December 2012
Volume 53, Issue 13
Free
Cornea  |   December 2012
Silica Hybrid for Corneal Replacement: Optical, Biomechanical, and Ex Vivo Biocompatibility Studies
Author Affiliations & Notes
  • Michael D. DiVito
    From the Departments of Mechanical Engineering and
  • Stephen G. Rudisill
    Chemistry, University of Minnesota, Minneapolis, Minnesota; and the
  • Andreas Stein
    Chemistry, University of Minnesota, Minneapolis, Minnesota; and the
  • Sanjay V. Patel
    Department of Ophthalmology, Mayo Clinic, Rochester, Minnesota.
  • Jay W. McLaren
    Department of Ophthalmology, Mayo Clinic, Rochester, Minnesota.
  • Allison Hubel
    From the Departments of Mechanical Engineering and
  • Corresponding author: Allison Hubel, Department of Mechanical Engineering, University of Minnesota, Minneapolis, MN 55455; Hubel001@umn.edu
Investigative Ophthalmology & Visual Science December 2012, Vol.53, 8192-8199. doi:10.1167/iovs.12-10561
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Michael D. DiVito, Stephen G. Rudisill, Andreas Stein, Sanjay V. Patel, Jay W. McLaren, Allison Hubel; Silica Hybrid for Corneal Replacement: Optical, Biomechanical, and Ex Vivo Biocompatibility Studies. Invest. Ophthalmol. Vis. Sci. 2012;53(13):8192-8199. doi: 10.1167/iovs.12-10561.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose.: To investigate compositions of silica–collagen hybrid materials as potential artificial corneal substitutes, how these components affect the optical and biomechanical properties of the hybrids, and their biocompatibility in an organ culture model.

Methods.: Hybrid materials were created from different proportions of collagen and silica precursors and manufactured to specific dimensions. The microstructure of the materials was determined by electron microscopy and mechanical strength was measured by using suture pullout tests. The refractive index and transmittance were measured by using an Abbe refractometer and a spectrophotometer. Materials were implanted into rabbit corneas to determine their epithelialization in organ culture.

Results.: Scanning electron microscopy demonstrated that the hybrid material consisted of silica-encapsulating collagen fibrils. The refractive index ranged from 1.332 to 1.403 depending upon the composition and manufacturing characteristics. The rupture strength of a 3:1 (silica:collagen ratio by weight) rehydrated xerogel was 0.161 ± 0.073 N/mm (n = 12), while the hydrogels and 9:1 xerogel were too fragile for suturing. Re-epithelialization of 5- to 6-mm-wide rabbit corneal epithelial defects was complete in 5.5 ± 2.4 days (n = 6), with evidence of epithelial stratification.

Conclusions.: Silica–collagen hybrid materials can be manufactured to specific dimensions to serve as a possible artificial corneal substitute. In preliminary studies, the materials had favorable optical, biomechanical, and biocompatibility properties necessary for replacing the corneal stroma.

