Biomechanical Strengthening of the Human Cornea Induced by Nanoplatform-Based Transepithelial Riboflavin/UV-A Corneal Cross-Linking

Abstract

Purpose: The purpose of this study was to investigate the biomechanical stiffening effect induced by nanoplatform-based transepithelial riboflavin/UV-A cross-linking protocol using atomic force microscopy (AFM).

Methods: Twelve eye bank donor human sclerocorneal tissues were investigated using a commercial atomic force microscope operated in force spectroscopy mode. Four specimens underwent transepithelial corneal cross-linking using a hypotonic solution of 0.1% riboflavin with biodegradable polymeric nanoparticles of 2-hydroxypropyl-β-cyclodextrin plus enhancers (trometamol and ethylenediaminetetraacetic acid) and UV-A irradiation with a 10 mW/cm² device for 9 minutes. After treatment, the corneal epithelium was removed using the Amoils brush, and the Young's modulus of the most anterior stroma was quantified as a function of scan rate by AFM. The results were compared with those collected from four specimens that underwent conventional riboflavin/UV-A corneal cross-linking and four untreated specimens.

Results: The average Young's modulus of the most anterior stroma after the nanoplatform-based transepithelial and conventional riboflavin/UV-A corneal cross-linking treatments was 2.5 times (P < 0.001) and 1.7 times (P < 0.001) greater than untreated controls respectively. The anterior stromal stiffness was significantly different between the two corneal cross-linking procedures (P < 0.001). The indentation depth decreased after corneal cross-linking treatments, ranging from an average of 2.4 ± 0.3 μm in untreated samples to an average of 1.2 ± 0.1 μm and 1.8 ± 0.1 μm after nanoplatform-based transepithelial and conventional cross-linking, respectively.

Conclusions: The present nanotechnology-based transepithelial riboflavin/UV-A corneal cross-linking was effective to improve the biomechanical strength of the most anterior stroma of the human cornea.

In the last decade, riboflavin/UV-A corneal cross-linking has been established as a method to slow down or halt the progression of keratoconus. The procedure consists of irradiating the cornea soaked with riboflavin, which acts as photosensitizer, with UV-A light in order to generate additional cross-linking bonds between stromal proteins that improve the corneal tissue mechanical stability and resistance to deformation.^1–5 It is well established that the adequate permeation of the corneal stroma with riboflavin before the UV-A exposure is the main condition for achieving effective outcomes of corneal cross-linking in patients.⁶ 

The main drawback of conventional corneal cross-linking treatment, which is also associated with the primary postoperative complications of the procedure, is the epithelial removal.^7,8 The understanding that dextran (or other high-molecular-weight molecules with high viscosity) limits the penetration of riboflavin in the stroma through the intact epithelium has driven the development of dextran-free, hypotonic, riboflavin solutions with enhancers (e.g., trometamol, ethylenediaminetetraacetic acid [EDTA], benzalkonium chloride, etc.).^9–11 Former laboratory and clinical outcomes^12–14 regarding novel strategies to deliver dextran-free riboflavin solutions for transepithelial corneal cross-linking are encouraging, though the conventional corneal cross-linking approach remains the gold standard for treating patients with progressive keratoconus. 

In previous studies,^15–17 we have shown the reliability of atomic force microscopy (AFM) to probe the biomechanical properties of human corneal tissues and to evaluate the changes induced by riboflavin/UV-A corneal cross-linking on the elastic and viscoelastic properties of the anterior stroma at the submicroscopic level. The results were consistent with the notion that riboflavin/UV-A corneal cross-linking induces the formation of additional cross-linking bonds between stromal proteoglycan core proteins and collagen, as postulated by other authors.^18,19 

In the present study, we evaluated the modulus of elasticity of the anterior stroma of human donor corneal tissues after nanoplatform-based transepithelial riboflavin/UV-A corneal cross-linking protocol using AFM. The results were compared with conventional riboflavin/UV-A corneal cross-linking. 

