WO2011154687A1 - Biomimetic corneal tissue - Google Patents

Abstract

This invention relates to the production of biomimetic corneal tissue by compression of gels, for example collagen gels, seeded with fibroblasts and subsequent seeding of epithelial cells onto the compressed surface. The biomimetic corneal tissue may optionally comprise multiple layers of compressed gel, an endothelial layer and/or embossed niches for epithelial stem cells. Biomimetic corneal tissue produced as described may be useful, for example, in therapeutic methods as well as in vitro modelling and screening methods.

Description

Biomimetic Corneal Tissue

This invention relates to biomimetic corneal tissue and methods of producing biomimetic corneal tissue.

The limbal epithelial stem cell (LESC) population of the cornea is responsible for maintaining the integrity of the outer epithelial surface. Destruction of LESCs may be caused by thermal or chemical injury, Stevens -Johnson syndrome, multiple surgeries, contact lens wear or microbial infection. This can lead to absence of an intact epithelial layer, conjunctival ingrowth, corneal nec asculari sat ion, chronic inflammation and discomfort and ultimately, impaired vision fl] . » 3L W

allogeneic transplant from a cadaveric or living relative donor. In this process, cells are commonly cultured on an amniotic membrane (AM) substrate before transfer to the injured eye ( 2 - 6 ] . Amniotic membrane has many favourable qualities for use in this context including antiinflammatory, anti-angiogenic and anti-scarring properties [7] .

Unfortunately, there are a number of drawbacks associated with its use including lack of a reliable supply of membranes, considerable biological variation amongst donors, costly screening processes and a lack of optimal transparency, which is a major issue when dealing with repair of the cornea. Also, LESCs do not always grow successfully on the amniotic membrane. LESC deficiency has also been treated using amniotic uteitibrane only, but, although this calms inflammation, it does not affect the

underlying deficiency. Limbal tissue transplantation may be helpful in restoring the stem cells, but can jeopardise the donor eye. furthermore, the use of allogeneic donor tissue requires long-terra immunosuppression ,

Many attempts have been made to develop an alternative to AM as a substrate for LESCs for use in corneal repair [8] . A substrate for a corneal equivalent or in vitro corneal model should be biocompatible, mechanically stable, and optically transparent and allow cell adhesion, migration and proliferation [11] . Collagen type I has been frequently reported as a corneal substrate, because it is the major component of the corneal stroma. However, this approach has had liraited success because hyperhydrated collagen gels are inherently weak due to their high water content. Chemical cross-linking has been used to improve the mechanical properties of collagen gels [12-16] . Although cross-linking enhances the mechanical properties, major drawbacks include the cytotoxicity of the cross-linker, reduced biomimicry and prevention of cell-based scaffold remodelling [17] .

Collagen vitrigel membranes have also been used to improve the mechanical properties of a collagen gel. Although this process offers superior optical and mechanical properties, the dehydration process required to produce the rigid, glass-like structure is lengthy and prevents the seeding of cells directly into the scaffold [18] .

This invention relates to the finding that synthetic corneal tissue may be produced from collagen constructs. This artificial tissue is highly biomimetic and may be useful, for example, in therapeutic methods as well as in vitro modelling and screening methods .

An aspect of invention provides a method of producing biomimetic corneal tissue comprising :

(i) providing a gel seeded with fibroblasts, (ii) compressing the gel to produce a compacted construct, and;

(iii) seeding a surface of the construct with ocular epithelial cells, thereby producing a biomimetic corneal tissue. In preferred embodiments, the gel is a collagen gel. A method of producing biomimetic corneal tissue may comprise:

(i) providing a collagen gel seeded with fibroblasts,

(ii) compressing the gel to produce a compacted collagen construct, and;

(iii) seeding a surface of the construct with ocular epithelial cells, thereby producing a biomimetic corneal tissue.

A method may further comprise the step of culturing the construct in culture medium to expand the epithelial cells on the surface.

Preferably, the ocular epithelial cells form an epithelial layer on the surface of the construct, most preferably a confluent epithelial layer. The layer may be at least one cell thick i.e. a monolayer. The cells may be seeded onto the surface in sufficient numbers to form an epithelial layer, or, more preferably the construct may be cultured in culture medium so that the epithelial cells form an epithelial layer or monolayer on the surface of the construct. The epithelial layer may be a confluent epithelial layer. The presence of a

epithelial or confluent epithelial layer may be determined, for example by light microscopy.

A collagen gel is a hydrogel comprising fibrils of collagen in an interstitial liquid. Collagen gels are generally isotropic and the collagen fibres are randomly orientated. Native fibril forming collagen types may be preferred in collagen gels including collagen types are I, I I , I I I , V , VI, IX and X I and combinations of these (e.g. I , I I I V or II, IX , XI ) . Preferably, native type I collagen is employed. In addition to collagen, a gel may further comprise one or more other natural gel-forming polymers, for example proteins such as laminin, silk, fibrin, fibronectin or elastin, glycoproteins such as

fibronectin, and polysaccharides such as chitin, or cellulose, or synthetic gel-forming polymers, for example organic polymers, such as polylactone, polyglycone, polycapryolactone or synthetic polypeptides and inorganic pol mers such as phosphate glass. The initial volume of the collagen solution used to produce a layer of compacted collagen construct will depend on the production methods and the design and intended use of the biomimetic corneal tissue. For example, a collagen solution may have a volume of 0.1 to 10 ml, for example 1, 2, 3, 4, or 5 ml, In some preferred embodiments, 2 to 3.5 ml of collagen solution may be used to produce a compressed collagen construct of about 100-150um thickness.

A collagen gel may be seeded with fibroblasts by admixing fibroblasts with a solution of collagen and then causing the collagen solution containing the cells to set to produce a seeded collagen gel.

The fibroblasts may be seeded into the collagen solution at a density of 1 x 10³ to 1 x 10⁶ cells/ml or 1 x 10* to 1 x 10⁶ cells/ml,

preferably about 1 x 10^s cells/ml.

The average collagen density of the seeded collagen gel before compression may be 0.5 to 5 mg/ml, preferably 1 to 4 mg/ml or 1.5 to 4 mg/ml. Suitable fibroblasts include ocular fibroblasts, such as lirabal or corneal fibroblasts, and dermal fibroblasts, for example neonatal dermal fibroblasts. In some embodiments, dermal fibroblasts may be preferred because they can be readily expanded in vitro using standard culture techniques. Preferably, the fibroblasts are human fibroblasts.

In some embodiments, fibroblasts which have been passaged in vitro no more than 6, 7, 8, or 9 times after explantation may be preferred.

Fibroblasts may be isolated or explanted from native tissue and, optionally, expanded before use. Techniques for the isolation and expansion of fibroblasts are well-known in the art.

In some embodiments, suitable fibroblasts may be obtained and expanded from a bank of fibroblast cells. The bank may contain samples of fibroblasts of different tissue antigen types (for example, different HLA (Human Leukocyte Antigen) types) . Cells may be identified which are tissue-matched for a particular individual, for example an individual in whom biomimetic corneal tissue is to be implanted.

Tissue matched cells may reduce or prevent an immune response against biomimetic corneal tissue implanted in the individual. Methods of identifying, tissue-typing samples are well-known in the art. Once suitable cells have been identified, they may be expanded and used in the methods described herein.

The fibroblast-seeded compacted collagen construct forms a stromal layer in the biomimetic corneal tissue.

In some preferred embodiments, the collagen solution, either before or after seeding with fibroblasts, may be added to a well, for example a well in a bioreactor. For example, the collagen solution may be added to the well, seeded with fibroblasts and then set to provide the seeded collagen gel in the well or the collagen solution may be seeded with fibroblasts, and then added to the well and set to provide the seeded collagen gel in the well. The well may be part of an array, for example, a well in a multi-well assay plate or an individual well held in a mounting plate or cassette with other wells. Optionally, as described below, the well may contain a suspension of endothelial cells onto which the collagen solution is added, either before or after it is seeded with fibroblasts. The endothelial cells may be retained in a matrix such as collagen gel or a compressed collagen layer. A layer of collagen or compressed collagen containing endothelial cells may be 10 to SOuiti thick, for example 20 to 30um thick. The well may contain 1, 600-4, 000 endothelial cells per mm² of the well bottom. As described below, the endothelial cells form an endothelial layer in biomimetic tissue produced by the described methods .

In preferred embodiments, the gel remains in the well for all compression, seeding and culturing steps i.e. steps (i) to (iv) are all performed on the gel in the well. After the biomimetic corneal tissue has bee produced in the well, it may be removed and used as required.

Preferably, the well is shaped and sized to mould the seeded collagen j ©1 3» ΓΊ t.C5 t -h€«■ Cite_* 2* · <KT€■ S3» 2 3 Fid S help 6 ?Jf» ί_·ΐΊ€ 3»omimetic corneal tissue.

