At its interface with ambient air the pre-corneal tear film is responsible for most of the dioptric power of the visual system and so must retain a spatially precise (to the quality of optical polishing) and stable surface during the inter-blink period despite an inherent instability.

The structure of the pre-corneal tear film is not established. Holly in American Journal of Optometry and Physiological Optics, 58, 324-330 (1981), reviewed the physiology of tear film and its interaction with contact lenses. He describes the structure of the film as consisting of three layers: the superficial lipid layer comprising about 1% of the total thickness; the middle aqueous tear layer representing over 98% of the tear film; and the mucous layer comprising less than 0.5% of the tear film. The mucous layer is said to comprise a hydrated layer of mucoproteins which are highly sialylated. The aqueous layer is described as containing inorganic salts, glucose, urea and surface active biopolymers, proteins and glycoproteins. The superficial, lipid layer apparently prevents evaporation of the underlying aqueous layer. The meibomian glands are known to supply these "lipids". Holly describes the occurrence of dry spot formation when blinking is prevented, the time between the last blink and the first dry spot being termed the tear film break up time (BUT) . Short BUTs have been used as a tool to detect tear film deficiencies. Holly proposes a mechanism of dry spot formation, in which lipids from the superficial defuse to the mucin layer, rendering it hydrophobic and thus unable to allow wetting of the underlying epithelial layer.

Based on this theory he suggested that there were many factors which could cause instability of the tear film, including excessive amounts of polar lipids in the superficial layer, reduced overall thickness of the tear film, abnormal epithelial surface morphology, effective mucous layer and the presence of surface active agents.

Trinkaus-Randall et al, in Investigative Ophthalmology and Visual Science, 31, 440-447 (1990), describe investigations of wound healing in corneal tissues. In these investigations corneal epithelial cells are cultured and the cells and medium analysed to detect the presence of fibronectin and integrins. They found that fibronectin was not detectable over a period of 72 hours of culture. Integrins were detectable, a proportion being incorporated into the cellular matrix, whilst a large proportion was released into the medium. Furthermore antibodies to integrins prevented cellular spreading. There is no suggestion by Trinkaus-Randall et al that adhesion proteins or any of the membrane receptors therefore have any role in the tear film.

Tervo et al in Cornea, 10, 461-465 (1991) describe the identification of various integrin sub units in normal and tissue-cultured human cornea, to determine whether storage of corneal tissue in culture medium affected the distribution of those proteins. They found various integrin sub units on the cell surfaces, proposing that these were involved in cell-cell adhesive interactions, as well as cell- extra cellular matrix (ECM) interactions and regulation of ECM synthesis. These mechanisms are said by Tervo et al to have implications for the mechanism of wound healing and normal tissue turnover as well as assisting in explanations for pathologic tissue reactions characteristic of a wide variety of diseases. There is no suggestion that the findings have any relevance to tear film. Meibomian gland dysfunction (MGD) has been reported due to contact lens wear and in polychlorinated biphenyl (PCB) poisoning. It has been postulated that

hyperkeratinization of meibomian glands causes MGD. Ong et al in Current Eye Research 10, 1113-1119 (1991) identify the presence of keratin using various anticytokeratin antibodies in normal and abnormal human meibum. The experiments indicate that keratins are present in larger quantities in abnormal meibum, which supports a hypothesis that increased keratinization of the meibomian duct epithelium may cause MGD.

Calculations on tear film suggest that a film with its rheological characteristics ought not to be able to retain its optically required structural stability for than a fraction of a second. We have conducted non-invasive measurements of tear film stability on healthy human volunteers showing that the break up time is an average of 35 ± 8 seconds, similar to the results reported by Mengher et al in Current Eye Research, 4, 1-7 (1985) . It has also been shown that tear film has anomalous shear dependent viscosity. We conclude that the structure of tear film is more complex than Holly's simple explanation. In previous investigations we have shown that fibronectrin is present in rabbit and human tear film. We have also shown that fibronectin is secreted from the meibomian glands and is not secreted from the lacrimal glands. Trinkaus-Randall et al (op cit) did not detect fibronectin in corneal epithelial cell cultures. It is our assumption therefore that the fibronectin in tear film is secreted primarily or exclusively from the meibomian glands. Our investigations have also shown that antibodies to a number of adhesion molecules cause tear film disruption in rabbits. We believe that these findings indicate that tear film comprises or consists of an extra cellular matrix and that its surface stability is maintained by virtue of the interactions between fibronectin and various proteins with which it cooperates or between the various proteins themselves.