Materials and Methods

Corneal Tissues

Twelve human donor sclerocorneal tissues from different donors, unsuitable for transplantation, were obtained from the Veneto Eye Bank Foundation (Venezia Zelarino, Italy). Inclusion criteria included an endothelial cell density (ECD) ≥1800 cells/mm²; exclusion criteria included any corneal or ocular surgery. All samples were used in accordance with the Declaration of Helsinki regarding research involving the use of human tissues, and the experimental protocol was approved by the National Research Council research ethics and bioethics advisory committee. The sclerocorneal tissues were explanted between 6 hours and 20 hours after death and immediately preserved at 4°C in dextran-enriched storage medium. All the tissues were cultivated at 30°C in corneal storage medium enriched with HEPES (25 mL/L), sodium pyruvate (10 mL/L), L-glutamine (10 mL/L), newborn calf serum (20 mL/L), antibiotics/antimicotics (1:100) for 15 to 20 days prior to shipment to the laboratory in 20% dextran-enriched storage medium. The tissues were used for experiments within 48 hours of shipment. 

Corneal Cross-Linking Protocols

The corneal specimens were divided in three groups; the study group (n = 4) underwent transepithelial corneal cross-linking using a novel 0.1% riboflavin solution (Nanotech; Sooft Italia SpA, Montegiorgio, Italia) and UV-A irradiation of the cornea with a 10 mW/cm² device for 9 minutes (370 ± 8 nm; Vega, CSO, Scandicci, Italy), according to the manufacturer's instructions. Each corneoscleral tissue was placed in an artificial anterior chamber (AAC) (Coronet; Network Medical Products, Ltd., Ripon, UK), pressurized with the AAC filled with 0.9% sodium chloride using a 5-mm syringe. The hypotonic solution consisted of 0.1% (phosphate-free) riboflavin that was kept in solution by 2-hydroxypropyl-β-cyclodextrin nanoparticles; EDTA and trometamol were added as enhancers (Table 1). The solution was instilled over the corneal epithelial surface every 20 to 30 seconds for 20 minutes using a silicone ring. Before UV-A irradiation, the epithelium was gently washed with balanced physiological solution. An irradiation area of 8.0-mm diameter was used in all cases; no riboflavin drop was administered over the epithelial surface during UV-A irradiation. Immediately after treatment, the epithelium was removed using an Amoils brush (Innovative Excimer Solutions, Inc., Toronto, ON, Canada), and the corneal specimens were kept in 20% dextran solution overnight. 

Table 1

 

Nanotechnology-Based 0.1% Riboflavin Solution Used for Transepithelial Corneal Cross-Linking

 

A positive control group (n = 4) underwent conventional corneal cross-linking; each corneal tissue was deepithelized using the Amoils brush and then was immersed in 20% dextran-enriched 0.1% riboflavin solution (Ricrolin; Sooft Italia SpA) for 30 minutes. Before UV-A irradiation, the specimens were gently washed with balanced physiological solution in order to remove the excess of riboflavin. Specimens were irradiated with a 3 mW/cm² UV-A device for 30 minutes (370 ± 8 nm; Vega, CSO). An irradiation area of 8.0-mm diameter was used in all cases; no riboflavin drop was administered over the stromal surface during UV-A irradiation. After treatment, the corneal specimens were kept in 20% dextran solution overnight. The entire procedure is schematized in Figure 1. Untreated sclerocorneal tissues were used as negative controls (n = 4). Data from control specimens were collected in previous studies.^15,17 

Figure 1

 

(A, B) Specimens in the study group underwent transepithelial corneal soaking with hypotonic dextran-free 0.1% riboflavin with 2-hydroxypropyl-β-cyclodextrin nanoparticles and enhancers for 20 minutes using a silicone ring. (C, D) After deepithelialization, specimens in the positive control group were immersed in 20% dextran-enriched 0.1% riboflavin solution for 30 minutes. (E) After corneal soaking, the study and control specimens were irradiated using a UV-A device with 10 mW/cm² for 9 minutes or with 3 mW/cm² for 30 minutes, respectively.