For example, the well may mould the seeded collagen gel into the size and shape of a human cornea for implantation . For example, the seeded collagen may be moulded into a round gel of at least 11 mm diaraeter. Preferably, the construct is larger than the cornea, to provide excess collagen to facilitate handling and attachment. For example, the seeded collagen may be moulded into a round gel of at least 18, 19, 20, 21 or 22 mm diameter. Typically, the seeded collagen may be moulded into a round gel of up to 26, 28, 30 or 32mm diameter, but larger constructs may also be produced and then trimmed afterwards to the size required for a particular application. The well may be further shaped or adapted to introduce functional topographical features to the gel which are retained in the biomiraetic corneal tissue. For example, the well may mould tags into the seeded collagen gel to facilitate handling and implantation of the biomimetic corneal tissue; structural asymmetry which enable the epithelial and endothelial surfaces of the biomimetic corneal tissue to be

distinguished? and/or increased collagen thickness around the

periphery of the gel to facilitate handling, for example a well may have a convex base.

After seeding, the collagen solution may be set by any convenient method- Typically, dissolved triple helical collagen monomers are induced to polymerise (aggregate) to fibrils by incubation at about 37° at neutral pH. As the fibrils polymerise, there is a phase change and the solid network of fibrils ^supports' the remaining interstitial liquid in approximately the same volume and shape - i.e. it gels.

In some preferred embodiments, one or more crypts may be embossed into the surface of the seeded collagen gel before compression. Embossing is described in more detail below.

After setting, and optional embossing, the seeded collagen gel may then be compressed to produce a compacted collagen construct. As described above, the gel is preferably compressed in the same well in which setting occurred.

Plastic compression causes the seeded collagen gel to deform and reduce its volume by expelling interstitial liquid from the gel. The gel retains or substantially retains its new volume, even after the compression is removed. Plastic compression is described in more detail in WO2006/03442, Brown RA et al {2005) Adv. Funct. Mat. 15: 176-177, and elsewhere. The surface of the gel through which liquid is expelled when

compression is applied to the gel is generally termed a fluid leaving surface (FLS) , The amount of liquid expelled through the FLS by plastic compression per unit of surface area of the FLS (i.e.

V_{e peUed} (mm³) A sfrani²) ) may be 2 to 14 mm, for example 3, 4, 5, 6, 7_» 8» 9, 10, 11, or 12 mm, and preferably 2.6 to 13 mm or 5 to 10 mm.

Plastic compression may reduce the volume of the gel by 50% to 99,9%, e.g. 801 to 99.5%. For example, the compressed gel may have 0,1» 0.5, 1, 2» 3, 4 or 5% of its original volume. In some preferred

embodiments, at least 95%, 961, 91%, 98%, 99% or 99.9% (w/w) of the liquid in the gel may be expelled.

In some embodiments, a weight may be used to compress the gel by gravitational force . For example, a weight of 1 to 200g (e.g. lOg, SOg, lOOg, or ISOg) may be applied to the gel for 1 to 15 mins {e.g. 2, 4, 6, 8 or 10 mins) .

In preferred embodiments, the gel is compressed by a porous or permeable rigid body, such as a plunger. Liquid which is expelled through a fluid leaving surface as the gel is compressed is then absorbed into the porous rigid body. Preferably, the gel surface which contacts the porous rigid body is the only surface of the gel through which liquid is expelled during compression (i.e. the only FLS) . For example, other surfaces of the gel may be confined by impermeable supports which prevent the expulsion of liquid through those surfaces. All liquid expulsion may thus be directed through the gel surface which contacts the porous body (the FLS) . Preferably, setting and compression are performed at about 37° (e.g. 36°~38°C) . Compacted collagen constructs which are produced at lower temperatures (e.g. temperatures below about 35°C, such as 21 °C or 10'C) may display lower break strength. After compression, the average collagen density of the compressed collagen construct may range from 20 to 260 trig/ml, for example 50 to 260 rag/ml . Collagen density may vary within the compressed collagen construct. For example, the collagen density at the FLS may be greater than elsewhere in the gel.

After plastic compression, the compressed collagen layer may have a thickness of 50 to 200um, preferably 100 to 150um. The compacted collagen construct may comprise a single layer of compressed f broblast-containing collagen or multiple layers of compressed fibroblast-containing collagen. The use of multiple layers of seeded collagen allows control of the thickness of the compacted collagen constructs. This may be useful, for example in producing a biomimetic tissue which mimics the whole cornea.

A multi-layer compacted collagen construct as described above may be produced by;

(a) providing a collagen gel seeded with fibroblasts,

(b) compressing the gel to produce a compressed collagen layer,

(c) introducing an additional collagen solution seeded with

fibroblasts onto the compressed collagen layer,

(d) allowing the additional collagen solution to set on the compressed collagen layer to form an additional gel seeded with fibroblasts,

(e) compressing the additional gel to produce a construct comprising multiple layers of compressed collagen.

Steps (c) to (e) may be repeated one or more times as required to form a compacted collagen construct comprising the desired number of layers of compressed collagen seeded with fibroblasts, and for example displaying the appropriate thickness for the intended application.

In other embodiments, a multi-layer compacted collagen construct as described above may be produced by; {a} providing a gel seeded with fibroblasts,

(b) introducing an additional collagen solution onto the gel ,

(c) allowing the additional collagen solution to set on the gel to form a multilayer gel,

(d) optionally repeating steps b) and c) one or more times to introduce one or more additional layers to the multi-layer gel, and;

( e ) compressing the multi-layered gel to produce a construct comprising multiple layers of compressed collagen,

In some embodiments, the initial collagen gel and the additional collagen gel are compressed by a rigid body, such as a plunger.

Preferably the rigid body is porous. As the gel is compressed by the porous rigid body, liquid is only expelled from the gel through gel surface which is in contact with the porous rigid body (i.e. the only FLS) . Liquid which is expelled through a fluid leaving surface as the gel is compressed is then absorbed into the porous rigid body {i.e. the rigid body is absorbent) . In some embodiments, an additional seeded collagen solution may be added to the compacted collagen construct and compressed to produce a two layered compacted collagen construct. Alternatively, 2, 3, 4, 5 6 or more cycles of collagen addition and compression may be performed to produce a compacted collagen construct having 3, 4, 5, 6, ? or more layers of compressed collagen which are seeded with fibroblasts. This may be useful in producing a stromal layer in the biomimetic corneal tissue which is sufficiently thick to mimic the whole cornea. For example, a multi-layered compacted collagen construct of 400-600um thickness may be produced for use as a biomimetic cornea.

Following compression, a surface of the compacted collagen construct may be seeded with ocular epithelial cells. These cells will form the epithelial surface of the biomimetic tissue which is produced by the method , Ocular epithelial cells may be seeded onto the surface by any convenient technique. For example, a suspension of epithelial cells may be pipetted onto the surface of the collagen.

5

The surface seeded with ocular epithelial cells may be a surface through which fluid was expelled during compaction (i.e. an FLS). In some embodiments, the seeded surface may be the only surface through which fluid was expelled during compaction. Compression may increase 10 the density of the compacted collagen construct at the FLS relative to other parts of the construct.

The ocular epithelial cells may be corneal or limbal epithelial cells. In some preferred embodiments, the ocular epithelial cells are a mixed

in cX tide s ø j t^ti^ <s 2. s t ein C6 X X s <

Preferably, the ocular epithelial cells are huma n ocular epithelial

20

Ocular epithelial cells may be isolated, explanted or separated from ocular tissue, for example limbal tissue, such as human limbal tissue. Suitable ocular tissue may be obtained from a suitable donor. For example, suitable ocular epithelial cells may be produced by

25 enzymatically separating a population of cells from donor limbal rims, for example using dispase, trypsin, therittolysin or tryple™~selec .

Alternatively, the ocular epithelial cells may be derived from cultured progenitor cells, such as ES or iPS cells, that have been

After isolating a population of epithelial cells from ocular tissue* the population of isolated cells may be expanded in vitro? for example on feeder cells. After expansion, some or all of the expanded

population may be seeded onto the surface of the compressed collagen. In other embodiments, the ocular epithelial cells may be seeded directly onto the surface without expansion.

Techniques for isolating and expanding ocular epithelial cells are well-known in the art (see for example. Short et al 2007 Survey in Ophthalmology 52 (5) : 483-502 ) . In some embodiments, the surface of the compacted collagen ma be coated with an extracellular matrix protein before seeding with epithelial cells. In other embodiments, the surface of the surface of the compacted collagen may not be coated with an extracellular matrix protein. For example, the surface of the compacted collagen may not be treated with laminin before the epithelial cells are seeded onto it.

