Lasers in Ophthalmology vo!. I no. 4 pp 177-183 (1987)Kugler Publications, Ghedini Editore, Amsterdam/Berkeley/Milano

Stromal remodelling following photo refractive keratectomy

Stephen Tuft, John Marshall and Stephen RotheryDepartment of Clinical Ophthalmology, Institute of Ophthalmology, Judd Street, London WC1H 9QS,UK

Keywords: collagen, cornea, fluorescent dyes, lasers

Abstract

An excimer laser (193 nm) was used to ablate discs 1.5 mm in diameter into the anterior surface of rabbitcorneas. The bed of each disc was then labelled with dichlorotriazinyl aminofluorescein and animalssacrificed at intervals of one or two months. This technique demonstrates the deposition of new collagenbeneath the epithelium during the process of corneal remodelling after deep keratectomy.

Introduction

A surgical technique has recently been proposedthat would alter the refractive power of the eye byusing laser energy to reprofile the anterior cornealsurface. n Most studies of this procedure, termedphoto refractive keratectomy (PRK), have used anargon fluoride (ArF) excimer laser (193 nm) be-cause of the clean wound surface that is produced.The properties of tissue photo ablation that are fun-damental to this type of surgery can be summarizedas follows: Firstly, the limited penetration of thecornea by ultraviolet radiation at 193 nm and thenature of the photoablative process restricts the ef-fect of each laser pulse to a superficial layer withonly a narrow zone of conductive damage to adja-cent tissue."? Secondly, the amount of tissue re-moved per laser pulse is energy dependent once theablation threshold has been attained. 8 Tissue canthus be removed from the target surface in predict-able increments. Thirdly, the ablated surface is ex-ceptionally smooth and sealed by a pseudo-rnern-brane.?: 10 leaving a suitable substrate to preservethe new optical properties of the eye.

Mathematical models predict that significantchanges in refractive power can be achieved byvolume ablation confined to Bowman's layer.PExperiments in animals have demonstrated thatphotoablation can effectively alter the refractiveproperties of the eye and suitable delivery opticshave been designed to flatten or steepen the curva-ture of the anterior surface of the cornea. For PRKto be a clinically acceptable procedure, a number ofrequirements must be satisfied. Epithelial migra-tion across the wound must proceed without signifi-cant delay or loss of adhesion to the. underlyingstroma. Secondly, the ablative process should notresult in visually significant stromal opacificationfrom either acute thermal effects or sub-epithelialscarring from tissue repair. Finally, the contour ofthe reprofiled surface must remain constant withinnarrow limits; a tendency for either the forwardmovement of thinned cornea or for epithelium a~d .new stromal collagen to fill ablated areas woulddefeat the long-term objective of a predictable andstable refractive change.

Studies of deep (150 /km) PRK wounds have notdemonstrated a significant latency of migration or

a b

178

c

Fig. 1. (a) Photograph of 3.5 mm disc ablated into the cornea of a rabbit. (b) With ultraviolet light the site of the disc is accentuatedby fluorescence from bound DTAF. (c) Light micrograph of a section at the edge of a disc from a rabbit one month post surgery. Hyper-plasia of the epithelium is marked over the edge and base of the disc, and the stroma beneath the disc has a disordered orientation (tolui-dine blue). (d) Photograph taken by fluorescence microscopy of a section adjacent to (c) demonstrates a band of collagen that has alower level of fluorescence than the epithelium or the labelled stroma.

abnormal epithelial adhesion, although epithelialhyperplasia overlying the disc has been reported. 11

However, there is some debate in relation toromal changes. A percentage of the ablation sites

de elop a sub-epithelial haze, 14 and a degree of dis-organisation of the underlying collagen has beenevident at histology. When care has been taken tolimit the maximum ablation depth to 50 jJ.m or lessthese problems are less apparent. The aetiology andnature of this opacity need to be defined.

Clinical experience indicates that the cornea hasa considerable ability to synthesise new tissue andin so doing to regain contour lost after injury or in-flammation. An alteration of the topography of theoptical zone following PRK would provide evi-dence of remodelling, but there is no suitable tech-nique to monitor changes in the shape of this smallcorneal area in vivo. Keratoscopic evaluation fol-lowing deep ablation of the monkey eye suggeststhat there is a tendency for the induced correction[0 regress.!" but similar studies following shallowPRK procedures are not available. Although cor-nealscarring may be evident clinically it is difficult[0 distinguish between normal stroma and scar tis-ue on conventional histopathological prepara-

tions. Clearly some remodelling must occur and atechnique to discern the original wound surface ina healed cornea would be of value for following thisreaction. Davison and Galbavy+ 6 have described atechnique whereby freshly exposed stromal tissueis covalently labelled with the fluorescent dye di-chlorotriazinyl aminofluorescein (DTAF). DTAFis retained in the stroma for at least a year and is notmobilised during remodelling.> This paper reportsour preliminary results with this technique in thestudy of collagen repair in the rabbit model forPRK.

