ZA200103884B - Agents and methods for treatment and diagnosis of ocular disorders. - Google Patents

Description

NOVEL AGENTS AND METHODS FOR TREATMENT AND

DIAGNOSIS OF OCULAR DISORDERS

FIELD OF THE INVENTION

The present invention relates, in general, to treatment and diagznosis of ocular disorders. More particularly, the invention is concerned with methods of treating and diagnosing ocular disord ers using sub-immunoglobulin antigen-binding moR ecules.

BACKGROUND OF THE INVENTION

Various ocular disorders are known but for many of these, there= are less than optimal methods of treatrment or diagnosis available. For example, acute anterior uveitis, an inflammation of the iris and ciliary body, is a common condition writh a lifetime cumulative incidence of 0.4% in the general population (Rothova er al, 1987, Am. J.

Opthalmol. 103: 137-1457). Disease tends to start in childhood or young adulthood, is frequently bilateral and is generally recurrent. Acute anterior uveitis is pairaful, is almost always associated with somme morbidity, and can be blinding. This disease is an antigen

I5 driven, delayed-type hypersensitivity response controlled by T lymphocytes (Smith et al, , 1998, Immunol. Cell Biol 76:497-512). Topical corticosteroids are usexd to control inflammation but are themmselves associated with significant complications including cataract, steroid-induced gl=ucoma and infection.

Corneal graft rejection is another ocular disorder that suffers from = paucity of efficacious treatments. The success rate for corneal transplantation in Austral ia is 62% at 10 years (Williams ef al, 1993, Aust. NZ. J. Opthalmol. 21: 1-48). Irreversibles rejection is the major cause of corneal graft failure, accounting for at least 30-50% of al 1 cases, and

N occurs despite universal use of topical glucocorticosteroids (Williams et al, 1993, Aust. . NZ. J. Opthalmol. 21: 1-48 ). Adjunctive treatment with systemic immunos uppressants such as cyclosporin A is used occasionally, but probably confers little bene fit on graft survival and can be associated with serious morbidity (Williams er al, 1993,

Transplantation Reviews.T: 4-4-64).

Herpetic keratitis is «caused by herpes simplex virus infection of the trigemina] ganglia (Streilein er. al. 1997 | Immunol. Today. 18: 443-9, Hendricks, R.L. 19997, Cornea. 16: 503-6). Herpetic eye disease is the most common infectious cause of bRindness ip developed countries (Streilein er. al. 1997, Immunol. Today 18: 443-9) and serious - sequelae including pain, iritis, bacterial superinfection, corneal perforation and blindness occur in 3% of affected individuals. Epithelial herpetic disease (eg. dendritic ulceration) is i readily diagnosed by polymerase chain reaction (PCR)-based tests, amongst others, and can be treated reasonably successfully with antiviral agents. Diagnosis and treatment are far more difficult for disciform endotheliitts, or for the stromal disease (eg. disciform keratitis, irregular stromal keratitis, kerato-uv-eitus) that occurs in one fifth of patients with ocular herpes. It is unusual to find virus particles in stromal biopsy specimens, and very difficult to culture virus from them. Diagnosis is usually made on clinical grounds and by exclusion. Topical acyclovir is the treatmemt of choice for herpetic keratitis (although emergence of drug-resistant strains is causirmg concern) and long-term oral prophylaxis halves the incidence of recurrences (Herpetic Eye Disease Study Group. 1998, N. Engl. J.

Med 339: 300-6). However, corneal damaage is not prevented by acyclovir. The pathogenesis of epithelial disease results from productive, lytic infection of comeal

I5 epithelial cells. In contrast, the pathogenesis of stromal or endothelial cell herpetic disease results from an immune reaction to viral antige=n expressed on corneal cells or released into the stroma (Streilein er. al. 1997, Immunol. Tocday 18: 443-9, Hendricks, R.L. 1997 Cornea 16: 503-6).

Reference also may be made to acanthamoeba keratitis, a very serious, painful disease that can result in loss of vision in the affected eye. The condition, caused by a common, free-living, soil and freshwater amoeba (Acanthamoeba), was once thought to be rare, but there has been an exponential increas e in the number of reported cases over the past 5-10 years. Over 90% of cases have occurred in contact lens-wearers who have not adhered to recommended lens cleaning and disimfection procedures. Of critical importance i. 25 in achieving a good visual outcome is early diagnosis. In the early stages of infection, amoebae are found in the corneal epithelium where they can be removed by debridement, “ often without the need for antimicrobial agents (Brooks et al. 1994, Cornea 13:186-189).

Unfortunately, the diagnosis is often delayed, as the clinical picture can resemble other forms of infectious keratitis, particularly herpetic keratitis. This allows the organisms to invade the corneal stoma where they attack keratocytes (Badenoch er al. 1995, [nr J

Parasitol. 25: 229-239) and cause irreversible tissue damage. In this situation, attempting to achieve a laboratory diagnosis by corneal s<rapings is usually unsuccessful and deep ; biopsy, itself damaging, is unreliable.

Immunoglobulins have gained widespread diagnostic and therapeutic applicat ion in . various medicinal flelds. For example, who le monoclonal and/or polyclonal antibodies have been used to suppress transplant rejection, to modulate different autoimmune

N diseases, to treat neoplasias, and to prevent and treat infectious diseases. Typically, systemic administration of antibodies is required for these purposes, which may cause serious systemic side effects. These side effect.s have prevented the systemic application of antibodies for treating diseases affecting non-vatal organs like the eye. However, there are many ocular diseases such as those described aBove, where antibody treatment would Fave advantages.

In view of the above, Whitcup er al (WO 93/06865) disclose the use of monoclonal antibodies against cell adhesion molecules to treat ocular inflammation. In particular,

Whitcup et al contemplate administration of s uch antibodies to the eye by intravenous injection, by intracameral or periocular injectiorn, by surgical implantation of a depot, and by topical administration using eye drops or ophthalmic ointment. Whitcup et al prov ide enabled methods for parenteral administration of anti-cell adhesion molecule antibodies in an animal model. However, they do not pr-ovide any enabled methods for topical administration of such antibodies and are wholly silent on any data establishing the utilaty of the claimed methods. In fact, the present invemtors have found that conventional whole antibodies alone cannot penetrate the cornea or thes sclera via topical routes.

Various anatomical barriers relating to thie eye may underlie the poor intraocular penetrance of whole antibodies. In this regard, the cornea is the principal barrier to entry of foreign substances. It has two distinct penetra_tion barriers, the corneal epithelium arad the comeal stroma (Burstein and Anderson, 19 85, J Ocul. Pharmacol. 1(3): 309-265;

Maurice DM, 1980, Int. Ophthalmol. Clin. 20(23):7-20; and Mishima S, 1981, Invesz.

Ophthalmol. Vis. Sci, 21(4): 504-41). The corneal epithelium is a major barrier for hydrophilic substances. Hydrophilic molecules less than 350 Da can enter via a paracellular route, but larger hydrophilic molecules are essentially barred. This barrier function can be modulated to an extent by the use of penetration enhancers. The second barrier, the corneal stroma, interferes with penetration of lipophilic drugs and large hydrophilic drugs. The comeal stroma is regularly quoted as being “transparent” to hydrophilic molecules of up to 500 kDa (Bartlett and Jaanus in Clinical Ocular

Pharmacology). This figure however is contradicted by studies of Olsen et af (1995, Inv».

Oph thal. Vis. Sci 36(9): 1893-1903) who showed in vitro that human scleral permeability - decrseased linearly with increasing molecular wesight of hydrophilic compounds up to about 40 k_Da after which permeability decreased sharply. Comparative studies with bovine and i rabb®t stroma and sclera suggest that stroma is similar though slightly more transparent (Maurice and Polgar, 1977, Exp. Eye Res. 25: 577-582). Accordingly, the finding of Olsen et al (1995, supra) would be expected to apply t© human comeal stroma.

Penetration of sclera or stroma of an intact eye by whole IgG occurs very slowly, if at all_, regardless of whether application is topic-al (Pleyer et al 1995, Invest. Ophthalmol.

Vis. =Sci. 36(1): 52-6), subconjunctival (Osusky er al 1993, Graefes Arch. Clin. Exp. OphtFralmol. 231(2): 122-128) or systemic (Vertmagen et al 1990, Invest. Ophthalmol. Vis.

Sci. 3 1(8): 1519-1525). Formulation of whole MgG in a lipid environment e.g., within a lipososme, can facilitate transfer across the corneall epithelium (Pleyer ef al 1995, supra) but does rot resolve the problem of slow passage acr-oss the stroma or sclera for molecules of this siz=ze.

Despite the encouraging data of Pleye:r er al (1995, supra) with liposomal encapssulation, the present inventors have shown that intracameral injection of monoclonal antibodies causes a harmful fibrinous reaction in tZhe anterior chamber of the eye (Williams et al. 1992, Transplantation, 54: 38-43). This pero-inflammatory side effect may reflect local c:omplement activation by the Fc-portion off the antibody. The eye contains many non-re_zZenerating structures that are irreversibly damaged by such inflammatory side effects. At least in one reported case, the inflammaatory side effect of intraocular antibodies had blinding consequences for the patient (case reported by M. Béhnke. Bern, presented at the inte=rnational conference on corneal transplantation, Heidelberg, September 1994).

From the foregoing, and despite a long-felt need for intraocular immunotherapeutic i 25 strategies, it would appear that antibodies have limited clinical potential for efficacious . treatme=nt of ocular disorders.

SUMMARY OF THE INVENTION

The present inventors have discovered that sub-immunoglobulin antigen-binding molecu les such as Fv immunoglobulin fragments, rminibodies and the like can penetrate the cornea to provide efficacious treatment and/or diagznosis of ocular disorders with markedly reduced side effects.

Accordingly, in one aspect, the invention provides a method of treating an ocular . disorder, comprising administerirg to a patient in need of such treatment, an effective amount of a sub-immunoglobulin antigen-binding molecule that is immumo-interactive ) with a target antigen associated with the disorder.

In another aspect of the irvention, there is provided a method of d iagnosing an ocular condition, comprising: - contacting a portion of an cye with a sub-immunoglobulin antigen-binding molecule that is immuno-interactive with a target antigen indicative of the condition; and detecting the presence of a «complex comprising the said antigen-binding molecule and the target antigen.

In yet another aspect of the i nvention there is provided a composition for treatment or diagnosis of an ocular disorder, comprising a sub-immunoglobulin antigen-binding molecule that is immuno-interactive with a target antigen associated with the disorder, together with a carrier.

In a further aspect, the invention resides in an ocular composition compri sing a sub- immunoglobulin antigen-binding mo lecule that is immuno-interactive with a target antigen in the eye.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1, taken from Pliickthhun et al (1996, Antibody engineering: A practical approach. 203-252), is a schematic representation of sub-immunoglobulin antigen binding molecules suitable for functional high level expression in E. coli and for use in accordance with the present invention. = Figure 2 illustrates a comeal perfusion chamber: A corneo-scleral button is dissected and tightly clamped into the perfusion chamber. A fine thread allows clamping adjustment without the risk of tissue damage. The fluid on the "recipient" corneal side circulates continuously. The total fluid volume is 3.5-4 mL. The elevated chamber : outflow into the reservoir causes a positive pressure inside the chamber mimicking physiological ocular conditions.

Figure 3 illustrates a chamber adapted for measuring penetrance of a drug into whole human or pig eyes. Whole eye s with intact conjunctival tissue are positioned onto the Teflon block such that the cornea is harboured: within a well formed in the Teflon ~ block. The well is filled with the drug such that th e cornea is the only tissue In contact with the drug. The whole eye is covered with an inve rted plastic container to form a humid ) charnber and the set-up is placed above a water bath to keep the eye at a temperature of 35 -37=C. -

Figure 4 is a graph showing antigen-binding a ctivity measured in a flow cytometry assaw. The scFv shows concentration dependent b inding to its target antigen, with a patte-rn similar to that obtained with the whole antibody.

Figure 5 is a graph showing scFv penetration tkarough pig corneal stroma (epithelial barrier removed) at pH 7.0.

Figure 6 is a graph showing scFv penetration tiarough pig comeal stroma (epithelial barriezr removed) at pH 8.0, pH 7.0 and pH 3.5. Loweering the pH from 7.0 to 3.5 has an adver se effect on penetration. In a cornea with intact epithelium, no penetration could be detected with scFv at a pH of 3.5.