In some preferred embodiments, one or more crypts may be embossed into the surface of the construct before it is seeded with epithelial cells The gel may be compressed and embossed in two separate steps. For example, the surface of the gel may be embossed with one or more crypts before the gel is compressed, as described above. A rigid body comprising one or more projections on its surface (i.e. an embossing tool) may be used to emboss the gel. The rigid body is pressed into the surface of the gel so that the projections are driven through the surface of the gel and emboss one or more reciprocal crypts into the surface of gel. The gel may then be compressed, for example using a separate porous body, such as a porous plunger. Alternatively, the gel may be compressed and embossed simultaneously. For example, the gel may be compressed using a rigid and preferably porous body which comprises one or more projections on its surface. The one or more projections are driven through the surface of the gel and emboss one or more reciprocal crypts into the surface of gel, as the gel is compacted.

The projections on the rigid body may be absorbent or non-absorbent.

A crypt is a recess or pocket in the surface of the compacted collagen construct. The crypt has an opening at the construct surface which allows the entry and exit of cells and reagents. Ά suitable crypt may have a width of about 60um to 250 urn, preferably ?0um to 150um, a length of 60um to 250 urn, preferably lOOum to 200umm, and a depth of 20um to 150 um, preferably 25um to 130um, for example about lOOum.

Multiple crypts may be embossed into the construct. For example, a construct may have 10 or more, 50 or more, 100 or more, 200 or more, 300 or more or 400 or more crypts embossed into its surface.

Preferably, an array of crypts is embossed into the surface of the construct. The array of crypts preferably surrounds a central zone of the construct, which lacks crypts. For example, an array of crypts may form a ring of crypts around a central zone of about 11 to 13mm diameter. The crypts may be spaced about 50 to 100 um apart, for example, 75um apart, around the periphery of the central cornea. For example, the array may contain 400 to 500 crypts. Preferably, the crypts are radially disposed around the central crypt-free zone. A stamp for embossing such an array is shown in figure 8.

In some embodiments, the crypts may be coated with extracellular matrix (ECM) surface proteins before seeding with ocular epithelial cells. Suitable ECM surface proteins include laminin, fibronectin and collagen IV. For example, the surface of the projections may be coated with extracellular matrix (ECM} surface proteins, such that the crypts are coated with extracellular matrix (ECM) surface proteins when they are formed by the projections. Ocular epithelial cells may be seeded onto the surface comprising the crypts after embossing. Ocular epithelial cells seeded onto the surface may enter the crypts without directed seeding of the crypts. Alternatively, the crypts may be specifically seeded with cells. For example, the surface of the projections used to emboss the crypts may be coated with ocular epithelial cells, such that the crypts are seeded with ocular epithelial cells when they are formed by the pro ections . Alternatively, ocular epithelial cells may be printed directly into the crypts after embossing, for example using ink jet pJT X FitXO †»€chu Jt«16s _*

In some embodiments, for example where the biomimetic tissue is intended to mimic the central cornea, no crypts may be required and no embossing step is needed.

In some preferred embodiments, a second surface of the compacted collagen construct may be seeded with endothelial cells, such that, following culturing, the endothelial cells form an endothelial layer, for example, a confluent endothelial layer, preferably a layer which is one cell thick (i.e. a monolayer! on the second surface of the

is"t_*-xrmc #

The construct may be seeded with endothelial cells at any stage in the production process. In some embodiments, the fibroblast seeded collagen gel which forms the stroma of the biomimetic tissue may initially be provided on a layer of endothelial cells , as described above. For example, the seeded collagen solution may be added onto a layer of endothelial cells before setting and compression. The layer of endothelial cells may be within a matrix such as a collagen gel or compressed collagen gel as described herein. After setting,

compression and culturing, the layer of endothelial cells forms an endothelium, preferably a confluent endothelium on the second surface of the compressed collagen construct. This may be useful, for example, when the biomimetic tissue is produced in a well, as described above.

Alternatively, the second surface of the construct may be seeded with endothelial cells after compression, either before or after the first surface is seeded with epithelial cells.

The second surface is preferably opposite the surface which is seeded with epithelial cells (i.e. the first surface) For example, the epithelial cells may be seeded onto the upper surface of the compacted collagen construct and the endothelial cells may be seeded on to the lower surface of the compacted collagen construct. After culturing , the fibroblast containing collagen layer is sandwiched between a layer of epithelial cells and a layer of endothelial cells. The biomimetic tissue may thus comprise layers which mimic the epithelium, stroma and endothelium of native corneal tissue.

Preferably, the compacted collagen construct seeded with epithelial and endothelial cells comprises multiple layers of compressed

collagen, such that it has a thickness of 400-600um. This mimics the thickness of the entire stromal layer of the cornea and allows the production of synthetic corneal tissue which is biomimetic for the complete cornea. The second surface may be treated with extracellular matrix (ECM) surface proteins before seeding with said endothelial cells. This may improve the adherence of the endothelial cells to the surface.

Suitable ECM surface proteins include I ami n.i n, fibronectin and collagen IV.

Ocular endothelial cells may be isolated, explanted or separated from ocular tissue, for example whole cornea or limbal tissue, such as the limbal rim. Ocular tissue may be obtained from a suitable donor. In some embodiments, the population of isolated ocular endothelial cells may be expanded before seeding onto the second surface. Suitable techniques for obtaining and expanding endothelial cells are well known in the art (for example, Joyce NC et al, 2004 Cornea 23 (8 suppl) : S8-S19) .

Alternatively, the ocular endothelial cells may be derived from cultured progenitor cells, such as ES or IPS cells, that have been driven towards a corneal ohenotvoe, In some embodiments, the ocular endothelial cells may be seeded onto the second surface as an explant or as a cell suspension.

Preferably the endothelial cells are human endothelial cells. Endothelial cells raay be seeded onto the second surface at a density of 1,600-4,000 /mm². Following seeding of the first and/or second surfaces of the compacted collagen construct, the construct may be cultured under conditions which support the growth and proliferation of the seeded cells. For example, the construct may be immersed in a suitable culture medium and incubated at a suitable temperature, typically 36 C to 38°C, most conveniently 37°C .

Suitable media include corneal epithelial cell media, such as Kpxlife medium, Invitrogen) and corneal Storage Media such as Gptisol™,

Optisol™ GS, Dexsol™, McCarey-Kaufman Media, (Bausch and Lomb) and Euscl-C™ (Means et al Arch Ophthalmol . 1995; 113 {6} : 805-809.

Suitable culture methods and techniques are well-known in the art (e.g. Pels et al Int Ophthalmol (2008 ) 28:155-163).

The construct may be cultured until epithelial cells are established and widely distributed on the first surface and/or endothelial cells on the second surface. Preferably, the construct is cultured until epithelial cells form a layer or monolayer, preferably a confluent 1? layer or monolayer, on the first surface and/or endothelial cells form a layer or monolayer, preferably a confluent layer or monolayer, on the second surface, as required. A construct comprising an epithelial layer and a fibroblast containing compressed collagen layer may be useful as a biomimetic corneal tissue. In some preferred embodiments, the biomimetic corneal tissue may also comprise a layer of endothelial cells, such that the central fibroblast-containing layer is located between the epithelial and endothelial layers.

In some embodiments, biomimetic corneal tissue produced by a method described above which comprises a layer of epithelial cells may be useful in therapy as an implant, for example for the treatment of diseased or damaged cornea in an individual . Implantation of the tissue into the eye of a patient exposes the epithelial cell layer of the tissue to an air/liquid interface which induces a differentiated epithelial phenotype which mimics native cornea. An epithelium with a differentiated phenotype is stratified and comprises multiple layers. Ά differentiated epithelium may include a population of self renewing progenitor cells (i.e. epithelial stem cells) capable of maintaining and regenerating the epithelial layer.

In other embodiments, a differentiated epithelial phenotype may be induced in vitro by "airlifting" the biomimetic tissue after a layer, preferably a confluent layer, of epithelial cells has formed on the collagen surface. Biomimetic corneal tissue may be useful as an implant in therapy, and may also be useful, for example, in modelling and screening methods.

The construct may be airlifted by culturing the construct under conditions in which the layer of epithelial cells, preferably the confluent layer of epithelial cells, is exposed to air e.g. the layer is positioned at an air/liquid interface, such that said layer forms a differentiated and stratified epithelium.

Suitable techniques for "airlifting" cell cultures are well known in the art. For example, the layer of epithelial cells may be positioned at an air/liquid interface by culturing the construct on a permeable support, such as a membrane (e.g. transwell™ permeable supports,

Costar™) . This allows the epithelial cells on the surface of the construct to be exposed to the air whilst culture medium accesses the construct through the permeable support.

The construct may then be cultured in the culture medium with the epithelial layer exposed to the air until the air-exposed epithelial layer adopts a d fferentiated epithelial phenotype comprising multiple layers of epithelial cells. For example, the construct may be cultured with the epithelial layer exposed to the air for 1, 2, 3 days or more. Differentiation may be determined for example by determining the expression of cytokeratin 3 and/or cytokeratin 12, which are markers of differentiated epithelial cells.

