Background of the Invention This invention relates to a plasmin activity reducing agent useful for reducing or preventing visual refractive regression following a wide area superficial ablation wound to the cornea. Such a wound may be created, e.g., by a surgical procedure. The invention also relates to screening and monitoring methods.

The healing of wounds, including eye wounds, involves complex interactions between two major systems controlling the degradation and removal of damaged tissue and the synthesis of cellular and extracellular elements. In the corneal stroma these systems are targeted on the degradation and synthesis of collagen and the glycosaminoglycan ("GAG") matrix. The first system is the plasminogen-activator/plasmin system, which is involved in the proteolysis, e.g., degradation and removal, of damaged extracellular materials by plasmin. Through this proteolytic activity, plasmin also modulates the growth of normal tissue and tissue damaged by trauma, infection, or inflammation. The second system is the activated keratocyte system, which is involved in the replacement of damaged collagen by the synthesis of new collagen and the collagen matrix of GAGs.

Any wound of the cornea activates several healing phases. The first healing phase involves the removal of cell debris and damaged extracellular materials, e.g., the GAGs which form a matrix that supports the collagen of the cornea, by the proteolytic activity of the plasminogen-activator/plasmin system. The second healing phase involves the repair of partially damaged tissues. The third phase involves the replacement of severely damaged extracellular materials by newly synthesized GAGs

and collagen, which is regulated by the activated keratocyte system, and the replacement of severely damaged cells by new epithelial cells which migrate to the wound site on a temporary fibrin/ fibronectin matrix. This cell proliferation is enhanced by growth factors, e.g., epidermal growth factor. Plasmin regulates re- epithelialization in this phase.

Proteolytically active plasmin is a serine protease derived from plasminogen, which is found in nearly all body fluids and tissues. The precursor plasminogen is activated by plasminogen-activators to become plasmin. Plasmin degrades many matrix proteins, such as GAGs, fibronectin, and laminin, and can activate other enzymes such as pro-collagenase and macrophage elastase. Fibronectin and laminin are important proteins in the extracellular matrix and facilitate cell healing by connecting the epithelium to the underlying tissue. In particular, fibronectin is an adhesive cell-surface glycoprotein which mediates cross-linking of collagen and epithelial cell migration, and laminin promotes cell adhesion.

Plasminogen-activators are classified in two main groups: tissue type plasminogen-activator ("tPA") , and urokinase type plasminogen-activator ("uPA") . The most prevalent form of PA in corneal wounds is uPA, whereas tPA is present in only small amounts. Corneas examined within 3 to 24 hours after healing starts show a clear increase in uPA in the anterior stroma and at the wound edges.

In man, the concentration of plasmin in the tear fluid of healthy, unwounded eyes is very low or zero, typically below 0.2 μg/ml of tear fluid. On the other hand, higher concentrations of plasmin, e.g., up to 100 μg/ml, are associated with various ocular surface wounds, disorders, or irritations. For example, it has been

found that mechanical wounding of the cornea, e.g., by keratectomy, or exposure to allergens, induces an increase in plasmin and a short-term decrease in plasminogen-activator in the tears. Increased plasmin levels are also associated with " icrobial ulcers of the cornea, corneal epithelial defects caused by contact lens wear or vernal keratoconjunctivitis, recent mechanically produced wounds, and lesions caused by corrosive chemicals." Tervo et al., Chapter 12, "Healing Processes of the Cornea", p. 151-163 (editors Beuerman, et al., Portfolio Publishing Co., The Woodlands, Texas (1989)).

Unlike plasmin, plasminogen-activators are normal components of tear fluid. However, like plasmin, higher activities of these plasminogen-activators also occur as wound healing progresses. The normal level of plasminogen activator is about 2.0 ±0.6 IU/ml, the level immediately after wounding is about 0.3 ±0.1 IU/ml, and the level during subsequent healing is about 2.1 ±0.3 IU/ml. The normal activation of the plasminogen- activator/plasmin system in corneal wound healing may therefore be evidenced by plasmin and plasminogen- activator levels in tear fluid. Van Setten, et al., Current Eve Res. , 8.:1293-1298 (1989).

Tervo et al. describe treating patients having corneal defects with topical aprotinin, a serine protease, to inhibit plasmin, to protect fibrinogen and laminin. In the context of such defects, treatment with aprotinin results in enhanced corneal resurfacing and healing (at p. 155-157) . In particular, Tervo et al. used aprotinin to treat non-healing corneal ulcers and other persistent epithelial lesions. The authors state that "aprotinin therapy is a valuable adjunct in the treatment of epithelial defects, and that monitoring of plasmin levels in the tears provides useful information

on the state of corneal healing and selection and timing of medical therapy in these eyes" (at p. 151) .

