Materials and Methods The animals used in this study were treated in accordance with the ARVO Resolution on the Use of Animals in Research. A total of 24 pigmented chinchilla rabbits (range, 2.5-3.0 kg) were used in the investigation. Sixteen of the rabbits received pretreatment intramuscular sedation by means of ketamine (25 mg/kg). The animals were anesthetized with intravenous ketamine (25 mg/kg) plus xylazine (2 mg/kg). Proparacaine hydrochloride was instilled into the right eye of the rabbits and the eyelids were held open with a speculum. Five millimeters of the central epithelium was removed with a crescent knife and a -5.0 diopter (D) (45 ,urn) PRK treatment was performed with the Aesculap Meditec MEL 60 (Heroldsberg, Germany) excimer laser. This ablation depth was chosen because epidemiologic and demographic data indicate that most patients with refractive errors belong to the low myopia range of 0.0 to -6.0 D. 9 During the PRK treatment, the eyes were fixed with a circular myopic vacuum mask around the limbal area and held in place with moderate suction. The operative energy and repetition rate were set at 250 mJ/cm2 and 20 Hz, respectively. The cut rate was 1.0 ,urn per pulse. After PRK treatment, the rabbits received topical antibiotics (oftoxacin) three times a day

376

until complete re-epithelialization (assessed biomicroscopically). The left eyes were untouched. The remaining eight rabbits were not treated with PRK. Twenty-one days post-PRK, 16 of the rabbits (8 previously PRK treated, 8 untreated) were placed in a standard dermatologic UV -light chamber (Waldmann UV 800 IK, Schwenningen, Germany). The wavelength was set at the UV-B range (range, 280-315 nm; peak, 306 and 313 nm). The right eyes were held open by a speculum, and 0.9% saline was instilled onto the eyes every 30 seconds to prevent drying of the corneas. The duration of UV exposure was 7 minutes, during which time the eyes received a total equivalent energy of 100 mJ/cm2 • The remaining eight previously PRK-treated rabbits did not receive UV-B radiation. Anterior stromal haze was assessed biomicroscopically, using the scale of Hanna et al, 10 every 2 weeks until the 16th week after UV-B exposure (Table 1). The results were subjected to statistical analysis with Student's t test. The animals were killed with an overdose of intravenous sodium phenobarbital at 4, 8, 12, and 16 weeks after UV irradiation. The eyes were fixed in 4% paraformaldehyde plus 1% glutaraldehyde. The eyes were dissected, and half of each eye was processed through graded alcohols and embedded in paraffin wax. Sections 4-,um thick were stained with hematoxylin-eosin, periodic acid-Schiff, or Alcian blue. Portions of corneal tissue from the remaining half of the eyes were post-fixed in 4% osmium tetroxide, dehydrated, and embedded in Epon resin. Semithin sections were stained with toluidine blue. Ultrathin sections were stained with uranyl acetate and lead citrate before being evaluated in an EM 9A Zeiss electron microscope.

Results The time to re-epithelialization after excimer PRK varied between 2 and 3 days (range, 2.3 ::!:: 0.85 days) in all

Nagy et al · UV-B and Excimer Laser

Figure I. A, a rabbit eye 1 month after -5.0 diopter photorefractive keratectomy treatment. Subepithelial haze graded 0.5 (the scale of Hanna et al' 0 ). B, a -5.0 diopter photorefractive keratectomized rabbit eye 1 month after ultraviolet-B ir· radiation. Note the subepi· thelia! haze graded 3.5.

rabbits. In rabbits treated with PRK alone, the subepithelial haze was on average 0.5 ::±:: 0.5 (range, 0 to 1.0) at the 21st post-PRK day (Fig lA); 0.25 ::±:: 0.10 at the 8th week; 0.1 ::±:: 0.03 at the 12th week, and less than 0.05 at the 16th week. Six to eight hours after UV-B, the irradiated eyes showed signs of photokeratitis, including conjunctival injection, discharge, photophobia, punctate epithelial erosions, corneal edema, and haze. In the eight eyes treated

only with UV-B, these signs resolved within 48 hours and the eyes exhibited no corneal haze or other biornicroscopically detectable disease. Therefore, the statistical evaluations were made only between rabbits treated with PRK and those treated with PRK plus UV-B. In UV-B-treated eyes that had been subjected previously to PRK, the signs of photokeratitis were more pronounced and prolonged. For example, corneal edema persisted on average for 4.73 ::±:: 1.22 days. Within the

377

Ophthalmology

Volume 104, Number 3, March 1997

Figure 2. A, light microa b scopic section of a -5.0 diopter photorefracti ve keratectomized rabbit eye 4 weeks after photorefractive keratectomy. The histologic section shows no abnormality (periodic acid-Schiff + hematoxylin-eosin staining; magnification, XlOO). B, light microscopic section of a -5.0 diopter photorefractive keratectomized, ultraviolet- B irra.diated eye 4 weeks after ultraviolet exposure. In the central cornea under the epithelium, the stromal architecture is disorganized. Within this area, there appears to be an increased keratocyte density (periodic acid-Schiff eosin; magnification, X 100).