Introduction
The cornea is the most commonly transplanted tissue in the United States, with over 40,000 transplants performed annually. 1 Successful corneal transplantation depends on the availability of good quality cadaveric donor tissue, the demand for which increased by 37% between 1997 and 2010. 2 Worldwide, the supply of donor corneas suitable for transplantation has never matched the demand, partly because of the lack of appropriate eye banking facilities. Even in the United States, several factors impact the future efficient provision of donor tissue and the need for increasing the donor supply. Threats to the donor supply include the prevalence of common viral infections (e.g., hepatitis B), increasingly stringent US Food and Drug Administration regulations towards testing for emerging infections, 2 and the large number of potential donors who have had cataract or refractive surgery, which can render tissue ineligible for donation. 3 In addition, the demand for donor corneas has increased because of an expanding population over age 65 years (baby boomers), and because improvements in surgical techniques have enabled safer and earlier intervention. 4 As a result, novel and alternative methods of expanding the donor pool must be investigated to meet future demand. 
Current artificial corneas (keratoprostheses) available for clinical use do not truly bio-integrate with surrounding tissues, nor do they allow for epithelialization of their surface, which can result in complications. 5,6 As a result, there is significant effort being directed towards the development of biocompatible corneal substitutes that will support tissue regeneration. 7 Regeneration of the cornea requires materials that can mimic the optical and mechanical properties of the native cornea, and that can be easily stored for use as needed, effectively eliminating the need for human donor tissue. Type I collagen-based structures have shown potential in meeting these design requirements with biomechanical strength obtained by chemical cross-linking. 811 Cross-linked collagen constructs have enabled tissue regeneration and have been piloted in human trials in vivo with moderate success. 7,12,13  
Seeking an alternative to cross-linked collagen constructs, we have investigated different compositions of silica–collagen bio-hybrid materials, which are known to show favorable optical and mechanical properties for corneal replacement. 14,15 Silica–collagen hydrogels have also been successful in hosting various fibroblast cell types. 16,17 In this study, silica precursor 3-aminopropyltriethoxsilane (APTS) and soluble type I bovine collagen were combined to create a hybrid material. We determined the mechanical and optical properties for two different silica compositions in hydrogel and xerogel states, and we developed a manufacturing process to construct implants of the appropriate size and geometry. We report the microstructural, biomechanical, and optical properties of these implants and initial biocompatibility studies in a cultured rabbit cornea model. 
Methods
Hybrid Preparation
For all tests and applications, the fabrication of the hybrid material followed a standard procedure. APTS (Sigma-Aldrich, St. Louis, MO) was mixed with acetic acid in a bath sonicator at 4°C. The molarity of the acetic acid was varied to adjust the pH of the hybrids. Soluble type I bovine collagen (Organogenesis Collagen, Canton, MA) was mixed with the diluted APTS at a 10:3 volume ratio for 10 minutes with a magnetic stir bar. To prevent premature gelation, the mixing vial was submerged in an ice bath. Hybrid composition was described by silica to collagen weight ratios. The two compositions investigated in this study were 3:1 and 9:1. The silica to collagen ratio was modulated with the volume of acetic acid used in the APTS dilution. 
Immediately after mixing, the gels were de-aerated by centrifugation and poured into polydimethylsiloxane negative molds. For manufacturing of the ex vivo implants, the gel was allowed to cure in a hemispherical mold, with curvature similar to that of the cornea, and then punched with a corneal trephine. Conventional methods of hybrid manufacturing involved formation of the hybrid under fully hydrated conditions (100% relative humidity). Hybrids manufactured by using this method will be referred to as hydrogels. Additional studies were performed on hybrids that were dehydrated (cured at 72% relative humidity for 24 hours), which are specified as “xerogels.” After complete formation and dehydration of the material, the samples were rehydrated in a phosphate-buffered saline (PBS; Sigma-Aldrich) solution. The rehydrated samples were kept in 100% relative humidity conditions prior to testing. Because all xerogels were rehydrated prior to testing, rehydrated xerogels will be referred to as simply xerogels. 
Optical Measurements
The transmittance, T, defined as the ratio of transmitted light intensity to incident light intensity, of each composition was measured in a 96-well plate by using a spectrophotometer (SpecroMax Plus; Molecular Devices, Sunnyvale, CA). A 50-μL volume of mixed gel was pipetted into each well. The gel was allowed to cure inside the well at 100% humidity. Each well was hydrated with 50 μL of PBS. Twelve wells were filled with 50 μL of PBS and used as blanks. The fraction of transmittance (ratio of the amount of light that passed through the sample to the amount of light that passed through the blank) was calculated for wavelengths of 300 to 800 nm in 10-nm intervals. With knowledge of the transmittance and thickness of a sample, d, a material constant αc for each composition was calculated from Equation 1. The material constant is an intrinsic material property. With a known αc and by using the same material thickness, d, transmittance of two different compositions could be directly compared. Transmittance can be considered an intrinsic property for a fixed thickness. 
Refractive indices of the materials were measured with an Abbe refractometer (Zeiss, Jena, Germany) at 25°C. The hybrids were individually immersed in saturated sucrose solutions. The solution that contained a sample was diluted dropwise with distilled water, until the sample became more transparent. An aliquot of the sucrose solution was removed, and its refractive index was measured. The solution was then diluted again by a few drops of distilled water, and the refractive index was remeasured. Changes in refractive index between dilutions were approximately 0.004. This procedure was repeated until the sample became more opaque, and the refractive index of this solution was recorded. 
Suture Pullout Measurements
The ability of the hybrid material to hold a suture was estimated by using a low force biaxial tester (Instron, Norwood, MA) to measure tensile strength of the sample. An 8-mm-diameter corneal trephine was used to punch out disc-shaped test implants, which were approximately 200-μm thick. One half of the disc was clamped in a spring-loaded grip. A 10-0 nylon suture (Ethicon, Somerville, NJ) was placed 2 mm from the edge of the opposite half of the implant, and the two free ends of the suture were placed in the other grip. All implants were tested in a PBS bath to prevent dehydration. The grip holding the suture was displaced at a rate of 0.1 mm/s relative to the opposing grip until the suture was pulled completely through the implant. The failure point was considered to be the first showing of ripping or tearing. Rupture strength was defined as the force at failure divided by the thickness of the disc. 
Scanning Electron Microscopy
The morphology and microstructure of the composite materials was assessed by scanning electron microscopy (SEM). Implants were fixed in formaldehyde (4 wt% in water; Sigma-Aldrich), dried sequentially in 20%, 40%, 60%, 80%, 95%, and finally 100% ethanol solutions (Pharmco-AAPER, Brookfield, CT), then transferred to a Tousimis samdri-780A CO2 critical point dryer (Tousimis, Rockville, MD). Samples (5-mm discs) were broken in half and mounted on carbon tape stubs, one half flat against the stub and the other perpendicular for cross-sectional imaging. The dried specimens were then coated in 5 nm of Pt and imaged on a JEOL 6700F SEM (JEOL USA, Peabody, MA) with a 5 kV beam at 10 μA. 
Ex Vivo Biocompatibility in a Rabbit Model
Preliminary biocompatibility of corneal re-epithelialization of the 9:1 hydrogel was studied in a rabbit cornea organ culture model similar to a method described by Evans et al. 18 Five New Zealand white rabbits were euthanized with an overdose of intravenous sodium pentobarbital immediately before operating on both corneas. All the procedures involving these animals adhered to the ARVO Statement for Use of Animals in Ophthalmic and Vision Research and were approved by the Mayo Clinic Institutional Animal Committee. Under sterile conditions, the center of the cornea was marked and the anterior stroma was incised with a guarded diamond blade to a depth of 200 μm. A 9-mm-diameter circular lamellar pocket was created at a depth of 200 μm from the anterior surface. The anterior lamella was excised centrally to create a keratectomy approximately 5 mm in diameter. 
After surgical preparation of the cornea for implantation, the eye was enucleated and the anterior segment was excised, including removal of the lens and iris. The anterior segment had an approximately 3-mm rim of sclera and was mounted to a modified petri dish 19 designed for rabbit anterior segment organ culture (Fig. 1). The corneoscleral rim was clamped with a locking ring and culture medium was infused to fill and maintain the anterior chamber. The culture medium was Dulbecco's modified Eagle's medium/Ham's F-12 with 20 mM L-glutamine (Sigma-Aldrich) with a 1:100 dilution of antibiotic/antimycotic suspension (penicillin G 100 U/mL, streptomycin 100 μg/mL, amphotericin B 0.25 μg/mL; Sigma-Aldrich) and a 1:100 dilution of insulin-transferrin-selenium (Gibco, Carlsbad, CA). From a sterile film of the hybrid, an 8.2-mm button was created by punching with a corneal trephine. The button was then implanted into eight corneas with the periphery of the implant enveloped in the lamellar pocket and the center of the implant exposed because of the keratectomy. The dimensions of the keratectomy (i.e., the de-epithelialized region over the implant) were measured by staining the region with sodium fluorescein and examining with a blue light (Fig. 2). Culture medium was placed externally to just cover the center of the cornea and implant. 
Figure 1. 
 