However, Tervo et al. also notes (at p. 158) :

As pointed out above, the generation of plasmin is evidently a normal and beneficial reaction during wound healing. Therefore, there is no reason to believe that a patient with a simple mechanically produced erosion would be better off if the routine therapy with antibiotic ointment and one-night's patching of the eye were replaced by aprotinin. In an unpublished preliminary study in rabbits, we saw no difference in healing rates of keratectomy wounds which were treated with either aprotinin or vehicle only.

Similarly, van Setten et al. notes (at pp. 1297) that "the clinically observed beneficial effect of topically applied aprotinin may be restricted to cases where excess plasmin is present in the tear [fluid] for several days or weeks and the normal regulation of the plasminogen activating system has gone awry. In absence of clear dysregu.lation [sic] of the plasminogen activator-plasmin system, the application of aprotinin probably does no have a marked effect on the rate of corneal wound healing (unpublished results by the authors) ."

The process of corneal wound healing is of critical importance in all ophthalmic methods of correcting myopia, astigmatism, etc. by surgical procedures. In these procedures, the cornea is cut or ablated, i.e., wounded, to alter or smooth the surface curvature of the cornea and thereby change, and hopefully improve, the visual acuity of the patient. If the wound created by the surgery heals with extensive tissue regrowth or scarring that cause refractive errors, the patient will experience a regression or decrease in

refractive power of the cornea, and the improvement gained by the surgery will be diminished or lost.

For example, in photorefractive keratectomy ("PRK") surgical procedures, a limited thickness of the cornea is ablated by exposure to laser irradiation, e.g., excimer laser irradiation. After the epithelial layer, which is typically less than about 50 μm thick, is removed, the laser is used to ablate very thin layers from the remaining corneal tissue to a depth of about 5 to 50 μm, through Bowman's membrane (thickness about 10 μm) , and into the stroma (thickness about 500 μm) , depending on the attempted correction. Re- epithelization, which is part of the wound healing process, is normally completed within 72 hours of the surgery.

Throughout the world, over 15,000 individuals have now undergone PRK with encouraging clinical results. However, in spite of these results, there are a number of problems associated with healing of the cornea after excimer laser PRK ablation procedures. For example, epithelial hyperplasia and scarring by new atypical collagen (Type III and VII) may occur after this form of ablation, which result in an undesired regression of visual acuity by a change in the intended refraction of the cornea.

In addition, PRK often results in a temporary loss of corneal transparency, often described as corneal scarring or "haze" that typically lasts from three months to a year after surgery. The worst aspect of haze occurs in the first 3 months and is due to activated keratocyte cells that migrate into the stromal collagen to effect repair. After this repair phase, these cells migrate out of the cornea and the scatter or haze due to these cells disappears. A secondary aspect of haze is caused by new collagen and vacuoles between intersected lamellae in the

stroma. These vacuoles are filled with cell debris and cause a high level of light scatter, which changes the refractive indices of the cornea and impairs its transparency.

Conventional corneal wounds caused by scalpels or thermal lasers result in severely damaged or dead (necrotic) tissue, i.e., cells and extracellular materials such as collagen, that remains in the cornea at the wound edge, e.g., the tissue adjacent to the tissue removed by the procedure. The scalpel causes crushing damage, whereas the laser causes thermal damage. It is the damaged or dead cells that trigger the wound healing process by releasing various factors.

The corneal wound created by excimer laser irradiation, on the other hand, leaves only a minute amount of damaged tissue at the wound edge. In other words, the tissue immediately adjacent the ablated tissue is largely unaffected by the excimer laser ablation. However, the minute amount of damaged or dead cells, a layer on the order of only a few Angstroms thick, is enough to release the factors that trigger wound healing and so-called "wound amplification," i.e., an enlargement of the wound area into those tissues adjacent the wound, which would normally be damaged in a conventional wound. In effect, in an excimer laser wound, the wound healing response is much greater than required by the amount of damaged tissue that exists at the wound site.

Summary of the Invention Applicants have discovered that the natural healing and amplification processes, which normally progress after a "wide area superficial ablation wound" occurs in external corneal tissue, should be reduced by pharmaceutical intervention to an absolute minimum and limited to the minute amount of damaged tissue at the

wound edge, because of the unique nature of this specific type of wound. Moreover, applicants have discovered that plasmin ^, s role in the normal healing response to such an ablation wound is actually detrimental to the eye and causes visual refractive regression.

As used herein, the term "wide area superficial ablation wound" itieans a wound which is very shallow, on the order of 5 to 150 μm deep, is associated with minimal or no damage of the tissue immediately adjacent to the ablated tissue, i.e., at the wound edge, and exhibits no inflammation of the adjacent tissue. Such a wound may be created, e.g, by excimer laser ablation, or by mechanical or chemical ablation.

