Abstract
Purpose:
Limbal epithelial stem cell (LESC) deficiency can lead to blindness. LESC may be transplanted on amnion but this is biologically variable and supply is unreliable. We have previously developed RAFT (a thin collagen construct) and shown that human limbal epithelial cells (hLE) can be expanded on RAFT. RAFT constructs may also be airlifted, enabling stratification of the epithelium, which is further improved by incorporating human limbal fibroblasts (hLF) into RAFT. However, hLF have the potential to activate to a scarring phenotype, incompatible with transplant. The purpose of this study was to determine if normal hLF in RAFT are safe for use.
Methods:
Normal hLF RAFTs were compared with RAFTs containing diseased fibroblasts (dFib) (isolated from scarring conjunctiva, ocular mucous membrane pemphigoid (Oc-MMP)). hLE were seeded onto half of each type of RAFT and cultured for 14 days (submerged) ± 7 days airlifting culture (AL). The scarring phenotype of each RAFT (i.e. hLF vs dFib, ±hLE and ±AL) was assessed using immunohistochemistry (α-sma), zymography (MMP-2 and -9), ELISA (MMP-1, TGF-β) and Sircol assay (collagen production).
Results:
Normal hLF and dFib displayed different phenotypes in RAFT. Very few normal hLF in RAFT expressed α-sma. However, in dFib RAFTs, α-sma expression was markedly increased, particularly in hLE- RAFTs. MMP-2 and -9 were expressed more highly on dFib cf. normal hLF on AL- RAFTs (no difference +AL). Patterns of MMP expression were similar for normal hLF and dFib RAFTs: MMP-9 was only seen on hLE+ RAFTs; MMP-2 was found in all conditions, but was increased on hLE- RAFTs and also by AL. There were no differences in MMP-1 expression between any conditions. Collagen production was greater from dFib cf. normal hLF RAFTs by ~3-fold (p<0.05 all except hLE-, AL-). Collagen production from normal hLF RAFTs was unaffected by hLE or AL, but collagen production from dFib RAFTs was decreased after AL (p<0.05). TGF-β production between normal hLF and dFib RAFTs did not differ but fell for both following AL (p<0.05).
Conclusions:
Normal hLF do not adopt a scarring phenotype in RAFT and so may be safe for transplant. Conversely, dFib maintain a scarring phenotype in RAFT, indicating that RAFT could be utilised as a 3D ex vivo disease model of Oc-MMP.
Keywords: 484 cornea: stroma and keratocytes •
482 cornea: epithelium •
474 conjunctiva