excimer laser-treated 5.0-mm diameter central area, stromal haze averaged 2.5 ::±:: 1.76 (range, 1.5-3.0, P < 0.01) at 4 weeks (Fig 1B); 1.5 ::±:: 0.84, P < 0.01 at 8 weeks; 1.0 ::±:: 0.56 (P not assessable, insufficient numbers at this interval) at 12 weeks; and 0.75 ::±:: 0.25 (P not assessable) at 16 weeks. Light microscopic evaluation of postirradiation UV-Bor PRK-only treated corneas showed that the epithelium was normal histologically at all intervals (Fig 2A). Four weeks postirradiation, two of the eyes treated with PRK and UV-B exhibited central epithelial irregularities (including focal thinning or thickening). At 8 weeks and subsequently, no epithelial irregularities were found in eyes treated with PRK plus UV-B. The stroma of the control- and UV-B-only-treated eyes were normal histologically at all intervals in the experiment. There appeared to be an increase in keratocyte density in the anterior stroma of PRK-treated eyes compared to the normal eyes at 4 weeks. However, this feature was not apparent at subsequent intervals. In all eyes treated with PRK and UV-B, stromal vacuolization was found in the PRK treatment area at 4 weeks. This vacuolization was confined to the anterior 25 to 30 p,m of the stroma (Fig 2B) and also was observed in one eye at 8 weeks. At 12 weeks, the vacuolization was no longer visible with the light microscope. The vacuolated stroma was hypercellular with a disorganization of normal collagen lamellae. In addition, unlike the normal stroma, the vacuolated area stained positively for mucopolysaccharides. No ultrastructural difference was observed in the epithelium of the control, the UV-irradiated, and PRKtreated eyes (Figs 3A, C, E) at any time in the experiment. In eyes treated with PRK plus UV-B, focally fragmented and ruptured epithelial basement membrane was found in the ablation area 4 weeks postirradiation. Hemidesmosomes were detected along the basal layer of the epithelial cells in all eyes, but tonofilaments were more prominent in the control- and UV-only-irradiated eyes than in the PRK plus UV-B-treated ones. The epithelium of eyes treated with PRK and UV-B appeared normal ultrastructurally at 8, 12, and 16 postirradiation weeks.

378

+

hematoxylin-

Control and UV-B-only-treated eyes showed no electron microscopic stromal abnormalities at 4 weeks. In one of the PRK-treated eyes, subepithelial vacuolization was found at 4 weeks. At 8, 12, and 16 weeks, none of the corneas treated with PRK alone showed stromal vacuolization or other stromal abnormalities (Figs 3C, E). Conversely, all eyes treated by PRK and UV-B showed altered stromal ultrastructure at 4 and 8 weeks (Figs 3B, D). The normal, regular pattern of collagen lamellae was not recognizable in the anterior stroma (Figs 3B, D, F) and was replaced by irregularly arranged, randomly oriented collagen fibers. The fibers were intermixed with electronlucent amorphous extracellular material. The keratocytes were swollen and showed abundant rough endoplasmic reticulum, a prominent Golgi apparatus, increased numbers of mitochondria, and dilated cisterns of rough endoplasmatic reticula containing amorphous material. The underlying posterior stroma was normal ultrastructurally. At 12 and 16 weeks, vacuolization and amorphous extracellular material were no longer observed. However, distorted collagen fibers of variable diameter were observed at 12 weeks (Fig 3F).

Discussion Our experimental results show that UV-B exposure exacerbates and prolongs detectable post-PRK haze in rabbits. Furthermore, we have shown that the haze is concomitant with profound morphologic changes in the anterior stroma. Post-PRK vacuolization has already been described by Marshall et al 11 in 1988 in photokeratectomized monkey eyes ablated to a depth of 150 p,m. Similar changes have been reported by Hanna et al 12 in rabbit eyes ablated to a depth of 60 p,m. Although new laser delivery systems allow better control of energy with resultant smoother corneal ablation and lesser stromal damage, we were able to detect vacuolization in one photokeratectomized nonUV-B-irradiated rabbit eye 4 weeks after PRK ablation to a depth of 45 p,m. However, this change could only be detected by electron microscopy and was not found

Nagy et al · UV-B and Excimer Laser

·Figure 3. Transmission electron micrographs of rabbit cornea. A, a rabbit cornea 4 weeks after a -5.0 diopter (D) photorefractive keratectomy (PRK) . The collagen is regular with a wavy appearance (magnification, X2.000). B, one month after ultraviolet-B (UV-B) irradiation in an eye treated with PRK. N ote the extracellular vacuoles filled with amorphous material and the keratocytes with prominent organelles. An inflammatory ce ll also is present (magnification , XZ.OOO). C, eight weeks after -5.0 D PRK. N o stromal vacuolization can be seen (magnification, XZ.OOO). D , eight weeks after UV-B irradiation in a previously -5.0 D photokeratectomized eye. A thick area of vacuolization can be observed (magnification, XZ.OOO). E, twelve weeks after -5.0 D PRK. The epithelium is normal. and the subepithelial collagen structure is normal. F, twelve weeks after UV-B exposure on a previously -5.0 D photokeratectomized eye. No subepithelial vacuolization can be detected, bur the anterior stromal collagen fibers are wavy, interwoven, and appear thicker than normal (magnification,

xZ.OOO).

