CROSS REFERENCE TO RELATED APPLICATIONS

[0001] The present application claims the benefit of United States Provisional Patent

Application 62/840,677, filed April 30, 2019, the disclosure of which is hereby incorporated by reference herein.

BACKGROUND

[0002] The cornea of the human eye includes a layer of cells referred to as the epithelium at the anterior surface; a thin layer referred to as “Bowman’s layer” immediately posterior of the epithelium, the collagenous comeal stroma constituting the major portion of the thickness of the cornea, a thin layer referred to as Descemet’s membrane posterior to the stroma, and a further layer of cells referred to as the endothelium forming the posterior surface of the cornea.

[0003] Corneal transplant surgery is often the only option for patients with a number of comeal maladies. Full thickness transplants (penetrating keratoplasty, or PK) is common for people with aggressive, drug resistant infections and very advanced keratoconus. Partial thickness procedures such as Descemet’s stripping automated keratoplasty (DSAEK) or Descemet’s and endothelial keratoplasty (DMEK) are commonly performed on patients with Fuchs' dystrophy and bullous keratopathy where the goal is to replace the patient’s endothelial layer. In many of these cases it is desirable to alter the mechanical properties of the graft tissue from the as-harvested state to improve the handling properties for the surgeon.

[0004] Corneal crosslinking (CXL) is a common procedure used to strengthen the comeal tissue in vivo for patients with progressive keratoconus. The stiffening provided by crosslinking can also be applied in vitro to harvested graft tissue to improve the handling properties. In vitro crosslinking procedures face a different set of challenges than in vivo crosslinking.

[0005] In vivo, instillation of the necessary riboflavin into the corneal stroma across the epithelial barrier is challenging. In one procedure commonly referred to as the “Dresden procedure” or “epi-off” CXL, the epithelium is removed prior to riboflavin instillation. Removing the epithelium is both painful and risky, as it increases the chance of infection. Trans- epithelial CXL (or ‘epi-on’ CXL), in which the epithelial barrier is disabled by treating the epithelium with a drug has been employed. No matter how the procedure is performed (epi-on or off), the next challenge is to crosslink as much of the stroma thickness as possible. The deeper the crosslinking, the greater the increase in the tissue stiffness, and the more efficacious the CXL procedure is at preventing the progression of keratoconus. However, this must be done without damaging the endothelial cells.

[0006] The depth of crosslinking depends on many factors: the wavelength and intensity of the applied light, the total dose of light energy applied, the thickness off the stroma and the concentration profile of the riboflavin within it, and the stromal oxygen content. Many of these variables are either hard to control or hard to know accurately or both. The endothelial cells are very sensitive to the applied crosslinking light energy in the presence of riboflavin. The established UV dose limit for the endothelium is 0.65J/cm ² . To be safe, standard in vivo CXL procedures err on the side of caution and intentionally do not attempt full thickness crosslinking. This conservative approach, combined with the unknown variables, means that crosslinking to a precise depth in vivo is very difficult.

[0007] However, for DMEK and DSAEK procedures, only the very posterior most regions of the cornea (just Descemet’s membrane and the single cell layer endothelium in the case of DMEK), are used, therefore, in order to improve the handling characteristics of these samples, it becomes important to crosslink the very deep layers of the corneal stroma, i.e., the layers which are close to Descemet’s membrane and the endothelium. To assess the difficulty of crosslinking the deep stroma without damaging the endothelium using the standard in vivo CXL procedures, it is valuable to evaluate the factors that control riboflavin mediated UV corneal crosslinking with UV.

[0008] Fick’s Second Law governs the rate at which the riboflavin concentration changes with time across the tissue thickness:

[0009] Where z is depth in tissue from the instillation surface (the anterior surface in the standard in vivo CXL procedure), t is the time after contacting this surface with the riboflavin solution and D is the diffusion coefficient. If a sufficiently large bolus of riboflavin solution is applied at t=0 so that the concentration can be assumed to be constant in this bolus, and the concentration at z= ¥ is zero, then the solution to this equation is the error function in the form:

[0010] Where the erfc is the error function and erfc(x) is given by:

[0011] The value of the stromal riboflavin diffusion coefficient D is known from literature (6.5xl0 ⁵ mm ² /s). To assess the dose of UV deposited in the endothelium by light applied from the front of the cornea, the intensity of the light as a function of depth in the cornea must be calculated. Ignoring scattering in the eye (which may or may not be valid for 375nm UV) Lambert-Beer law gives:

/ = 7 ₀ e ^{_ tz}

[0012] Where m is the linear attenuation coefficient, which is the sum of the attenuation in the corneal tissue and the absorption in the riboflavin. The attenuation coefficient of the riboflavin is concentration dependent, with the axial and temporal distribution given above as c(z,t).

