3D Functional Corneal Stromal Tissue Equivalent Based on Corneal Stromal Stem Cells and Multi-Layered Silk Film Architecture

Abstract

The worldwide need for human cornea equivalents continues to grow. Few clinical options are limited to allogenic and synthetic material replacements. We hypothesized that tissue engineered human cornea systems based on mechanically robust, patterned, porous, thin, optically clear silk protein films, in combination with human corneal stromal stem cells (hCSSCs), would generate 3D functional corneal stroma tissue equivalents, in comparison to previously developed 2D approaches. Silk film contact guidance was used to control the alignment and distribution of hCSSCs on RGD-treated single porous silk films, which were then stacked in an orthogonally, multi-layered architecture and cultured for 9 weeks. These systems were compared similar systems generated with human corneal fibroblasts (hCFs). Both cell types were viable and preferentially aligned along the biomaterial patterns for up to 9 weeks in culture. H&E histological sections showed that the systems seeded with the hCSSCs displayed ECM production throughout the entire thickness of the constructs. In addition, the ECM proteins tested positive for keratocyte-specific tissue markers, including keratan sulfate, lumican, and keratocan. The quantification of hCSSC gene expression of keratocyte-tissue markers, including keratocan, lumican, human aldehyde dehydrogenase 3A1 (ALDH3A1), prostaglandin D2 synthase (PTDGS), and pyruvate dehydrogenase kinase, isozyme 4 (PDK4), within the 3D tissue systems demonstrated upregulation when compared to 2D single silk films and to the systems generated with the hCFs. Furthermore, the production of ECM from the hCSSC seeded systems and subsequent remodeling of the initial matrix significantly improved cohesiveness and mechanical performance of the constructs, while maintaining transparency after 9 weeks.

Fig 1. 3D functional corneal stromal tissue equivalent preparation.

Upon hCSSC extraction from human donor, cells were expanded in culture until seeding. Silk film topography, porosity and surface chemistry were optimized. hCSSCs were seeded on the silk films and cultured in 2D until confluent. Afterwards, 7 cellular silk layers were stacked in an orthogonal fashion to mimic the 3D physiological architecture. After 9 weeks in culture, the functional organization of the stroma equivalent was assessed based on cell organization, keratocyte-specific ECM production, optical and mechanical properties.

Fig 2. Morphological and structural characterization of silk film.

A. SEM micrographs of the silk films, where high magnification images show the details of the surface pattern to guide cell alignment and porosity. B. Maximum intensity projection of CLSM analysis of silk films show 5 μm pores passing through the film. C. ATR-FTIR spectra of samples, where beta sheet crystalline content was induced upon water-annealing treatment. Samples were characterized by an Amide I absorbance split into two peaks centered at 1643 and 1621 cm^-1.

Fig 3. 3D hCSSC and hCF tissue equivalent distribution at 9 weeks.

Maximum intensity projections and side views of CLSM analysis of Calcein AM-stained hCSSCs and hCFs in the stroma equivalent. H&E staining of histological sections of stroma equivalent seeded with hCSSCs and hCFs showed evidence of ECM production at 9 weeks in culture. Scale bar = 100 and 200 μm.

Fig 4. Immunofluorescent staining of corneal-specific proteins.

3D and 2D silk film constructs seeded with hCSSCs in comparison to hCFs were stained at 9 weeks in culture against keratocan, keratin sulfate, and lumican. The proteoglycans displayed preferential alignment along the silk film surface grooves (arrows). Scale bar = 100 μm.

Fig 5. mRNA expression within hCSSC and hCF tissue equivalent in comparison to 2D culture.

Changes in keratocyte gene expression within hCSSC and hCF in 3D culture at 9 weeks in comparison to culture on single silk films relative to hCSSC and hCF at day 0. At each time point, RNA was extracted and reverse-transcribed for RT-qPCR. RNA expression of each gene was first normalized against an endogenous reference gene (18S) and then related to the normalized expression level of the target gene at day 0 per each cell type. Keratocan, lumican, PDK4, PTDGS, and ALDH3A1 were significantly up-regulated for 3D hCSSC, suggesting the enhanced keratocytic phenotype in 3D cultures in comparison to 2D culture and hCF. * Significant effect of cell type (p<0.05); * significant effect of culture (p<0.05).

Fig 6. Tissue equivalent functional properties at 9 weeks in culture.

A. Mechanical properties: Uniaxial tensile properties of hCSSC-seeded stroma equivalent in comparison to hCF-seeded and as made stroma equivalent. B. Optical properties: Transmittance measure of of hCSSC-seeded stroma equivalent in comparison to hCF-seeded and as made stroma equivalent. C. Macro images of each cornea equivalent are reported against written text to show differences in transparency.