As used herein, the term "visual refractive regression" means any detrimental change in refractive power of the cornea that results from the natural healing process which occurs after a wide area superficial ablation wound is made to the cornea. Such wounds are typically made to change the refractive power and thereby enhance eyesight, but are hampered by the problem of regression.

In particular, applicants have discovered that after a wide area superficial ablation wound to the cornea is created, e.g., by excimer laser PRK, the normal wound healing reaction that results in plasmin degradation of the GAGs surrounding the collagen fibers and collagenase degradation of the collagen fibers remaining in the cornea at the wound edge is a critical precursor to subsequent visual regression. Applicants have discovered that by limiting or inhibiting the first phase of wound healing, i.e., limiting the removal of extracellular materials and cell debris by plasmin, the subsequent wound healing phases of repair and replacement that cause visual regression can also be inhibited or eliminated.

Applicants recognize that epithelial hyperplasia, which is one aspect of regression, results from plasmin's role in the regulation of the growth of normal tissue and re-epithelization, and have found that inhibiting plasmin activity inhibits epithelial hyperplasia. Furthermore, applicants have found that by inhibiting plasmin activity, and thereby impeding the removal of healthy collagen remaining in the cornea, the replacement of the ablated collagen (Type IV and Type V) with new atypical Type III and VII collagen, which is another aspect of regression, is also inhibited. Ideally, no Type III or VII collagen should be added to the eye to "heal" the wide area superficial ablation wound.

Applicants have also discovered that a tear fluid plasmin level greater than 0.2 μg/ml, e.g., pre- operatively, indicates the likelihood of postoperative visual regression in a prospective patient for any surgical procedure which creates a wide area superficial ablation wound. In essence, applicants have recognized that high pre-operative plasmin levels are a contraindication for a surgical procedure which creates a wide area superficial ablation wound, e.g., excimer laser PRK. Similarly, high levels of plasminogen and plasminogen-activator (over about 1.5 IU/ml) are also an indication of potential postoperative visual regression.

Based on these discoveries, applicants have established the applicability of a method of preoperatively testing a tear fluid sample from a potential patient for the risk of postoperative visual refractive regression associated with a wide area superficial ablation wound on the anterior surface of the cornea resulting from the operation, by collecting a sample of tear fluid from the patient, and determining the level of a proteolytic agent in the sample as an indicator of potential postoperative visual regression.

This method may also be carried out without a sample collecting step if a patient sample is available.

As used herein, the term "proteolytic agent" means plasmin, plasminogen, and/or plasminogen-activator ("PA") .

The invention also features a method of monitoring tear fluid from a patient for proteolytic agents after a wide area superficial ablation wound has been created on the anterior surface of the cornea, for the risk of post- wound visual refractive regression, by collecting a sample of tear fluid from the patient after the ablation wound has been created, and determining the level of a proteolytic agent in the sample as an indicator of potential post-wound visual regression.

In these methods, the ablation wound may be the result of a photo-, mechanical, or chemical ablation procedure, e.g., photorefractive keratectomy. If the tested proteolytic agent is plasmin, a level of plasmin greater than 0.2 μg/ml of tear fluid is an indictor of potential post-operative visual regression. If the tested proteolytic agent is plasminogen-activator, a level of plasminogen-activator greater than about 1.5 IU/ml of tear fluid is an indictor of potential post¬ operative visual regression. The method also includes testing the level of plasminogen in the tear fluid.

Applicants have also developed a pharmaceutical treatment for reducing, in an eye of a patient, visual refractive regression associated with a wide area superficial ablation wound of the anterior corneal surface, by administering to the eye a plasmin activity reducing agent.

As used herein, the term "plasmin activity reducing agent" means any inhibitor of plasmin, plasminogen, or PA. The term includes all such "inhibitors" regardless of whether they degrade,

inactivate, disrupt the synthesis of, or otherwise render ineffective, plasmin, plasminogen, or PA, as long as they ultimately reduce or eliminate plasmin activity. In addition, a plasiin activity reducing agent may include any combination of plasmin, plasminogen, and/or PA inhibitors and any other drug that may have a therapeutic effect on the eye. For example, in addition to including a plasmin inhibitor, the plasmin activity reducing agent may include a steroid or other drug to inhibit inflammation or infection.

The plasmin activity reducing agent may be a plasmin inhibitor, e.g., aprotinin, or a plasminogen inhibitor or plasminogen-activator inhibitor, e.g., plasminogen-activator inhibitor-2.