Anterior segment organ culture system. Left: schematic. Right: photograph of modified petri dish with mounted anterior segment. Culture medium could be perfused into the artificial anterior chamber for medium exchange, and medium was filled external to the anterior segment to just cover the vertex of the cornea (dashed line).
Figure 1. 
 
Anterior segment organ culture system. Left: schematic. Right: photograph of modified petri dish with mounted anterior segment. Culture medium could be perfused into the artificial anterior chamber for medium exchange, and medium was filled external to the anterior segment to just cover the vertex of the cornea (dashed line).
Figure 2. 
 
Assessment of closure of corneal epithelial defect. Left: the epithelial defect created by central lamellar keratectomy stained with sodium fluorescein. The size of the epithelial defect was determined daily until closure. Right: the same cornea after closure of the epithelial defect, devoid of central staining.
Figure 2. 
 
Assessment of closure of corneal epithelial defect. Left: the epithelial defect created by central lamellar keratectomy stained with sodium fluorescein. The size of the epithelial defect was determined daily until closure. Right: the same cornea after closure of the epithelial defect, devoid of central staining.
Two control eyes received an identical surgical procedure but without implantation of a bio-hybrid. In all cases, the epithelial defect was measured daily until closure, and time to closure was recorded. Culture medium was changed daily. Anterior segments were incubated for 7 to 18 days at 37°C in humidified air containing 5% CO2, after which the corneoscleral rims were fixed in paraformaldehyde 4% for examination by light and fluorescence microscopy. The fixed tissue was embedded in paraffin, cut into 4-μm-thick sections, and stained with hematoxylin and eosin or deparaffinized and rehydrated for immunostaining with antigen retrieval. Sections were incubated overnight at room temperature with keratin AE1/AE3, 1:250 (Millipore, Billerica, MA). After washing, sections were incubated at room temperature with a secondary antibody (Alexa Fluor 488 Goat anti-Rabbit IgG [H+L]; Invitrogen, Molecular Probes, Eugene, OR) for 45 minutes, then washed and stained with Vectashield/DAPI (Vector Laboratories, Burlingame, CA) prior to examination. Positive and negative controls were stained at the same time as the staining of experimental tissue. 
Tissue sections were also prepared for transmission electron microscopy (TEM) by dehydration in ascending concentrations of ethanol and embedding in epoxy resin. Thin sections were mounted on copper grids, stained with uranyl acetate (saturated solution in 50% ethanol) and 0.1% lead citrate and examined on a JEOL 1400 transmission electron microscope (Peabody, MA). 
Statistical Analysis
All mean values in this paper are presented with the standard deviation. 
Results
Microstructural Properties
Under cross-sectional SEM, the 3:1 and 9:1 compositions in both the hydrogel and xerogel state consisted of a network of evenly sized fibers (Fig. 3a). Nonfibrillar structures observed in the fibrillar network (Fig. 3b) were assumed to be silica aggregates. A denser structure was seen at the surface, with some thicker fibers (Fig. 3c); however, on closer examination (Fig. 3d) these fibers appeared to consist of several smaller, intertwined fibers. It should be noted that the structures observed have been reduced in size by 10% to 25% from the ethanol dehydration and critical point drying procedure 20 ; however, this shrinkage is typically uniform throughout the sample and does not result in collapse (as air-drying does). 
Figure 3. 
 
Cross-sectional SEM of a 9:1 hydrogel at (a) lower magnification and (b) higher magnification. The lower magnification indicates a fibrillar network of consistently sized fibrils. The higher magnification shows the same fibrillar microstructure as well as aggregates formed on the surface of the sample. Surface SEM of the same hydrogel at (c) lower magnification and (d) higher magnification reveals a denser structure with thicker fibrils. It can be seen at the higher magnification that these larger structures actually consist of several intertwined smaller fibrils.
Figure 3. 
 
Cross-sectional SEM of a 9:1 hydrogel at (a) lower magnification and (b) higher magnification. The lower magnification indicates a fibrillar network of consistently sized fibrils. The higher magnification shows the same fibrillar microstructure as well as aggregates formed on the surface of the sample. Surface SEM of the same hydrogel at (c) lower magnification and (d) higher magnification reveals a denser structure with thicker fibrils. It can be seen at the higher magnification that these larger structures actually consist of several intertwined smaller fibrils.
Optical Properties
Transmittance increased as wavelength increased from the ultraviolet to the infrared for both compositions and hydration states. In general, the hydrogels had a higher transmittance than the xerogels (Fig. 4a), and the lower silica composition had a higher transmittance. The transmittance spectra for all compositions and hydration states are shown in Figure 4b (n > 30). The transmittance of the rabbit cornea was from data published by McLaren and Brubaker. 21  
Figure 4. 
 