A new use according to the present invention of a monoclonal antibody to an adhesion molecule is in the

manufacture of a product for use in a method of diagnosis of disorders of the eye.

The present invention has particular utility where the disorder of the eye is dry eye syndrome or is a complication relating to contact lens wear.

In this specification the term adhesion molecule is used in the normal sense. The term therefore includes cell-surface including intrinsic cell membrane proteins and extra cellular matrix molecules which regulate adherence. Cell surface adhesion receptors are preferably of the integrin family which are intrinsic membrane proteins. ECM adhesion molecules preferably contain the tripeptide arginine-glycine-aspartic acid (RGD) as their recognition site. Ruoslahti et al in Science, 238, 491-497 (1987) describe the common RGD sequence found in many adhesion proteins including fibronectin, vitronectin, osteopontin, polygenes, thrombospondin, fibrinogen and Von Willebrand factor. They indicate further that the RGD sequence of all of those proteins are recognised by integrins. Of these proteins, as discussed above, fibronectin has been identified in the eye, specifically in tear film, and so antibody to fibronectin is of particular utility in the present invention, however antibodies to other extra cellular matrix adhesion proteins also present in tear film may also be used in the invention. Likewise antibodies to the fibronectin receptor family of integrins will be of particular utility in the present invention, although antibodies to other integrins, that is receptors to other ECM adhesive proteins present in tear film meibum, lacrimal fluid or corneal epithelial cells, will also be of utility.

The method of diagnosis of the present invention may be carried out by addition of the monoclonal antibody to an in vitro sample or to the eye in vivo. For instance an in vitro method may involve adding the antibody to fluid selected from tear film, with optional purification or removal of contaminating cells, with meibomian fluid, with lacrimal fluid, with corneal epithelial cells, extracts of

such cells or cultured corneal epithelial cells or extracts thereof, including culture medium therefrom. Such in vitro methods will include the necessary steps for identifying antibody/adhesion molecule complexes which may be any of those commonly used in immunological investigations. Of course the antibody may be added to the sample or the sample added to a test portion including the antibody. The term "addition to" as used herein does not necessarily imply any particular order of mixing of components. In vivo methods may include non invasive methods for observing tear film stability, for instance as described by Mengher et al (op cit) by recording the image of a grid reflected from the tear film surface and observing the discontinuities created on administration of the monoclonal antibody and the time taken for recovery of tear film stability. In these techniques, the monoclonal antibody is for instance administered directly into the tear film in the form of a liquid composition including pharmaceutically acceptable liquid carrier and dissolved monoclonal antibody of an appropriate concentration. Other ways of observing the effect of monoclonal antibody on tear film stability in vivo are based on the diagnostic methods for detecting dry eye conditions carried out after administration of monoclonal antibody. These methods (described further by Holly, op cit) include slitlamp examination, the Schirmer test, staining by various dyes, including rose bengal, fluorescein and/or lissamine green, and determination of lysozyme content.

Where the method of diagnosis is for the investigation of contact lens complications, the method of diagnosis carried out in vivo will usually but not always be carried out in the presence of the contact lens. The monoclonal antibody may be administered to either or both of the pre- lens and post-lens films. According to a further aspect of the present invention there is provided a new method of diagnosis of eye disorders by an in vitro investigation in which a fluid

selected from meibum, tear film, lacrimal fluid, corneal epithelial tissue or extract thereof (including cultures thereof) is added to a monoclonal antibody to an adhesion molecule. Although integrins are generally known as cell-surface receptors, Tervo et al (op cit) showed that corneal epithelial cells in culture secreted integrin sub units into the medium. It may be therefore that integrin sub units are secreted by corneal epithelial cells into the extra cellular matrix which forms tear film and that they may contribute to tear film stability, for instance by associating with fibronectin or other adhesion molecules normally present in the extra cellular region. Furthermore such proteins may be secreted by the meibomian glands from which they may become part of the tear film.

In a further aspect of the invention therefore there is provided a method in which tear film or meibum is added in vitro to a monoclonal antibody to an intrinsic membrane protein which is an adhesion molecule receptor, usually integrin or a sub unit thereof. Usually the integrin or sub unit is a receptor for fibronectin, although receptors for other normally extra-cellular adhesion proteins present in tear film may also usefully be detected in this aspect of the invention. The protein may be embedded in cell membrane when being tested but it is more usual to extract it or to use fluid extracted which does not contain cells.

The use of antibodies in the invention may be used as a technique for determining abnormal levels of adhesion proteins which may be used as an indication of abnormal tear film stability, which in turn may contribute to dry eye syndrome and contact lens complications.