The plasmin activity reducing agent may be administered topically to the cornea prior to, during, or after the occurrence of the ablation wound, or may be administered in a series of administrations after a wide area superficial ablation wound has been created.

The superficial ablation wound may be the result of a photo-, mechanical, or chemical ablation procedure, e.g., photorefractive keratectomy. Furthermore, the ablation wound may be the result of a procedure to change the curvature of the cornea a desired degree to produce a refractive correction, or a procedure to produce corneal smoothing without substantial change to the refractive properties of the eye.

The pharmaceutical treatment may also include administering to the eye an anti-inflammatory drug, e.g., a steroid, in combination with the plasmin activity reducing agent.

The invention also features a plasmin activity reducing agent useful for the pharmaceutical treatment described herein, and a plasmin activity reducing agent for use in the manufacture of a medicament for use in

this pharmaceutical treatment. In either case, the treatment may involve the administration of the plasmin activity reducing agent to the surface of the cornea prior to, during, or after, the occurrence of the ablation wound, and this ablation wound may be the result of a photo-, mechanical, or chemical ablation procedure.

Other features and advantages of the invention will be apparent from the following description of the preferred embodiments in conjunction with the claims and the drawings.

Brief Description of the Drawings

Fig. 1 is a graph showing the average plasmin level in tear fluid of patients before and after a wide area superficial corneal ablation procedure.

Fig. 2 is a graph showing the average pre- operative plasmin level in patients compared to the refractive change three months after surgery.

Fig. 3 is a graph showing mean change in refraction in patients over time after surgery.

Fig. 4 is a graph showing mean change in refraction in steroid (solid line) and placebo (dotted line) treated patients over time after a -3.0 diopter correction.

Fig. 5 is a graph showing mean change in refraction in steroid (solid line) and placebo (dotted line) treated patients over time after a -6.0 diopter correction.

Description of the Preferred Embodiment(s) Screening Patients for Wide Area Superficial Ablation

The amount of visual regression that a patient develops varies greatly between individual patients and the amount of ablation or attempted correction, and the severity of regression typically increases with the degree of correction attempted. For example, a -6.0

diopter correction typically results in little regression, whereas any correction above -7.0 diopters causes at least a 0.5 diopter regression. Certain patients are also more susceptible to regression even with lower order corrections or removal of very thin layers of corneal tissue.

The screening test of the invention, i.e., pre- operatively measuring tear fluid levels of plasmin, plasminogen, and/or plasminogen-activator, is used to determine whether a patient is susceptible to regression, especially in those cases in which the attempted correction is over -6.0 diopters. Based on the test results, the patient is counseled whether to attempt the surgery at all, or whether to use the pharmaceutical treatment for reducing visual regression before, during, and/or after the surgical procedure to alleviate the regression.

Tear fluid collection

Tear fluid samples are collected from one eye of each prospective patient. In each case, a sample of 10 μl of tear fluid may be collected by capillary action using either two blunted 5 μl glass microcapillary tubes or a single 10 μl tube (Brand, FRG) . The glass tube is held at 10 to 30 degrees over the horizontal axis and at 10 to 40 degrees to the surface of the lower fornix. The tip of the tube is brought into contact with the tear fluid meniscus and then slowly moved on the inferior fornix from the middle towards the lateral canthus. All tear fluid samples are immediately frozen on dry ice and stored at -70°C until the plasmin concentration can be determined.

Measuring Proteolytic Agent Levels in Tear Fluid

Plasmin, plasminogen, and plasminogen-activator ("PA") levels in tear fluid of potential candidates for a wide area ablation wound to the cornea, e.g., by PRK, may be measured pre-operatively, or monitored postoperatively, by various techniques. For example, as described in Salonen et al., Acta Ophthalmol. 65:3-12 (1987) , the plasmin and PA levels in tear fluid may be measured as follows (p. 4) :

Tear fluid was collected into a glass capillary. Proteolytic activity, using an 8 μl specimen of tear fluid, was measured by the radial caseinolysis procedure (Saksela, Anal. Bioche . , 111:276-282 (1981)), using agarose gel and bovine milk casein as substrate. Human plasmin (20 casein units per mg; Kabi Diagnostica, Stockholm) was used as standard. The results are expressed as micrograms of plasmin-like activity per ml tear fluid. The advantages of this assay include, the small sample volume needed (minimum 5 μl) , small intra-assay variation (< 5%) and sensitivity (0.1 μg plasmin per ml). Repeated freezing and thawing of tear fluid specimens was found to decrease the enzyme activity. Plasminogen activator levels were determined according to Saksela (1981) using plasminogen- containing casein-agarose gels and urokinase (50,000 Ploug units mg; Calbiochem) as standard. Rabbit antibodies to human plasminogen and albumin (DAKO, Copenhagen) were used in the identification.