Following airlifting, the biomimetic corneal tissue comprises a

multilayered differentiated epithelium, a stromal layer comprising fibroblasts and optionally an endothelium. Following production, biomimetic corneal tissue may be subjected to a range of tests. For example, it may be tested for viability (e.g.

live/dead assay) and functionality (e.g. CFE assay of removed

epithelial cells) . In some embodiments, testing may be non-destructive and does not damage or alter the biomimetic corneal tissue. Alternatively, tests may be carried out on a sample from a batch of biomimetic tissue to determine the properties, such as the histology, physical and

functional properties, of the batch of biomimetic tissue. For example, the number, shape, metabolism or condition of viable cells in the tissue, and/or the opacity, electrical resistance or permeability of the tissue or the wound healing response after injury may be determined.

The histology of the biomimetic corneal tissue may be determined. For example, the cell size, shape and number of epithelial cell layers and basement membrane deposition may be determined. Suitable biomimetic tissue mimics the histology of native cornea. Histology may be determined by standard microscopy techniques.

The biomimetic corneal tissue may be tested for the expression of epithelial markers, such as cytokeratins 3, 12 and 15, epithelial stem cell markers such as p63a and ABCG2, and basement membrane markers such as laminin, collagen IV and betal integrin (Schlotzer-Schrehardt U Experimental Eye Research (2005) 81(3) : 247-264 ) .

The angiogenic properties of the biomimetic corneal tissue may be tested. Suitable biomimetic corneal tissue is non-angiocenic, preferably anti-angiogenic .

The physical properties of the biomimetic corneal tissue may be tested. For example, the ability of the tissue to conform to the curve of the eye and to withstand the physical stresses of the ocular environment may be determined.

The biomimetic corneal tissue may also be tested for proteomic and genomic signatures, cytokine release and gene regulation relative to native corneal tissue.

Another aspect of the invention provides a biomimetic corneal tissue produced by a method described above. Preferably, in addition to a compressed collagen layer containing fibroblasts (which corresponds to the corneal stroma) , the biomimetic corneal tissue comprises at least a monolayer of epithelial cells and most preferably a stratified epithelium. In some embodiments, the tissue may further comprise an endothelial layer.

A biomimetic corneal tissue produced by a method described above may be used in a method of treatment of the human or animal body. Biomimetic corneal tissue produced by a method described above may be used in the manufacture of a medicament for use in the treatment of ocular disease or damage.

Ocular disease or damage may include limbal epithelial stem cell deficiency, diseased or damage cornea, impaired vision or blindness.

Blindness or impaired vision may be associated with LESC deficiency or lack of function.

Biomimetic corneal tissue may be useful in patients where the limbal epithelial stem cell population has been damaged or destroyed. This may arise as a result of chemical or thermal injury, Stevens-Johnson syndrome, multiple surgeries, contact lens wear, microbial infection, aniridia, or radiation injury. Treatment may include surface repair or replacement of the outermost corneal layer, for example with tissue comprising epithelial and

In these embodiments, the biomimetic corneal tissue may comprise epithelial and stromal layers without an endothelial layer (i.e.

biomimetic corneal surface tissue) .

Biomimetic corneal tissue may also be useful in patients in which the cornea has been damaged or destroyed. For example, corneal damage may arise from Fuchs dystrophy, keratoconus , penetrating injury, and other corneal dystrophy, infection and scarring.

Treatment may include surface repair or replacement of the whole

In these embodiments, the biomimetic corneal tissue may comprise epithelial, stromal and endothelial layers. Treatment may include repair or replacement of the whole cornea, for example with biomimetic tissue comprising epithelial, stromal and endothelial layers

A method of treatment of an individual may comprise;

implanting biomimetic corneal tissue produced by a method described above into the eye of the individual.

For example, the biomimetic corneal tissue may replace the outermost corneal layer in the individual or the whole cornea.

The implant may be fixed in place by a surgeon using standard

techniques. For example, an implant may be fixed in place with adhesive (e.g. fibrin glue) . In some embodiments, the implant may be covered by a protective contact lens to prevent damage and hold the implant in place.

Biomimetic corneal tissue of the invention may also be useful in methods of modelling corneal tissue and screening for ocular irritants and agents which may be useful in ocular therapy.

A method o f screen i ng may compr i se :

contacting a test compound with biomimetic corneal tissue produced by a method described above

and determining the effect of the test compound on the tissue. The effect of the compound on the functional properties of the tissue or the viability or condition of cells in the biomimetic tissue may be determined.

For example, the effect of the compound on the histology, opacity, permeability, electrical resistance, or wound healing response of the tissue may be determined. Cell viability may be determined by any convenient method, for example MTT assay, bioluminescent ATP assays or BrdU labelling may be used.

The effect of the compound on the expression of cell markers in the tissue may be determined, for example epithelial markers, such as cytokeratins 3, 12 and 15, epithelial stem cell markers such as p63a and ABCG2, and basement membrane markers such as laminin, collagen IV and β1 integr i .

A test compound which reduces the cell viability or functional properties of the biomimetic tissue may be identified as an ocular irritant. The extent of irritation may be determined from the extent of functional inhibition or loss of viability. A test compound which displays no effect on cell viability or functional properties of the biomimetic tissue may be identified as a non-irritant or non-toxic. This may be useful in the development of cosmetics and other consumer and personal care products, reducing the need for animal testing. This may also be useful in safety and toxicological testing of chemicals and pharmaceuticals. A test compound which increases the cell viability or functional properties of the biomimetic tissue may be identified as a potential ocular therapeutic. This may be useful in the development of new ocular druqs and druq formulations. Suitable compounds may be tested further.

Non-collagen gels may also be used to produce biomimetic corneal tissue in accordance with the invention,

An aspect of the invention provides a method of producing biomimetic corneal tissue comprising:

(i) providing a non-collagen gel seeded with fibroblasts,

(ii) compressing the non-collagen gel to produce a compacted gel construct, and;

(iii) seeding a surface of the construct with ocular epithelial cells, thereby producing a biomimetic corneal tissue.

A non-collagen gel may comprise one or more biocompatible non-collagen gel forming polymers, Suitable non-collagen gel forming polymers include natural gel-forming polymers, for example proteins such as laminin, silk, fibrin, fibronectin or elastin, glycoproteins such as fibronectin, and polysaccharides such as chitin, or cellulose, or synthetic gel-forming polymers, for example organic polymers, such as polylactone, polyglycone, polycapryolactone or synthetic polypeptides and inorganic polymers such as phosphate glass -

Non-collagen gels may be used in the same way as collagen gels to produce biomimetic corneal tissue and all features and aspects of the methods and constructs described above for collagen gels apply mutatis mutandis to non-collagen gels.

Various further aspects and embodiments of the present invention will be apparent to those skilled in the art in view of the present disclosure.

All documents mentioned in this specification are incorporated herein by reference in their entirety. "and/or" where used herein is to be taken as specific disclosure of each of the two specified features or components with or without the other. For example "A and/or B" is to be taken as specific disclosure of each of (i) A, iii) B and (iii) A and B, just as if each is set out individually herein.

Unless context dictates otherwise, the descriptions and definitions of the features set out above are not limited to any particular aspect or embodiment of the invention and apply equally to all aspects and embodiments which are described.

Certain aspects and embodiments of the invention will now be

illustrated by way of example and with reference to the figures and tables described below.

Figure la shows the production of biomimetic corneal tissue in transverse section and figure lb shows biomimetic corneal tissue produced by the methods described herein. Figure 2 shows plastically compressed collagen constructs . (A)

Constructs are relatively transparent after PC. (B] Constructs are able to withstand manipulation and lie flat on a porcine eye. (C) Fibroblasts in constructs remain viable immediately after PC and (D) after 1 week in culture as indicated by LIVE/DEAD stain (alive, light; dead, dark) . {E} The percentage of viable fibroblasts 1 hour after compression and 1 week after culture.

Figure 3 shows epithelial layer morphology. Comparative light micrographs of histological cross sections of human corneal epithelium and LECs on PC constructs stained with H&E (A, C, E) or semi- thin sections stained with toludine blue (B, D, F) . (A, B) Human central cornea sections. (C, D) Sections of Fib+ PC constructs. (E, F)

Sections of Fib- PC constructs . Scale bar 100mm. Figure 4 shows a quantitative comparison of basal epithelial cell density. Statistical significance indicated by *p«3.05, ** p < 0.01 as determined by oneway ANOVA and Tukey* s post-hoc test (CC, central cornea, Fib+, constructs, Fib-, constructs) , Scale bars 50mm,

Figure 5 shows a schematic diagram of an experimental setup for PC of pre-formed collagen gels including loading and blotting elements, and nylon meshes to prevent adhesion.