Optical characterization of the hybrid material. (a) 5-mm buttons submerged in PBS on top of a ruler. The 3:1 hydrogel button (left) was more transparent than the 3:1 xerogel button (right). (b) Light transmission curves in the visible range for 3:1 and 9:1 compositions in the hydrogel and xerogel states. The rabbit cornea curve was made from tabulated data of McLaren and Brubaker. 21 All of the hybrid curves were adjusted to a common thickness of 0.3 mm.
Figure 4. 
 
Optical characterization of the hybrid material. (a) 5-mm buttons submerged in PBS on top of a ruler. The 3:1 hydrogel button (left) was more transparent than the 3:1 xerogel button (right). (b) Light transmission curves in the visible range for 3:1 and 9:1 compositions in the hydrogel and xerogel states. The rabbit cornea curve was made from tabulated data of McLaren and Brubaker. 21 All of the hybrid curves were adjusted to a common thickness of 0.3 mm.
The refractive index was 1.334 and 1.332 for the 9:1 and 3:1 compositions, respectively. Creation of a xerogel increased the index of refraction of both compositions to 1.403 and 1.389, respectively. 
Suture Pullout Strength
The hydrogel samples and the 9:1 rehydrated xerogel were too fragile for suture pullout tests. However, an example of a rupture strength-displacement curve for a 3:1 xerogel specimen is shown in Figure 5. The material displayed strain-hardening behavior prior to failure, and then proceeded to rupture in a ductile manner. The rupture strength of the 3:1 xerogel was 0.161 ± 0.073 N/mm. 
Figure 5. 
 
Force-displacement curve for the 3:1 rehydrated xerogel from a suture pullout test. The material displayed strain hardening in the initial part of the curve followed by a linear region. The material then failed (denoted by asterisk) and continued to rip until complete fracture.
Figure 5. 
 
Force-displacement curve for the 3:1 rehydrated xerogel from a suture pullout test. The material displayed strain hardening in the initial part of the curve followed by a linear region. The material then failed (denoted by asterisk) and continued to rip until complete fracture.
Ex Vivo Biocompatibility in a Rabbit Model
Eight rabbit corneas received surgery with implantation of a hybrid implant, and two corneas received surgery without implantation (controls). In all corneas, epithelial defects immediately after surgery were approximately 5 to 6 mm at the widest diameter (Fig. 2). In the two control corneas, complete re-epithelialization over exposed bare stroma took 8 and 15 days. In experimental corneas, two corneas became contaminated during the culture period and were excluded from analysis; the contamination was unrelated to sterility of the implants. In the remaining six implanted corneas, complete re-epithelialization over the exposed implant took 5.5 ± 2.4 days (range, 3–10 days; Fig. 2, Table). 
Table. 
 
Results of Ex Vivo Organ Culture Studies
Table. 
 
Results of Ex Vivo Organ Culture Studies
Epithelial Defect Time in Organ Culture, d Histologic Analysis
Size, mm Time to Closure, d Mean No. of Epithelial Cell Layers (Range) Keratin AE1/AE3 Staining Basement Membrane, TEM
Implanted corneas
 1 5.5 by 4 6 9 1 (1–2) +
 2 5 by 4 5 10 2 (1–3) +
 3 5 by 5 10 10 1 (1–2) +
 4 5 by 5 5 7 3 (1–5) +
 5 5 by 5 3 7 3 (2–6) +
 6 5 by 6 4 7 5 (4–8) +
Controls
 1 7 by 7 15 17 3 (1–5) + +
 2 5 by 5 8 10 2 (1–4) + +
Epithelialization of the implant surface was confirmed by light microscopy in all cases (Fig. 6). An early stratified epithelium was noted with longer duration of culture, but there was no evidence of stromal cell migration into the implants at the end of the culture period (Fig. 7). Regenerated epithelium stained positive for keratin AE1/AE3 in all cases (Fig. 7, Table). There was no evidence of epithelial basement membrane deposition over the hybrid surface by TEM, although the basal epithelial cell membrane was intimately associated with the hybrid surface (Fig. 8). 
Figure 6. 
 
Light microscopy images of implanted rabbit cornea. The peripheral implant (arrows) was enveloped in a paracentral lamellar pocket of rabbit stroma, and the center of the implant was exposed in the region of the anterior lamellar keratectomy (denoted by asterisks). A thin layer of epithelium covered the implant and was continuous with the epithelium at the edge of the keratectomy. Epithelial closure took 5 days, and culture was maintained for an additional 5 days. Inset: higher magnification image of the area denoted by the box showing an intact thin layer of epithelium covering a thin implant. Vacuolization of basal cells was noted over the surface of the implant, but similar changes were also noted in the peripheral native cornea where the epithelium had not been disrupted, and thus were most likely a manifestation of cellular changes because of organ culture. Hematoxylin and eosin.
Figure 6. 
 
Light microscopy images of implanted rabbit cornea. The peripheral implant (arrows) was enveloped in a paracentral lamellar pocket of rabbit stroma, and the center of the implant was exposed in the region of the anterior lamellar keratectomy (denoted by asterisks). A thin layer of epithelium covered the implant and was continuous with the epithelium at the edge of the keratectomy. Epithelial closure took 5 days, and culture was maintained for an additional 5 days. Inset: higher magnification image of the area denoted by the box showing an intact thin layer of epithelium covering a thin implant. Vacuolization of basal cells was noted over the surface of the implant, but similar changes were also noted in the peripheral native cornea where the epithelium had not been disrupted, and thus were most likely a manifestation of cellular changes because of organ culture. Hematoxylin and eosin.
Figure 7. 
 