Figure 7 is a graph showing that in cat corneas without epithelial barriers pure scFv penetrated rapidly with increasing binding activity on t_he donor side after 3 hours. Feline comeas are 20-40% thinner than porcine corneas explafning the faster penetration. Whole antibo dies (mAb) did not penetrate within an observatio n period of § hours.

Figure 8 is a graph showing scdium flucrescei= penetration through an intact pig comeam in the perfusion chamber. On the recipient side of the pig cornea, sodium fluorescein became detectable only after 3 hours with a subsequent exponential increase during the rest of the experiment

Figure 9 is a graph showing the relationship between penetration of scFv in corneal i perfusion and whole globe penetration experiments. The arrow indicates scFv-binding activitsy in the anterior chamber of the whole eye (withoiat epithelium).

Figure 10 is a graph showing OX38 scFv penetr ation through intact pig comeas in perfusion chambers in the presence and absence of 0.5% capric acid.

Figure 11 is a graph showing penetration of miniantibodies and whole IgG- antibocdies through pig corneal stroma in the perfusion clamber.

Figure 12 is a graph showing penetration of miniantibodies throtagh corneal stroma - and intact corneas in whole pig eyes. . Figure 13 is a graph showing penetration of OX38 miniantibodiess through an intact human comea in the who le eye penetration set-up.

Figure 14 is a graph showing titration of miniantibody binding acEivity in eye drops and after penetration in the anterior chamber.

DETAILED DESCRIPTION OF THE INVENTION

1. Definitions

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the art to which the invention belongs. Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, preferred methods and materials are described. For the purposes of the present invention, the following terms are defined below. ts By “affinity” is me ant the strength of the interaction between an individual antigen binding site on an immunoreactive molecule and its corresponding site on the antigen.

By “associated with the disorder’ is meant that the target antigen may be linked directly or indirectly with tine disorder. For instance, the target antigen may be an epitope of a signal molecule that indirectly causes, contributes to or is indicative of the ocular disorder.

The term “diagnosis "is used herein in its broadest sense to include detection of an . antigen reactive to a sub-immmunoglobulin antigen-binding molecule. Also i ncluded within its scope, 1s the analysis of disorder mechanisms associated with ocwlar disorders. - Accordingly, the term "diagmosis” includes the use of sub-immunoglobulin antigen binding molecules for research purposes as tools to detect and understand mechanisms associated with ocular disorders.

The term “effective eamount” as used herein is an amount of sub-immunogiobulin antigen binding molecule that will treat or prevent an ocular disorder. The effective amount may be an amount that can inhibit an action of a target antigen for which the sub- immunoglobulin antigen-binding molecule is reactive thereagainst. Method_s of inhibiting

. RN 8 - an action may include saturating the target antigen in the eye with aa sub-immunoglobulin : antigen-binding molecule thereby neutralising the target antigen. ~ Reference herein to “immuno-interactive” includes reference to any interaction, reaction , or other form of association between molecules and in particular where one of the molecules 1s, or mimics, a component of the immune system.

The term “ocular condition” is used herein to describe any condition of the eyes.

It may include a healthy condition or an unhealthy condition such as a disorder of the eye (ocular disorder).

The term “patient” refers to patients of human or other anima. origin and includes ~ any individual it is desired to examine or treat using the methocds of the invention.

However. it will be understood that “patient” does not imply that symptoms are present.

By “pharmaceutically acceptable carrier’ is meant a solid or liquid filler, diluent or encapsulating substance that may be safely used in systemic administration. .

By “sub-immunoglobulin antigen binding molecule” is mean t a molecule that is smaller thhan an immunoglobulin molecule and has binding affinity for— a target antigen. It will be understood that this term extends to non-immunoglobullin derived protein framewor ks which exhibit antigen binding activity. It will also be wminderstood that this term is not limited by the source from which any nucleotide sequenc e encoding the said antigen-bi nding molecule is obtained, or by the source from which the said antigen-binding molecule 3s expressed or isolated.

By~ “target antigen” is meant an antigen that is associated with an ocular disorder for which treatment or diagnosis is sought.

Th e term “treating” is used herein in its broadest sense to incliade both therapeutic i and prophylactic (i.e., preventative) treatment designed to ameliorate thes ocular disorder. - 25 Thoroughout this specification, unless the context requires otherwise, the words “comprises”, “comprises” and “comprising” will be understood to impRy the inclusion of a stated inte=ger or group of integers but not the exclusion of any other integer or group of integers.

Wa 00/40262 PCT/AU99/01163 2. Sub-immunoglobulin antigen binding molecules ) The subject invention stems from the unexpected discovery that sub- . immunoglobulin antigen-binding molecules can pen etrate the cornea to significant levels when applied topically to the eye. In light of this discovery, the present inventors believe that penetrance of such molecules into the eye can also be achieved by alternate routes in cluding parenteral administration and periocular injection.

To allow ticatment of ocular disorders, it is preferable that sub-immunoglobulin antigen-binding molecules of the invention fulfil the following conditions. Firstly, it is de sirable that they should allow for application to the eye to avoid or at least partially all eviate some systemic side effects. A second condition is that these binding molecules must not cause any potentially harmful pro-inflammatory side effects inside the eye.

Suitable sub-immunoglobulin antigen bindieg molecules include, but are not res tricted to, Fv, Fab, Fab' and F(ab’), immunoglobu lin fragments. Preferably, the sub- immxmunoglobulin antigen-binding molecule does not comprise the Fc portion of an imrmunoglobulin molecule.

Preferably, the sub-immunoglobulin antigem-binding molecule comprises a synthetic Fv fragment. Suitably, the synthetic Fv Eragment is stabilised. Exemplary syn thetic stabilised Fv fragments include single chain Fv fragments (sFv, frequently termed scFv) in which a peptide linker is used to bridges the N terminus or C terminus of a

Vy domain with the C terminus or N-terminus, respecti-vely, of a V; domain. ScFv lack all constant parts of whole antibodies and are not able to activate complement. Suitable peptide linkers for joining the Vy and V, domains are those which allow the Vj; and V; domains to fold into a single polypeptide chain having an antigen binding site with a three . dimensional structure similar to that of the antigen binding site of a whole antibody from which the Fv fragment is derived. Linkers having the desired properties may be obtained - by the method disclosed in U.S. Patent No 4,946,778, the disclosure of which is incorporated herein by reference. However, in some cases a linker is absent. Exemplary synthetic Fv fragments are illustrated in Figure 1.

ScFvs may be prepared, for example, in accordance with methods outlined in Kreber er al (Krebber er al. 1997, J. Immunol. Methods; 201(1): 35-55), the disclosures of which are incorporated herein by reference. Alternatively, they may be prepared by methods described in U.S. Patent No 5,091,513, Eumopean Patent No 239,400 or the articles by Winter and Milstein (1991, Nature 349:293) and Plickthun er al C1996, In > Antibody engineering. A practical approach. 203-252), which are incorporated erein by reference.

Alternatively, the synthetic stabilised Fv fragment comprises a disulphide stabilised

Fv (dsFv) in which cysteine residues are introduced into the Vy and V; domains such that in the fully folded Fv molecule the two residues will form a disulphide bond theretoetween.

Suitable methods of producirmmg dsFv are described for example in (Glockscuthesr er al.

Biochem. 29: 1363-1367; Reiterer al. 1994, J. Biol. Chem. 269: 18327-18331; Reiter et al. 1994, Biochem. 33: 5451-54599; Reiter et al. 1994. Cancer Res. 54: 2714-2718; W.ebber er al 1995, Mol. Immunol. 32: 24-9-258), which are incorporated herein by reference.

Also contemplated as sub-immunoglobulin antigen binding molecules are single variable region domains (terrmeed dAbs) as for example disclosed in (Ward et a”. 1989,

Nature 341: 544-546; Hamers -Casterman et al. 1993, Nature. 363: 446-448; Daavies &

Riechmann, 1994, FEBS Lett. 339: 285-290), which are incorporated herein by reference.

Suitably, the sub-immu noglobulin antigen-binding molecule is a “"minibod_y". In this regard, minibodies are small versions of whole antibodies, which encode in a_ single chain the essential elements of a whole antibody. Suitably, the minibody is comprised of the Vy and V, domains of a native antibody fused to the hinge region and CH3 dormain of the immunoglobulin molecule as, for example, disclosed in U.S. Patent No 5,837,821, which is incorporated herein by reference.

In an altemate embodirment, the sub-immunoglobulin antigen binding molecule may comprise non-immunoglobeulin derived, protein frameworks. For example, ref erence may be made to (Ku & Schultz, 1995, Proc. Natl. Acad. Sci. USA, 92: 652-6556) which - discloses a four-helix bundle pzotein cytochrome b562 having two loops randomi sed to create complementarity determiring regions (CDRs), which have been selected for a_ntigen i binding.

The sub-immunoglobuli n antigen-binding molecule may comprise a modifying moiety. In one form, the modifying moiety may modify the effector function =f said molecule. For example, the modifying moiety may comprise a peptide for detect-ion of said antigen-binding molecule, for example in an immunoassay. Alternative] y, the modifying moiety may facilitate purification of said antigen binding molecule. Mn this instance, the modifying moiety includes, but is not limited to, glutathione-S-transsferase

(GST), maltose binding protein (MBP) and hexahistidine (HISq), which are particularly - useful for isolation of the antigen-binding molecule by affinity chromatography. For the purposes of purification by affinity chromatography, relevant matrices for affinity ) chromatography are glutathione-, amylose-, and nickel- or cobalt-conjugated resins respectively as is well known in the art. The modifying moiety may comprise a therapeutic agent to be targeted to at least a portion of the eye. Alternatively, the modifying moiety may comprise an enzyme such as an enzyme of potential therapeutic benefit e.g., fibrolytic agents (Holvoet ef al., 1991, J Biol Chem. 36G: 19717-19724; Yang et al. 1994, Biochem. 33: 606-612) or prodrug activators (Bosslett et al, 1992, Br. J Cancer, 65: 234-238;

Goshorn et al., 1993, Cancer Res., 53: 2123-2127; Rodrigues et al., 1995, Cancer Res., 35: 63-70) and enzymes capable of driving colorimetric assays for use in immunoassays. In another form, the modifying moiety may increase intraocular penetrance of said antigen- binding molecule. In this instance. the modifyi ng moiety may comprise a lipid moiety or lipid tail as for example described by Barr er cal (1998, Adv. Drug Delivery Reviews 32: 247-271, incorporated herein by reference). Alternatively, the modifying moiety may comprise the HIV tat protein as described for e-xample by Schwarze er al (1999, Science 285: 1569-1572), which is incorporated herein by reference.

The sub-immunoglobulin antigen binding molecule may be multivalent (i.e., having more than one antigen binding site) providing it i s capable of penetration into the eye when administered parenterally or topically or injected into the subconjunctival, peri- or retrobulbar space, or injected directly into the eye. Such multivalent molecules may be specific for one or more antigens. Multivalent molecules of this type may be prepared by dimerisation of two antibody fragments through a cysteinyl-containing peptide as, for example disclosed by (Adams et al., 1993, Caracer Res. 53: 4026-4034; Cumber et al., " 25 1992, J. Immunol. 149: 120-126), which are incorporated herein by reference.

Alternatively, dimerisation may be facilitated by fusion of the antibody fragments to . amphiphilic helices that naturally dimerise (Pack P. Plinckthun, 1992, Biochem. 31: 1579- 1584, incorporated herein by reference), or by use of domains (such as the leucine zippers jun and fos) that preferentially heterodimerise (XKostelny et al., 1992, J Immunol 148: 1547-1553, incorporated herein by reference). In an alternate embodiment, the multivalent molecule may comprise a multivalent single chain antibody (multi-scFv) comprising at least two scFvs linked together by a peptide linker. In this regard, non-covalently or covalently linked scFv dimers termed "diabodiess" may be used. Multi-scFvs may be bispecific or greater depending on the number o f scFvs employed having different antigen ) binding specificities. Multi-scFvs may be prepared for example by methods disclosed in i U.S. Patent No. 5,892,020, which is incorporated herein by reference. 3. Compositions

The invention also extends to a compositi=on for use in treatment or diagnosis of ar: ocular disorder wherein the composition comprises a sub-immunoglobulin antigen-bindingz molecule that is immuno-interactive with a targ et antigen associated with the disorder. together with a carrier.