Material and methods

ew Zealand white rabbits of approximately 2.5 kgwere used throughout this study. Anesthesia was in-duced by intramuscular injection of Fentanyl 0.75mllkg (Hypnorm) and local application ofproparacaine hydrochloride. A circular disc of 3.5mm diameter was ablated into the centre of the cor-

179

nea of the right eye of each animal using a Questecseries 2000 excimer laser filled with ArF to the man-ufacturers' specifications. The beam of the laserwas cropped by a 5 mm mask before it was passedthrough a plano-convex lens (focal length 35 mm)of UV grade quartz (Spectracil 11), and the maskimaged on the cornea. The beam energy wascalibrated at the image plane on the cornea with aJoulemeter (GenTec ED 200) and was adjusted to100 m L'cm- with a repetition rate of-l0 Hz. It wasfound empirically that 1000 pulses were required toproduce a disc of approximately 75 jJ.m in the stro-ma after ablation of the corneal epithelium. At thecompletion of irradiation the wound was irrigatedwith a 0.5070 solution of DTAF dissolved in 0.2 Msodium bicarbonate. After 30 sec the excess dye waswashed from the wound with phosphate bufferedsaline and a drop of 0.5% chloramphenicol instill-ed. As re-epithelialisation progressed, the epithelialcells masked the fluorescence from the underlyingstroma which permitted the process of re-epithelial-isation to be followed by illuminating the corneawith an ultraviolet source (UVS-ll). Animals wereexamined twice daily until re-epithelialisation wascomplete and twice weekly thereafter.

Animals were killed at an interval of one or twomonths after the procedure by intravenous Pento-barbitone (Euthetal). Following enucleation theepithelium was fixed for 30 min in a solution of 1%glutaraldehyde in 0.1 M phosphate buffer, pH 7.2,a sclerotomy was then performed and the globereturned to the fixative. After 24 h the cornea wasdissected from the globe and divided into two seg-ments. Tissue for light microscopy was dehydratedthrough an alcohol series and embedded in JB4 re-sin. Sample sections of each disc were cut with aglass knife and stained with toluidine blue for histo-logical examination and adjacent 1 jJ.m sectionsphotographed using an epi-fluorescent microscope.

Samples for TEM were first fixed for one hour in0.1 % osmium tetroxide before being dehydratedand embedded in Araldite. Sections were cut witha diamond knife on a Reichart OMU4 microtome,stained with lead citrate and uranyl acetate beforebeing viewed in an JEOL lOOS microscope.

180

a b

c d

Results

Following the ablation of a 3.5 mm disc in the rab-bit cornea all wounds had re-epitheJialised by threedays with no evidence of subsequent wound break-down during the period of follow-up. Sub-epithe-lial haze beneath the area of the disc was noted atboth the one and two month observation periods inall animals, and this could be easily distinguishedusing a slit lamp microscope from the fluorescencedue to bound DTAF.

Histological examination of stained sections con-firmed the presence of an ablated disc in the stromaof the cornea (Fig. 1). The edge of the wound ap-peared wider at the lip than at the base. The epi-thelium that covered the wound was hyperplastic\ ith an accentuated columnar appearance to thebasal cells. The thickened epithelium graduallymerged into the normal epithelium at the edge ofthe disc. Immediately beneath the epithelium theorientation of the lamellae was irregular. Kerato-c tes appeared to be present in normal numbers andno inflammatory cells were noted.

Fluorescence microscopy demonstrated a back-ground fluorescence throughout the specimen butfluorescence due to DT AF bound to the stroma ofthe cornea was clearly demonstrated over thewound bed. A zone of collagen with a reduced levelof fluorescence was observed beneath the epithe-lium and this gradually reduced in thickness as itascended the edge of the wound. At the interfacebetween the two zones of labelled and unlabelledcollagen there was some undulation of the fluores-cent border. We interpret the junction between thefluorescent and the non-fluorescent collagen asrepresenting the site of the original wound surface.