[0013] To find the UV dose deposited in the endothelium (dose = intensity x time) during a period of UV irradiation, it would seem a reasonably straight forward exercise to integrate the intensity over time and over the thickness in z of the endothelium. However, there are some additional complexities that make this challenging. First, the diffusion coefficient D that was assumed to be a constant is not. As the stomal tissue crosslinks, it increases in density; this density increase reduces the diffusion coefficient. So, because the amount of crosslinking is a function of the deposited UV dose in riboflavin containing tissue, D becomes a function of the deposited dose, which is a function of time and position z. Additionally, riboflavin, when exposed to UV, undergoes photobleaching, where some fraction of the riboflavin molecules break down into non-photoreactive species. The photobleaching rate is a function of the UV intensity and the concentration of the riboflavin, thus also dependent on time and axial position. [0014] The assumptions made to this point, coupled with the inevitable sample-to-sample variations in the critical coefficients make titrating the UV dose to align the posterior extent of the crosslinked tissue at Descemet’s membrane extremely difficult using the standard in vivo CXL procedure. Moreover, in the standard in vivo CXL procedure, the crosslinking typically is strongest in the anterior portion of the corneal stroma, with little or no crosslinking in the posterior portion of the stroma, near Descemet’s membrane.

SUMMARY

[0015] One aspect of the invention provides methods of crosslinking a cornea in vitro. A method according to this aspect of the invention desirably includes contacting the posterior surface of the cornea with a liquid containing a crosslinking catalyst so as to instill the crosslinking catalyst into the cornea from the posterior surface. The method desirably further includes applying light to the cornea through the anterior surface of the cornea to cause crosslinking of the stroma. The light- applying step desirably commences after commencement of the contacting step. The contacting step may be continued during the light-applying step, or may be terminated before the light- applying step. The cornea treated in the process desirably includes an intact endothelium, and the catalyst is instilled into the stroma of the cornea through the endothelium and through Descemet’s membrane. Most typically, the crosslinking catalyst is riboflavin, and the light applied is ultraviolet (“UV”) light.

[0016] A further aspect of the present invention provides apparatus for corneal crosslinking. Apparatus according to this aspect of the invention desirably includes means for contacting the posterior surface of the cornea with a liquid containing a crosslinking catalyst. The apparatus desirably also includes means for applying light to the cornea through the anterior surface of the cornea.

BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a diagrammatic sectional view depicting apparatus according to one embodiment of the invention in conjunction with a cornea.

[0018] Figure 2 is a fragmentary, diagrammatic sectional view on an enlarged scale of the cornea depicted in Figure 1.

[0019] Figure 3 is a graph schematically depicting catalyst distributions at different times during a method according to an embodiment of the invention.

DET AIDED DESCRIPTION

[0020] Apparatus according to one embodiment of the invention (Fig. 1) includes a sample holder 10. The sample holder includes a ring-shaped carrier 12 having an opening 13. Carrier 12 has a set of ring-shaped corrugations or teeth 14 on an upper surface. A handle 16, partially depicted in Fig. 1, projects from carrier 12. Sample holder 10 further includes a ring- shaped upper element 18 having an opening 20. The carrier 12 and upper element 18 are adapted to grip a tissue sample between them, with the weight of upper element 18 bearing on the tissue sample and holding the sample in engagement with the upper element and with the ring, and with the openings 13 and 20 aligned with one another. In this condition, the tissue sample extends across the opening 20 in the upper element and engages the lower surface of upper element 20. As depicted in Fig. 1, the tissue sample includes an intact cornea 30 surrounded by a ring 32 of scleral tissue which is used for handling the sample but which will not form a part of the transplant. The scleral tissue ring 32 is engaged between the elements of the sample holder, whereas the cornea 30 is aligned with the openings so that the cornea projects downwardly through opening 13 of carrier 12. The elements of the sample holder do not touch the cornea. As depicted, the posterior surface 34 of the cornea faces upwardly towards upper element 18, and is aligned with opening 20 in upper element 18. The apparatus further includes a source 36 of a liquid containing a cross-linking catalyst such as riboflavin. Source 36 is arranged to introduce the liquid into the opening 20 of the upper element, so that the liquid contacts the tissue sample. This liquid contacts the posterior surface 32 of the cornea, but desirably does not contact the opposite, anterior surface of the cornea. The scleral tissue may form a reasonably liquid-tight seal with the upper element 20. The upper element may include a soft gasket (not shown) on its lower surface so as to improve the seal.