Preferably, five concentrations between 50 and 3.125 μg/ml made up from a stock solution of human plasmin (20 casein units/mg, Sigma) are used as positive plasmin standards and distilled water is used as a control. Each assay uses an aliquot of 7 μl tear fluid and an incubation time of 48 hours. The plasmin levels of each sample are proportional to the radial displacement of degraded casein which is determined by

staining the gel with Amido black (1 mg/ml in 2% acetic acid) . Concentrations are determined by direct comparison with the plasmin standards. The proteolytic activity is then confirmed to be human plasmin in a further series of samples in which the gels are reacted with monoclonal human plasmin antibodies (anti-pig 1) .

Tear fluid plasmin levels can also be measured by various standard immunofluorescence techniques that can easily be adapted to detect plasmin in tear fluid.

If the plasmin, plasminogen, and/or PA levels are above certain threshold levels, e.g., above 0.2 μg/ml for plasmin, and above 1.5 IU/ml for plasminogen-activator, the candidate is informed of an increased risk of postoperative visual regression. If the patient chooses to undergo the surgery anyway, a plasmin activity reducing agent is applied to the eye before, during, and/or after the wide area superficial ablation procedure.

If the plasmin, plasminogen, and PA levels in the potential patient are normal, the ablation procedure may be performed without the ancillary pharmaceutical treatment, but the patient's tear fluid levels are carefully monitored for proteolytic agents during the postoperative healing process, and the plasmin activity reducing agent is applied at the first increase of plasmin, plasminogen, or PA in the tear fluid.

Clinical Results of Screening

Tear fluid samples were collected pre-operatively from one eye each of 21 patients. Samples were taken immediately befoire the PRK procedure, and then 15 minutes and 1 week after the ablation wound was created by the procedure. Relatively high pre-operative levels of plasmin were detected in only 3 of the 21 patients. The mean value for the pre-wound plasmin level of the group,

including all the zero or negligible values and the three elevated values, was 4.3 μg/ml. Fig. 1 shows that by fifteen minutes excimer laser ablation, the average plasmin level was almost 60 μg/ml. Plasmin was found in the tear fluids of all 21 patients. Post-wound, i.e., postoperative, values were higher than pre-wound values in every patient, and gave a mean value of 38.2 μg/ml. After re-epithelialization, plasmin levels decreased significantly in the tear fluid samples with a mean value of 8.3 μg/ml at one week postoperatively.

As shown in Fig. 2, applicants found a strong correlation between the pre-operative (pre-wound) plasmin levels and the postoperative refractive outcome at three months, in that only the three patients who had plasmin in the tear fluid prior to surgery (greater than 15 μg/ml) showed a significant myopic regression of -1.0 to -2.0 diopters, whereas the majority of patients (~ o μg/ml plasmin) showed a slight overcorrection.

Methods of Creating a Wide Area Superficial Ablation Wound

The wide area superficial ablation wound described above may be created with a laser, e.g., an excimer laser, or by various mechanical or chemical techniques.

Excimer laser PRK may be carried out with a Summit

Technology UV200 ExciMed excimer laser with an emission wavelength of 193 nm (Summit Technology, Inc. , Watertown,

Massachusetts) . This laser can be adjusted to have a fixed pulse repetition rate of 10 Hz and a fixed radiant exposure of 180 mJ/cm ² at the cornea. The maximal ablation diameter is preferably adjusted to about 4.0 to

5.0 mm. This laser and its method of operation are described in, e.g., Marshall et al., Lasers Ophthalmol..

1:21-48 (1986), which is incorporated herein by reference. The ExciMed laser system and others, and their methods of use, are also described in Marshall et

al., U.S. Patent No. 4,941,093, which is incorporated herein by reference.

Other lasers and wavelengths may also be used to create a wide area superficial ablation wound of the cornea. For example, very high repetition rate, short pulsed, infrared lasers should also create wide area superficial ablation wounds of the cornea. In particular, a near-infrared wavelength of about 2.9 microns is the maximum wavelength that can be absorbed by water and is suitable to create a wide area superficial ablation wound. Optimal wavelengths for lasers are 2.7 to 2.9 microns in the infrared region and 193 nanometers in the vacuum ultraviolet.

Mechanical ablation may be carried out by freezing the cornea and removing, e.g., grinding away, very thin layers of tissue with a high speed rotating burr. In addition, mechanical ablation can be achieved by rotating or oscillating a sharpened knife edge tangentially to the top of the cornea to scrape away very thin layers of corneal tissue as described in Kilmer et al., U.S. Patent No. 5,063,942, which is incorporated herein by reference.