The carrier 1s preferably a pharmaceutically acceptable carrier. Depending upon the particular route of administration, a variety of pharmaceutically acceptable carriers,” well known in the art may be used. These carriers may be selected from a group including sugars, starches, cellulose and its derivatives, malt, gelatine, talc, calcium sulfate, vegetable oils, synthetic oils, polyols, alginic acid, phosphate buffered solutions, emulsifiers, 1sotonic saline, and pyrogen-free water_

Any suitable route of administration may bes employed for providing a patient with the composition of the invention. For example, ora_l, rectal, parenteral, sublingual, buccal, intravenous, intra-articular, intra-muscular, intra-dermal, subcutaneous, inhalational, intraocular, intraperitoneal, intracerebroventricular-, transdermal and the like may be employed. Preferably, a parenteral or topical rosute is employed. Alternatively, the composition is suitably administered by injection imto the subconjunctival, peribulbar or retrobulbar space.

Dosage forms include tablets, dispersions, suspensions, injections, solutions,

CL syrups, troches, capsules, and suppositories, aerosols, collagen shields, transdermal patches and the like. These dosage forms may also include injecting or implanting controlled - 25 releasing devices designed specifically for this pourpose or other forms of implants modified to act additionally in this fashion. Contreolled release of said antigen binding molecule may be effected by coating the same, for =example, with hydrophobic polymers including acrylic resins, waxes, higher aliphatic alcolaols, polylactic and polyglycolic acids ] and certain cellulose derivatives such as hydroxyperopylmethyl cellulose. In addition, controlled release may be effected by use of other polymer matrices, liposomes and/or microspheres.

Pharmaceutical compositions of the present inventio n suitable for oral or parenteral i administration ray be presented as discrete units such as c.apsules, sachets or tablets each ) containing a pre-determined amount of one or more therape=utic agents of the invention, as a powder or granules or as a solution or a suspension in an aqueous liquid, a non-aqueous liquid, an oil-in—water emulsion or a water-in-oil liquid emu. Ision. Such compositions may be prepared by any of the methods of pharmacy but all methods include the step of bringing into as sociation one or more immunogenic agentss as described above with the carrier which co nstitutes one or more necessary ingredients. In general, the compositions are prepared by uniformly and intimately admixing the ant-igen binding molecule of the invention with liequid carriers or finely divided solid carriers or both, and then, if necessary, shaping the prodwmuct into the desired presentation.

The abov-e compositions may be administered in a manner compatible with the dosage formulatieon, and in such amount as is therapeutically effective to alleviate patients from ocular diso rder. The dose administered to a patient, “in the context of the present invention, should be sufficient to effect a beneficial response in a patient over time such as a reduction or cessation of blood loss. The quantity of the artigen binding molecule(s) to be administered rmay depend on the subject to be treated incl usive of the age, sex, weight - and general health condition thereof. In this regard, precise armounts of the antigen binding molecule(s) for @dministration will depend on the judgem ent of the practitioner. In determining the effective amount of the antigen-binding molecule to be administered in the treatment of an ocular disorder, the physician may evaluate th e progression of the disorder over time. In any event, those of skill in the art may readily dletermine suitable dosages of the antigen-bindirag molecule of the invention. Such dosagzes may be in the order of nanograms to mill igrams of the antigen-binding molecule. ) 25 The composition suitably includes a penetration enhancer such as benzalkonium } chloride, digitonir, dihydrocytochalasin B, capric acid, increeasing pH from 7.0 to 8.0.

Preferably, the pesnetration enhancer is directed to enhancingz penetration of the antigen binding molecule( s) through the corneal epithelium. :

Suitably, the composition includes liposomes in which the antigen-binding molecules of the Invention are encapsulated. Any suitalble procedure of liposome encapsulation of active is contemplated. In this regard, exem.plary methods are disclosed by Frasta et al (1 999, J Pharm. Pharmacol. 51(5): 565-5765) and Masuda et al (1996,

Invest. Ophthalmol. Vis. Sci. 37(9): 1 914-1920), which are incorporated herein by i reference. - The half-life of the antigen-bindin_g molecule of the invention may be prolonged by any suitable procedure if desired. Preferably, such molecules are chemically modified with polyethylene glycol (PEG), including maoenomethoxy-polyethylene glycol, as for example disclosed by Chapman et al (1999, Nature Biotechnology 17: 780-783), which is incorporated herein by reference.

Suitable target antigens include, bumt are not limited to, an MHC molecule including

MHC molecules that are class I- or class II-restricted, a co-stimulatory molecule (eg,

CD80, CD86 and CD152), an adhesiorh molecule (eg, CDllble, CD18, CID54 and

CD62L), a receptor-associated molecule (eg., CD3, CD4, CDS, CD28, CD40. CD<40L and

CTLA4), a cytokine receptor (eg., the interleukin 2 receptor (IL-2R), or subunit thereof such a CD25, and the interferon y receptomr (IFN-yR) and a viral surface antigen (eg., gD2 and gB2 antigen of herpes simplex virus or a surface antigen of the herpes virus causing herpetic keratitis).

The invention also provides =n ocular composition comprising a sub- immunoglobulin antigen-binding molecule that is immuno-interactive with a target antigen in the eye. 4. Methods of administration

A further feature of the inventiorn is a method of treating an ocular disorder comprising administering to a patient in need of such treatment an effective amount of a sub-immunoglobulin antigen binding molecule that is immuno-interactive with a target . antigen associated with the disorder.

The ocular disorder may be any disease or disorder associated with the eye. ) 25 Generally the disorder will be a disease of whe eye which directly affects the eye. Diseases or disorders associated with the eye may bee those diseases or disorders that originate from other parts of the body and that indirectly a_ffect the eye as a side effect of the main disease or disorder.

Typical disease or disorders inclu de corneal graft rejection, uveitis, any ocular infection, inflammatory and infectious diseases such as viral (e.g, herpetic keratitis and adenoviral), bacterial or chlamydial conjunctivitis, ocular tumours, neowascular proliferative da seases such as diabetic retinopathy, neovascular maculopathies, rheumatoid } corneal melting disorders or autoimmune disorders such as ocular penthi goid. ) The me thod includes administering an effective amount of the su b-immunoglobulin antigen-binding molecule to the patient. The sub-immunoglobulima antigen binding molecule may bye administered locally, either topically to a portion of the eye or be injected into the eye for instance into the subconjunctivital, peri- or retrobulbar space or directly into the eye. Alternatively, said antigen binding molecule may be administered systemically by parental administration.

It is momst preferable that said antigen-binding molecule is applied to the eye topically and as an eye drop. The eye drop may be applied to the comea (clear part in the centre of the ewe) thereby allowing the molecules to permeate into the eye. For the treatment of a dxsease affecting the posterior of the eye, it may be most «desirable that the antigen-binding molecule penetrates the sclera when injected under the conjunctiva or around the globe .

The adm inistering of the antigen-binding molecule may be pesrformed after a preliminary step of modulating the surface of the eye to improve pe netration of the molecules. Preferably, the epithelial layer such as the comeal epithelium is modulated by a penetration enharacer to allow for a sufficient and rapid penetration of the amolecules as for example described above.

Alternativ-ely, the step of administration is characterised by subjecting at least a portion of the eye to iontophoresis such that the antigen-binding molecule penetrates into a desired intraocula r region. Suitable methods that may be employed in this regard include those disclosed, f<or example, by Behar-Cohen et al (1998, Invest. Ophth almol. Vis. Sci. 39(6): 897-904), and Frucht-Pery er al (1996, Graefes Arch. Clin. Exp. Ophthalmol 234(12): 765-9), vwhich are incorporated herein by reference. i The portion of the eye into or onto which the antigen-bindirag molecule is preferably administered is the portion that allows for penetration of the antigen binding molecule. As des.cribed above, administration is preferably performed on the cornea and conjunctiva or the substance may be injected into the subconjunctival, peri - or retrobulbar spaceto reach the inside of the eye after penetration through the sclera.

S. Methods of diagnosis

The invention also extends to a method of diagnosis of an ocular condition, . comprising contacting a portion of an eye with a sub-immunoglobulin antigen binding molecule that is immuno-interactive with an antigen indicative of the condition, and detecting the presence of a complex comprising the sub-immunoglobulin antigen binding molecule and the antigen.

In a preferred aspect, the invention provides a method of” diagnosis of an ocular disorder, compris ing contacting a portion of an eye with a sub-ianmunoglobulin antigen binding molecule that is immuno-interactive with an antigen associated with the disorder, and detecting the presence of a complex comprising the sub-irmnmunoglobulin antigen binding molecule and the antigen.

The antigen that is indicative of the condition may be an aritigen that is associated directly or indirectly with the ocular condition. It may be a target antigen or an antigen of a molecule that results as a by-product of the ocular condition.

The preserace or absence of the target antigen in the eye nay be determined by administering the antigen binding molecule to the eye as for example described in Section 4 above. Preferably, the antigen-binding molecule is administered to the eye as an eye drop or by periocu lar injection.

Detection of the complex is subsequently determined. Any suitable technique for determining formation of the complex may be used. For example, a sub-immunogiobulin antigen-binding molecule according to the invention having a dete ctable label associated therewith may be utilised in immunoassays applicable to in vivo detection of target antigen.

Immunoassays may include competitive assays as understood in the art.

The label associated with the antigen-binding molecule may mnclude the following: : 25 (1) dire ct attachment of the label to the antigen-binding molecule; (i1) indirect attachment of the label to the antigen binding molecule; ie, attachment of the label to another assay reagent which subsequently binds to the antigen binding molecule; and (iif) attachment to a subsequent reaction product of the antigen binding molecule. The label may be selected from a group including a chromogen, a catalyst,

an enzyme, a fluorochrome, a chemL luminescent molecule, a lanthanide 1on such as . Europium (Eu). a radioisotope and a d irect visual label. ] In the case of a direct visual Jabez), use may be made of a colloidal metallic or mon- metallic particle, a dye particle, an enzyme or a substrate, an organic polymer, a latex particle, a liposome, or other vesicle con taining a signal producing substance and the like.

A large number of enzymes suit=able for use as labels is disclosed in United States

Patent Specifications U.S. 4,366,241, U.S. 4,843,000, and U.S. 4,849,338, all of which are incorporated herein by reference. Suitable enzyme labels useful in the present inventmon include alkaline phosphatase, horseradish peroxidase, luciferase, [B-galactosidase, glucose oxidase, lysozyme, malate dehydrogenasse and the like. The enzyme label may be ussed alone or in combination with a second en=zyme that is in solution.

Suitably, the fluorochrome is selected from a group including fluoresce=in isothiocyanate (FITC), tetramethylrhodarmine isothiocyanate (TRITL) or R-Phycoerythrin (RPE).

A preferred isotope for imaging applications is ¥mTe, because of its physic al properties, cost availability and safety. Imm order to facilitate labelling with this radiometal, antigen-binding molecules such as scFv mave been fused with a short peptide designed t=o chelate ®™Tc within a N3S co-ordination site (Huston er al. 1996, Quarterly Journal of

Nuclear Medicine, in press; George et af’., 1995, Proc. Natl. Acad Sci. USA, 92: 835- 8363). The resulting molecule is readily labelled in a single step reaction, yielding a highly stable product that retains the immmogenicity of the original scFv.

The present invention will now Wbe more fully described with reference to th e following examples. It should be understood, however, that the description following i s - illustrative only and should not be taken in. any way as a restriction on the generality of thee invention described above.

EXAMPLES

. EXAMPLE 1

Intraocular penetration of scFv antibody fragmerats and divalent miniantibodies

Engineering and expression of scFv antibody fragments and divalent miniantibodies.