[0021] The apparatus further includes a light- applying device 40. The light-applying device includes a light source 42. The source may include a lamp or laser adapted to emit light such as UV light at a wavelength which will interact with the catalyst to cause cross-linking of comeal stroma. In this embodiment, the light-applying device includes a fiber optic 44 connected to the light source and a collimating lens 46 aligned with the fiber optic, as well as a light scattering element 48 commonly referred to as a diffuser. The diffuser 48, lens 46 and the end of the fiber optic 44 adjacent the lens are aligned with the opening 13 of carrier 12. For example, the handle 16 of carrier 12 may be connected to a frame (not shown) which is also connected to the diffuser, lens and fiber optic. In operation, light emitted by the source spreads from the end of the fiber optic and is collimated by the lens and redirected through the diffuser. The light emanating from the diffuser desirably has substantially uniform intensity. [0022] The apparatus in this embodiment further includes a reservoir 50 having an open top and having a bottom wall 52 formed from a material which is transparent or translucent to the light from source 42. For example, the bottom wall 52 may be formed from borosilicate glass which is transparent to UV light at about 375 nm wavelength. Reservoir 50 may be mounted to the same frame (not shown) as the other elements. The reservoir, sample holder 10 or both may be slidably mounted on the frame (not shown) so the positions of these elements may be adjusted during use.

[0023] In a method according to one embodiment of the invention, a sample including an intact whole cornea 30 and scleral ring 32 is engaged in sample holder 10 as discussed above. A liquid 38 including a catalyst such as riboflavin at the desired concentration is added to the concave posterior well formed by the cornea 30. Care must be taken to not touch the endothelium at the posterior surface 34 of the cornea. Thus, liquid 38 is contacted with the posterior surface of the cornea, without contacting the anterior surface 35. Another fluid 56 such as balanced saline, cornea preservation medium, or perfluorocarbon is placed into reservoir 50. Fluid 56 typically contains dissolved oxygen, but desirably does not contain the catalyst so that fluid 56 is transparent to the UV light which will be applied by source 42. Sample holder 10 may be lowered or reservoir 50 may be raised to bring the anterior surface 52 of the cornea into contact with the fluid 56. During this process, some of the fluid 56 may be displaced from the reservoir. The top edge of the reservoir may be provided with channels (not shown) to allow fluid 56 to spill even if carrier ring 12 contacts the top edge. A catch basin (not shown) may be provided around reservoir 50 to catch spilled fluid.

[0024] At this stage of the method, the posterior surface 34 of cornea 30 is in contact with the catalyst in liquid 38, whereas the anterior surface 35 of the cornea is in contact with the fluid 56. A small section of cornea 30 in this condition is schematically depicted in Fig. 2. Unlike the epithelium 60 at the anterior surface, the endothelium 62 at the posterior surface does not have tight junctions between the cells. Thus, a catalyst such as riboflavin can be instilled readily through the endothelium on the posterior surface of the graft tissue. This configuration ensures that the posterior part of the stroma will have a higher riboflavin concentration than the anterior portion. Figure 3 schematically depicts the distribution of catalyst in the cornea at different times after the beginning of contact with the catalyst-bearing liquid. The vertical axis denotes position within the cornea, with the various layers of the cornea labelled with the same reference numerals shown in Figure 2.

[0025] Diffusion of the catalyst into the cornea from the posterior surface is governed by

Fick’s Laws. The catalyst concentration at a time shortly after the beginning of contact is shown by solid line 66. At this time, the catalyst concentration in Descemet’s layer 63 and in the adjacent portion of the corneal stroma 64 is approximately equal to the catalyst concentration in liquid 38. As also shown by solid line 66, the catalyst concentration declines to nearly zero at a location within the stroma 64. The catalyst concentration at a later time is schematically shown by curve 68, and the catalyst distribution at a still later time is shown by curve 70. At these later times, the catalyst concentration at and immediately adjacent the posterior surface 34 remains at or near the catalyst concentration in liquid 38, and declines to nearly zero at a location within the stroma. However, this location moves in the anterior direction as time passes. Thus, a catalyst field is diffusing from the posterior surface toward the anterior surface.

[0026] Desirably, the catalyst-bearing liquid 38 remains in contact with the posterior surface of the cornea for a time referred to as the “soak time” before the light-applying device 40 (Fig. 1) is actuated to apply light to the cornea, and the catalyst bearing liquid remains in contact with the posterior surface during light application. The soak time is selected to that by the time light application begins, the catalyst field in the stroma and Descemet’s layer contains sufficient catalyst to substantially absorb the light, and thus protect the endothelium 62 from the light. The depth of penetration of the catalyst from the posterior toward the anterior can be calculated by Fick’s second law.