AFM Data Acquisition

Specimens were trephined to 9-mm diameter using a Barron donor punch (Coronet; Network Medical Products, Ltd.). Each corneal tissue was gently placed on a specially designed Teflon environmental cell with the endothelial side facing downward and kept in place without the use of glue.^15–17 The mechanical properties of the anterior corneal stroma were measured using an AFM (Multimode 8 Atomic Force Microscope with Nanoscope V controller; Bruker Italy Srl, Milano, Italy) in the force spectroscopy mode according to standardized protocol used in previous studies.^15,17 Measurements were performed at 27°C with the specimens immersed in 20% dextran solution, using phosphorus-doped rectangular silicon cantilevers of nominal elastic constant between 20 and 80 N/m (TESPA; Bruker Italia Srl). Dextran was used as imaging medium in order to keep the tissues at the same osmolarity as overnight storage. The nominal value of the tip radius of curvature was 10 nm. Force curves were obtained at three different locations at the center of the anterior stroma of each sample, and hundreds of force curves (30 curves for each rate) were recorded at four different approach speeds at each location: 1.7 μm/s, 3.5 μm/s, 8.2 μm/s, and 12.3 μm/s. No adhesion between the tip and the sample surface was detected in any sample. We made sure to avoid any possible alteration of the tissue during measurement; any rupture in the material would have induced a discontinuity in the force curve profile and therefore would have been clearly identified. 

AFM Data Analysis

The force curves were analyzed using the Hertz-Sneddon model for a conical indenter, as previously described.^{15–17,20,21} The optical sensitivity and spring constant of each cantilever were determined for accurate measurement of the elastic modulus. Optical sensitivity was measured as the slope of the force curve taken in 20% dextran solution when the tip was in contact with a rigid surface such as a mica sheet. The optical sensitivity was used to convert cantilever deflection in volts to deflection in nanometers. 

A purpose-developed postprocessing algorithm was developed using MatLab (ver. 2013; The Mathworks, Inc., Natick, MA, USA). The software automatically detected the contact point and extracted the Young's moduli values using the Sneddon model.^16,17 

This model relates the loading force to indentation, which for a conical indenter is  

  ,

where F is the loading force in Newtons (N), ν is Poisson's ratio (0.49 for the corneal tissue), δ is the indentation depth, E is the Young's modulus in Pascals, F₀ and δ₀ are the loading force at baseline and the indentation depth at the contact point, respectively. The fit was applied to the approach curve.  

Microscopy Imaging of the Corneal Epithelium

Two-photon emission fluorescence (TPEF) images of the corneal tissues were collected in order to assess the structural integrity of the epithelium before and after transepithelial corneal cross-linking. The two-photon microscopy setup used in this study was based on an upright microscope (Leica DM6000CS; Leica Microsystems GmbH, Wetzlar, Germany) and was described previously.^16,22 Each tissue was placed upward on a quartz microscope slide under the microscope and illuminated by the laser tuned to 810 nm. Images were collected before treatment and both after corneal soaking and UV-A irradiation of the cornea. The laser power was 15 mW before entering the water immersion objective. The reflected TPEF light was first filtered by a short-pass infrared filter (λ < 680 nm, SP680: Leica Microsystems GmbH) and then passed through a 525/50 band-pass filter (Semrock, Inc., Rochester, NY, USA) entering a nondescan detector (NDD) for reflected light. The NDD sensitivity settings for collecting the TPEF light were the same for all tissue specimens. The second harmonic generation (SHG) signal emitted by the stromal collagen was simultaneously collected in forward direction by a NDD in the transmission path of the microscope. The incoming light was filtered by the short-pass filter (SP680) and then by a 10-nm full-width at half maximum band-pass filter centered at 405 nm (FF01-405/10-25; Semrock, Inc.). The SHG signal was used to detect the anterior and posterior stromal interfaces and hence to align the TPEF intensity profile with stromal thickness. 

Statistics

Data were given as mean ± standard deviation. Each value of E refers to an average on three different neighboring areas of each corneal tissue. The Student's t-test for paired data was used to statistically compare E between groups. Difference with a P value of 0.05 or less were considered statistically significant. A commercial software program (KyPlot; KyensLab, Inc., Tokyo, Japan) was used for statistical testing. Sample size was calculated to determine a mean difference of 1.2 megapascals (MPa) (±0.6 MPa) between different groups at a statistical significance of 5% and a power of 80%. 

Results

In the study group, donors' age ranged between 69 and 72 years old, the central corneal thickness (CCT) ranged between 510 μm and 560 μm, and the ECD ranged between 2000 cells/mm² and 2400 cells/mm². In control groups, donors' age ranged between 68 and 70 years old, CCT between 501 μm and 547 μm, and ECD between 1900 cells/mm² and 2600 cells/mm². Immediately after transepithelial and conventional corneal cross-linking treatments, the average central stromal thickness was 561 ± 12 μm and 518 ± 20 μm, respectively. The central stromal thickness was 547 ± 12 μm in the untreated control group. 