Starting material for the antibody engineering was «0X38, a hybridoma cell line that produces whole antibodies with binding specificity against rat CD4 antigen. The enginee ring process for the scFv antibody fragment follow ed exactly a technique described by Kreber et al (1997, J Immunol Methods 201(1): 35-55) and Pliickthun er al (Producing antibodies in Escherichia coli: from PCR to fermentation. In: McCafferty J, Hoogenboom

H, Chiswell D, eds. Antibody engineering: A practical approach. Oxford: Oxford -

Univers ity Press, 1996:203-252. (Hames B, ed. The practical approach series)). In brief, mRNA was extracted from a growing hybridoma cell lime and the mRNA was reverse transcribed into cDNA. The genetic information for the variable part of the light and.the heavy chain of the antibody was separately amplified by polymerase chain reactions (PCR) using ar extended primer set as described by Kreber et al (supra) and Pliickthun et af (supra). The DNA products for both chains were spliced together and inserted into the pAK10C: vector that was used to transform £. coli JM83 (gift from A. Plickthun). ScFv fragments were expressed in E. coli and collected from #he culture supernatant. These fragments were tested for their binding activity in a flow cy tometer using rat thymocytes as a source of the rat CD4 antigen. } For large-scale production, the insert coding for a functional scFv was cut out of the pAK_1100 vector and integrated into a pHB400 vector that was derived from integrating : the skp co-expression cassette from vector pHB110 (Bothmann er al, 1998, Nar. Biotechriol. 16(4): 376-80) into the vector pAK400 (Kresber et al. supra). The vector pHB400 contains a strong translation initiation region (Shine-Dalgamo SDT7gl0 sequences) for optimised scFv expression, a 6-histidine tag for purification by an immobilised metal ion-affinity chromatography (IMAC) and a cassette for co-expression of skp, a_ chaperon that improves correct scFv folding in the periplasmic space.

For the expression of scFv fragment s as divalent miniantibodies, the insert with the scFv sequence was ligated into vector pHE3540, which vector was constructed by adding . the scFv dimerisation cassette from pAK500 (Kreber er af, supra) into vector pHB400. By this means, scFv fragments can be expressed with an additional double helix motif derived from an antibody hinge region that allows two monovalent fragments to dimerise non- covalently to form a divalent miniantibody with a size of 66 kDa.

Fermentation of £. coli

E. coli HB2151 or E. coli BL21 were transformed with the above vectors for scFv or miniantibody expression. Bacterial cultures were grown in a 20 L bench-top fermenter similar to the one described by Pliickthum et al (supra) using terrific broth complex medium. Cultures were induced for protei n expression at 25°C and an optical density (ODgoonm) of 20 by addition of 0.1 mM IPTG (isopropyl B-D-thiogalactopyranoside). After three to four hours of expression, cultures were harvested in a Beckman J21 centrifuge (4000xg for 25 min at 4°C) and bacterial pel lets either processed immediately or stored at -70°C.

Purification of scFv and miniantibodies

To obtain correctly folded and functional scFv or miniantibodies, the bacterial peliets were resuspended in 50 mM sodium borate (NaBo), 0.15 M NaCl, pH 8.0 using 2 mL extraction buffer per gram of pellet and bacteria disrupted by sonication in a Manton- Gaulin flow-through homogeniser. The lysate was centrifuged at 13,000xg for 30 min at 4°C before the scFv or miniantibody-containing supernatant was passed through a 0.2 pm filter and supplemented with 1% Triton X-100, 1% Tween-20 and 20mM imidazole. The ’ clarified lysate was loaded onto an IMAC column with nickel-nitrilotriacetic acid (Ni-

NTA) resin (Qiagen. QIAexpressionist. (Qiagen, Germany, 1999). The column was washed with 20 column volumes (CV) of equilibration buffer (S0OmM NaBo, 1% Triton X- 100, 1% Tween-20, 20 mM imidazole, pH 8.0) and 5 CV of 50 mM NaBo, 20 mM imidazole, pH 8.0. Bound protein was elutecd from the column using a linear 20-500 mM imidazole gradient over 10CV. Protein-conta.ining fractions were pooled and loaded onto a

Q-Sepharose-HP anion exchange column and equilibrated with 20 mM HEPES buffer (N- [2-hydroxyethyl]piperazine-N’-[4-butanesulfonic acid]), pH 8.0. Unbound protein was washed with 3 CV of equilibration buffer "before bound protein was eluted in 20 mM

-20 - oo

HEPES pH 8.00 with a linear 0-1 M NaCl gradient over 77 CV. Fractions containing the i purified scFv o r miniantibodies were pooled, sterile filterexd and dialysed against 2 x 150 i volumes of 20 rmM HEPES pH 8.0, 0.15 mM NaCl at 4°C . Samples were taken along the purification process and analysed using SDS-PAGE (sodium dodecyl sulfate polyacrylamide gel electrophoresis). In a preferred embosdiment, the purified molecules are chemically modified with monomethoxy-polyethyleene glycol (PEG) using the procedure of Chapman et al (1999, Nature Biotechnology 17: 780-783) to prolong their half-life.

Testing secFv for antigen binding

All testirig on scFv fragments for antigen bindi ng were performed by flow cytometry on ra-t thymocytes. Thymocyte cell suspensioris were prepared by standard protocols on fresZhly harvested rat thymus tissue. Thymocytees (5x10%) suspended in 50 pL

RPMI culture medium were used per assay tube. High sensitivity flow cytometric assays were carried out at 4°C using 30 minutes incubation steps for scFv (50 pL sample), anti- poly-His mouse antibody (50 pL diluted 1:200, Sigma), biotinylated goat anti-mouse antibody (50 pL diluted 1:50, Vector) and Streptavidin R-Phycoerythrin conjugate (50 pL diluted 1:50, Sig-ma). Cells were washed with 30 volumes of PBS-azide in between incubations. Asszays were fixed by adding 50 pL fixative (FPBS, 0.5% formaldehyde) and read in a flow cytometer.

Corneal pesnetration experiments . Corneal tissue : . Penetratior of scFv, miniantibodies and whole IgG antibodies was tested in eyes of normal 8-14 months old pigs that were killed for human consumption at a local abattoir 2-5 ) hours prior to the e=xperiments. A single human eye that was mot suitable for clinical use in transplantation waas obtained from the local eye bank and used within 2 hours after enucleation. Four corneas from cats that were housed and killed at the animal house of the

Institute for Med ical and Veterinary Sciences (Adelaide, SA) for other experimental reasons were used within 3 hours of enucleation.

Corneo-scl eral buttons were dissected using standard e2ye bank techniques. Special care was taken to prevent distortion of the corneas during tissue preparation or mounting onto the chambers. as this might affect the integrity of the bamrier function. In those eyes ) that we=re used to investigate only stromal penetration, the corrmeal epithelium was carefully ] remove=d with a no. 11 scalpel and a dry cotton bud befores preparation of the comea- scleral Bbutton. ~Corneal penetration set-ups

Corneal perfusion chambers

Pn vitro penetration of scFv through cat and pig corneas was studied in a corneal perfusio n chamber (Figure 2). The perfusion chamber consists of 3 parts: the main body, a clamping sleeve and a fixation ring, all parts made out of polycarbonate (manufactured by the Cematre of Biomedical Engineering, Flinders Medical Centre, Adelaide). The “shoulde=rs” of the main body that form the posterior side of the clamp are angled by 30° to support the natural curvature of corneas from different species such as human, pigs and cats. Thae central cavity that forms together with the mounted -cornea the chamber on the receiving side, is 12 mm in diameter and 3 mm in depth. The mmain body also harbours the channels for in- and outflow for the perfusion lines. The clampoing sleeve that clamps the cornea from the epithelial side to the main body has two stainless steel guide pins to direct the vertical movement without twisting, to prevent possible sheeer stress and edge damage to the cornea during assembly. The clamping sleeve also contai ns a fluid channel through which applied eye drops can be sucked off, mimicking lacrimal drainage. The cornea and the clamping sleeve on its surface are pressed against the main chamber body by the fixation r-ing, which has a fine thread (1-mm pitch) to allow a ‘very precise adjustment of clamping power. Once the.cornea is mounted properly, the whol e chamber is attached to a . heat exch ange block through which warm water is circulating fro m a water bath. ; Assessme=nt of corneal tissue

Tissue viability was routinely assessed every 30 min during all experiments by measuring corneal thickness with a hand-held pachymeter (BVI. pocket pachymeter, B.V.

Internatiosnal, 63100 Clermont-Ferrand, France). At the beginnimg of each experiment the comneas wwere perfused for one hour and 3 or more different measurements during the second 30 minutes established baseline values of the comeal tThickness. An increase in

“2 2- thickness of more than 10% from baseline values during the subsequent experiments was ) regarded as endothelial dysfunction and the coormea was discarded. : The experimental drugs such as scF-v, miniantibodies, whole IgG antibodies or fluorescein, in a pure form or supplemented with different penetration enhancers, wer-e administered as eye drops every 20 to 30 mimutes to the epithelial side of the cormea. Om the recipient side of the cornea, the perfusion chamber formed an artificial intraocular space that was filled with BSS-Plus (Alcon). The total volume of the BSS Plus solutior that continuously circulated through the artificial eye and the reservoir was 4 mL-.

Elevating the solution reservoir 25 cm above the level of the artificial eye created ax positive pressure of 12 mmHg inside the ch amber, simulating physiological conditions (Figure 2).

To assess drug penetration, hourly samples of 100 pL were taken from the circulating fluid on the recipient side and replaced by the same volume of fresh BSS-Plus_

ScFv, miniantibody and whole IgG antibody penetration was determined by standard fluorocytometry (Zola et al, 1990, J Immumol Methods 135(1-2): 247-55) using rat thymocytes as a CD4-antigen source and a mouse anti-polyHIS-tag secondary antibody (Sigma). Penetration of sodium fluorescein was assessed by measuring absorption at 493 nm using an ELISA reader with the appropriate filters.

Whole eye penetration set-up

Whole human or pig eyes with intact «conjunctival tissue were positioned onto a

Teflon block with a depression tooled into thee Teflon surface, such that the downward facing cornea could be placed into a well (Figuare 3). The well was partly filled with 100 ] pL of scFv or miniantibodies. Care was taken that no other tissue except the comea came in contact with the drug. The whole eye was covered with an inverted plastic container to . 25 form a humid chamber and the set-up was placed above a water bath to keep the eye at a temperature of 35 - 37°C. At the end of the incubation period, the whole globe was washed several times with BSS, dried with a p aper tissue and washed again. Samples of anterior chamber fluid were collected with a 277G syringe needle that was passed through the cornea into the anterior chamber. The concentration of penetrated scFv and miniantibody was assessed in a FACS assay as 1n the perfusion chamber set-up.

At the end of all experimemits. corneas were fixed in buffered formalin ( 10% : formalin in PBS) and embedded in paraffin. Standard 10 um sections were stained wwith ] PAS and examined under a microscope.

Results

ScFv and miniantibody purification

The eluate from the nickel-NT.A purification column had a protein concentration of 0.5mg/mL with purity of 52% for th.e OX38 scFv and a protein concentration of 0. 74 mg/mL with purity of 21% for the €X38 miniantibody respectively. The subsequent purification over a Q-Sepharose™ anion exchange column resulted in a final scIv concentration of 3.98mg/mL with a p urity of 98% and a miniantibody concentration of 1.83 mg/mL and a purity of 75%. The two step purification protocol over a nickel and a subsequent ion exchange column resulted in a clinical grade purity with endotoxin leve=1Is of only 1.08 endotoxin units (EU)/mg o fscFv and 36 EU/mg of miniantibodies.

After the first purification step over the nickel column, the main contaminants En the scFv batch were rotamase (peptidyl—propyl cis-trans isomerase SLYD) and the histone- like protein HLPl. Rotomase is a Thistidine-rich protein that has likely co-purifie d independently of the scFv fragments on the nickel column. It was effectively removed by "the subsequent purification over the Q-Sepharose column. The second contaminant HLP 1 contains neither histidine, cysteine nor tryptophan. It was therefore hypothesised that

HLPI was attached to either the scFv or to rotamase to permit its co-purification in thes nickel column. Based on HLPI havirig an isoelectric point of 9.5, we expected thi=s contaminant to be removed by Q-Sephar-ose™ anion exchange chromatography at pH 8.0 . ] However HPL1 was also found in the final product after this chromatographic step , indicating that it must have been carrmed through the purification process bound to =a : 2S portion of the OX38 scFv fragment. HL.PI has a size of 15.6 kDa increasing the effective: size of contaminated scFv molecules to 4 3 kDa.