[0027] The light- applying device 40 directs light from the source 42 into the anterior surface 35 of the cornea. Because the crosslinking light energy is applied through the anterior side of the cornea, no crosslinking occurs until the light reaches the front of the catalyst field. Only the constant stromal UV attenuation coefficient affects the incident intensity on the catalyst front within the tissue. This can be easily accounted for. By instilling the riboflavin from the posterior side of the epithelium, the thickness of the cornea, which can vary from graft to graft as much as 100 microns or more, is not a factor in the depth of the crosslinked section of the tissue. Additionally, as the diffusion coefficient decreases in the crosslinked tissue beginning at the anterior extent of the diffusing riboflavin front, anterior movement of the catalyst is impeded, helping to keep the concentration of the catalyst high in the posterior region, which is needed to protect the endothelium. The proper values for variables such as concentration of riboflavin in the endothelial bolus, the soak time prior to light application, the light intensity and duration can be selected by experimentation and modeling; however the posterior application of riboflavin greatly simplifies the problem of how to crosslink the deep stroma and Descemet’s without damaging the endothelium.

[0028] In the procedure discussed above, instillation of the riboflavin through the posterior surface of the cornea, i.e., through the endothelium, provides a concentration profile with the maximum riboflavin concentration at and near Descemet’s membrane and the deepest layers of the comeal stroma, and with minimal or zero riboflavin concentration in the anterior layers of the stroma. This favors crosslinking of the deepest layers and Descemet’s membrane, while providing enough attenuation of the applied ultraviolet to keep the dose to the endothelium below a safe level. Maintaining the riboflavin-bearing liquid in contact with the endothelium during the crosslinking procedure can assure the Descemet’s layer and the deepest layers of the stroma remain at a riboflavin concentration at or near the concentration in the riboflavin-bearing liquid during UV application, despite diffusion of riboflavin during the crosslinking procedure. [0029] The procedures and structures discussed above can be varied. For example, the details of the sample holder and the device used to apply ultraviolet light can be modified. For example, the apparatus depicted in Figure 1 directs light from a single light source onto a single cornea. However, light from a single source could also be scattered and diffused over a larger area, to impinge on multiple corneas. In the sample holder 10 depicted in Figure 1, the weight of the upper element 18 serves to clamp the scleral ring of the sample against carrier 12. However, the sample holder may include other force-applying elements such as springs for urging the upper element and carrier ring together.

[0030] In a further modification, the epithelium may be removed or the epithelial barrier to riboflavin diffusion may be disabled by application of a drug. In this variant, the fluid 56 (Fig. 1) in contact with the anterior surface of the cornea will tend to extract riboflavin from the anterior surface. As disclosed in commonly owned United States Provisional Patent Application 62/825,388, filed March 28, 2019, the disclosure of which incorporated by reference herein, extraction of riboflavin through the anterior surface helps to reduce riboflavin concentration in the anterior layers of the cornea. The fluid in contact with the anterior surface may be continually replaced by fresh fluid to keep the concentration of catalyst in this fluid at or near zero, and thus minimize attenuation of the light by the fluid. In a further variant, the container 50 and the fluid 56 may be omitted, and the anterior surface of the cornea may be exposed to the atmosphere or to another gas containing oxygen so as to maintain oxygenation of the cornea during light application.

[0031] Riboflavin is the most commonly used crosslinking catalyst. Where riboflavin is used as the crosslinking catalyst, the light applied typically is UV at about 375nm wavelength. The term “riboflavin” as used herein should be understood as including riboflavin 5’ phosphate sodium salt and another pharmaceutically acceptable forms of riboflavin. However, the present invention may be employed with crosslinking catalysts other than riboflavin. Also, light other than UV light can be used.

[0032] In a further variant, antioxidants such as those commonly used in corneal preservation fluid can be applied to the epithelium to provide additional protection for the endothelium from oxidative damage. This raises the endothelial cells’ ability to survive exposure to UV light. The antioxidant can be incorporated in the catalyst-bearing liquid 38 (Fig.l) which is applied to the posterior surface of the cornea. Alternatively or additionally, the catalyst-bearing liquid can be removed and replaced by corneal preservation fluid after instillation of the catalyst, such as after the “soak time” discussed above, but before or during application of the crosslinking light. The replacement fluid may also contain the catalyst, at the same concentration used for instillation, or at a different concentration.

[0033] The apparatus and methods discussed above can be used to prepare corneas for use in procedures such as Descemet’s stripping automated keratoplasty (DSAEK) or Descemet’s and endothelial keratoplasty (DMEK), where the epithelium, Descemet’s layer or both will be separated from the stroma of the sample and transplanted into the patient’s eye. However, the apparatus and methods discussed above also can be used to prepare corneas for other procedures in which part or all of the stroma from the sample will be implanted.