In untreated samples, the average Young's modulus of the anterior stroma ranged between 1.32 ± 0.17 MPa and 2.92 ± 0.81 MPa from 1.7 μm/s to 12.3 μm/s scan rates, respectively. After transepithelial corneal cross-linking, the average Young's modulus ranged from 4.10 ± 1.15 MPa to 5.48 ± 1.65 MPa between 1.7 μm/s and 12.3 μm/s scan rates, respectively. After conventional riboflavin/UV-A corneal cross-linking, the average Young's modulus ranged between 2.76 ± 0.98 MPa and 3.84 ± 1.05 MPa, respectively. The Young's modulus increased with increasing application rates, as expected for viscoelastic materials; the indentation depth after corneal cross-linking treatments was lower than in untreated samples, ranging from an average of 2.4 ± 0.3 μm in untreated samples to an average of 1.2 ± 0.1 μm and 1.8 ± 0.1 μm after transepithelial and conventional cross-linking respectively. The results collected from all specimens are summarized in Table 2. 

Table 2

 

Young's Modulus (M ± SD) of the Most Anterior Stroma in Each Corneal Specimen as a Function of AFM Tip Scan Rate

 

After conventional and transepithelial riboflavin/UV-A corneal cross-linking protocols, the Young's modulus of the most anterior stroma increased on average by 1.7 (P < 0.001) and 2.5 (P < 0.001) times in comparison with untreated corneal specimens, respectively (Fig. 2). After nanoplatform-based transepithelial protocol, the Young's modulus of the anterior stroma was on average 46% greater than specimens treated by conventional corneal cross-linking (P < 0.001). 

Figure 2

 

Average values of the Young's modulus of the most anterior stroma as a function of scan rate in untreated corneas (gray squares), corneas treated by conventional riboflavin/UV-A corneal cross-linking protocol (white circles), and corneas treated by nanotechnology-based transepithelial corneal cross-linking protocol (black triangles). Bars indicate standard deviation.

No damage to the corneal epithelium was observed in any specimen during transepithelial corneal cross-linking (Fig. 3). 

Figure 3

 

Two-photon emission fluorescence images of the basal epithelial layer before (left panel) and after (right panel) transepithelial corneal cross-linking using the nanoplatform-based riboflavin solution. The protocol did not induce any damage to epithelial cells, which retained normal morphology and arrangement after transepithelial soaking and UV-A irradiation of the cornea. Scale bar: 25 μm.

Discussion

In this work, we investigated the modulus of elasticity of the most anterior stroma after performing a nanoplatform-based transepithelial corneal cross-linking protocol on human eye bank donor corneas. The protocol consisted of using a 0.1% riboflavin-2-hydroxypropyl-β-cyclodextrin hypotonic solution with enhancers and a 10 mW/cm² UV-A device. The biodegradable polymeric 2-hydroxypropyl-β-cyclodextrin nanoparticles act as carrier molecules to enhance the solubility and bioavailability of active molecules, such as riboflavin.^23–26 Because β-cyclodextrins are hydrophobic inside and hydrophilic outside, they form complexes with riboflavin and facilitate its diffusion through the epithelium. In addition, EDTA and trometamol were added to the solution as enhancers in order to loosen the epithelial tight junctions and further improve the penetration of riboflavin in the stroma through the intact epithelium.⁶ The results were compared with those of control samples, which were treated using the conventional corneal cross-linking protocol or were left untreated. Atomic force microscopy was used to quantitatively assess the biomechanical strengthening effect induced by corneal cross-linking in the anterior corneal stroma at the level of molecular interactions.^17,27–29 

After nanoplatform-based transepithelial and conventional riboflavin/UV-A corneal cross-linking treatments, the average Young's modulus of the most anterior stroma was 2.5 times and 1.7 times greater in comparison with untreated controls, respectively. The results of conventional riboflavin/UV-A corneal cross-linking were in accordance with previous work,¹⁷ in which the elastic modulus of the anterior stroma increased on average by 1.3 times with respect to baseline values. 