Figure 6 shows representative transmission electron (TE) micrographs of corneal epithelial cells and epithelial cells on Fib+ and Fib- constructs. (A) TE micrograph of human corneal epithelium. (B) Higher magnification TE micrograph showing microvilli on epithelial surface. (C) TE micrograph of a transverse section through stratified

epithelial cells on a Fib+ PC construct . (D) Higher magnification micrographs revealing the presence of desmosomes (white arrows) and associated intracellular keratin cytoskeletal filaments (KF) . (E) TE micrograph of a transverse section through stratified epithelium on Fib- PC construct. (G) Higher magnification micrograph of the apical surface showing microvilli on Fib- PC construct. Scale bars A, C, E, 2mm; B, lmm; D, 200nm; F, 0.8mm,

Figure 7 shows representative scanning electron (SE) micrographs of corneal epithelial cells and epithelial cells on Fib+ and Fib- constructs. (A) SE micrograph of corneal epithelial cells. (B) Higher magnification SE micrograph of surface features of corneal epithelial cells. (C) SE micrograph of epithelial cells on Fib+ constructs (D) Higher magnification SE micrograph of surface features of epithelial cells on Fib- constructs, (E) SE micrograph of epithelial cells on Fib+ constructs (F) Higher magnification SE micrograph of surface features of epithelial cells on Fib- constructs. Scale bars Ά, C, E, 10mm; B, lmm; D, F, 2mm. Figure 8 shows a schematic representation of the head of an embossing tool for embossing an array of crypts onto a collagen gel. The array of projections, which correspond to crypts, surround a central area which corresponds to the central crypt-free zone of the cornea.

Figure 9 shows P63 alpha and Phalloidin wholemount staining of epithelial cells on fibroblast containing collagen constructs.

Experiments

Methods and Materials

Donor tissue

Cadaveric donor corneal rims with appropriate research consent were obtained from Moorfields Lions Eye Bank (UK) . Ethical permission for this study was obtained from the Research Ethics Committee (UK) (Ref No. 08/H0715/83) . Corneas were stored at 4°C in Optisol (Chiron

Ophthalmics Inc. Irivine, California) after enucleaction and prior to LEG and fibroblast isolation.

Isolation and culture of human limbal epithelial cells

Human donor corneal rims were cut into quarters and incubated with 1.2 U ml-1 dispase II (Roche Diagnostics GmbH, Mannheim, Germany) in phosphate buffered saline (PBS; Invitrogen Ltd, Paisley, UK) for 2 hours at 37°C or overnight at 4°C. The limbus was gently scraped with the point of forceps to isolate limbal epithelial cells in that region. A single cell suspension was obtained by trituration in corneal epithelial cell media (CECM) containing DMEM : F12 basal medium, 10% fetal bovine serum, II antibiotic antimycotic, EGF (lOng/ml;

Invit ogen Ltd, Paisley, UK), hydrocortisone (0. mg/ml ) , insulin (5mg/ml) , adenine ( 0.18mM), transferrin (5mg/ml), T3 (2nM) , cholera toxin (O.lnM; Sigma-Aldrich, Dorset, OK) . A single cell suspension was then seeded onto feeder layers of 3T3-J2 cells that had been growth a ested with 4mgml^"1 mitomycin C for 2 hou s . CECM culture media was changed three times a week and additional growth arrested 3T3s were added when required. After 12-14 days of growth, cells were passaged 2? using 0,5% trypsin-EDTA (Invitrogen Ltd, Paisley, UK) and further expanded to passage 1 before being transferred onto collagen

constructs . isolation and culture of human 1 imbal fibroblasts (H Fs)

After isolation of LECs from donor rims, the remaining sclera 1-1 imbal quarters were placed in explant culture with ibroblast culture medium consisting of DMEM-Glutamax containing 1% anti-anti (Invitrogen Ltd, Paisley, UK) , and 10% adult bovine serum (Sigma-Aldrich, Dorset, UK) . Cultures were maintained for > 1 month to allow fibroblast out-growth. Fibroblasts could be expanded and passaged using 0.05% trypsin-EDTA {Invitrogen Ltd, Paisley, UK) and were seeded in constructs up to passage six. Preparation of plastic compressed collagen constructs

Collagen gels were prepared by sodium hydroxide (Sigma- drich,

Dorset, UK) neutralisation of 8ml of sterile rat-tail type I collagen (2.06 mg ml-1 First Link, Birmingham, UK) and lml of 10X Minimum Essential Medium (Invitrogen Ltd, Paisley, UK) . After neutralisation, 1 ml of media containing HLFs was added at a concentration of

100,000 cells/ml (lml of media alone, if casting acellular gels) and the solution left on ice for 30 minutes before casting. Gels were cast in four, 20mm diameter circular ring moulds and left to stabilise at 37 °C in a 5% C0₂ incubator for 30 minutes. The hyperhydrated collagen gels were then subjected to unconfined compression by applying a weight for 5 min to gels which were sandwiched between layers of nylon and metal mesh. The resulting collagen constructs were transferred to the wells of a 12 well plate and placed in fibroblast media until the addition of the epithelial cells. Epithelial cells were seeded onto the surface of collagen constructs at a concentration of approximately 650,000 cells/construct and constructs maintained in culture at 37°C in 5% CC½ for 12 days before being subjected to airlifting. Culture medium was replenished three times a week. In other embodiments, 8mi of rat tail collagen was mixed with 1 ml of MEM medium and kept on ice. Human fibroblasts were prepared and resuspended to appropriate concentration in fibroblast medium. The collagen solution was neutralized using sodium hydroxide and 1ml of fibroblast suspension at a concentration of 880,000 cells/ml was added to get a final concentration of around 220,000 cells/gel. The seeded collagen solution was left on ice in the fridge for 30 mins for the

After 30 mins, 2.5ml of the seeded collagen solution was pipetted into each of 4 wells and lids placed on them. The 4 cassettes were placed on a large round petri dish in the incubator and left for 30 mins to set at 37oC. An alignment tube and plunger were carefully placed in the wells and weights added on top. The wells were left for 15 mins to compress the collagen.

The plungers were then removed from the wells and a sterile plastic ring added to the well to hold down the compressed gel. 650,000 epithelial cells were added to each compressed gel in 1.5ml of corneal epithelial culture medium and the wells returned to the incubator until airlifting on day 13.

After a further 5 days (on day 18), the wells were processed for analysis by either fixation in 4% PFA for 30 mins at RT for wax embedding and sectioning, followed by storage in PBS at ^CC and wax embedding must be done within 48hrs of fixation; or fixation in

Karnovsky' s fixative for SEM and TEM processing.

Airlifting to induce stratification

After 12 days in culture, constructs were transferred to a cell culture insert (Miilipore PICM03050, West Lothian, UK) in a 6 well plate containing CECM and maintained at an air-liquid interface for a further 5 days to induce stratification. Constructs were then fixed for immunochemical analysis with 4% PFA for 1 hour at room tempe a ure .

Normal human central cornea was used as a control tissue throughout and was fixed and processed using protocols identical to that for collagen constructs.

Transfer of collagen constructs to the porcine eye

Whole porcine eyes were obtained from a commercial meat producer and washed with PBS, rinsed with 2% povidone iodine for 5 minutes, washed with PBS and then immersed in PBS containing 2% antibiotic- antimycotic . The corneal epithelium was debrided using a scalpel and the surface was coated with 500 ml fibrinogen (4.0 mg/mL in PBS, Sigma) . The collagen construct was dried using a sterile cotton bud and coated with 500 ml of thrombin ( 50u7ml solution in PBS, Sigma) . It was then transferred to the surface of the cornea and any creases removed by gentle manipulation of the membrane. As a control, the porcine corneal surface and the RAFT construct were moistened with PBS at room temperature. Adherence of the RAFT construct was assessed by tugging at the construct margin.

Live/dead viability assay

The viability of limbal fibroblasts within the collagen constructs was assessed using a LIVE/DEAD® viability kit (Invitrogen, Ltd, Paisley, UK), 1 hour and 1 week after compression and culture in CECM media. The kit uses a two-colour fluorescence system, labelling live cells green as intracellular esterase activity converts non-fluorescent calcein AM to fluorescent calcein. Ethidium homodimer-1 enters dead cells with damaged membranes where the red fluorescence is enhanced upon binding to nucleic acids. A dead cell positive control was produced by treating fibroblast containing collagen constructs with 70% methanol for 30 rain before staining with the LIVE/DEAD® viability Wholemount immunofluorescence of collagen constructs

Collagen constructs and central corneal specimens were processed forimmunochemistry by first blocking in 5% normal goat serum with 0.25% Triton X-100 (Sigma-Aldrich Ltd, Dorset, UK) in PBS and with 5% normal goat serum without Triton X-100 for ABCG2 (mouse) . Samples were then incubated in primary antibody to differentiated corneal

epithelial marker cytokeratin 3 (CK3) , ABCG2 (mouse) , p63 (rabbit) or CK15 (mouse) at dilutions shown in table 1, overnight at 4 °C.