On examination of transmission electron micro-graphs of the interface of normal basal epitheliumand stroma a number of features are apparent(Fig. 2). The epithelial cells are separated from the

181

superficial stroma by their basement membrane (50to 70 nm). There is a slight undulation of the mem-brane, with an excursion of up to 100 nm and a peri-odicity of 300 nm. The basal cells adhere to the base-ment membrane at hemidesmosomes distributedevery 300 to 400 nm. The superficial 2 /km of stromadoes not have the organized collagen arrays of thedeeper lamellae but is formed of a random array offibers 30 to 35 nm in diameter with an interfibrillarmatrix of a similar density to the basement mem-brane of the epithelium. Keratocytes populate thestroma immediately beneath the basement mem-brane.

In the base of the photoablated discs the base-ment membrane was less regular in thickness (50 to150 nm) with less marked undulations. There werenormal numbers of basal hemidesmosomes of nor-mal morphology. The underlying stroma did notshow the 2 /km layer of disorganised collagen, buta much deeper layer of poorly organised material25 /km thick. The collagen fibers in this area aver-aged 25 to 30 /km in diameter. There was electrondense interfibrillar material throughout the depthof the remodelled zone. In this study we have notcorrelated TEM of the ablated region with the inter-face defined by fluorescence.

Discussion

The potential benefit of a method to preciselymanipulate the curvature of the corneal opticalzone would be considerable. The feasibility of PRKhas been demonstrated in animals and the acutehistological changes have been documented. Thispaper addresses the nature of corneal repair in arabbit model following 193 nm excimer laser PRK.

Collagen fibers deposited during embryogenesisare arranged in a regular pattern of lamellae and itis thought that both the restriction in fiber cross-sectional diameter and the regular inter fibrillary

Fig. 2. Survey electron micrographs of normal rabbit cornea (a) and cornea following irradiation (b). The basal portion of the epithelialcells (E) is separated from the superficial stroma of randomly orientated collagen (closed arrow) by the basement membrane (open ar-row). In contrast, a wider zone of disorganized collagen is clearly demonstrated in (b). The distribution of keratocytes (K) and hemi-desmosomes appears similar. Higher power micrographs of similar areas (c and d) show the comparison between the diameters of thecollagen fibrils and thickenings of the basement membrane (open arrow) after irradiation. Bar markers are I ,.m.

182

spacing is necessary for corneal transparency+. 13

thus the stereo-spatial morphology of collagen inpart determines corneal transparency and conver-sely an increased fiber thickness and interfibrilspacing or a decreased density of fibers are asso-ciated with a reduction in transparency. The regu-larity of the collagen fibrils may reflect the patternof their component collagen. To date, of the 11types of collagen that have been characterized,types I, 1I, V and VI have been identified as majorcomponents of human corneal stroma. 16, 18 Bow-man's layer differs from the rest of the stroma inthat it is acellular and contains mainly type V colla-gen, while the epithelial basement membrane isformed mainly by type IV collagen.l

Repair of corneal wounds does not follow the se-quence of normal embryonal fibrogenesis, withdifferences in both keratocyte organisation and col-lagen deposition.I- 3 The physical properties ofcorneal stromal scars also differ significantly fromnormal cornea but it is uncertain if this is the resultof differences in the types of collagen fibers laiddown by keratocytes, which change during develop-ment, or an altered cross-linkage between the colla-gen fibers.2,3 In the rabbit there is an increasedtransparency of the wound after prolonged periodswith a gradual process of fiber rearrangement to aregular morphology and a reduced cellularity.'?Tensile strength of the wound, however, remainspermanently impaired.

It is usual to distinguish between wounds of dif-ferent depth when considering corneal healing. Amechanical abrasion removes the epithelial layer,but the basement membrane and the attachmentcomplexes remain intact and available for the mi-grating epithelium to use, therefore healing is usu-ally rapid and complete: a keratectomy removespart of the anterior stroma in addition to the base-ment membrane, while deep lacerations may passthrough Descemet's membrane and enter the an-terior chamber. Injuries in which corneal stroma islost are usually associated with an attempt to re-form the original contour. This may be achieved byan epithelial plug in superficial wounds, whereasmore severe stromal loss is invariably followed bystromal scarring and permanent corneal thinning.The majority of stromal repair studies have used

deep linear incisions. Following this type of injury,Matsuda and Smelser l? noted degeneration ofdamaged fibroblasts within a zone of 300 !lm fromthe wound edge followed by fibroblast re-popu-lation within 24 h and the onset of collagen deposi-tion. While keratocyte and epithelial cell inter-action is important for scar formation in thesuperficial portion of the wound, the authors sug-gested that the origin of collagen in the deep part ofthe wound was predominantly of an endothelial cellorigin.V