High magnification of regenerated epithelium. Upper Left: stratified rabbit corneal epithelium covering the acellular and amorphous implant. This section was located at the center of the exposed stromal implant within the lamellar keratectomy. Epithelial closure took 4 days, and culture was maintained for an additional 3 days. Hematoxylin and eosin. Upper Right: normal rabbit anterior cornea with stratified epithelium attached to cellular stroma. Hematoxylin and eosin. Lower Left: positive staining with keratin AE1/AE3 indicating the epithelial phenotype of the regenerated cell layer over the acellular implant. Lower Right: positive epithelial staining with keratin AE1/AE3 in normal rabbit anterior cornea.
Figure 7. 
 
High magnification of regenerated epithelium. Upper Left: stratified rabbit corneal epithelium covering the acellular and amorphous implant. This section was located at the center of the exposed stromal implant within the lamellar keratectomy. Epithelial closure took 4 days, and culture was maintained for an additional 3 days. Hematoxylin and eosin. Upper Right: normal rabbit anterior cornea with stratified epithelium attached to cellular stroma. Hematoxylin and eosin. Lower Left: positive staining with keratin AE1/AE3 indicating the epithelial phenotype of the regenerated cell layer over the acellular implant. Lower Right: positive epithelial staining with keratin AE1/AE3 in normal rabbit anterior cornea.
Figure 8. 
 
Transmission electron microscopy of the basal epithelium. Left: the basal epithelial cell membrane was intimately associated with the surface of the hybrid material, although basement membrane was not visible. Right: normal rabbit basal epithelium with basement membrane (arrows) between the basal epithelial cell and the fibrillar stroma.
Figure 8. 
 