after 4 weeks. Nevertheless, this finding indicates that even low degrees of myopic correction may have an electron microscopically detectable effect on stromal morphology. By contrast, morphologic changes were detectable at the light microscopic level in eyes treated with UV -B after PRK. Furthermore, the duration and extent of the changes were much greater than those seen in the PRKonly treated eyes. We interpret the changes as an augmentation of the wound healing response as evidenced by the highly activated keratocytes and electron microscopically detectable effects on stromal morphology. This morphologic change may reflect synergistic effects of UV-B range of electromagnetic spectrum and 193-nm excimer

laser (UV -C region) on the anterior stroma and probably explain the haze seen clinically. In our experimental model, the energy of UV irradiation was set higher than was the radiant UV -B threshold 2 energy, which, in humans (between 0.12 mJ/cm and 0.56 2 13 mJ/cm , causes photokeratitis. According to Pitts et al,Z the threshold radiant exposure of reversible corneal damage in rabbits is 0.070 J/cm2 at 305 nm and 2.250 J/cm2 at 315 nm. Therefore, to compensate for the higher peak wavelength of the delivery system used, the total exposure as 100 mJ/cm2 was defined. This regimen in practice differs substantially from that to which patients are exposed environmentally. For example, patients are likely to receive UV-B in smaller, repeated energy doses over a

379

Ophthalmology

Volume 104, Number 3, March 1997

longer time span. However, despite this constraint, our results indicate that UV-B exposure above an as-yet undetermined threshold is likely to modulate post-PRK stromal repair. Our results justify more sophisticated studies with different exposure energy and longer, perhaps fractionated, exposure. The rabbit cornea does not have a Bowman layer, which might have a protecting role in humans against development of subepithelial haze. Conversely, Bowman is destroyed during PRK in humans and, thus, PRK could render the stroma more susceptible to UV-B. With laser in situ keratomileusis, Bowman layer is intact after PRK and substantial haze is not reported, even after deep ablation. 14 Nevertheless, there have been no studies concerning the effect of UV-B on the outcome of laser in situ keratomileusis. Until these issues are resolved fully, patients should be advised to avoid excessive UV exposure within the time span of post-PRK stromal healing. Moreover, it is possible that UV-B may modulate the response of tissues other than the cornea to laser.

References 1. Seiler T, Holschbach A, Derse M, et al. Complications of myopic photorefractive keratectomy with the excimer laser. Ophthalmology 1994; 101:153-60. 2. Pitts DG, Cullen AP, Hacker PD. Ocular effects of ultraviolet radiation from 295 to 365 nm. Invest Ophthalmol Vis Sci 1977; 16:932-9. 3. Pitts DG, Bergmanson JP, Chu LW. Ultrastructural analysis

380

4. 5. 6.

7. 8. 9. 10.

11.

12.

13. 14.

of corneal exposure to UV radiation. Acta Ophthalmol Scand Suppl 1987;65:263-73. Cullen AP, Chou BR, Hall MG, Jany SE. Ultraviolet-B damages corneal endothelium. Am J Optom Physiol Opt 1984;61:473-8. Ringvold A. Cornea and ultraviolet radiation. Acta Ophthalmol Scand Suppl 1980;58:63-8. Ringvold A. Damage of the cornea epithelium caused by ultraviolet radiation. A scanning electron microscopic study in rabbits. Acta Ophthalmol Scand Suppl 1983;61:898907. Ringvold A, Davanger M. Changes in the rabbit corneal stroma caused by UV -radiation. Acta Ophthalmol Scand Suppll985;63:601-6. Nagy ZZ, Siiveges I, Nemeth J, Fiist A. Experience with excimer laser photorefractive keratectomy [in Hungarian]. Orv Hetil 1995; 136:1035-41. Curtin BJ. The Myopias: Basic Science and Clinical Management. Philadelphia: Harper & Row, 1985;39-59. Hanna KD, Pouliquen YM, Waring GO III, et al. Corneal wound healing in monkeys after repeated excimer laser photorefractive keratectomy. Arch Ophthalmol 1992; 110:1286-91. Marshall J, Trokel SL, Rothery S, Krueger RR. Long-term healing of the central cornea after photorefractive keratectomy using an excimer laser. Ophthalmology 1988; 95:1411-21. Hanna KD, Pouliquen Y, Waring GO III, et al. Corneal stromal wound healing in rabbits after 193-nm excimer laser surface ablation. Arch Ophthalmol 1989; 107:895901. Blumthaler M, Ambach W, Daxecker F. On the threshold radiant exposure for keratitis solaris. Invest Ophthalmol Vis Sci 1987;28:1713-6. Pallikaris IG, Siganos DS. Excimer laser in situ keratomileusis and photorefractive keratectomy for correction of high myopia. J Refract Corneal Surg 1994; 10:498-510.