The scFv fragments used for peretration experiments in this study were purified over a nickel column only with a final scFv concentration of 0.2 — 0.25 mg/mL. The miniantibodies used for the penetration experiment underwent the full two-step purification protocol with a final miniantibody conceratration of 1.8 mg/mL.

ScFv specificity” for rat CD4 antigen

ScFv binding to rat CD4 antigen was tested in a flow cytormetry assay using rat ) thymocytes. Greater thman 95% of all thymocytes expressed CD4 antigen on their surface as shown in a flow cytometry assay with OX38, a whole monoclomal antibody. OX38 binds specifically to rat CD4. The CHRI-r4-Fv] scFv bound in an iclentical manner to all thymocytes. The maxirmnal fluorescence intensity of samples with CHIRI-r4-Fvl scFv was about two-thirds the intensity found with OX38. Iowcver, the fluosrescence intensity in dilution series decreased in an identical pattern in OX38 antibodies and CHRI-r4-Fvl scFv, indicating similar binding properties (Figure 4).

To exclude CD4—independent or non-specific binding of CHR I-r4-Fv1 scFv to cell membranes, flow cytometry assays with Jurkat cells (a human T-cell line) and J5.CD4 oo (transfected Jurkat cell line that expressed full length rat CD4 o n its surface) were performed. CHRI-r4-F-v! scFv did not exhibit any non-specific bending as shown by negative flow cytometry assays using Jurkat cells. Binding specificity to rat CD4 was 1S proven in a flow cytometry assay on transfected J5.CD4, where scFu bound in a similar manner as 0X38.

Corneal penetratieon of scFv (28 kDa)

ScFv penetration through the corneal stroma was tested in a first set of experiments using the corneal perfusion chamber. The comeal epithelium was removed prior to the experiment. As shown ir Figure 5, the scFv penetrated successfully therough the stroma of pig corneas and retained its specific antigen binding activity. The re was a four-hour interval between the first drop of scFv on the donor side and the detection of a clear . binding activity on the recipient chamber side.

The pH of the Buffer in which a drug is dissolved can often influence drug solubility and corneal peretration. The OX38 scFv has an isoelectric p oint of 6.1, at which pH the scFv penetration ds expected to be worst. We therefore investi gated penetration of the OX38 scFv fragment through pig corneal stroma at pH of 8.0, 7.08 and 3.5. While an increase in pH from 7.0 t 0 8.0 resulted in an accelerated penetration, lo-wering the pH away from the isoelectric point to 3.5 caused a reduction in penetration (Figu-re 6).

To prove speciess-independent stromal penetration, the exper:iment was repeated with a normal cat cormea. Results from the perfusion chamber showed a similar penetration of scFv in cats a s in pigs. ScFv binding activity after penetration through the 25% thinner cat cornea becarme detectable after 2 hours (Figure 7). :

A To prove that these positive penetration results were not caused by tissue damage at the clamped edges of the co mea or by a leakage around the mounted cornea that could have simulated stromal pene€ration, whole OX38 IgG antibodies were used as- a negative control. Whole antibodies are not expected to penetrate the stroma becau se of their molecular size of about 150kMDa. This clearly exceeds the molecular cut off ramge of 60 - 90 kDa for stromal penetration. Monoclonal antibodies at a concentration that allowed positive detection at a dilutiora of | part in 2 million were administered as eye Erops every minutes. No binding activ-ity was found on the recipient side during sampli ng periods of up to 14 hours (Figures 6 arad 10).

Penetration experiments in the perfusion chambers allow individual cormeas to be studied over an extended period of time, with repetitive sampling of the recipient side to assess the kinetics of the pemetration process. However, as the fragments that have

I5 penetrated across the cornea into the perfusion chambers become diluted in a largze volume (4 mL), the results underestimate the concentration of scFv and miniantibodie s that are achievable in a real eye with =n anterior chamber volume of 0.3 - 0.5 mL. Teo allow a rough comparison of the penetr ation results obtained in the perfusion chamber ex goeriments with the expected values in an eye of a living human being, additional penetration 20 experiments in the perfusion «chambers were carried out using sodium fluorescein eye drops. Sodium fluorescein is a 332 Da molecule, the penetration of which into the human eye has been studied extensivel y in vivo. After a single drop of 2% (w/v) fluorescein, the drug becomes detectable in thee human eye after 30 minutes (Araie er al, 1983, Jpn J

Ophthalmol 27(3): 421-33). Respeating the administration of fluorescein causes a further ) 25 substantial increase in the amount to be found in the anterior chamber (Adler er al, 1971, .

Exp Eye Res 11(1): 34-42). To be able to compare the scFv penetration results obstained in pig eyes, a pig cornea with noramal epithelium was treated with fluorescein 2% eye drops every 20 minutes for 14 hours. One hundred-microlitre samples were removed from the chamber reservoir in the same manner as for the scFv and miniantibody experiments.

Sodium fluorescein concentration was determined in an ELISA reader at 492 nr with a detection threshold of 2 parts per million. On the recipient side of the pig comea., sodium fluorescein became detec table only after 3 hours with a subsequent expon ential increase during the rest of the experiment as shown in Figure 8. : To allow a better estimation of the scFv concentration that is achievatole in a normal eye, scFv penetration was also tested in the whole eye model. In this mode 1 each eye can

S be sampled only once, re quiring many eyes with different exposure times 10 give some kinetic information. However, as the anterior segment of the eye is, in a physiological sense, a well defined struc ture with only minimal pharmacological interactiom with the rest of the body, results from such whole eye experiments reflect more closely ®he amount of penetration that is achievable in vivo.

For this experiment, the corneal epithelium of a whole pig eye was removed and the cornea immersed in O=X38 scFv-containing Teflon well for 7 hours. After thoroughly cleaning the surface, a 100-pL sample of anterior chamber fluid was obtaired through a syringe with a 27G needle. The binding activity was tested in a FACS assay on rat thymocytes with a detection threshold of 1 part per two thousand. The anterior chamber fluid that was sampled aftesr 7 hours revealed a strong binding activity that w=as equivalent to a scFv concentration of 1 part in 250 (Figure 9).

The kinetics of scF v penetration through the comeal epithelium was swudied more closely in the perfusion ch ambers using intact corneas. There was no detectable binding activity over a penetration period of 8 hours when scFv in pure PBS solution (pH 7.2-7.4) were applied as an eye drop. When the scFv / PBS solution was supplemented with 0.01% benzalkonium chloride, a commonly used preservative for eye drops, weak bimding activity was observed at the end of the eight hour sampling period in pig corneas (data not shown).

The same experiment was repeated with scFv in PBS at a pH of 8.0 (&adjusted with ) 1M NaOH). Under these conditions scFv penetrated even without a penetration enhancer through a healthy epitheli um and stroma, showing increasing binding ac®ivity on the recipient side after 8 hours. The addition of 0.5% capric acid resulted in a facilitated corneal penetration with increasing binding activity detectable after 6 hours (Figure 10).

Corneal penetration of OX38 miniantibodies (66.8 kDa)

The penetration of (OX38 miniantibodies with a molecular size of 66.8 kDa through the corneal stroma was tested in an identical way as that of the monovalent sc. Fv fragments in the perfusion chamber. INo penetration of whole OX38 IgG antibody, used as a negative control. could be detected during the whole experimental period of 14 hours. In contrast, . OX38 miniantibody binding became positive after 3 hours with a further expomential increase until the fluocytometer assay reached saturation with the 8-hour and subsequent samples (Figure 11).

Penetration of OX38 mini-antibody without penetration enhancer through a pig comnea with an intact epithelial barrier was much slower and became weakly detectable only after 11 hours.

As with the monovalent scFw, the achievable concentration of miniantibodies £ n the anterior chamber of a normal eye w as investigated in the whole eye model. Pig eyes with the corneal epithelium removed were first sampled after 3 hours and revealed already half the maximal signal intensity. Samples after S and 7 hours of miniantibody exposure showed binding activity within the mange of a maximal test signal. Subsequently six pig eyes with normal epithelial barrier were exposed to miniantibodies supplemented with 0.5% capric acid. Miniantibody Binding activity became clearly detectable in eyes sampled after 5 hours exposure time and activity after 7 and 8 hours was only minimally lower than in eyes without an epitheli al barrier (Figure 12). :

The most challenging conditions in all the experiments above were the penetration of the relatively large miniantibodies through corneas with an intact epithelial barrier. All these results were obtained in young pig eyes that have a 25% thicker stroma than a hurcan cornea. However, the most importan t penetration barrier is the epithelial surface and the knowledge of the ultrastructure of this barrier in pig eyes is limited. To compare the above penetration results obtained in pig and cat eyes with potential penetration of su ib- immunoglobulin fragments through a human cornea. the most challenging experimermtal conditions were used in a fresh human. eye. This eye with an intact epithelium was treated } 25 with OX38 miniantibody in the whole eye set-up. The cornea was exposed for five hours } to 100 pL of the standard miniantibo-dy concentration of 1.8 mg/mL supplemented with 0.5% capric acid at pH 8.0. The imtraocular sample obtained after 5 hours reveal ed maximal signal intensity in the fluocy-tometer, completely saturating 1.5x10° thymocytes per mL. A titration of binding activity in this sample revealed a miniantibo«dy concentration of 0.8 pg/mL in the antesrior chamber of this human eye examined (Figures 13 and 14).

Discussion ) This study demonstrates th at engineered recombinant antibody fragments can . successfully penetrate the comea to reach inside the eye after topical administration as eye drops. Moreover these fragments r etain their full antigen binding activity, offering the possibility of an antibody based therapeutic and diagnostic intervention in the eye. At: the same time this study has, as many others before (Verhagen et al, 1990, Invest Ophthal mol

Vis Sci 31(8): 1519-25; Osusky et al, 1993, Graefes Arch Clin Exp Ophthalmol 231 (2): 122-8), shown that naturally-occurr-ing whole antibodies do not, or a least not im a therapeutically useful manner, penetrate the cornea and that the cut-off for the maximmal molecular size that enables penetration through the cornea is much lower than the 500 kzDa found in a frequently cited but never repeated study (Maurice D, Mishima S. Ocia/ar

Pharmacokinetics. Springer-Verlag Berlin. 1984. Sears M, ed. Pharmacology of the Eve ).

The comea contains two important penetration barriers, the superficial corn eal epithelium and the comeal stroma. The comeal epithelium forms a strong penetratmon barrier for all but small lipophilic molecules. Larger molecules have to penetrate through the intercellular space that is normally~ blocked by tight junctions (Maurice DM., 1980, _/nt

Ophthalmol Clin 20(3): 7-20; Burstein. er al, 1985, J Ocul Pharmacol 1(3): 309-26). Pomes within these tight junctions allow penetration of molecules up to a size of 1-5kHDa (Edelhauser er al, 1998, Adv Exp _Med Biol 438: 265-71). The corneal epithelial penetration barrier depends on healt®iy cellular structures. It is well known that t he epithelial barrier function in humans iss significantly reduced by ageing (Chang er al, 19993

Cornea 12(6): 493-9 [see comments]) and many common diseases such as diabetes or d_ry eyes (Gobbels er al, 1989, Graefes Arch Clin Exp Ophthalmol 227(2): 142-4; Yokoi er «al, 1998, Br J Ophthalmol 82(7): 797-8040). In addition to that, several studies have shovsvn ’ 25 that surgical interventions such as cormeal transplantation (Shimazaki ef al, 1999, Cornea 18(5): 559-64; Chang et al, 1994, Ophthalmic Res 26(5): 283-9) or trabeculectomi es (Tanihara er al, 1997, Am J Ophthal mol 123(4): 487-93) to treat glaucoma result in a breakdown of the barrier function for months or even years.

The second barrier, the corneal stroma, allows penetration of larger fragments amd has been cited to allow penetration of rnolecules up to 500 kDa or even larger molecules if the ussue becomes oedematous (Burstein er al, 1985, supra). However the figure of 5&0 kDa is derived from a single study that was published in the non-peer reviewed literatimre

(Maurice er al, 1984, supra). In fact, several studies have addressed the penetration of IgG ) antibodies with. a size of 150 kDa. These studies have repeatedly shown that molecules of ] this size do neot penetrate the cornea, or penetrate on ly at an exceedingly slow rate requiring a timee interval of 1-2 weeks to cross the corneal stroma (Verhagen es al, 1990,

S supra: Osusky eet al, 1993, supra). In contrast to the epithelial barrier there are no known penetration enhancers to improve drug delivery though the corneal stroma.