Although the photo ablated discs in our experi-ments were square edged in profile and passeddeeper into the stroma than is anticipated for clini-cal PRK, 1 5 the epithelium was able to migraterapidly over the wound edge and to cover the base.However, the regenerated epithelial layer was thick-er than normal producing an apparent levelling ef-fect, a phenomenon that has previously been notedto occur following lamellar keratectomy in the rab-bit.? The observed deposition of new collagenbeneath the thickened epithelium would tend to en-hance this filling effect and produce sub-epithelialhaze on clinical examination. It has been notedfrom clinical specimens that collagen deposition isless marked after superficial than deep keratecto-mies, and whereas the rabbit does not have akeratocyte-free Bowman's layer, it is possible thatwounds limited to this layer of the primate corneawill not demonstrate collagen deposition. Clearlythe influence of edge profile, wound depth andother variables upon this reparative response needsto be further investigated.

Acknowledgements

This study was supported by a grant from the TFC FrostCharitable Trust (SJT). We would like to thank Summit Tech-nology for making the laser available to us. The fluorescentmicroscope was purchased from a grant from the Clothworker'sGuild.

References

I. Ben-Zvi, A., Rodriquez, M.M., Krachmer, J.H. et al. Im-munohistochemical characterisation of extracellular matrix

in the developing human cornea. Curr. Eye Res. 5: 105-11 , 1986.

_. Cintron, e., Hassinger , L.e., Kublin, C.L. et al. Biochemi-cal and ultrastructural changes in collagen during cornealwound healing. J. Ultrastruct. Res. 65: 13-22, 1978.

'. Ciruron, e., Covington, H., Kublin, e.L. Morphogenesisof rabbit corneal stroma. Invest. Ophthalmol. Vis. Sci. 24:-43- 556, 1983.

-4. Cox, J.L., Farrell, R.A., Hart, R. W. The transparency ofthe mammalian cornea. J. Physiol. 210:601-616, 1970.

-. Davison, P.F., Galbavy, E.J. Fluorescent dyes demonstratethe uniform expansion of the growing rabbit cornea. Invest.Ophthalmol. Vis. Sci. 26: 1202-1209, 1983.

6. Daviscn, P.F., Galbavy, E.J. Connective tissue remodellingin corneal and scleral wounds. Invest. Ophthalmol. Vis. Sci.27: 1478-1484, 1986.

7. Khodadoust , A.A., Silver stein, A.M., Kenyon, K.R. et al.dhesion of regenerating corneal epithelium: The role of the

basement membrane. Am. J. Ophthalmol. 65: 339-348,1968.

. Krueger , R.R., Trokel, S. Quantitation of corneal ablationby ultraviolet laser light. Arch. Ophthalmol 103: 1741-1742, 1985.

9. Marshall, J., Trokel, S., Rothery, S. et al. An ultrastructur-al study of corneal incisions induced .by anexcimer laser.at193 nm.oOphthal~olo~y 92: 749-758,1985. . . .

183

10. Marshall, 1., Trokel, S., Rothery, S. et al. A comparativestudy of corneal incisions induced by diamond and steelknives and two ultraviolet radiations from an excimer laser.Brit. J. Ophthalmol. 70: 482-501, 1986.

11. Marshall, J., Trokel, S., Rother y, S., Krueger, R.R. Pho-toablative reprofiling of the cornea using an excimer laser:Photorefractive keratectomy. Lasers in Ophthalmol. 1:21-48,1986.

12. Matsuda, H., Smelser , G.K. Electron microscopy of cornealwound healing. Exp. Eye Res. 16: 427-442, 1973.

13. Maurice, D.M. The structure and tran~parency of the cor-nea. 136: 263-286, 1957.

14. Mandel, E., Krueger , R.R., Puliafito, C. et al. Excimer laserlarge area ablation of the cornea. Invest. Ophthalmol. Vis.Sci. (Suppl.) 28: 275,1987.

15. Munnerlyn, e., Koons, S., Marshall, J. Photorefractivekeratectomy: A technique for laser refractive surgery. J.Cataract. Ref. Surg. 1987 (in press).

16. Nakayasu, K., Tanaka, M., Konorni, H. et al. Distributionof types I, II,IlI, IV and V collagen in normal and keratoco-nus corneas. Ophthalmic. Res. 18: 1-10, 1986 .

17. Trokel, S., Srinivasen, R., Braren, B. Excimer laser surgeryof the cornea. Am. J. Ophthalmol. 96: 710-715, 1983.

18. Zimmermann, D.R., Trueb, B., Winterhalter, K.H. et al.TYPe VI. collagenis amajor component of the human cor-nea. FEBS Lett.197: 55-58, 1986.