Multimodal imaging quality control of epithelia regenerated with cultured human donor corneal limbal epithelial stem cells

Abstract

Current imaging techniques for the characterization of differentiated corneal limbal stem cells are destructive and cannot be used in eye bank for monitoring the regenerated epithelium in culture. We presented a minimally invasive, multimodal, marker-free imaging method for the investigation of epithelia regenerated with cultured human donor corneal limbal epithelial stem cells. Two-photon fluorescence and harmonic generation signals were collected from specimens in culture and used for evaluating the structure and morphology of epithelia cultured on two different bio-scaffolds; in addition, donor human corneal tissues were used as controls. The method provided reliable information on the organization of cellular and extracellular components of biomaterial substrates and was highly sensitive to determine differences between the density packing arrangement of epithelial cells of different biomaterials without relying on inferences from exogenous labels. The present minimally invasive standardized quality control methodology can be reliably translated to eye banks and used for monitoring harvested corneal limbal stem cells growth and differentiation in bioengineered materials.

Figure 1

Pseudo-palisade of Vogt in the hemicornea. (a) Three-dimensional (3D) reconstruction of the hemicornea (TPEF signal) showing the pseudo-palisade of Vogt (scale bar 50 µm). Cross-sections (b and c; 400 × 60 µm) and enface (d and e; 400 × 400 µm) images of the hemicornea at different depths highlighting the intrastromal digital invasion by basal epithelial cells (arrows). This structure was randomly observed in scattered regions the hemicornea. The palisades of Vogt are distinctive features of the adult corneoscleral limbus and represent the specialized microenvironment (niche) of corneal stem cells.

Figure 2

Basal nerve plexus in the hemicornea. (a) 3D reconstruction of the hemicornea (TPEF signal) showing the basal nerve plexus (scale bar 50 µm). (b) The nerve fiber bundles (arrows) run parallel to each other forming thin branches that penetrate into the basal epithelial plane. The corneal epithelium is one of the most highly innervated structures in the human body; proper innervation is necessary for maintenance of normal corneal functions. There is strong evidence from laboratory studies that the nerve fibers provide trophic support to the corneal epithelium.

Figure 3

Corneal stroma keratocytes in the hemicornea. (a and b) 3D reconstruction of the hemicornea (TPEF signal) showing the stromal keratocytes at different depths (scale bar 50 µm). The white box in panel A encloses the anterior stroma ranging between 40 and 90 µm depth (c and e); the white box in panel b encloses the stroma ranging between 120 and 170 µm depth (d and f). The panels c, d and e, f show the cross-section and en face images of the hemicornea stroma respectively. The stromal keratocytes show a dendritic-like morphology with processes connecting the cells to each other forming cellular networks (syncytium; arrows).

Figure 4

Organization of anterior stromal collagen fibers in the human cornea. (Upper row) 3D reconstruction of the anterior 250 µm stroma of the human cornea (forward SHG signal) showing the depth-dependent organization and arrangement of collagen fibers (scale bar 100 µm). Middle row) cross-section images of the corneal stroma; from the left to right, the boxes enclose regions of the stroma at varying depths, ranging from the most anterior stroma underlying the Bowman’s layer to 250 µm depth. (Lower row) corresponding enface images of the stroma. In the most anterior stroma (a), the collagen fibers are arranging in tiny and short bundles densely intertwined at different planes; they arrange in thin and densely packed lamellae intersecting each other across 100- and 150-µm depth (b, c and d); these lamellae become increasingly wider and thicker with increasing depth. From ≥200 µm depth, the collagen lamellae shows a grid-like structure, crossing each other at almost vertical angles (e and f).

Figure 5

Regenerated epithelium on fibrin gel. (Upper row) 3D reconstruction of the fibrin gel (TPEF signal) with cultivated limbal corneal epithelial cells (scale bar 100 µm). The white boxes highlight the characteristics of regenerated epithelium, which are shown in the middle (cross-section images) and lower (enface images) rows at different depths respectively. Middle rows) abnormal stratification of the regenerated epithelium was found in all specimens. (Lower rows) the cells show also variations in shape and size across depth and are not confluent even at the basal plane (arrows). Neither TPEF nor SHG signals could be collected from fibrin scaffolds.

Figure 6

F/B ratio of the hemicornea and control anterior stromal lenticule. Averaged measured forward/backward SHG intensities as a function of depth (from the Bowman’s layer to 50 µm depth) for the hemicornea (black curve) and control anterior stromal lenticules (grey curve). The vertical lines represent ± SD.