Chemical ablation to create a wide area superficial ablation wound of the cornea may be carried out by applying an alkaline substance to the anterior surface of the cornea in a carefully controlled manner to limit damage to tissue adjacent to the ablated tissue to a few nanometers. Chemicals such as ammonium, sodium hydroxide, and calcium hydroxide may be used to degrade the corneal tissue, and may be applied to the cornea by standard techniques.

Reducing Visual Regression in a Patient

Patients who undergo a surgical procedure that creates a wide area superficial ablation wound of the cornea may experience subsequent visual regression,

depending on their individual reaction to the procedure and the attempted level of correction. To prevent or inhibit such regression, these patients may be treated with a plasmin activity reducing agent that inhibits plasmin activity in the cornea. This treatment is carried out independently of the ablation procedure and may be performed before, during, or after the ablation, depending on the individual patients' needs. These needs may be determined, for example, by the screening test described above.

Plasmin Activity Reducing Agents

The purpose of the plasmin activity reducing agents of the invention is to decrease or eliminate plasmin activity in the cornea. This can be achieved by degrading the plasmin, plasminogen, and/or PA, or by rendering them inactive or ineffective. Suitable plasmin inhibitors include aprotinin (20,000 IU/ml Trasylol®, Bayer) , α^-antitrypsin, α ₂ -antiplasmin, α ₂ -macroglobulin, and monoclonal antibodies to plasmin.

Aprotinin is a polypeptide derived from animal organs and acts as a serine proteinase, and is the presently preferred inhibitor for use in the present invention. Alpha ₂ -antiplasmin operates by binding to the lysine active site which controls the interaction of plasmin with fibrin, thereby inactivating the plasmin.

Plasminogen-activator inhibitors ("PAI") include PAI-1, which is derived from endothelial cells, and PAI- 2, which is derived from placental cells. Monoclonal antibodies that bind to PA, or to plasminogen, are also suitable for carrying out the treatment of the invention if they bind in such a way as to prevent any activation of the plasminogen by the PA. PAI-2 is the preferred PAI for use in the present invention.

In addition, the plasmin activity reducing agent may include any combination of plasmin inhibitors and/or

PA inhibitors, and may be combined with any other drug that has a therapeutic effect on the eye. For example, the plasmin activity reducing agent may be combined with an antibiotic, e.g., chloramphenicol, a non-steroid anti- inflammatory drug, e.g., flurbiprofen, or an anti- inflammatory steroid, e.g., dexamethasone, prednisolone, fluorometholone, or others as described, e.g., in Robertson et al., U.S. Patent No. 4,939,135, which is incorporated herein by reference.

Steroids inhibit the keratocyte invasion of the collagen in the stroma which causes inflammation. The combined use of steroids and aprotinin, or other plasmin inhibitors, should create a synergistic effect that substantially prevents the replacement of the ablated collagen by Types III and VII atypical collagen. By using a plasmin inhibitor to limit the first phase of wound healing, and a steroid or other anti-inflammatory drug to limit the second phase of wound healing, the collagen replacement aspect of the third phase of wound healing should be substantially limited so as to avoid visual regression.

Treatment with a Plasmin Activity Reducing Agent

The plasmin activity reducing agent, e.g., aprotinin (Bayer, Leverkusen, FRG) , is administered topically to the anterior surface of the cornea. Aprotinin is diluted from stock preparations in sterile saline or commercially obtained wetting agent (e.g., Liquifilm Tears; Allergan) to achieve a concentration of 20 or 40 IU/ml, and is administered in 1-2 drops (50 μl each) per eye, 4 times per day during waking hours for the first two weeks after surgery. Although this is the preferred administration, any combination of plasmin, plasminogen, and/or PA inhibitors, and other therapeutic agents, may be used in, or in combination with, the plasmin activity reducing agent of this invention.

If the plasmin activity reducing agent is applied before the ablation wound is created, it should be applied in the same concentrations as described above, and the single application, or last in a series of applications, should be no more than a few hours prior to the surgery, so that the plasmin activity is effectively inhibited at the time the wound is created.

Clinical Results of Treatment

Twenty-seven patients (one eye each) were divided into two groups on the basis of their pre-operative refraction. Fourteen patients received a -3.0 diopter correction and thirteen a -6.0 diopter correction. The ablation depth over this dioptric range was 36 μm or 62 μm, respectively. All patients were treated postoperatively with a single dose of topical antibiotics (e.g. , chloramphenicol) and with aprotinin (Bayer, Leverkusen, FRG) 4 times per day for the first two weeks. Thereafter no medication was given.