The datam presented above demonstrate that recombanant antibody fragments at least up to a size of 6 6 kDa can rapidly penetrate the corneal stroma and retain their full antigen- binding capacit-y inside the eye. Penetration of these amtibody fragments seems to be species indeperadent as penetration could be document ed in pigs, cats and humans.

Stromal drug penetration has an indirect linear correlatiorn to stromal thickness and drug size and a direct correlation to drug concentration (Bursteir es al. 1985, supra). The faster penetration of scFv through the 25% thinner cat and hum an cornea than through the pig cornea seems to be in good agreement with the finding aboeve. The fact that the two and a half-fold larger aminiantibody was detected on the recipierat side of the corneal perfusion chamber after a similar time interval as the small scFv fragment is probably due to the 7-8 fold higher concentration of miniantibodies in the eye drops used in this study. An increase of the scFv concentration as we have demonstratecd with the two step purification process for cliniecal grade material is therefore likely to re=sult in even higher intraocular concentrations than the ones found in the present study.

As mentieoned above, the superficial epithelium is tine stronger barrier against drug penetration and mt is therefore no surprise that the penetration of scFv and miniantibody fragments throug h corneas with an intact epithelial barrier iss considerably slower. In fact it is rather surprisirag that scFv and miniantibodies could be shown to penetrate, albeit at a ) 25 very low rate, tlmrough an intact epithelial barrier at all. MHowever, the epithelial barrier . interferes even wvith much smaller molecules to an extent ®that most of the currently used ophthalmic drug s are supplemented with some form of penetration enhancer. These penetration enharcers work by loosening the tight junctionss of the most superior epithelial cells (Burstein NL. 1985, Trans Ophthalmol Soc U K 104(P 14): 402-9; Ashton er al, 1991,

J Pharmacol Expo Ther 259(2): 719-24; Green er al, 1971, Am J Ophthalmol 72(5): 897- 905). The most commonly used penetration enhancer is b enzalkonium chloride (Tang er al, 1994, J Pharm Sci 83(1): 85-90; Burstein NL. et al, 1980, Invest Ophthalmol Vis Sci

19(3): 308-13), which also works as preservative against microbial contamination. It is : usually added to a final concentratiom of 0.01 - 0.05%. A future pharmaceutical preparation of recombinant antibody fragments in the form of eye drops will need ’ preservatives independently of their epithelial penetration properties. We therefore supplemented some of the scFv and the miniantibody fragments in this study with 0.5% capric acid, a normal dairy food constitiient that results in a reversible increase in tight

Junction permeability in the intestinal mucosa as well as in the corneal epithelium (Morimoto er al, 1989, Arch Int Pharmacodyn Ther 302: 18-26; Soderholm er al, 1998,

Dig Dis Sci 43(7): 1547-52; Sasaki et al, 1995, Pharm Res 12(8): 1146-50). Capric acid improved scFv and miniantibody penetration through an intact epithelium considerably. In contrast to other penetration enhancers such as benzalkonium chloride, digitonin or

Tween20. corneas treated with 0.5% capxic acid showed no epithelial alterations when examined with conventional histology.

The single human eye that could be obtained for these penetration experiments

I5 received miniantibodies supplemented with 0.5% capric acid. The miniantibody binding activity found in the human anterior charmber fluid after 5 hours greatly exceeded the : penetration found in pig eyes after the sarme time. This could reflect a primarily lower penetration barrier of this 77-year old human cornea in comparison to the 8-14 months old pig eyes or a generally better effect of capric acid in human corneas. However, this remains a speculation until more human eyes can be studied. :

In summary this study shows that scFv antibody fragments can penetrate into the eye when given as eye drops. This makes t hem a potential candidate for a novel class of drugs for the treatment of many ophthalmic diseases, especially prevention and treatment of corneal transplant rejection, inflammatory or infectious keratitis, conjunctivitis and : 25 uveitis, as described hereinafter. As scFv can be labelled with different tags they are also considered to have diagnostic potential.

EXAMPLE 2 } Experimental animal models

Animals

Male and female Fischer 344, DA, Lewis and Wistar-Furth inbred rats and Porton outbred rats aged 6-16 weeks are bred within the Flinders University of South Australia (M.delaide, Australia). Rats are housed at 21°C and 50% humidity in a 12-hour light and 12-hour dark cycle, and are fed water and dried ration (New Joint Stock, Ridley

A griproducts, Murray Bridge, SA. Australia) ad /ibitum. Adult male and female BALB/c or OLA mice are housed at 21°C and 50% humidity in a 12-hour light and 12-hour dark cycle. and fed water and dried ration (New Joint Stock, ‘Ridley Agriproducts, Murray

Bridge, Australia). All procedures and euthanasia are carried out under inhalation general anaesthesia (halothane; Zeneca Lid, Macclesfield, United Kingdom). Experimental protocols were developed in accordance with the National Health and Medical Research

Council of Australia “Guidelines for the Use otf Animals in Research”, and are always submitted to the Animal Welfare Committee at Flt nders University of South Australia for approval.

Induction of endotoxin-induced uveitis (EIT)

Rats are injected in one hind footpad with 200 microgram of E. coli 055:B55 lipopolysaccharide (Sigma Chemical Company) solubilised in 100 microgram normal saline, according to the method of Rosenbaum er al (1980, Nature 286: 611-613).

Infl ammation becomes apparent at 12 hours post-injection, is maximal at 24 hours and is ] resolving by 48 hours. Animals are examined and scored at the slit-lamp at 24 hours. We hav e extensive experience with this model (Smith et al, 1998, Invest Ophthalmol Vis Sci - 39: 658-661; Immunol Cell Biol 76: 497-512).

Induction of experimental melanin-induced uveitis (EMIU)

Experimental melanin-induced uveitis (EM IU) is a robust and relevant model of acute anterior uveitis (Smith er al, 1998, Immunol. Cell. Biol 76: 497-512; Smith et al., 1999, Clin. Exp. Immunol. 115: 64-71). Ocular melanin is extracted from bovine choroids according to the protocol described by Broekhuyse and colleagues (1984, Curr Eye Res 3:

1405-1412), &and quantified as a dry weight. EMIU is induced in Fischer 344 rats by : immunization with 250 pg purified bovine choroidal melanin emulsified in Hunter's } adjuvant, with: pertussis toxin co-adjuvant (Broekhuyse and collea_gues, 1984, supra; 1995,

Ocul Immuno Inflamm 3: 149-155; 1996, Jpn J Ophthalmol 40: 459-468). Specifically, rats receive 1225 microgram of bovine ocular melanin in a 1:1 e mulsion of sterile, non- pyrogenic 0.9 % normal saline and Hunter's TitreMax™ adjuvant (Sigma Chemical

Company, St. Louis, MO, USA) in a total volume of 60 microlitr-e by right hind footpad injection. Im mediately afterwards, they are injected intraperit=oneally with the same quantity of mmelanin mixed with 1 microgram of pertussis toxin (Sigma Chemical

Company) in a total volume of 40 microlitre normal saline. Uveit 1s is generally bilateral, is apparent 10-720 days after injection, resolves after 2-3 weeks and is recurrent in 20% of animals.

Monitoring of uveitis and collection of uveitic eyes

Animalss are examined daily at the slit-lamp to identify tlae onset and course of uveitis. Severity is recorded according to a well-validated system _in which aqueous cells and flare, preserace of fibrin, synechiae, pupil reactivity, iris hyperacmia and empyema are all scored. Th.e scale used is as follows: no inflammation: normal appearance; mild disease: no swel ling, but obvious infiltration of anterior uvea, + celltalar exudate in anterior or posterior chambers; moderate disease: mild or moderate swellang and infiltration of antertor uvea, m oderate cellular exudate in anterior and posterior chambers; severe disease: massive swellin_g and infiltration of anterior uvea, large cellular exudate in anterior and posterior chamb ers with vitritis. Animals are killed by overdose of~ anaesthetic halothane and the eyes are immediately removed for further processing. ’ 25 The preseznt inventors were the first to develop the widely used model of orthotopic corneal transplartation of the inbred rat (Williams & Coster, 1985, Jnvest Ophthalmol Vis

Sci 26: 23-30; WJ ilhams er al, Corneal transplantation in small aninaals. In: Experimental

Transplantation Models in Small Animals, eds. MK Green, TE Mandel. Chur, Switzerland:

Harwood Acadezmic Publishers, 1995, chapter 5, ppl107-132). Inbre=d Fischer 344 (F344; major histocompatibility complex (MHC) haplotype RT"! and inbred Wistar-Furth (WF;

MHC haplotype RT") rats are used for comeal transplantation. Aedult F344 rats are the recipients of corneal grafts from either adult F344 rats (isografts) or adult W/F rats ) (allografts). Grafts are not performed across a gender difference, but allogra fis are } performed across MHC class 1 and II and minor antigen barriers.

A unilateral right penetrating 3-mm diameter corneal graft is performed under general anaesthetic as previously described. Chloramphenicol ointment 1% is applied to the graft post-operatively, but no immurosuppression other than as specified in a particular experiment is administered. Graft sutures are not removed. Recipients are examinecd daily at the slit-lamp for corneal clarity, anterior segment inflammation, synechiae, and ceorneal neovascularisation. Specifically, clarity and oedema are each scored on a scale o-f 0-4, with 0 being completely transparent (for clarity) or thin (for oedema). and 4 being completely opaque (for clarity) or maxi mally thick (for oedema). Any rat that develops post-operative cataract at any stage or th at has corneal opacities on the fifth post-ope rative - day is deemed a technical failure and is e xcluded from analysis.

Rejection appears as increasing comeal opacity in a previously thin. clear coomeal graft and is always associated with graft neovascularisation. The day of rejecti«on is defined as the first day that the pupil mar gin is no longer clearly visible through the grafted cornea. In the inbred strain combinatiom WF to F344, rejection occurs at a median of 16 days post-graft. Rats that reach 100 days with clear grafts are considered long-survivo rs.

Mouse models of HSV stromal kezatitis

BALB/c mouse corneas are scari fied with a needle and 4 pL containing 10° pfu

HSV-1 (clinical isolate) is applied topically (Thomas er al, 1998, J Immunol 160: 3 965- 70). Disease progresses to stromal keratitis over a three-week period. A clinical scoring system is used to follow disease at the slit-lamp: 0 - clear cornea, 1 - mild haze_ 2 - moderate opacity or scarring, 3 - severe opacification but iris margin visible, 4 - iris not ; 25 visible, 5 - necrotising stromal keratitis. A well-characterised model of recurrent THSV keratitis in the NIH/OLA mouse (Harland) is also used (Laycock er al, 1991, [ravest

Ophthalmol Vis Sci 32: 2741-2746). One cornea is scarified and inoculated as above . At the same time, 0.5 mL of pooled human serum containing antibody to HSV (Chema con,

Temecula, CA) is given intraperitoneally to protect the eye from damage during the cute phase of infection. The mice are then left for 5 weeks to establish latency. “Wiral reactivation is achieved using a UVB transilluminator (FS40 Westinghouse larmps).

i Anaestheetised mice are screened so that only the eye is irradiated. A dose of 248 ml/em’ over 55 seconds is optimal for HSV reactivation.

Rat model of Acanthamoeba keratitis

Acanthamoeba keratitis is induced by inoculation of the outbred rat cornea with acanthamnoebae and a Corynebacterium sp.. according to a model developed in our laboratory (Badenoch ef al, 1990, Arch Ophthalmol 108: 107-112). Adult female Porton rats are zanaesthetized. A 2-mm partial-thickness incision is made with a scalpel blade in the peripheral cornea, approximately 1 mm from the limbus. Under the operating microsco pe, a 1 pL volume 10° Acanthamoeba castellamii (>90% trophoxoites; >90% viability) mixed with 10° Corynebacterium xerosis is inoculated using a microsyringe fitted wit h a 31-gauge needle. The needle is inserted thro ugh the wound into the central corneal lmmellac. All animals are observed daily after inocumlation. Acanthamoeba keratitis develops in all animals, with leucocytic infiltration visible on day 5 and suppurative keratitis v=vell established by day 7.