After nanoplatform-based transepithelial corneal cross-linking, the Young's modulus of the most anterior stroma was on average 1.4 times greater than the conventional cross-linking treatment. This result provides the evidence that the nanoplatform was able to enrich the anterior stroma with riboflavin, which in turn induced effective formation of additional cross-linking bonds between stromal proteins after UV-A irradiation with 10 mW/cm² for 9 minutes. In a preliminary study³⁰ assessing the concentration of riboflavin in the stroma using two-photon optical microscopy, the present nanotechnology-based drug delivery platform was able to achieve a peak concentration of riboflavin in the most anterior 50 μm stroma of human donor corneal tissues higher than standard stromal soaking with 20% dextran-enriched solution (0.038% vs. 0.025%; Supplementary Fig. S1). However, the concentration of riboflavin dropped to almost 50% of that achieved by standard stromal soaking in the middle and posterior stroma. The differences of concentration between treatments in the most anterior stroma were likely related to the washing procedure with balanced physiological solution of the corneal surface (which is the stroma or epithelium in conventional and transepithelial protocol, respectively) before UV-A irradiation. 

In this study, the biomechanical measurement of the corneal stroma was limited to the most 3 μm anterior depth (i.e., limited to the most anterior collagen bundles under the Bowman's layer); therefore, it does not provide a full representation of the effect of the nanoplatform-based transepithelial protocol to strengthen the biomechanical properties of the whole corneal stroma. In order to assess the depth-dependent stiffening effect of emerging corneal cross-linking procedure in comparison with the conventional approach, further measurements should be performed across depth, as shown previously.^16,28,29 

The samples in the study and control groups were treated by using different UV-A irradiation protocols, however, delivering the same energy dose to the cornea (5.4 J/cm²) in accordance with the Bunsen-Roscoe law of reciprocity.^6,31,32 The presence of epithelium filtered 20% of the total UV-A energy; therefore, the stiffening effect on the anterior stroma of study group's samples could be greater than that found in the present study if the irradiation protocol for transepithelial corneal cross-linking was enhanced according to this notion.⁶ The novel procedure was performed according to the instructions of the manufacturer; therefore, it was not the aim of this study to compensate for the UV-A filtering effect of the epithelium. 

Bottos et al.³³ have previously demonstrated that 0.5% riboflavin-5-phosphate nanoemulsion could penetrate efficiently in the stroma of rabbit eyes through the intact epithelium, however, after 2 hours of soaking time. Although the results were comparable to standard stromal soaking, the long soaking time makes the methodology unsuitable for clinical use; in addition, the authors did not provide any information on the stromal stiffening induced by UV-A irradiation of the rabbit cornea. In the present work, the transepithelial soaking time with 0.1% riboflavin-2-hydroxypropyl-β-cyclodextrin nanoparticles was 20 minutes and was followed by UV-A irradiation of the cornea in order to collect information that may be clinically relevant. 

The corneal storage protocol was effective to provide tissues with intact epithelium,^34,35 and no damage to the epithelial cells was found after transepithelial corneal cross-linking. In order to avoid any bias due to changes in the hydration state during the experiment, the corneal tissues were kept in 20% dextran solution overnight before AFM testing.^{15–17,36–39} Based on the studies by Hamaoui et al.³⁶ and according to our standardized study protocol,^15–17 it was found that 20% dextran solution was effective in avoiding tissue swelling and maintaining corneal hydration during experimentation of riboflavin/UV-A corneal cross-linking. In this study, we indented less than 3 μm anterior stromal depth (which corresponds to the less hydrated part of the cornea; <1% total stromal thickness), and the changes of hydration as well as of indented stroma were therefore considered negligible.⁴⁰ 

In conclusion, as new and emerging corneal cross-linking protocols are entering the clinical scene, it is of great importance to test and validate their efficacy in the laboratory by using appropriate methodological approaches and by comparing the results with the conventional procedure. In this study, a nanotechnology-based drug delivery platform for transepithelial riboflavin/UV-A corneal cross-linking was shown to generate effective biomechanical strengthening of the anterior corneal stroma in human donor tissues. 

Acknowledgments

The authors thank Dario Rusciano and Sooft Italia Spa for generously providing the riboflavin solutions used in the present work. 

Supported by the National Framework Program for Research and Innovation Programma Operativo Nazional (PON) Research Grant 0100110 (CL, MPDS, GL), the Italian Ministry of Health (ML), and Fondazione Roma (ML). 

Disclosure: C. Labate, None; M. Lombardo, None; G. Lombardo, None; M.P. De Santo, None