Where a disorder is due to excessive quantities of an adhesion molecule in tear film, in a further aspect of the invention the administration of a monoclonal antibody to that molecule or to a receptor for the molecule in the tear film may be used as a method of treatment for the disorder. In a further aspect of the invention disorders due to

reduced quantities of adhesion molecule in tear film are treated.

A new use according to a further aspect of the present invention of an adhesion molecule is in the manufacture of a product for use in a method of treatment of a disorder of the eye.

The disruption of tear film by administration to the eye of antibodies to fibronectin and integrin sub units known to be fibronectin receptors indicates that dry eye may be due to abnormal levels of adhesion molecules in tear film. The identification of fibronectin in meibum and the possible association of meibomian gland disfunction with dry eye would suggest that the successful treatment of dry eye may well be carried out by the administration of fibronectin. The administration of other adhesion proteins, especially proteins containing the RGD sequence, and/or receptors for these, especially integrin and sub units thereof, is also of utility in the present invention.

The composition which is administered for treatment of eye disorders may be of a general type used for administration of active ingredients directly into the eye. It is thus an aqueous based fluid, including a pharmaceutically acceptable aqueous carrier vehicle. The composition may contain components to assist dissolution of the adhesion molecule or its dispersion into the eye. Suitable ingredients are, for instance, surfactants, buffers etc.

The adhesion molecule which is administered in the invention may be a natural product, for instance which is isolated from natural sources, for instance from human or animal sources. Alternatively the adhesion molecule may be produced by bacteria, by cloning the necessary genes into a suitable bacterial or microbial cell. Alternatively the molecule may be made by synthetic techniques, for instance by protein synthesis techniques.

The method of diagnosis of this invention may be used in conjunction with a method of treatment of the present invention.

The invention is illustrated in the following examples:

Example 1 - Detection of Fibronectin

Tear film samples were collected by careful application of a microcapillary tube tip to the lower tear film prism (rate of collection less than 1 μl/minute) over a 15 minute period. Potential cellular contamination was eliminated by centrifugation at 9,500 g for 20 minutes. Meibum was collected by pouching the lower lid, expressing meibum by gentle digital pressure of meibomian glands and harvesting with a chalazion curette using the technique described by Ong et al (op cit) . Fibronectin and total protein in the samples were assayed using monoclonal antibodies of two types (see below) . Total protein was determined by conventional methods.

The determination was carried out on 8 healthy volunteers with normal tear film stabilities.

The determination on tear film from the tear prism was repeated after stimulation of the lacrymal glands chemically or physically.

The results are shown in figure 1. The results indicate that the concentration of fibronectin decreased disproportionately when compared with total protein indicating that fibronectin is not secreted into the tear film via the meibomian gland. Effective mixing between the meibum and the tear film is increased during eyelid blink. Similar concentrations of fibronectin were estimated using monoclonal antibodies to either cellular fibronectin or plasma and cellular fibronectin.

Example 2 - Effect of Monoclonal Antibody to Adhesion Molecules on Rabbit Tear Film

10 μl of phosphate buffered saline including 1 μg of monoclonal antibody to 3 adhesion molecules (α ₁ integrin

sub unit (CD 11a - Cymbus Bioscience) , β ₂ integrin sub unit (CD18 - Cymbus Bioscience) and cellular fibronectin, were administered separately into the eye of a rabbit.

The localised tear film disarray was observed using the apparatus described by Mengher et al, photographing the reflection of the grid pattern from the tear film surface at regular intervals after administration of the monoclonal antibody, until the tear film had returned to normal. The results are shown in figures 2a-f, figure 3a-f and figure 4a-l. The photographs are analysed for the relative shape distortion of the tear film and the results plotted and shown in figure 5.

The results indicate that there is first a latent period of more than 30 seconds during which there is no response, followed by increasing disarray which peaks around 8 minutes. Without any perceptible plateau, the disarray then diminishes over the next 20 minutes, to return to the fourth phase which is consistent with complete recovery of optical integrity. The pattern and the extent (as estimated by counting the number of graticule elements distorted by the administration of monoclonal antibody) shows a similar pattern for each of the three antibodies.

The above results indicate that abnormal tear film stability may be due to abnormal levels of adhesion proteins, and that these may in turn be identified by using monoclonal antibodies to one or more of those adhesion molecules. It is expected that tear film stability would be improved by supplying adhesion molecule to supplement endogenous adhesion molecule so that the normal levels of molecules are present in tear film.