Statistical analyses

Exc perimental groups contain 10 animals. Graft surv-ival and disease incidence data are analyssed with the Mann-Whitney U-test corrected for ties. Categorical data (eg. disease severity scores) are analysed using the two-way Fish er’s exact test.

Co-1mneal photography. immunocvtochemistry. histopaathology and flow cytometry

Co rneal photography is performed on anaesthet-ized animals using a Wild

Heerbruggs Photoautomat MPS 45 camera attached to the operating microscope (for rats, ) mice). Imnmunohistochemistry is performed using an immunoperoxidase technique on ] cryostat-cuat sections of PLP-fixed tissue; a wide range of mmnonoclonal antibodies directed against rat and mouse leucocyte and other cell surface m arkers that will allow precise 95 identification of the composition of comeal leucocytic _infiltrates. Histopathological correlatiomms are made on H&E sections of formalin-fixed, pearaffin-embedded whole eyes.

Flow cytometry is performed using a standard single-colowar indirect immunofluoresence and a high--sensitivity technique.

Additional antibodies used in antibody engineering - Anti-rat B7-1 (CD80: 3H5; IgGl) and anti-B7-2 (CD86: 24F; IgG1) hybradomas were the kind gift of Dr H Yagita (Juntendo University, Tokyo, Japan). The anti—~CD28 ] hybridoma (JJ319; 1gG1) was the generous gift of Dr T Hunig (Institute for Virology and

Immunology, University of Wurzburg, Germany). Anti-poly-histidine antibodies (EdIS-1; mouse; 1gG2a) was purchased froma Sigma Chemical Company (St Louis, MO, TSA).

FITC-conjugated goat anti-mouse immunoglobulins were purchased from Silenus

Laboratories (Melbourne, VIC). Horseradish peroxidase-conjugated streptavidira and biotinylated goat anti-mouse immunoglobulins were purchased from DAKO Corporation (Carpinteria, CA, USA).

Administration of antibody rea gents

Systemic administration can be performed by intraperitoneal (ip) injectiom of antibody in buffered saline with 1% foetal calf serum (FCS) added as a carrier. The regimen of injection depends upon the experiment. Topical administration cara be performed by directly application to the cornea of antibody in ophthalmic Balanced Salt

Solution and supplemented with 1% FCS and one or more penetration enhancers.

Ascites production

To generate monoclonal antibody as ascites, hybridoma cells are injected intraperitoneally into separate groups of pristane-treated BALB/c mice according to standard methods. Ascites is collected aseptically after the mice have been euthanas ed.

The resultant ascites is diluted with steri le, pyrogen-free PBS containing 1% v/v foetal =alf serum. . Mouse monoclonal antibodies used in the immunohistochemical identification of rat or mouse cell surface antigens are produced as undiluted supernatants from stationary phase hybridomas. Hybridomas X63 (unknown specificity; IgG1 isotype negative contro 1),

OX-1 (anti-CD45 [leucocyte-common antigen]; IgGl isotype), OX-35 and OX-38 (an_ti-

CD4; IgGl and IgG2a isotypes, respectively), OX-8 (anti-CD8; IgGl isotype), NDS&I (anti-CD25 [Interleukin-2 receptor (IL-2R)]; IgGl isotype), R73 (anti-T cell receptor (TCR); IgGl isotype), OX-26 (anti-CD71 [transferrin receptor}; IgG2a isotype), OX-=42

(anti-CD11b/CD18 [C3b receptor]; Ig(32a isotype), ED-1 (anti-rat monocyte, macrophagze and dendritic cell cytoplasmic antigen; IgGl isotype) and OX-6 (anti-MHC monomorphic . class II determinant; IgGl isotype) were obtained from the European Collection of Animal’

Cell Cultures, and hybridoma SALS (anti-Salmonella antigen; IgG2a isotype negativ-e control) was a gift from Dr. L. Ashman (University of Adelaide, Adelaide, Australia).

Hybridomas producing monoclonal ant ibody to various mouse surface markers (mouse

CD4 (GK 1.5), CD8 (TIB150), CD11b/c (TTB213, HB8466) and IL-2 receptor (PC61) were the gifts of Dr P Macardle and Dr PH Ha rt, both of Flinders University.

Immunohistochemical staining fo r cell surface antigens

After euthanasia of the rat or m.ouse, eyes were enucleated and were punctured posterior to the pars plana, pre-fixed by immersion at 4°C in 2% weight/volume (wiv - paraformaldehyde-0.07S M lysine-0.0375 M sodium phosphate-0.01 M sodium periodate= for 4 hours and dehydrated in sequemmtial changes of 7% and 15% w/v sucrose in

Dulbecco's A phosphate-buffered saline (PBS), embedded in OCT compound (Miles,

Elkhart, IN, USA) and snap-frozen in liquid nitrogen. Ocular cross-sections were cut by cryostat at 5-8 micron thickness. Sec tions were incubated for 10 minutes at room ‘ temperature with 10% v/v normal swirae serum (Commonwealth Serum Laboratories,

Melbourne, Australia) in Dulbecco’s A phosphate-buffered saline (PBS), then with the primary antibody for 18 hours. At this foint and subsequently the sections were washed with PBS containing 0.2% w/v gelatin . They were incubated for 30 minutes with biotinylated affinity-isolated goat anti-mouse immunoglobulin (DAKO Corporation,

Carpinteria, California, USA) diluted 1 in 500 in PBS containing 1% v/v normal rat serum, washed, and then incubated for a further 30 minutes with horseradish peroxidase- ’ conjugated streptavidin (DAKO Corporation) diluted 1 in 1000 in PBS. Sections were washed and developed for 5 minutes in 9 mM Tris-HCI buffer at pH 7.6 with 40 mM sodium azide, 20 mM 3,3'-diaminobenzi dine tetrahydrochloride, 9 mM imidazole, and 0.07% v/v hydrogen peroxide (Sigma Chemical Company). Reactivity and specificity of antibodies were checked by flow cytometry on appropriate rat or mouse target cells (normal cells and/or transfectants) prior to their use in immunohistochemistry, and on fixed sections of non-ocular tissue prior to their -use in the eye.

Histological assessment of rat and mouse eyes

Enucleated. eyes are fixed for a minimum of 24 hours dn 10% buffered formalin, ; dehydrated to 100 % ethanol, further fixed for 18 hours in chloroform, and embedded in paraffin wax. Tissue cross-sections are cut 5 micron thick and =stained with haematoxylin and eosin.

Monitoring of rat peripheral T cell subsets in vivo

Peripheral T cell subsets were monitored by flow -cytometry of tail blood preparations. Samples were collected from rats just prior to the fi rst antibody injection and on days 7, 14 and 2 8 after disease induction. Up to 1 mL of peripeheral blood was collected into a lithium heparin-coated tube, and diluted to 5 mL with HEFPES-buffered RPMI 1640 medium (ICN Biomaedicals Inc., Aurora, Ohio, USA) containing 1 00 IU/mL penicillin, 100 microgram/mL streptomycin sulphate, 2 mM glutamine and 10924 v/v foetal calf serum.

Lymphocytes were separated using Lymphoprep™ (Nycomed Phaamma As, Oslo, Norway) as described by the manufacturer, except that 40 mM meglumine diatrizoate (Angiografin,

I5 Schering AG, Berlin, Germany) was used to adjust the specific gravity of Lymphoprep to + 1.09. The recovere d cells were suspended in PBS-0.02 M sodium azide (PBS-azide) at 1x10” cells/mL, and to S50 microlitre of this lymphocyte susspension was added 50 microlitre of monocl onal antibody. This mixture was incubated oma ice for 30 minutes, then washed in PBS-azid e and centrifuged at 400xg for 5 minutes at 4°C. The pellet was re- © 20 suspended in 25 mic Tolitre of normal rat serum and 50 microlitre of FITC-conjugated goat anti-mouse immunogzlobulin (Silenus Laboratories, Melbourne, A ustralia) diluted | in 50 with PBS-azide. A 30-minute incubation on ice was followed by ttwo consecutive washes.

The cell pellet was r-e-suspended in 50 mL of fixative containing 10 mM glucose, 5% v/v : formaldehyde and 5 mM sodium azide in PBS. Antibody binding was measured by flow cytometry using a standard fluorescein filter set (FACScan™, Bectson Dickinson, Mountain

View, California, US A).

Endotoxin assay

Endotoxin levels were measured by the Multi-test Limulus Amebocyte Lysate assay (Whittaker Bioproducts, Walkersville, MO, USA) (sensitzivity = 0.125 EU/mL) according to the man-ufacturer’s instructions.

Protein estimation ) Total protein in reagents was measured by a modified Lowry procedure tasing a ] commercial protein estimation kit (Sigma Diagnostics, St Louis, MO, USA) followmng the manufacturer’s instructions. Absorbance woas measured at 750 nm.

Rat mixed lymphocyte reaction

Mononuclear cell fractions were isolated by density centrifugation on Lymplaoprep (Nycomed Pharma, Oslo, Norway). Twen ty-thousand viable responder splenocytes were incubated with 2x10° mitomycin C-inacti-vated (250 ng/ml, 37°C for 30 mins; Klyowa

Hakko Kogyo Co Ltd, Kyowa, Japan) stimulators in a final volume of 250 ul. whole RPMI media (20% v/v FCS, 100 IU/mL penicillin . 100 pg/mL streptomycin, 2 mM L-gluta mine, 5x10 M 2-mercaptoethanol) in 96-well reound-bottomed microtitre plates. Monoclonal antibody reagents or fragments were addled and cultures incubated with 5% C€O- in humidified air at 37°C for 6 days. To meassure cell proliferation, 1 pCi tritiated thymidine (Amersham, Buckinghamshire, UK) was added to the cultures for the final 18 hours.

Following harvesting, the impregnated glass fibre papers were placed into masks ard 20 nL scintillation fluid added to each well. Papers were counted m a Topcount™ cotanter.

All cultures were performed in triplicate. ’

EXAMPLE 3

Evaluation of engineered antibody fragments in vitro - 20 Functional activity of monovalent or divalent antibodies in vitro 1s compared with the parent whole monoclonal antibody and wwith irrelevant control scFvs. Fragments may . be tested singly and in combination. If thee scFvs contain the "tag" sequence poly -His, detection for flow cytometry and immunohizstochemistry can be accomplished by uses of a ) monoclonal antibody (HIS-1, Sigma) to the tag. However, it will be understood that «other identification tags can similarly be used. To confirm appropriate specificity of the engineered constructs, binding is assessed by~ flow cytometry on unfixed normal rat ce 1ls or transfectants, and/or immunohistochemistr-y on cytocentrifuge preparations of fixed transfectants, normal rat or mouse cells, miteogen-stimulated rat spleen cells, HSV-infected

Hep-2 cells and MRC-5 fibroblasts, or cultured Acanthamoeba cells. Activity is titratezd by flow cytometry where appropriate. Tissues binding is assessed on 5-8 pm cryostat-cut sections cf PLP-fixed normal rat eye, lymph nodes, spleen, kidney, heart and liver by immunoperoxidase staining. Flow cytometry on pe ripheral blood collected from the tail . vein is useed to determine whether fragments have b ound to circulating cells in vivo scFv administraation. ScFv affinity on/off rates will be examined with an optical biosensor (BlAcore™™) using purified antigen as the solid phase.

EXAMPLE 4

Mo dulation of EMIU by scFv

We will first test anti-CD4 scFv, as we have already shown that whole anti-CD4 (but not CID8) monoclonal antibody blocks de novo disease development in EMIU and may block recurrent disease (Smith er al., 1999, Clin_ Exp. Immunol. 115: 64-71). Anti-

ICAM-1 sc Fv will then be examined. Controls will imclude groups receiving no treatment other than disease induction, disease induction plus negative control scFv, and disease induction pelus whole monoclonal antibody. For sy stemic administration, rats will be injected ip with >] mg/mL protein scFv (I mL vo lumes) at 3 days prior to disease induction, o nthe day 0 and on days 3, 5, 7, 9 and 11 thaereafter, and incidence and severity of uveitis mmeasured by slit-lamp examination. Experi ments to determine the appropriate therapeutic dose will need to be performed: in a preliminary experiment, systemic administration of 1 mg/mL total protein of anti-CD4 scFv prevented EMIU in over 30% animals testeed (see Table 1).

