The contribution of blue and white cells in the peripheral corneal epithelium of each chimeric eye (B
outer) was measured to estimate the proportion of the limbus that was populated by active
LacZ + LSCs. These data are summarized in
Table 2 . For each eye, B
outer was compared with the global composition of the chimera, as estimated by mean %GPI1-B. It was reasoned that if the distribution of
LacZ + cells to chimeric corneal epithelia was unbiased and not affected by
Pax6 genotype, then B
outer minus %GPI1-B would approximate 0. This was true for
Pax6 +/+↔
Pax6 +/+ chimeras in which mean B
outer minus %GPI1-B was 6.81 ± 4.35 (
n = 16). The contribution of
LacZ + cells to
Pax6 +/+↔
Pax6 +/− corneas, however, was significantly less than expected: mean B
outer minus the %GPI1-B was −25.46 ± 3.11;
n = 24 (
t-test comparing control and experimental chimeras:
P < 0.0001). A similar underrepresentation was demonstrated by comparing the mean B
outer/%GPI1-B ratios
(Table 2) . This underrepresentation of
Pax6 +/− cells is evidence for an autonomous requirement for correct
Pax6 dosage in LSCs during normal replenishment of the corneal epithelium.
The underrepresentation of Pax6 +/− cells in the corneal epithelium of chimeras suggests that Pax6 +/− LSCs are autonomously depleted or defective, yet the heterozygous Pax6 +/− X-inactivation mosaics had larger coherent LSC clones than wild-type Pax6 +/+ mosaics. This apparent paradox could be reconciled if Pax6 +/− LSCs are at a competitive disadvantage, compared with wild-type cells, in chimeras, but that this defect does not manifest in XlacZ, Pax6 +/− X-inactivation mosaics because all LSCs in these eyes are heterozygous. Pax6 +/− eyes may have larger LSC clones than wild-type Pax6 +/+ mosaics as a result of expansion of a smaller number of specified LSCs to fill the limbal epithelium and/or reduced cell mixing. The possibility that Pax6 +/− LSCs are at a competitive disadvantage in chimeras predicts that coherent clones of Pax6 +/− LSCs may be smaller than wild-type clones in Pax6 +/+↔Pax6 +/− chimeras. Our analysis of clone size and number produces average values for blue and white clones in each eye. Consequently, any differences between mutant and wild-type cells would be masked, and we cannot test directly whether the mutant clones are smaller. The Pax6 +/+↔Pax6 +/− experimental chimeras had slightly more (smaller) LSC clones per eye (82.92 ± 7.97; n = 19) than the Pax6 +/+↔Pax6 +/+ control chimeras (65.11 ± 7.65; n = 13) of equivalent diameter, but the difference was not significant (t-test: P = 0.09). Thus, although our data suggest that LSC function is controlled autonomously by Pax6 dosage within the LSCs, the data do not yet demonstrate whether Pax6 +/− LSC-coherent clones are fewer or smaller than wild-type clones in chimeras.