Traditionally testing of hPSC lines for gross chromosomal ch

Traditionally, testing of hPSC lines for gross chromosomal changes employed karyology by chromosome banding of metaphase cftr inhibitor (Amps et al., 2011; Baker et al., 2007). Although karyotyping allows examination of the entire cell genome in a single assay, an often overlooked issue in evaluating hPSC cultures is the sensitivity of karyotyping in the detection of a low-grade mosaicism. Clinical cytogeneticists have an established set of criteria for the number of metaphases that need to be screened to detect the presence of variant cells in clinical samples with a certain level of confidence. For example, analysis of 30 metaphases excludes 10% mosaicism, whereas analysis of 50 metaphases excludes 6% mosaicism, both with 95% confidence (Hook, 1977). However, such calculations are based on a statistical random sampling of a homogeneous population and there is a question of whether this is applicable to hPSCs, given that they grow in adherent cultures as colonies and may not behave as a homogeneous population when dissociated. Even with the appropriate numbers of metaphases sampled, karyotyping has additional shortfalls. The analysis is limited by the fact that only mitotic cells can be assessed, and it is also labor intense and relatively expensive. Furthermore, karyotyping has a limited resolution of about 5–10 Mb (Shaffer et al., 2013). Although there are occasions when abnormalities of less than 5 Mb could be detected, this is limited to very specific karyotype changes in which the chromosome band size, location (small, clear separation from neighboring bands), and staining intensity allows for such a small change to be visible. The limited resolution may present a problem for assessing the genetic status of hPSC cultures, as some of the most common changes in hPSCs are present at a subkaryotype level. Common structural variants are a gain of 20q11.21 copy-number variant (CNV) (Amps et al., 2011; Lefort et al., 2008; Martins-Taylor et al., 2011) or loss of 10p13-pter, 18q21-qter, and 22q13-qter (Amps et al., 2011). In the International Stem Cell Initiative study, 20q CNV was identified in more than 20% of the 120 lines analyzed, and 22 of the lines harboring 20q CNV appeared normal by karyology (Amps et al., 2011). Thus, detecting this particularly frequent change usually requires the use of alternative methods such as fluorescence in situ hybridization (FISH) on interphase cells. The use of such a labor-intensive and expensive method is limiting the frequency of assessing genetic stability of hPSCs during routine maintenance. Hence, there is a need for a rapid, cost-effective assay that could be employed in common laboratory practice for regular screening of hPSCs for common genetic changes. qPCR offers a rapid alternative method to karyotyping or FISH for detecting copy-number changes (Hoebeeck et al., 2007). Unlike karyotyping, which provides a view of the whole genome of a cell, qPCR and FISH-based methods are target specific and hence typically serve as complementary analyses to genome-wide methods (Gekas et al., 2011; Olde Nordkamp et al., 2009). D\'Hulst et al. (2013) have employed qPCR as a way of rapid detection of common karyotypic changes in murine PSCs. Their assays for a gain of chromosome 8 or a loss of chromosome Y were able to detect abnormal cells when they were present in 10% or more of cells in culture (D\'Hulst et al., 2013). Similarly, Avery et al. (2013) have used a qPCR assay for detecting a gain of chromosome 20q11.21 in hPSCs. However, the limitations of this assay in respect of detecting low-level mosaicism in hPSCs remain unclear. More recently, digital PCR has been developed to allow absolute quantification of DNA and to afford higher sensitivity and precision in detecting mutant alleles or copy-number variation (Vogelstein and Kinzler, 1999).
Results
Discussion For karyotyping, we questioned whether the statistical assumptions currently used in clinical cytogenetic practice also apply to hPSCs due to the possible distorting effects of culturing and sampling conditions of hPSCs on the outcome of these tests. Our results from the mixing experiments confirmed that overall the observed numbers of abnormal hPSCs fall within the expected statistical assumptions, and scoring abnormal variants was not substantially distorted by different growth characteristics of the variant cells. In some experiments at higher percentages of abnormal cells in mosaic cultures (>13%), a few of the observed numbers of abnormal cells were higher than the statistically predicted ones. This overestimation of the number of abnormal cells may be due to a difference in cell-cycle time between normal and variant cells. Indeed, we have previously reported a faster cycling time of aneuploid cells compared with their diploid counterparts (Barbaric et al., 2014). As karyotyping relies on cells arrested in metaphase, a different proliferative activity of cells within a mosaic population may bias the analysis toward the more proliferative cells (Gohring et al., 2011). This interpretation is supported by results from FISH on interphase cells from the same samples, where detected numbers of abnormal cells do not fall outside the upper limit of statistically expected numbers (data not shown). Thus, karyotyping of mosaic cultures entailing variant cells with a high proliferative activity would tend to overestimate, rather than underestimate, the presence of variant cells.