TAB LE 1

Controls 27/28 (96%)

Anti CD8 whole Mab | 8/8 (100%)

Anti CD4 whole Mab | 8/20 (40%)

Anti CD4 scFv 16/23 (70%)

Topical administration of scFv will be startesd on the day of disease induction and. : will be continued for 28 days. In both sets of expeximents, some animals will be killed at - specific time points after immunization with melanim, and the eyes taken for histology and immunohistochemistry to determine the compositior of any cellular infiltrate and wheter it has been altered by treatment. Should either test scFv significantly reduce the inciderm ce of uveitis, then attempts will be made to reinduce disease by further immunization wath melanin, to determine whether non-responsiveness has been induced We will examine t he effect of treatment on existing disease by administermng scFv on the first day that cells amd flare are observed in the anterior chamber, and dail y thereafter for 5 days. Rats will be : followed at the slit-lamp and some eyes collected for histology.

EXAMPLE 5 oo

Modulation of corneal allograft rejection by scFv

The efficacy of scFvs directed against CD4, CCD54, IL-2R, CD28, CD80 and CD& 6 in the prevention of corneal graft rejection in WF to F-344 allografts will be examined. For the co-stimulatory molecules CD80 and CD86, a scFw cocktail will be used. Controls will i nclude rats receiving no treatment, treatment with a scFv of irrelevant specificity, with anti-CD8 scFv, and with whole antibody. A small ntimber of isografts will be performed t o check for toxicity or other side effects. Systemics administration will be tested using protocols known to be effective for whole antibodies. Rats will be injected ip on days -3,

OO. 3,5 and 7 with respect to transplantation. Estimnation of the amount of engineered fragment required for an effect in vivo will be established in preliminary experiments, t>ased on the amount of whole antibody needed for sirmilar effect, together with titration o f _ s cFv activity. ScFvs will then be tested topically for= their ability to prevent the onset o r-ejection (direct application to the graft daily from the day of graft for 28 days), and to treat ) 25 ongoing rejection (direct application daily for 5 days from the first day that comeal blood wessels first cross the graft-host junction or from thes day that signs of anterior segment inflammation are observed, which ever occurs first). Some animals will be killed at s pecific time points after graft, and the eyes taken for histology and immunohistochemistrsy teo determine the composition of any cellular infiltrate.

Whole monoclonal antibody directed against wat CD4 when given systemically as described above at approximately 0.2 mg/mL can proL ong corneal graft survival, as showm in Table 2 below and by Ayliffe er al (1992, Brit J Ophthalmol 76: 602-6), and He a r al ' (1991, Invest Ophthalmol Vis Sci 32: 2723-8). : TABLE 2

Antibody Graft Survival (days)

Control 13,14, 17,23

W3/25 9,12,20,>100

OX35+0X38 | 29,29, >100,>100

EXAMPLE 6

Diagnosis and treatment of herpetic keratitis by antibody fragments

To investigate the potential of scFvs as therapeutic agents for stromal herpetic keratitis, both the necrotising and recurrent models of disease will be used. Mice willl receive topically applied scFv (an ti-CD4, CD8, CDI11b/c, IL-2R or HSV gpD antigen) im eye-drops or scFv of irrelevant specificity, | drop four times daily. An additional control group will receive topical 0.5% ophthalmic prednisolone phosphate and 3% acyclovir ophthalmic ointment (Zovirax, Glaxo-Wellcome), 4 times daily. Animals with necrotising disease will be treated from day 14 to day 21 after inoculation; animals with recurrent disease from day 2 to day 4 post-re activation. Systemic administration of scFv wil] also be investigated. Efficacy will be measured clinically at the slit lamp and by endpoint immunohistochemistry and histology to determine the composition of the leucocytic ) infiltrate. For diagnostic purposes, the scFv to HSV gpD will be coupled to fluorescein, a . fluorochrome already used in topical ophthalmic preparations that can readily be visualised using the blue light-source at the slit-lamp. After administration of the fluorochrome- tagged scFv as an eye-drop to the inafected cornea, we will attempt to detect the presence of viral antigen in the cornea in situ by direct examination and by slit-lamp photography with the blue filter. In the necrotising HSV keratitis model, the potential of scFv for rapid diagnosis will be investigated in the third week after inoculation. In the recurrent keratitis model, eye-drops will be given on the third day post-reactivation. The results of direct ' examination will be correlated with histopathology and viral culture.

EXAMPLE 7

Diagnosis and treatment of A canthamoeba keratitis by antibody fragments

Fluorochrome-labelled scFvs and dimerised fragments directed to a 40 kDa protein known to be present on the surface o f both trophozoites and cysts will be generated. The appropriate IgG hybridoma is available from the ATCC. Porton rats will receive scFv eye- drops on day 4, 7 or 10 after inoculation with Acanthamoeba castellanii Neff. At these stages, there will be most trophozoites, trophozoites and cysts, or mostly cysts in their corneas. respectively. On each day, eye-drops containing various concentrations of the - . conjugated scFv will be given and the eyes will be examined and photographed using the slit lamp. The animals will then be killed and their eyes processed for histology. This will allow us to correlate the slit lamp observations with pathology. To investigate the potential 7 therapeutic use of fragments, infected rats will be treated with unconjugated dimeric fragment topically every hour through the day from day 6 to day 10 after inoculation.

Other rats will receive a control fragment only, or a combination of propamidine and polyhexamethylene biguanide, the cuurent treatments of choice for amoebic keratitis.

Treatment efficacy will be assessed by~ clinical score, histopathology, and quantification of viable amoebae by limiting dilution from homogenised corneas.

Throughout the specification the aim has been to describe the preferred embodiments of the invention without limiting the invention to any one embodiment or specific collection of features. Those of skill in the art will therefore appreciate that, in . light of the instant disclosure, various modifications and changes can be made in the particular embodiments exemplified without departing from the scope of the present i 25 invention. All such modifications arid changes are intended to be included within the scope of the appendant claims.

Claims (1)

1. Use of a sub-immunoglobulin antigen-bimding molecule in the preparation of a medicament for treating an ocular disorder whewrein said medicament is formulated for S topical administration.

2. A use of a sub-immunoglobulin antigen-birding molecule as claimed in Claim 1, in which the sub-immunoglobulin antigen-binding molecule is immuno-interactive with a target antigen associated with the disorder.

3. A topical composition for treatment or diagnosis of an ocular disorder, comprising a sub-immunoglobulin antigen-binding molecule that is immuno-interactive with a target antigen associated with the disorder, together with a carrier. 4 An ocular composition formulated for topical administration to the eye, comprising a sub-immunoglobulin antigen-binding molecule that is immuno-interactive with a target antigen in the eye.

S. A sub-immunoglobulin antigen-binding mo lecule that is immuno-interactive with a target antigen indicative of an ocular condition for use in a method of providing an indication useful in the diagnosis of the ocular condition, the method comprising: (1) contacting a portion of an eye with the antigen-binding molecule by topical administration; and (it) detecting the presence of a com plex comprising the antigen-binding molecule and the target antigen.

6. A use according to Claim 1 or 2, a composition according to Claim 3 or 4, or an antigen-binding molecule according to Claim 5, in which the sub-immunoglobulin antigen- binding molecule does not comprise the Fc portion of an immunoglobulin molecule. Amended Sheet 08/07/2002

7. A use, composition or antigen-binding molecule according to any preceding claim, in which the sub-immunoglobulin antigen-binding molecule comprises a synthetic Fv fragment, a synthetic stabilised Fv fragment, or a minibody. & A use, composition or antigen-bindingg molecule according to Claim 7, in which the synthetic stabilised Fv fragment is a single chain Fv fragment.

9. A use, composition or antigen-binding molecule according to any preceding claim, in which the target antigen is selected from the group consisting of an MHC molecule, a co-stimulatory molecule, an adhesion molecule, a receptor-associated molecule, a cytokine receptor and a viral surface antigen.

10. A use, composition or antigen-bindings molecule according to Claim 9, in which the co-stimulatory molecule is selected from the group consisting of CD80, CD86 and CD152.

11. A use, composition or antigen-bindirag molecule according to Claim 9 or 10, in which the adhesion molecule is selected from the group consisting of CD11, CD18, CD54 and CD62L.

12. A use, composition or antigen-binding molecule according to Claim 9, 10 or 11, in which the receptor-associated molecule is s elected from the group consisting of CD3, CD4, CDS, CD28, CD40, CD40L and CTLAA.

13. A use, composition or antigen-bindin_g molecule according to any of Claims 9 to 12, in which the cytokine receptor is selected from the group consisting of the interleukin 2 receptor, a subunit of the interleukin 2 receptosr, and the interferon y receptor.

14. A use, composition or antigen-bindin g molecule according to any of Claims 9 to 13, in which the viral surface antigen comprises a herpes virus surface antigen.

15. A use, composition or antigen-binding molecule according to Claim 14, in which the herpes virus surface antigen is gD2 or gB2.. Amended Sheet 08/07/2002

16. A use, composition or antigen-binding molecule according to any preceding claim, in which the sub-immunoglobulin antigen-binding molecule is modified to prolong its half— life.

17. A use, composition or antigen-binding molecule according to Claim 16, in which the sub-immunoglobulin antigen-binding molecule is chemically modified witha polyethylene glycol.

18. An antigen-binding molecule according to Claim 5 or any of Claims 6 to 17 as dependent thereon, in which the step of contacting is characterised in that the sub— immunoglobulin antigen-binding molecule is delivered into the portion by modulating thes surface of the eye to improve penetration of the sub-immunoglobulin antigen-bindings molecule.

19. An antigen-binding molecule according to Claim 18, in which the corneal epithelium of the eye is modulated by a penetration enhancer.

20. A use according to Claim 1 or 2 or any of Claims 6 to 17 as dependent thereon, or a composition according to Claim 3 or 4 or any of Claims 6 to 17 as dependent thereon, ix which the medicament or composition is formulated with a penetration enhancer tc- improve penetration of the sub-immunoglobulin antigen-binding molecule into the eye or the comeal epithelium.

21. An antigen-binding molecule according to Claim 19, or a use or composition according to Claim 20, in which the penetration enhancer is selected from the groups consisting of benzalkonium chloride, capric acid, DMSO, dihydrocytochalasin B, digitonin, detergents and any combination thereof.

22. An antigen-binding molecule according to Claim 5, any of Claims 6 to 19 as dependent thereon, or Claim 21, in which the step of contacting is characterised in that thes sub-immunoglobulin antigen-binding molecule is delivered into the portion by subjecting: at least a portion of the eye to iontophoresis. Amended Sheet 08/07/2002

23. A use according to Claim 1, 2, any of Claims 6 to 17 as dependent thereon, or any of Claims 20 to 21 or a composition accordin g to Claim 3, 4, any of Claims 6 to 17 as dependent thereon, or any of Claims 20 to 21, in which the medicament is formulated for iontophoresis.

24. An antigen-binding molecule accordin g to Claim 5, any of Claims 6 to 19 as dependent thereon, Claim 21 or 22, in which thae step of contacting is characterised in that the sub-immunoglobulin antigen-binding molecule is delivered into the portion in the form of an eye drop or a collagen shield.

25. An antigen-binding molecule accordin g to Claim 5, any of Claims 6 to 19 as dependent thereon, Claim 21, 22 or 24, in which the sub-immunoglobulin antigen-binding molecule has a detectable label associated therewith.

26. An antigen-binding molecule according to Claim 25, in which the label comprises a fluorochrome.

27. An antigen-binding molecule according to Claim 26, in which the fluorochrome is fluorescein.

28. A use, composition or antigen-binding molecule according to any preceding claim, in which the medicament or composition is in thae form of an eye drop or a collagen shield.

29. A use, composition or antigen-binding rmolecule according to any preceding claim, in which the ocular disorder is selected from thes group consisting of corneal graft rejection, uveitis, inflammatory and infectious diseases, ocular tumours, neovascular proliferative diseases, neovascular maculopathies, rheurmatoid corneal melting disorders and autoimmune disorders.

30. A use, composition or antigen-binding molecule according to Claim 29, in which the ocular disorder is selected from the group «consisting of viral, bacterial or chlamydial conjunctivitis or keratitis, and ocular penthigoid . Amended Sheet 08/07/2002