In all patients re-epithelialization took between 48 and 72 hours. However, completion of re- epithelialization occurred earlier in patients who were treated with aprotinin compared with control patients. In addition, aprotinin treated patients reported less postoperative pain. The mean change in refraction over time for the -3.0 diopter and -6.0 diopter groups is shown in Fig. 3.

As shown in Fig. 3, the aprotinin treated patients showed an initial small overcorrection (of about +1.0 diopters) at one week, followed by rapid stabilization at the desired refractive correction by 6 to 12 weeks. This initial overcorrection, i.e., hyperopic shift, is thought to be caused by edema of the cornea due to disturbed water relationships arising after ablation. This initial hyperopic shift is common to all PRK patients, but was significantly smaller in patients treated with aprotinin.

The plasmin inhibitor aprotinin helps reduce this edema because less correal tissue is removed. Untreated patients typically show a marked overcorrection (e.g., +2.0 to +3.0 diopters) followed by 3 to 5 months of regression to a stable, typically myopic correction.

These same 27 PRK patients were also measured objectively for the time course and magnitude of anterior stromal haze for both reflected and scattered light using a digital video system which includes a Charge Coupled Device (CCD)-camera fixed to a slit-lamp microscope and connected via a frame grabber to a computer. This device measures haze by measuring corneal light scattering and can discriminate between reflected and scattered light. The device and its method of operation are described in Lohmann et al., Refr. Corneal Surg.. :114-121 (March/April 1992) , which is incorporated herein by reference. The aprotinin treated group of patients exhibited significantly less haze and a reduction in the period of time that vision was degraded by scattered light compared to the control group of patients.

As described in Lohmann et al., Eur. J. Ophthalmol.. 1:173-80 (1991), objective measurements of the cornea show haze in the corneal stroma within one week of an ablation procedure which progressively increases until the second postoperative month and then declines. This first peak of corneal haze is caused mostly by scattered light, though reflected light makes up part of this haze. Most patients return to the normal range within three months, but some show a secondary phase in which visual acuity again declines from three to four months postoperatively, but improves steadily thereafter to return to the normal range by about six months. This second phase is likely due mostly to reflected light caused by an increase in new atypical

collagen and the presence of vacuoles or GAG-filled spaces.

Patients that receive no aprotinin treatment typically show significant light scatter in the first two to three months after surgery as the major contributor of the haze phenomenon. The patients treated with aprotinin had some haze, but this haze is caused only by reflected light. Applicants believe that the absence of scattered light is explained by the reduction of vacuoles in the stroma in these patients caused by the aprotinin treatment. The reflected light is caused by new collagen, which can be inhibited by steroid treatment. These hypotheses were confirmed by the rabbit studies described below.

Clinical Controls

Two groups of patients were treated with PRK to receive a correction of -3.0 (57 patients) and -6.0 (56 patients) diopters, respectively. Within each of the two groups, some patients were treated only with a steroid, dexamethasone (no plasmin inhibitor) , at a dosage of 1 drop 5 times per day for the first two months. This was followed by a stepwise reduction in the number of applications such that no medication was given after the end of the third month, i.e., 4 times per day for the next two weeks, 3 times per day for the next week, and 2 times per day for the last week. The remaining patients received the same treatment regimen, but with a placebo containing neither steroids nor aprotinin.

As shown in Fig. 4, the patients receiving the - 3.0 diopter correction showed a hyperopic shift in the first two weeks after surgery of about 2.0 diopters greater than the desired correction in the steroid treated group (solid line), and about 1.0 diopter greater in the placebo treated group (dotted line) . Both groups then regressed through the first three months after

surgery. The steroid treated group stabilized at about 1.0 diopter below the desired correction at three months and improved only very slightly by six months. The placebo treated group regressed to a point more than 1.0 diopter below the desired correction at three months, but then improved gradually to almost the same level as the steroid treated group, i.e., almost 1.0 diopter below the desired correction.

As shown in Fig. 5, the patients receiving the - 6.0 diopter correction showed a hyperopic shift in the first two weeks after surgery of more than 3.0 diopters greater than the desired correction in the steroid treated group (solid line), and about 2.0 diopters greater in the placebo treated group (dotted line) . Both groups then regressed through the first three months after surgery. The steroid treated group regressed to about 2.0 diopters below the desired correction at three months and regressed slightly more by six months. The placebo treated group regressed to a point more than 3.0 diopters below the desired correction at three months, but then improved gradually to about 3.0 diopters below the desired correction.

