Identifying stem cell locations in the mouse corneal and limbal epithelia by lineage tracing

Abstract

Purpose: The location of stem cells that maintain the corneal epithelium during normal homeostasis remains controversial. Most evidence supports the limbal epithelial stem cell (LESC) hypothesis, which proposes that the corneal epithelium is maintained entirely by LESCs. However, labelled mouse limbal epithelial tissue, transplanted to the corneal limbus of unlabelled mice, failed to colonise the cornea unless the host corneal epithelium was later removed (Majo et al., Nature 456, 250-5, 2008). This is inconsistent with the LESC hypothesis and led the authors to suggest that, during normal homeostasis, stem cells distributed throughout the corneal epithelium itself maintained the tissue. We used an inducible lineage tracing system to test the hypothesis that, without surgery, labelled limbal cells would produce clones of labelled cells that colonised the corneal epithelium during normal homeostasis, as predicted by the LESC hypothesis.

Methods: CAGG-CreER;ROSA26-LacZ and CAGG-CreER;ROSA^mT/mG transgenic mice were injected with tamoxifen at 12 weeks, to label some cells (and their clonal progeny) with b-galactosidase or green fluorescent protein, respectively. CAGG-CreER-mediated labelling is ubiquitous, so not restricted to stem cells. However, only stem cell-derived, labelled clones are expected to be long-lived because other, more differentiated, labelled founder cells will be shed during a chase period. Labelling patterns were compared after chase periods of 0-30 weeks by β-galactosidase staining or confocal microscopy. Clone sizes and their distances from the limbus were measured.

Results: Within a week of labelling, there were labelled cells throughout the corneal epithelium but numbers declined over 10 weeks. After a 6-week chase, stripes of labelled cells emerged from the limbus and extended centripetally, reaching the centre by 14-16 weeks. The proportion of stripes with their most central end at least 1 mm from the limbus increased with chase time. (Preliminary results for 6, 10 and 14 weeks chase: 1/73 vs. 3/89 vs. 25/79; P < 0.001 by chi square test.) Conversely, the percentage of stripes with one end at the limbus did not increase. (42/73 vs. 36/89 vs. 34/79.) Both observations are consistent with the LESC hypothesis.

Conclusions: Labelled clones emerged from the limbus, colonised the corneal epithelium and extended towards the centre. This supports the LESC hypothesis.