After washing in PBS, samples were incubated for 1 hour at room temperature with secondary goat anti-mouse 594 Alexa Fluor antibodies (1:500; Invitrogen Ltd, Paisley, UK) and double stained with FITC- label led phalloidin (1:1000, Sigma-Aldrich Ltd, Dorset, UK) in order to visualise fibroblasts and epithelial cytoplasm. Negative control sections were treated identically except for omission of primary antibody. Nuclei were counterstained with propidium iodide (Sigma- Aldrich Ltd, Dorset, UK) for 10 minutes at room temperature before mounting and coverslipping for confocal analysis on a Zeiss LSM 510 microscope.

Immunofluorescent and histological staining of parafin sections

Collagen constructs and central corneal quarters were fixed in 41 PFA for 1 hour at room temperature before embedding in parafin for sectioning (5mra) with a microtome. Sections were subjected to trypsin (Invitrogen Ltd, Paisley, UK) pretreacment for 20 mins at 37 " before treatment with 0.5% Triton X-100 in PBS for 5 mins and then blocking in 10% normal goat serum (Sigma-Aldrich Ltd, Dorset, UK) for 1 hour at room temperature. Primary antibodies were added in the

concentrations shown in table 1, and incubated overnight at 4°C. After washing with PBS, secondary Alexa Fluor 488 antibodies were applied at a concentration of 1 : 500 in PBS for 1 hour at room temperature .

Nuclear counterstaining was pe formed with propidium iodide and slides moun ed using vectcrshield mounting medium (Vector Laboratories Ltd, Peterborough, UK) for confocal analysis on a Zeiss LSM 510 microscope. After rehydration through a series of alcohols to water,

representative sections were stained with haematoxylin and eosin (H & E) and then mounted and coverslipped usi g DPX . Sections were imaged using an Olympus BX50 light microscope and Evolution MP colour camera (Media Cybernetics) ,

Electron microscopy

Cellular collagen constructs were examined by scanning electron microscopy (SEM) and transmission electron microscopy (TEM) . specimens were fixed in Karnovsky' s fixative, rinsed with sodium cacodylate buffer and then post-fixed in osmium tetroxide. They were then washed with distilled water before being passed through a graded ethanol series. For SEM preparation samples were dried, mounted on stubs and splutter coated with gold before examination on a digital scanning electron microscope (Zeiss Sigma FESEM) . For TEM preparation, specimens were embedded in araldite epoxy resin mixture before 0.7mm semi-thin sections were cut and stained with alcoholic toludine blue, Oltrathin (70nm) sections were collected on coppe grids and stained with uranylacetate and lead citrate prior to examination on a transmission electron microscope (JEOL) .

Determin tion of epithelial basal cell density

Epithelial basal cell density was calculated by analysis o Z-stack images taken at the basal cell level. Three images were taken at random in the x-y plane using a confocal microscope with a 16x objective for each of three specimens in each condition; constructs containing fibroblasts (Fib+) , without fibroblasts (Fib-) or central corneal specimens.

The numbe of propidium iodide nuclei per field was counted and the mean density per mm² was determined for each preparation. This experiment was performed in triplicate. Statistical analysis

Epithelial basal cell density and cell viability values are expressed as mean ± SEM. A one-way A OVA was performed to determine significance between cell density calculations !p<0.05) followed by post-hoc analysis using the Bonferroni test to test for significant difference between groups (p<0 . 05 ) . All experiments were performed in triplicate.

Collagen constructs

Cellular collagen gels could be cast and compressed rapidly and simply to produce thin, (lOO-lSOram) transparent constructs {Fig. 2A) that were easy to handle in liquid and maintained sufficient mechanical strength to withstand manipulation on the surface of a porcine eye. The constructs were able to lie lat on the concave cornea surface with no evidence of folding or puckering and could be secured in place using fibrin glue (Fig .2B) . LECs adhered to the collagen surface of

constructs within 24 hours and were able to survive and expand in culture. A high proportion of human limbal fibroblasts within the collagen constructs remained viable immediately after compression

(85.110.11; Fig. 2C) and also after culture for 1 week (85.6±0.1%; Fig. 2D) as indicated by a LIVE/DEAD viability assay (Fig. 2E) .

Epithelial cell morphology

Healthy human central corneal epithelium is a stratified, squamous, multicellular layer that ated superior to the f broblast- containing collagen stroma (Fig. 3A) . The squamous cells are located on the apical surface with the subjacent wing cells overlying

basal cuboidal cells. The cuboidal cells are in contact with the basement membrane, clearly indicated in toludine blue stained semi- thin sections (Fig. 3B) . After 12 days in submerged culture and a further 5 days at the air-liquid interface, human LECs on the surface of PC collagen constructs adopted a conformation very similar to human central cornea (Fig. 2C-F) . We compared the epithelial cell layers on Fib+ PC constructs and Fib- constructs with human central cornea. On Fib+ constructs, at the apical surface cells appeared flattened and squaraous-like, whereas cells adjacent to the basement membrane displayed a more cuboidal appearance (Fig. 3C) . Toludine blue staining revealed a similar cell morphological pattern but also more clearly showed a number of wing-like cells present in the intermediate layers and visible nucleoli, indicative of cells that are not in division {Fig. 3D) . On Fib- PC constructs, epithelial cells also adopted a stratified conformation with squamous and wing-like cells above cuboidal /columnar shaped cells (Fig. 3 E ) .

However, basal cells appeared enlarged compared to those on Fib+ constructs. This was particularly evident in toludine blue stained sections where superficial squamous and wing-like cells also displayed very apparent nucleoli not seen in the basal layer of cells (Fig. 3F) .

Basal epithelial cell density

Wholemount PC constructs and human central corneal specimens were stained with FITC-labelled phalloidin and cytokeratin 3 (CK3)

antibody, a differentiated corneal epithelial cell marker, and representative z-stack images were analysed. Due to storage and processing conditions of the cadaveric donor cornea, a number of the superficial surface epithelial were lost or damaged. Of the remaining cells, many were positive for CK3 and showed a polygonal morphology typical of a healthy corneal epithelium (Fig. 4Ά) . The basal layer showed tightly packed cells with extremely high nuclear to cytoplasm (N/C) ratio (Fig. 4B) . In the sub-basal region, fibroblast type cells stained with phalloidin were visible in the stroma (Fig. 4C) . On the superficial surface of Fib+ constructs, a number of squa ous- like cells were shown to express CK3 (Fig. 4D) . Moving through the z- stack to the basal level, cells had a lower N/C ratio when compared to cells from the central cornea but also did not express CK3 (Fig. 4E) . Cells beneath the basal layer, within the collagen construct, showed the typical morphology of a healthy fibroblast with extended processes {Fig. 4 F ) .

A limited number of superficial surface LECs expressed CK3 on Fib- PC constructs and a number of cells had a more elongated appearance compared to those on constructs with fibroblasts (Fig, AC) . Cells at the basal level appeared to have a greatly decreased N/C ratio compared to both cells in Fib+ constructs and central cornea (Fig. 4H) . There were no cells present in the body of the Fib- PC constructs (Fig. 41).

The basal epithelial cell density was significantly higher in central cornea (6681 ± 1063 cells/mm²) compared to confluent areas in both Fib+ constructs (1640 ± 248.1 cells/mm2; p<0.01 } and Fib- constructs (1326 i 215.0 cells/mm2; p<0.01) . Although the basal cell density was higher in the Fib+ condition compared to Fib-, this was not found to be a significant difference (p>0.05) (Fig. J) ,

Basement membrane protein and putative stem and differentiated cell marker expression profiles.

The expression patterns of differentiated corneal epithelial marker CK3, putative stew cell markers P63a, CK15 and bl-integrin and extracellular matrix basement membrane proteins collagen IV, laminin, collagen VII and perlecan in the human central cornea were compared with those on collagen constructs. CK3 expression was seen throughout all layers of the human central cornea and this expression pattern was replicated in the multilayered epithelium on Fib*- and Fib- PC

constructs. Equally, putative stem cell marker CK15 was expressed cytoplasmically in all cell layers of the central corneal epithelium and collagen constructs, with some high intensity staining in a number of squamous cells on the superficial surface, particularly obvious in the Fib- section. Another putative stem cell marker, P63a, was predominantly expressed in the nuclei of basal cells of the central corneal epithelium but was not detected in cells on either of the collagen construct types in the paraffin samples, βΐ-integrin is a putative progenitor and

proliferative cell marker, which was strongly expressed in the basal cell membranes of the central corneal epithelial layer and was also strongly expressed in the membranes of the basal cells on collagen Fib+ constructs. Expression in Fib- constructs was detectable but was considerably weaker than in central cornea and Fib+ constructs.

