Rates of acquisition of de novo mutations in human pluripotent stem cells under different culture conditions

The presence of mutations in human pluripotent stem cells (PSC), whether embryonic stem (ESC) cells, or induced pluripotent stem (iPSC) cells, is a concern for their safe use in therapeutic applications. Indeed, in one case, a potential trial of retinal pigment cells from an autologous iPS cell line was abandoned because the cells carried a mutation of unknown significance (Mandai et al. 2017). Certainly some such variants are likely to have been present in the embryos or somatic cells from which particular PSC were derived and can be classed as 'variants of origin' (Amps et al. 2011). However, the propensity of PSC to acquire genetic variants on prolonged passage poses additional concerns, not only because of the difficulty for their early detection (Baker et al. 2016), but also their selective growth advantages might presage malignant potential (Andrews et al. 2017). In this study, to determine the rate and types of mutation that occur during culture of human PSC, we have used a cloning strategy coupled with whole genome- and RNA-sequencing to compare two well characterized, clinical grade human ES cell lines, as well as the effects of culture in the presence of a Rho kinase inhibitor, now commonly used in routine culture of human PSC, and of culture under low oxygen (5% O2) conditions. We have tested whether the mutations are randomly distributed throughout the genome or clustered in regions related to chromatin structure and gene expression. Our results indicate that the mutation rate is low compared to estimates in somatic or cancer cells, and can be further reduced by culture under low oxygen conditions. Furthermore, the mutational signature of human PSC is similar to that of other cultured human cells, but low oxygen culture does alter the frequency of some types of base pair change associated with oxidative damage. Overall design: For these experiments, we chose two male human ES cell lines, MShef4 and MShef11 that we derived under GMP-like conditions suitable for their potential clinical use in regenerative medicine. During their initial derivation, banking, and characterisation, no karyotypic variants of MShef4 had been observed, and it was designated, a priori, as a genetically stable line. By contrast, several cultures of MShef11, though not those used to produce master and working banks, were found to contain chromosomal aberrations and so this line was designated, a priori, as genetically less stable. To assess the mutation rates in these cells, the strategy we adopted was to produce a clonal subline that was karyotypically diploid, grow that clone for a defined period, and then produce a series of subclones that were subjected to whole genome genetic, epigenetic and transcriptional analysis. In the case of MShef4, we initially isolated clones by single cell deposition. Following cytogenetic and SNP array assessment for the absence of any gross genetic or karyotypic changes at the earliest possible time during their expansion, one clone was selected for further analysis. This was grown under our standard growth-conditions (see Methods for details) for a total of 109 days following initiation, after which 20 subclones were isolated and expanded to provide DNA and RNA for sequencing. In the case of MShef11 we compared mutation rates of cells grown under the standard conditions used for MShef4, with mutation rates of cells grown in either the presence of the Rho Kinase inhibitor, Y27632, or under low oxygen (5% O2) conditions. To achieve this, we produced an initial MShef11 clone that was screened by SNP array and karyotyping to exclude gross genetic variants, and then re-cloned as soon as possible, with single cells being deposited directly for growth under one of the three conditions (Figure 1). These clones were again screened by SNP array and karyotyping as early as possible to exclude gross variants, and genetically normal clones carried forward for further analysis. Two such clones were grown under standard conditions, and two were grown in the presence of the Rho Kinase inhibitor at each passage, for 111 days and 115 days respectively. We then produced 10 subclones from each parent clone for sequence analysis – i.e. 20 subclones for each culture condition but in two cohorts. In the case of low oxygen, two clones were grown for 111 days but we were only able to produce nine viable subclones from one of these. Therefore, a further cohort of 12 subclones was derived from a continuation culture of the first low oxygen clone, but only after it had been removed from low oxygen at 111 days and maintained under standard conditions for a further 28 days Bulk samples are defined as cultures of MShef4 and MShef11 that have been expanded from the same cell bank as was used to create the clones and subclones, but are not themselves clonal cultures, and have been expanded under 'standard' conditions as described in the methods.