Transmission electron microscopy of the basal epithelium. Left: the basal epithelial cell membrane was intimately associated with the surface of the hybrid material, although basement membrane was not visible. Right: normal rabbit basal epithelium with basement membrane (arrows) between the basal epithelial cell and the fibrillar stroma.
Discussion
In this study, we successfully engineered a silica–collagen bio-hybrid material as an artificial corneal stroma for implantation into an animal model. The material was transparent and could be manufactured to specific dimensions for surgical procedures. Basic biocompatibility was established by showing corneal re-epithelialization in an ex vivo rabbit model. 
Material Properties of Hybrid
Qualitatively, the hybrid materials fabricated by using the given manufacturing method were ductile and relatively easy to handle when compared to pure collagen gels made from soluble collagen of the same concentration. SEM images displayed the hybrid's fibrillar morphology. The absence of D-banding indicated that the collagen fibrils were coated with silica particles. Ono et al. 22 displayed this assumed microstructure with similar fibril diameters for TEOS-collagen hybrids, where TEM images showed contrast between the silica coating and collagen fibril center. Additional studies are necessary to quantify important microstructural properties such as fibril diameter, spacing, and orientation. These properties have shown to be critical for the high optical quality in the native cornea. 
The hydrogels had a higher transmittance than did the native cornea but were too weak to be sutured without fracture. The increase in strength and stiffness from the hydro- to xerogel state was likely because of the higher concentration of collagen in the xerogels. Even after rehydration, the hydrogel volume decreased by approximately 90%. The 3:1 composition was stronger and more transparent than the higher silica composition. The silica aggregates in the 9:1 composition are indicative of structural inhomogeneities, which may account for the differences in biomechanical and optical properties. Some of the 9:1 xerogel samples were strong enough to receive a suture but failed while being loaded into the tensile test apparatus. The 3:1 rehydrated xerogel had a 0.161 ± 0.073 N/mm rupture strength. Liu et al. 23 reported suture strengths in the range of approximately 0.07 to 0.40 N/mm for collagen hydrogels cross-linked with carbodiimide. The human cornea has a rupture strength that exceeds the tensile strength of the nylon sutures used; that is, the sutures fail before the cornea ruptures in suture tests. Menovksy et al. 24 reported a tensile strength of 0.35 ± 0.02 N for 10-0 nylon sutures. This indicates that the suture strength of the human cornea is significantly higher than the silica–collagen hybrid. 
The indeces of refraction of the hydrogel samples were both approximately the same as that of water. After the samples were dehydrated, the refractive indices increased for both silica compositions. This was to be expected because the index of refraction of silica without water is approximately 1.45. 25 Additionally, samples with the higher silica content yielded a higher index of refraction, one closer to that of silica when compared to sample of lower silica content. 
The epithelium of the human cornea has a refractive index of approximately 1.40, and the refractive index of the stroma ranges from 1.37 to 1.38. 26 The 3:1 and 9:1 xerogels have refractive indices of 1.389 and 1.403, respectively, which fall close to the human cornea indices. The small difference between the 3:1 sample and that of the human stroma could be eliminated by minor adjustments in silica content. 
The mechanical properties of the hybrid were significantly improved when compared to mechanical properties of a control made solely with collagen. However, compositions analyzed in this study did not reach the strength of the native cornea. 
The transmittance characteristics of the hybrid material have a distinct advantage over many other materials that have been investigated (Fig. 4). The 3:1 xerogel mimics native cornea closely in both the visible and UV spectrum. 27 The 3:1 hydrogel has a transmittance spectrum in the visible that is similar to that of rabbit cornea in vivo, 21,28 although transmittance does not drop as quickly at wavelengths below 350 nm as it does in the living cornea. The high transmittance to 400 nm is consistent with measurements in human corneas where the corneas were not removed from the globe and were minimally disturbed. 21 To optimize the mechanical and optical properties, parameters such as pH, silica:collagen composition, synthesis and manufacturing methodology, collagen concentration, and hydration state are currently being investigated in more detail. The relationship between optical properties and strength and initial collagen concentration should be carefully investigated. Dehydration of the gels led to improved mechanical properties, but this dehydration resulted in a decreased transmittance. An increase in initial collagen concentration could result in a denser fibril network, yielding a transparent hydrogel with similar mechanical properties as the xerogels presented in this paper. Also, higher concentrations of collagen have been shown to assemble into highly organized fibrils, 29 which could potentially decrease light scattering. 
Biocompatibility
For the initial biocompatibility studies, our goal was to show that corneal epithelium would regenerate and cover the surface of the implant. Without a continuous epithelial covering and cellular attachment, the implant would be prone to degradation and infection in vivo. With our organ culture method and a surgical technique established by Evans et al. 18 we found that the implanted bio-hybrid consistently supported complete re-epithelialization within an average of 1 week. While the control corneas in this study took longer to re-epithelialize, they were anatomically different from the experimental corneas with a significant discontinuity at the edge of the 200-μm-deep keratectomy; we would have expected a simple epithelial defect to have closed within 3 days if we had not performed a keratectomy. By immunohistochemistry, we confirmed the epithelial phenotype with keratin AE1/AE3 staining, but we were unable to detect visible basement membrane formation by TEM. The lack of visible basement membrane might be explained by the short duration of culture after closure of the epithelial defect, 18 and longer-term studies in vivo will be required to convincingly show cell attachment and formation of focal adhesion complexes. Epithelialization of any artificial corneal material is critical to its success because the material is more likely to be retained. Fagerholm et al. 7 showed that slow epithelialization of a cross-linked collagen hydrogel implanted into human corneas resulted in initial thinning of the corneas. While our initial results are encouraging and represent a major advantage compared to current plastic keratoprostheses, which do not support epithelium, further studies are needed to determine epithelial outcomes in vivo. 
With initial favorable characteristics of the silica–collagen bio-hybrid, in vivo studies are planned to further assess biocompatibility over a longer duration of follow-up. These studies will enable assessment of corneal cellular changes, 30 transparency and backscatter, 21,31 and oxygen transmissibility. 32 Although the initial implants did not require suturing to the recipient cornea, based on the suture pullout tests, an overlay suturing technique will be required in vivo, similar to that used by Fagerholm et al. 7 in their experiments. 
Conclusions
In summary, we have described the manufacture and basic characteristics of a novel artificial corneal substrate that might have potential in corneal regeneration in the future. In vitro and ex vivo studies indicate that the material has favorable optical properties and biocompatibility for use as an artificial cornea. Further studies are required to optimize the biomechanical strength, and to assess the optical properties and biocompatibility in vivo. 
References
EBAA. Eye Bank Association of America. 2010–2011, A Year in Review . Washington, DC: Eye Bank Association of America; 2011.
EBAA. Eye Bank Association of America. Annual Statistical Report . Washington, DC: Eye Bank Association of America; 2010.
Beck RW Gal RL Mannis MJ Is donor age an important determinant of graft survival? Cornea . 2000;19:503–510. [CrossRef]
Tan D Mehta J. Future directions in lamellar corneal transplantation. Cornea . 2007;26:S21–S28. [CrossRef] [PubMed]
Hicks CR Crawford GJ Lou X Corneal replacement using a synthetic hydrogel cornea, AlphaCor: device, preliminary outcomes and complications. Eye . 2003;17:385–392. [CrossRef] [PubMed]
Ma JJK Graney JM Dohlman CH. Repeat penetrating keratoplasty versus the Boston keratoprosthesis in graft failure. Int Opthalmol Clin . 2005;45:49–59. [CrossRef]
Fagerholm P Lagali NS Merrett K A biosynthetic alternative to human donor tissue for inducing corneal regeneration: 24-month follow-up of a phase 1 clinical study. Sci Transl Med . 2010;2:46ra61. [CrossRef]
Orwin EJ Borene ML Hubel A. Biomechanical and optical characteristics of a corneal stromal equivalent. J Biomech Eng . 2003;125:439–444. [CrossRef] [PubMed]
Crabb RAB Hubel A. Influence of matrix processing on the optical and biomechanical properties of a corneal stroma equivalent. Tissue Eng Part A . 2008;14:173–182. [CrossRef] [PubMed]
Duncan TJ Tanaka Y Shi D Flow-manipulated, crosslinked collagen gels for use as corneal equivalents. Biomaterials . 2010;31:8996–9005. [CrossRef] [PubMed]
Liu Y Gan L Carlsson DJ A simple, cross-linked collagen tissue substitute for corneal implantation. Invest Ophthalmol Vis Sci . 2006;47:1869–1875. [CrossRef] [PubMed]
Crabb RAB Chau E Evans M Barocas V Hubel A. Collagen film-based corneal stroma equivalent. Tissue Eng . 2006;12:1565–1575. [CrossRef] [PubMed]
Fagerholm P Lagali NS Carlsson DJ Merrett K Griffith M. Corneal regeneration following implantation of a biomimetic tissue-engineered substitute. Clin Transl Sci . 2009;2:162–164. [CrossRef] [PubMed]
Coradin T Helary C Livage J. A novel route to collagen-silica biohybrids. Mater Res Soc Symp Proc . 2002;726:1–5.
Coradin T Allouche J Boissière M Livage J. Sol-gel biopolymer/silica nanocomposites in biotechnology. Curr Nanosci . 2006;2:219–230. [CrossRef]
Brasack I Bottcher H. Biocompatibility of modified silica-protein composite layers. J Sol-Gel Sci Technol . 2000;19:479–482. [CrossRef]
Desimone MF Hélary C Mosser G Fibroblast encapsulation in hybrid silica–collagen hydrogels. J Mater Chem . 2010;20:666–668. [CrossRef]
Evans MDM McFarland GA Xie RZ The use of corneal organ culture in biocompatibility studies. Biomaterials . 2002;23:1359–1367. [CrossRef] [PubMed]
Brunette I Nelson LR Bourne WM. A system for long-term corneal perfusion. Invest Ophthalmol Vis Sci . 1989;30:1813–1822. [PubMed]
Wells OC. Scanning Electron Microscopy . McGraw-Hill: New York; 1974:325–331.
McLaren JW Brubaker RF. Measurement of transmission of ultraviolet and visible light in the living rabbit cornea. Curr Eye Res . 1996;15:411–421. [CrossRef] [PubMed]
Ono Y Kanekiyo Y Inoue K Preparation of novel hollow fiber silica using collagen fibers as a template. Chem Lett . 1999;6:475–476. [CrossRef]
Liu Y Griffith M Watsky MA Properties of porcine and recombinant human collagen matrices for optically clear tissue engineering applications. Biomacromolecules . 2006;7:1819–1828. [CrossRef] [PubMed]
Menovsky T Beek JF van Gemert MJ. Effect of the CO2 milliwatt laser on tensile strength of microsutures. Lasers Surg Med . 1997;20:64–68. [CrossRef] [PubMed]
Haynes W. Optical properties of selected organic and inorganic materials. In: CRC Handbook of Chemistry and Physics. 92nd ed. Boca Raton, FL: CRC Press/Taylor and Francis; 2012:163–164.
Patel S Marshall J Fitzke FW. Refractive-index of the human corneal epithelium and stroma. J Refract Surg . 1995;11:100–105. [PubMed]
van den Berg TJTP Tan KEWP. Research note light transmittance for different ages of the human cornea from 320 to 700 nm. Vision Res . 1994;34 :1453–1456. [CrossRef] [PubMed]
Walsh JE Bergmanson JPG Koehler LV Fibre optic spectrophotometry for the in vitro evaluation of ultraviolet radiation (UVR) spectral transmittance of rabbit corneas. Physiol Meas . 2008;29:375–388. [CrossRef] [PubMed]
Besseau L Giraud-Guille M. Stabilization of fluid cholesteric phases of collagen to ordered gelated matrices. J Mol Biol . 1995;251:197–202. [CrossRef] [PubMed]
Erie JC McLaren JW Patel SV. Confocal microscopy in ophthalmology. Am J Ophthalmol . 2009;148 :639–646. [CrossRef] [PubMed]
McLaren JW Bourne WM Patel SV. Standardization of corneal haze measurement in confocal microscopy. Invest Ophthalmol Vis Sci . 2010;51:5610–5616. [CrossRef] [PubMed]
McLaren JW Dinslage S Dillon JP Roberts JE Brubaker RF. Measuring oxygen tension in the anterior chamber of rabbits. Invest Ophthalmol Vis Sci . 1998;39:1899–1909. [PubMed]
Footnotes
 Supported by the Minnesota Partnership for Genomics and Biotechnology; Research to Prevent Blindness, New York, (unrestricted grant to the Department of Ophthalmology at Mayo Clinic, and SVP as Olga Keith Wiess Special Scholar); and Mayo Foundation, Rochester, Minnesota.
Footnotes
 Disclosure: M.D. DiVito, None; S.G. Rudisill, None; A. Stein, None; S.V. Patel, None; J.W. McLaren, None; A. Hubel, P
Figure 1. 
 