Animal Studies

Eight New Zealand white rabbits of approximately 2.5 kg each were subjected to wide area superficial corneal ablation with an excimer laser. One eye of each animal was exposed to the excimer laser to ablate a -6.0 diopter change, i.e., about 62 μm of the cornea were ablated. General anaesthesia was induced by intravenous injection of Fentanyl 0.75 ml/kg (Hypnorm) supplemented by topical application of proparacaine hydrochloride. The animals were placed on a positioning board with the target eye held open with a lid speculum. Animals were killed with an intravenous overdose of phenobarbital at either 4 days or 3 months postoperatively. Immediately

after death, the eyes were removed and processed for light and scanning electron microscopy.

Two of the rabbits were used in the histological study described below to determine the presence and distribution of plasminogen-activator in the wounded area. One was treated postoperatively with a steroid solution (no aprotinin) and the other was treated postoperatively with aprotinin for four days using the dosage regimens described below. These two animals were sacrificed four days after surgery. In the remaining six animals, three different postoperative regimens were employed with two rabbits in each set. All six animals had a basic postoperative regimen consisting of topically applied antibiotics (chloramphenicol) twice a day over one week.

The first set was a control which had no further medication. In the second set, aprotinin (40 IU/ml) was administrated topically 1 drop 4 times a day over a two week period. In the final set, a similar regimen was employed for aprotinin, but was supplemented by topically applied steroids, e.g., prednisolone acetate (Allergan), and a non-steroid anti-inflammatory, e.g., flurbiprofen (Allergan) at a dosage of 1 drop 12 times per day for the first week, followed by 5 times per day for the first two months. This was followed by a stepwise reduction in the number of applications such that no medication was given after the end of the third month. After four months of observations, the rabbits treated with aprotinin exhibited limited clinical signs of haze.

Histopathology

1) Immunofluorescence

Both eyes of two rabbits were enucleated immediately after death and the corneas were removed and immediately fixed at 4°C in 96% ethanol for 1 to 2 hours, rehydrated, and immersed overnight in 0.1 M phosphate-

buffered saline (PBS), pH 7.4, containing 25% w/v sucrose. The corneas were hemisected with a razor blade quench frozen by rapid immersion in partially frozen isopentane precooled in liquid nitrogen. The frozen samples were mounted on a microtome chuck before being sectioned in a Reichart Frigocut. Sections of about 10 ^Λ to 15 microns in thickness were mounted on gelatinized slides and dried for at least 30 minutes and stored in a cryostat.

An indirect immunofluorescence method was used to demonstrate the presence or absence of urokinase type PA ("uPA") . Non-specific binding sites were blocked with diluted normal rabbit serum. Polyclonal goat anti-human urokinase (uPA-PAI-1 and -2 complexes, Cambio, UK) served as the primary antibodies and were incubated on the samples overnight at 4°C. Fluorescein isothiocyanate- conjugated rabbit anti-goat IgG (Sigma, UK) served as the secondary (labelled) antibody and was applied to the sample and incubated for 1/2 to 1 hour at room temperature. Control sections omitted either the primary or secondary antibodies. Samples were viewed on a Leitz fluorescence microscope under UV illumination.

2) Light and electron microscopy

The enucleated eyes were immediately immersed in a fixative solution of 2.5% glutaraldehyde buffered in 0.1M sodium cacodylate containing 10 mg/ml calcium chloride with a final pH of 7.4. After 5 minutes in this solution, the corneas were isolated by a circumferential incision and replaced in the initial fixative solution for a period of one hour. The corneal specimens were hemisected through the area of ablation and were processed for light (LM) and transmission electron microscopy (TEM) . These specimens were post-fixed for two hours in 2.0% osmium tetroxide buffered in 0.2 M sodium cacodylate before being dehydrated through an

ascending series of alcohol concentrations and embedded in Araldite via epoxy-propane. For SEM the specimens were post-fixed overnight in the osmium solution and then dehydrated through an acetone series before being critical point dried (Emscope CPD750) and sputter coated (Emscope SC500) with a 30 nm layer of gold. Sections for LM were cut at 1 μm with glass knives using an ultramicrotome (Reichart, FRG) and stained with toluidine blue. Sections for TEM were cut ultrathin with diamond knives on a Reichart ultramicrotome and stained with lead citrate and uranyl acetate before examination in an AEI 801 transmission electron microscope.

3) Results

Microscopic evaluation revealed that the rabbit corneas treated with aprotinin show a regular and smooth epithelium with some disturbances in the anterior stroma, whereas the corneas treated with a steroid without aprotinin show epithelial hyperplasia, rough and undulated epithelium, and somewhat fewer disturbances in the corneal stroma.

Other Embodiments It is to be understood that while the invention has been described in conjunction with the preferred embodiments thereof, the foregoing description is intended to illustrate and not limit the scope of the invention, which is defined by the scope of the appended claims. Other aspects, advantages, and modifications within the scope of the invention will be apparent to those skilled in the art to which the invention pertains.

What is claimed is: