We studied four independent hESC lines HES H

We studied four independent hESC lines: HES3, H1, ESI-035, and Shef5. Cultures from these four cell lines that had been reported to contain a 20q11.21 CNV were paired with cultures from previous passages that did not appear to be positive for the CNV by SNP array (Amps et al., 2011). Hereafter, we refer to these lines as CNV and control lines, respectively. Upon receipt of these cell lines, we confirmed the presence and length of the CNV by quantitative PCR (qPCR; Figures 1A and 1B). HES3-CNV contained the longest CNV, followed by H1-CNV, ESI-035-CNV, and Shef5-CNV in descending order. Reference banks of these PF2341066 were prepared and karyotype analysis was performed. Each cell had a normal karyotype when it was received, with the exception of Shef5-CNV, in which one of the X chromosomes was absent (Table S1 available online). Fluorescence in situ hybridization (FISH) analysis for the BCL2L1 locus indicated the presence of the amplicon in all CNV lines and multiple extra copies in HES3 and H1 CNV cells. However, the control HES3 and H1 lines that we received also displayed a degree of mosaicism for the CNV, most likely reflecting the propensity of cells to acquire this CNV and gain a selective advantage (Amps et al., 2011). Nevertheless, as a population, the dosage was much lower than that of CNV cells (average 20q11.21 copies: HES3 control 2.2, HES3-CNV 3.5, H1 control 2.5, and H1-CNV 4.2), enabling culture comparisons (Table S1). All of the cell lines formed teratomas when injected into immunocompromised mice, with no apparent differences in differentiation potential. ESI-035 and HES3 control cells were transfected with HM13, ID1, or BCL-XL expression constructs to generate individual, constitutively overexpressing sublines reflecting the three hESC-expressed genes located within the minimal CNV. The gene BCL2L1 encodes two splice variants: the antiapoptotic BCL-XL and the proapoptotic BCL-XS. Since RNA sequencing data show that BCL-XL is the dominant isoform expressed in hESCs and the only isoform in which protein was detected (Figures S1A and S1B), BCL-XS-overexpressing cells were not generated. BCL-XL serves to relocate the proapoptotic protein BAX away from mitochondria and back to the cytosol, thereby preventing cellular apoptosis (Edlich et al., 2011). In addition, BCL-XL also promotes cell survival by binding to and inhibiting Beclin-1 to inhibit stress-induced autophagy (Maiuri et al., 2007). HM13 is a minor histocompatibility antigen that influences anchorage-independent growth of SW480 cells (Sillars-Hardebol et al., 2012b), whereas the basic-helix-loop protein ID1 has a role in maintaining the self-renewal of mouse ESCs (Ying et al., 2003) and promotes tumor metastasis (Gumireddy et al., 2009). To determine whether the 20q11.21 CNV provides a selective advantage, we compared growth rates for the paired cell lines by counting the total number of cells 4 days after seeding at a density of 8 × 104 cells/cm2 (Figure 2A), a density that reflects the typical seed density during routine cell passage. In all cases, CNV cells displayed a higher population growth rate than control cells, with a collective average of three times as many cells. In addition, CNV cells appeared to be less dependent upon the initial seed density, with average cell counts 5-fold and 7-fold greater than control cells following seeding at the lower densities of 4 × 104 cells/cm2 and 2 × 104 cells/cm2, respectively (Figure S2). This increased growth rate was also mimicked by ESI-035 cells overexpressing BCL-XL, but not by those overexpressing either HM13 or ID1. To further assess the kinetics of the growth advantage in mixed cultures, we performed competition assays using HES3 cells. For this purpose, we mixed GFP-expressing HES3 control cells at a ratio of 9:1 with HES3 control, CNV, or HES3 overexpressing each of the three candidate genes (Figure 2B). The HES3-CNV and HES3-BCL-XL-overexpressing cells rapidly outcompeted the GFP-expressing control cells, with the 9:1 ratio reversed to 1:9 by passage 10, whereas the ratio of green/nongreen cells was maintained in cultures containing control, HES3-HM13, and HES3-ID1 cells. This again highlights the similar characteristics of CNV and BCL-XL-overexpressing cells and demonstrates how rapidly these cells are selected for in mixed cultures. By contrast, small-molecule inhibition of BCL-XL with the compound ABT-263 reduced the high cloning efficiency of CNV cells to levels comparable to those observed in the untreated control cells at a concentration of 250 nM (Figure 2C). In addition, when BCL-XL protein levels were reduced in H1 CNV cells (by induced shRNAi) to levels comparable to those in control cells, we observed a strong reduction in the growth rate of these cells (Figures 2D and S2B). These data further support the notion that BCL-XL confers the growth advantage observed in CNV cells. We also confirmed that CNV cells, which have increased copies of BCL2L1, had higher levels of BCL-XL protein than the control cells, as would be expected if BCL-XL were responsible for the selective advantage (Figure S1B).