Anterior segment organ culture system. Left: schematic. Right: photograph of modified petri dish with mounted anterior segment. Culture medium could be perfused into the artificial anterior chamber for medium exchange, and medium was filled external to the anterior segment to just cover the vertex of the cornea (dashed line).
Figure 1. 
 
Anterior segment organ culture system. Left: schematic. Right: photograph of modified petri dish with mounted anterior segment. Culture medium could be perfused into the artificial anterior chamber for medium exchange, and medium was filled external to the anterior segment to just cover the vertex of the cornea (dashed line).
Figure 2. 
 
Assessment of closure of corneal epithelial defect. Left: the epithelial defect created by central lamellar keratectomy stained with sodium fluorescein. The size of the epithelial defect was determined daily until closure. Right: the same cornea after closure of the epithelial defect, devoid of central staining.
Figure 2. 
 
Assessment of closure of corneal epithelial defect. Left: the epithelial defect created by central lamellar keratectomy stained with sodium fluorescein. The size of the epithelial defect was determined daily until closure. Right: the same cornea after closure of the epithelial defect, devoid of central staining.
Figure 3. 
 
Cross-sectional SEM of a 9:1 hydrogel at (a) lower magnification and (b) higher magnification. The lower magnification indicates a fibrillar network of consistently sized fibrils. The higher magnification shows the same fibrillar microstructure as well as aggregates formed on the surface of the sample. Surface SEM of the same hydrogel at (c) lower magnification and (d) higher magnification reveals a denser structure with thicker fibrils. It can be seen at the higher magnification that these larger structures actually consist of several intertwined smaller fibrils.
Figure 3. 
 
Cross-sectional SEM of a 9:1 hydrogel at (a) lower magnification and (b) higher magnification. The lower magnification indicates a fibrillar network of consistently sized fibrils. The higher magnification shows the same fibrillar microstructure as well as aggregates formed on the surface of the sample. Surface SEM of the same hydrogel at (c) lower magnification and (d) higher magnification reveals a denser structure with thicker fibrils. It can be seen at the higher magnification that these larger structures actually consist of several intertwined smaller fibrils.
Figure 4. 
 