The major structural basement membrane (BM) protein, collagen IV, was strongly expressed in central corneal epithelial BM and was also strongly expressed in collagen construct sections, however, in Fib- constructs, staining was almost undetectable. Laminin, a major non- collagenous BM constituent was strongly expressed in the central cornea and in some areas of Fib+ constructs. In Fib+ and Fib- constructs there was also some intracellular expression of laminin but there was weak staining of the BM in Fib- constructs. Both collagen VII, a major structural component of anchoring fibrils, and perlecan, a major BM heparan sulfate proteoglycan, were strongly expressed in central cornea but not expressed in sub-epithelial regions on either Fib+ or Fib- collagen constructs.

Ultrastructure of epithelial cells

SEM and TEM revealed that LECs cultured on collagen constructs at the air-liquid interface produced an epithelial layer with similar ultrastructural features to human central cornea (Figs. 6 and 7) » Representative TE micrographs of human central corneal epithelium show a multilayered epithelium with distinct cell borders visible (Fig. 6A) and on the superficial surface of the corneal epithelium, numerous microvilli were present (Fig. 6B) . Representative micrographs show a stratified epithelium on Fib+ collagen constructs with cuboidal-shaped cells on the surface of PC constructs and squamous cells at the superficial margin (Fig. 6C) . At higher magnification, desmosomes and associated intracellular cytoskeletal keratin filaments were visible between highly interdigitated cells, a feature common to wing-like cells in the human corneal epithelium (Fig. 60) . Cytoplasmic keratin filaments, abundant in differentiated corneal epithelial cells, were also clearly visible in close proximity to cell nuclei (Fig. 6D) .

Stratification was also seen at the ultrastructural level in sections of Fib- collagen constructs. In a representative specimen, at least five cell layers could be seen (Fig. 6E) . The superficial layers contained squamous cells with abundant keratin fi laments surrounding central nuclei (Fig . 6E) and at higher

magnification numerous surface microvilli became apparent on the super icial squamous layer ( Fig 6F) . SEM of human central cornea showed tightly apposed, polygonal surface epithelial cells with distinct borders (Fig . 7A) . At higher

magnification numerous microvilli were clearly visible on the cell

55 XTf<3C¾5 L

-Tf C XB> JL X t- €»1 X<¾JL C€»JL-LS

(Fig. 7B) . SEM images of cells on Fib+ constructs also showed tightly opposed polygonal cells, however, in some areas the intracellular margins were less distinct (Fig. 7B) . At higher magnification the numerous microvilli and microplicae were visible and at this level the cell borders did become apparent (Fig. 7C) . In Fib- constructs polygonal epithelial cells on the surface were clearly visible.

Additionally, visible depressions left by detached cells were seen in some areas ( Fig . 7D) . Numerous microvilli were seen on

the cell surface at higher magnification and cell borders were obvious (Fig. 7E) . A numbe of cells positive for stem cell marker P63 alpha were observed in the basal epithelial layer of a multilayered epithelium on a compressed collagen construct using confocal microscopy analysis on stained wholemount samples (figure 9} . This study demonstrates that PC collagen supports the expansion and stratification of human corneal epithelial cells. The resulting cellular constructs display many of the typical characteristics of human corneal epithelium. The process of plastic compression described here is attractive as it produces collagen constructs with superior mechanical properties and also allows cells to be seeded directly into the scaffold { 19] . We have shown that limbal fibroblasts can be rapidly and simply seeded into the collagen gel, that the cells survive the compression process and can be maintained within the scaffold in culture for at least 4 weeks. This process eliminates the delay normally encountered while cells populate a synthetic scaffold. Furthermore the construct remains optically transparent and can withstand manipulation on the surface of the eye, which are important requirements if this material is to be considered as a replacement for amniotic membrane.

The Draize test [20] for ocular toxicology has been heavily criticised for its inhumane treatment of animals as well as the irreproducibility of its subjective scoring method (21] , but no single replacement has yet been validated [22] . Many three-dimensional models of cultured corneal epithelial cells, corneal equivalents and organotypic models have been proposed as an alternative to these current tests for ocular irritancy (reviewed in [23]), Human corneal epithelial cells cultured on PC collagen constructs form an intact epithelial layer, which very closely resembles the human cornea, both in terms of structure and protein expression profiles using markers of corneal epithelial cells and basement membrane proteins. The corneal constructs can be rapidly produced and

replicated suggesting that this system has potential to form the basis of an in vitro ocular toxicology or irritancy test.

The data presented here provide indication that the presence of fibroblasts in the collagen construct influenced the behaviour of the epithelial cells in the overlying layer. Although not significantly different, there was a trend towards a decrease in basal cell density in Fib- constructs, providing indication that the N/C ratio of cells in constructs without fibroblasts was decreased. It has been reported that small cell size and a large N/C ratio are typical characteristics of LESCs [24-26] . In this case, the larger N/C ratio on the Fib+ compared to Fib- constructs is likely to be indicative of poorly differentiated cells rather than LESCs as no expression of the putative LESC marker, P63a, was detected on either construct type. We did detect a number of P63a positive cells in the human central cornea and this is likely to be as a result of post mortem changes that have been described previously (28] . The lack of ?63o expression in collagen constructs without crypts is unsurprising as these constructs mimic an intact central corneal epithelium which has, to date, not

Epithelial layers on the surface of Fib+ constructs appeared deeper than on Fib- constructs and it was observed that cell sheets removed from Fib+ constructs required longer incubation with trypsin to dissociate to a single cell suspension than cells on a Fib- construct .

This provides indication that intercellular junctions may be more abundant in Fib+ than in Fib- constructs as previously reported by others [29] . Similarly, expression levels of basement membrane proteins larainin and collagen IV and its ligand, bl integrin were noticeably higher in Fib+ constructs compared to Fib- constructs. Both perlecan and collagen VII were not detected in either Fib+ or Fib- constructs as seen in central cornea, however, this could be as a result of the airlift culture period of 5 days. The differences in basement membrane protein deposition we have seen between Fib+ and Fib- constructs were also reported in 3-D organotypic cultures using skin keratmocytes either in the absence or presence of dermal fibroblasts. In this case, basement membrane assembly was delayed in the absence of fibroblasts [31] . Taken together, these results provide indication that incorporating fibroblasts in our corneal epithelial constructs increases biomiraicry but also the potential for use of this construct as a model to study epi helial -mesenchymal crosstalk.

The human corneal epithelial constructs produced using PC collagen as a substrate can be easily standardized and replicated to the required specifications. The resulting multilayered epithelial has many characteristics of human central corneal epithelium, including typical epithelial marker and basement membrane marker expression as well as ultrastructural similarities, and so could easily form the basis of an in vitro model of the corneal epithelium. Furthermore, the biomimetic tissue described herein has the potential to offer an attractive treatment of LESC deficiency.

1 . Tseng SCG. Mol Biol Rep 1996;23:47-58,

2. Kenyon KR, Tseng SC. Ophthalmology 1989;96:709-722.

3. Schwab IR, et al. Cornea 2000 Jul ; 19 (4) : 21- 426.

4. Tsai RJ-F, Li L-K, Chen J-K. New Engl J Med 2000;343:86-93.

5. Shortt A et al Biomaterials 2009 Feb 1; 30 (6) : 1056-1065.

6. Grueterich M et al Ophthalmology 2002 Aug; 109 (8) : 1547-1552.

7. Gomes JAP et al Curr Opin Ophthalmol 2005 2005; 16(4) : 233-240.

8. Levi H , et al Curr Opin Biotechnol 2009 Oct; 20 (5) :593-597.

9. Neel EAA et al Soft Matter 2006 Jan 1.

10. Brown RA et al Adv Funct Mater 2005 Jan 1.

11. Shah A et al. Pediatr Res 2008 May 1 ; 63 (5) : 535-544.

12. Liu Y et al Invest Ophthalmol Vis Sci 2006 May 1 ; 4 (5) : 1869-1875.

13. Rafat M, et al Biomaterials 2008 Oct 1;29{29) : 3960-3972.

14. Liu W et al Biomaterials 2009 Mar 1 ; 30 (8) : 1551-1559.

15. Orban JM et al. J Biomed Mater Res Pt A 2004 Mar 15; 68 (4) : 756-762.

16. Vrana E et al J Biomed Mater Res Pt A 2008 Feb 1 ; 84 (2) : 454-463.

17. Neel E et al Soft Matter 2006 Jan 1.

18. Mcintosh Ambrose W, et al J Biomed Mater Res Part B Appl Biomater 2009 Mar 12.

19. Brown R et al Adv Funct Mater 2005 Jan 1.

20. Draize JH, et al J Pharmacol Exp Ther 1944;82:377-389.

21. Xu KP et al . Toxicol Sci 2000 Dec 1 ; 58 { 2 ) : 306-31 .

22. Jester JV et al Toxicol In Vitro 2010 Jan 1 ; 24 (2 ) : 597 60 .

23. Barile FA. J Pharmacol Toxicol Methods 2010 Jan 1; 61 (2) : 136-145.

24. Arpitha P et al Invest Ophthalmol Vis Sci 2005 Oct 1; 46 (10) : 3631-

25. Barrandcn Y et al PNAS USA 1985 Aug 1; 82 (16) : 5390-5394.

26. Romano AC et al Invest Ophthalmol Vis Sci 2003 Dec 1; 44 (12) : 5125- 5129.

27. Majo F et al Nature 2008 Nov 13; 56 (7219) :25C-25 .

28. Di lorio E et al. PNAS USA 2005 Jul 5; 102 (27) : 9523-9528.

29. Ko J-A et al Invest Ophthalmol Vis Sci 2008 Ja 1;49(1) : 113-119.

30. Li W et al Invest Ophthalmol Vis Sci 2006 Jun 1; 7 ( 6) : 2381-2389. 31. Andriani F et al J Invest Dermatol 2003 Jun 1; 120 {6) : 923-931.

Antibody Clone Concentration Supplier β-integriti HBl.l 1:50 Millipore

Collagen IV 1:500 Abeam

Collagen VII LH7.2 1:100 Ch em icon

Cytokeratin 3 AE5 1:500 Millipore

Table 1

Claims

Claims :

1. A method of producing biomimetic corneal tissue comprising:

(i) providing a gel seeded with fibroblasts,

(ii) compressing the gel to produce a compacted gel construct,

(iii) seeding a surface of the construct with ocular epithelial cells, thereby producing biomimetic corneal tissue. 2. A method according to claim 1 further comprising;

Civ) culturing the construct to expand the population of ocular epithelial cells on the surface.

3. A method according to claim 1 or claim 2 whe ein the ocular epithelial cells form a layer on the surface.

4. A method according to claim 3 wherein the ocular epithelial cell layer is confluent . 5. A method according to any one of claims 1 to 4 wherein the compacted construct comprises a single layer of compressed gel .

6. A method according to any one of claims 1 to 4 wherein the compacted construct comprises multiple layers of compressed gel .

7. A method according to claim 6 wherein the compacted construct comprises multiple layers of compressed gel produced by;

(a) providing a gel seeded with fibroblasts,

(b) compressing the gel to produce a compressed gel layer,

(c) introducing an additional gel solution onto the compressed gel layer,

(d) allowing the additional gel solution to set on the compressed gel layer to form an additional gel, (e) compressing the additional gel to produce a construct

comprising multiple layers of compressed gel,

8, A method according to claim 7 wherein steps (c) to (e) are repeated one or more tiroes to produce a construct comprising multiple layers of compressed gel.

9. A method according to claim 6 wherein the compacted construct comprises multiple layers of compressed gel produced by;

(a) providing a gel seeded with fibroblasts,

(b) introducing an additional gel solution onto the

jjiίϋ_{* f}

(c) allowing the additional gel solution to set on the gel to form a multilaye gel ,

(d) optionally repeating steps b) and c) one or more times to introduce one or more additional layers to the multi-layer gel, and;

(e) compressing the multi-layered gel to produce a construct comprising multiple laye s of compressed gel . 10. A method according to claim 9 wherein steps b) and c) are repeated one or more times to produce a construct comprising multiple

II . A method according to ny one of claims 7 to 10 wherein the additional gel solution is seeded with fibroblasts,

12. A method according to any one of the preceding claims comprising embossing one or more crypts into the surface of the compacted gel construct before the surface is seeded with epithelial cells.

13. A method according to claim 12 wherein the gel is compressed a d embossed with one or more crypts simultaneously .

14. A method according to claim 13 wherein the gel is compressed by a rigid body which comprises one or more projections, said projections embossing the one or more crypts into the first surface of the gel. 15. A method according to claim 14 wherein the rigid body is porous, such that liquid expelled from the gel during compression is absorbed into the porous body.

16. A method according to any one of claims 12 to 15 wherein each crypt has width of about 60um to 250 um and a length of 6Gum to 250 um.

17. A method according to any one of claims 12 to 16 wherein each crypt has a depth of 20um to 150um

18. A method according to any one of claims 12 to 17 wherein an array of crypts is embossed into the surface of the construct.

19. A method according to claim 18 wherein the array of crypts is surrounds a central zone of the construct.

20. A method according to any one of the preceding claims wherein the gel is compressed by a porous body, such that liquid expelled f om the gel during compression is absorbed into the porous body.

21. A method according to claim 20 wherein the gel surface which contacts the porous body is the only fluid leaving surface of the gel during compression . 22. A method according to any one of the preceding claims wherein the gel is produced by seeding a gel solution with fibroblasts and setting the seeded gel solution to produce a seeded gel .

23. A method according to claim 22 further comprising adding or introducing the gel solution or the seeded gel solution to a well.

24, A method according to claim 23 wherein the well is adapted to mou

25 A method according to claim 23 or claim 24 wherein the well is adapted to mould tags onto the seeded gel to facilitate handling of the biomimetic corneal tissue.

26. A method according to any one of the preceding claims wherein the average gel density in the seeded gel before compression is 0.5 to 5 mg/ml . 27. A method according to any one of the preceding claims wherein the average cell density in the seeded gel before compression is from 1 x 10³ to 1 x 10⁶ fibroblasts per ml.

28 , A method according to any one of the preceding claims wherein compression expels liquid from a fluid leaving surface { FLS) of the gel and the amount of liquid expelled through the FLS of the gel per unit of the surface area of the FLS (expressed as a drop in height) is 2 mm to 14 mm. 29. A method according to any one of the preceding claims wherein the ocular epithelial cells are produced by separating a population of cells from donor ocular tissue and, optionally expanding the

population of separated cells. 30. A method according to any one of the preceding claims comprising seeding a second surface of the compacted gel construct with

endothelial cells.

31. A method according to claim 30 comprising culturing the construct in culture medium to expand the population of ocular endothelial cells on the second surface. 32. method according to claim 30 or claim 31 wherein the

endothelial cells form an endothelial layer on the second surface.

33. A method according to claim 32 wherein the endothelial layer is confluent .

34. A method according to any one of the preceding claims wherein the gel is provided on top of a layer of endothelial cells.

35. A method according to claim 34 wherein the layer of endothelial cell is contained within a gel or a compressed gel layer.

36. A method according to any one of claims 30 to 35 wherein the endothelial cells are produced by separating a population of

<Πϊ ^ ^ JL <9l €5JL JL S J-J ^IT^ J-T 5 -* ! _* S S J _*

37. A method according to any one of the preceding claims wherein the construct is cultured by immersing the construct in culture medium to produce a layer of epithelial cells. 38. A method according to any one of the preceding claims comprising culturing a construct comprising a layer of epithelial cells on its surface under conditions in which the layer is exposed to air, such that said epithelial cells form a differentiated epithelium. 39. A method according to claim 37 or claim 38 wherein the layer of epithelial cells is confluent.

40. A method according to claim 38 or claim 39 wherein the construct is cultured on a support which is permeable to culture media in order to position the layer of epithelial cells at an air/liquid interface, 41. A method according to any one of the preceding claims wherein the biomimetic corneal tissue comprises an epithelial layer and a stromal layer comprising fibroblasts.

42. A method according to claim 41 wherein the biomimetic corneal tissue further comprises an endothelial layer.

43. A method according to any one of the preceding claims wherein the gel is a collagen gel and the compacted gel construct is a compacted collagen construct.

44. A method according to claim 43 wherein the collagen is type I collage .

45. A method according to claim 43 or claim 44 wherein the collagen gel is produced by seeding a collagen solution with fibroblasts and setting the seeded collagen solution to produce a seeded collagen gel.

46. A method according to claim 45 further comprising adding the collagen solution or the seeded collagen solution to a well.

47. Biomimetic corneal tissue produced by a method according to any one of claims 1 to 46.

48. Biomimetic corneal tissue according to claim 47 for use in a method of treatment of the human or animal body.

49. Biomimetic corneal tissue according to claim 48 for use in a method of treatment of diseased or damaged cornea.

50. Use of biomimetic corneal tissue according to claim 41 in the manufacture of a medicament for use in the treatment of diseased or damaged cornea. 51. A method of treatment of diseased or damaged cornea comprising; implanting into an individual biomimetic corneal tissue according to claim 47.

52. A method of screening comprising:

contacting a test compound with a biomimetic corneal tissue according to claim 41, and

determining the effect of the test compound on the tissue.

53. A method according to claim 52 wherein the effect is determined by determining the viability of cells in the tissue in the presence and absence of the test compound.

54. A method according to claim 52 wherein the effect is determined by determining the functional properties of the tissue in the presence and absence of the test compound.

55. A method according to claim 52 wherein the effect is determined by determining the expression of cellular markers by cells in the tissue .

56. A method according to any of claims 52 to 55 wherein an increase in cell viability or functional properties in the presence relative to the absence of the test compound is indicative that the test compound is useful in ocular therapy .

57. A method according to any of claims 52 to 55 wherein a decrease in cell viability or functional properties in the presence relative to the absence of the test compound is indicative that the test compound is an ocular irritant.