Optical characterization of the hybrid material. (a) 5-mm buttons submerged in PBS on top of a ruler. The 3:1 hydrogel button (left) was more transparent than the 3:1 xerogel button (right). (b) Light transmission curves in the visible range for 3:1 and 9:1 compositions in the hydrogel and xerogel states. The rabbit cornea curve was made from tabulated data of McLaren and Brubaker. 21 All of the hybrid curves were adjusted to a common thickness of 0.3 mm.
Figure 4. 
 
Optical characterization of the hybrid material. (a) 5-mm buttons submerged in PBS on top of a ruler. The 3:1 hydrogel button (left) was more transparent than the 3:1 xerogel button (right). (b) Light transmission curves in the visible range for 3:1 and 9:1 compositions in the hydrogel and xerogel states. The rabbit cornea curve was made from tabulated data of McLaren and Brubaker. 21 All of the hybrid curves were adjusted to a common thickness of 0.3 mm.
Figure 5. 
 
Force-displacement curve for the 3:1 rehydrated xerogel from a suture pullout test. The material displayed strain hardening in the initial part of the curve followed by a linear region. The material then failed (denoted by asterisk) and continued to rip until complete fracture.
Figure 5. 
 
Force-displacement curve for the 3:1 rehydrated xerogel from a suture pullout test. The material displayed strain hardening in the initial part of the curve followed by a linear region. The material then failed (denoted by asterisk) and continued to rip until complete fracture.
Figure 6. 
 
Light microscopy images of implanted rabbit cornea. The peripheral implant (arrows) was enveloped in a paracentral lamellar pocket of rabbit stroma, and the center of the implant was exposed in the region of the anterior lamellar keratectomy (denoted by asterisks). A thin layer of epithelium covered the implant and was continuous with the epithelium at the edge of the keratectomy. Epithelial closure took 5 days, and culture was maintained for an additional 5 days. Inset: higher magnification image of the area denoted by the box showing an intact thin layer of epithelium covering a thin implant. Vacuolization of basal cells was noted over the surface of the implant, but similar changes were also noted in the peripheral native cornea where the epithelium had not been disrupted, and thus were most likely a manifestation of cellular changes because of organ culture. Hematoxylin and eosin.
Figure 6. 
 
Light microscopy images of implanted rabbit cornea. The peripheral implant (arrows) was enveloped in a paracentral lamellar pocket of rabbit stroma, and the center of the implant was exposed in the region of the anterior lamellar keratectomy (denoted by asterisks). A thin layer of epithelium covered the implant and was continuous with the epithelium at the edge of the keratectomy. Epithelial closure took 5 days, and culture was maintained for an additional 5 days. Inset: higher magnification image of the area denoted by the box showing an intact thin layer of epithelium covering a thin implant. Vacuolization of basal cells was noted over the surface of the implant, but similar changes were also noted in the peripheral native cornea where the epithelium had not been disrupted, and thus were most likely a manifestation of cellular changes because of organ culture. Hematoxylin and eosin.
Figure 7. 
 
High magnification of regenerated epithelium. Upper Left: stratified rabbit corneal epithelium covering the acellular and amorphous implant. This section was located at the center of the exposed stromal implant within the lamellar keratectomy. Epithelial closure took 4 days, and culture was maintained for an additional 3 days. Hematoxylin and eosin. Upper Right: normal rabbit anterior cornea with stratified epithelium attached to cellular stroma. Hematoxylin and eosin. Lower Left: positive staining with keratin AE1/AE3 indicating the epithelial phenotype of the regenerated cell layer over the acellular implant. Lower Right: positive epithelial staining with keratin AE1/AE3 in normal rabbit anterior cornea.
Figure 7. 
 
High magnification of regenerated epithelium. Upper Left: stratified rabbit corneal epithelium covering the acellular and amorphous implant. This section was located at the center of the exposed stromal implant within the lamellar keratectomy. Epithelial closure took 4 days, and culture was maintained for an additional 3 days. Hematoxylin and eosin. Upper Right: normal rabbit anterior cornea with stratified epithelium attached to cellular stroma. Hematoxylin and eosin. Lower Left: positive staining with keratin AE1/AE3 indicating the epithelial phenotype of the regenerated cell layer over the acellular implant. Lower Right: positive epithelial staining with keratin AE1/AE3 in normal rabbit anterior cornea.
Figure 8. 
 
Transmission electron microscopy of the basal epithelium. Left: the basal epithelial cell membrane was intimately associated with the surface of the hybrid material, although basement membrane was not visible. Right: normal rabbit basal epithelium with basement membrane (arrows) between the basal epithelial cell and the fibrillar stroma.
Figure 8. 
 
Transmission electron microscopy of the basal epithelium. Left: the basal epithelial cell membrane was intimately associated with the surface of the hybrid material, although basement membrane was not visible. Right: normal rabbit basal epithelium with basement membrane (arrows) between the basal epithelial cell and the fibrillar stroma.
Table. 
 
Results of Ex Vivo Organ Culture Studies
Table. 
 
Results of Ex Vivo Organ Culture Studies
Epithelial Defect Time in Organ Culture, d Histologic Analysis
Size, mm Time to Closure, d Mean No. of Epithelial Cell Layers (Range) Keratin AE1/AE3 Staining Basement Membrane, TEM
Implanted corneas
 1 5.5 by 4 6 9 1 (1–2) +
 2 5 by 4 5 10 2 (1–3) +
 3 5 by 5 10 10 1 (1–2) +
 4 5 by 5 5 7 3 (1–5) +
 5 5 by 5 3 7 3 (2–6) +
 6 5 by 6 4 7 5 (4–8) +
Controls
 1 7 by 7 15 17 3 (1–5) + +
 2 5 by 5 8 10 2 (1–4) + +
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×