Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Mar 23;11(1):1528.
doi: 10.1038/s41467-020-15271-3.

Low Rates of Mutation in Clinical Grade Human Pluripotent Stem Cells Under Different Culture Conditions

Affiliations
Free PMC article

Low Rates of Mutation in Clinical Grade Human Pluripotent Stem Cells Under Different Culture Conditions

Oliver Thompson et al. Nat Commun. .
Free PMC article

Abstract

The occurrence of repetitive genomic changes that provide a selective growth advantage in pluripotent stem cells is of concern for their clinical application. However, the effect of different culture conditions on the underlying mutation rate is unknown. Here we show that the mutation rate in two human embryonic stem cell lines derived and banked for clinical application is low and not substantially affected by culture with Rho Kinase inhibitor, commonly used in their routine maintenance. However, the mutation rate is reduced by >50% in cells cultured under 5% oxygen, when we also found alterations in imprint methylation and reversible DNA hypomethylation. Mutations are evenly distributed across the chromosomes, except for a slight increase on the X-chromosome, and an elevation in intergenic regions suggesting that chromatin structure may affect mutation rate. Overall the results suggest that pluripotent stem cells are not subject to unusually high rates of genetic or epigenetic alterations.

Conflict of interest statement

W.R. is a consultant and shareholder of Cambridge Epigenetix but all other authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Schematic representation of the experimental design.
Schematic representation of the experimental regime, in which MShef4 and MShef11 human ES cell lines were cultured for an extended period under different growth-conditions. Stock cultures of MShef4 and MShef11 were cloned by single-cell deposition to ensure a homogeneous starting point. A clone of each was then recloned to generate Parent Clones that were then cultured under the different growth-conditions for a period exceeding three months—MShef4 clone B8, standard conditions, 109 days (19 passages); MShef11 clones E4 & E7, standard conditions, 111 days (25 passages); MShef11 clones F1 & F4, standard conditions + Y27632, 115 days, (25 passages); MShef11 clone G8 low oxygen, 111 days (25 passages) to give subclone cohort N, and 111 days + 28 days in standard oxygen (25 + 6 passages) to give subclone cohort Q. At the end of the expansion period cultures were recloned again into standard conditions to obtain subclone cohorts (J, O, P, K, L, N and Q, respectively, as shown), from which genetic material was extracted as early as possible (5 passages) for sequencing analysis.
Fig. 2
Fig. 2. Mutation rate and genomic distribution of single-nucleotide variant mutations in different growth-conditions.
Box and whisker plots showing haploid mutation rates and genomic distribution of single-nucleotide variant mutations in subclones of MShef4 (N = 20) and MShef11 (N = 19) grown under standard conditions, and subclones of MShef11 cultured with the rho-kinase inhibitor Y27632 at passage (N = 20) or cultured in low oxygen (N = 21). Boxes represent 25th–75th percentiles; whiskers show the min and max ranges; horizontal lines indicate the median values. a The median mutation rates (shown below each plot) for MShef4 and MShef11, cultured in standard conditions, and MShef11 cultured with the rho-kinase inhibitor Y27632 were not significantly different. However, MShef11 cultured in low oxygen exhibited a ~54% lower mutation rate (unpaired two-tailed t-tests with Welch’s correction; P = < 0.0001). ns: P > 0.05; *P ≤ 0.05; **P ≤ 0.01, ***P ≤ 0.001; ****P ≤ 0.0001. b Across all chromosomes only the X-chromosome mutation rate deviated from the genome-wide rates, being elevated in all growth conditions, although only significant in MShef4 and MShef11 under low oxygen, assessed by independent pairwise two-tailed Mann-Whitney tests comparing the median genome-wide rate in each condition with that of each chromosome (Bonferroni-adjusted P = < 0.007; P = < 0.034). When data from all growth conditions were combined, the elevated X-chromosome mutation rate remained significant (P = < 0.05). ns: P > 0.05; *P ≤ 0.05; **P ≤ 0.01, ***P ≤ 0.001; ****P ≤ 0.0001. c Across coding and non-coding regions, normalised to the genomic DNA content of each class of region, under each growth-condition intergenic DNA showed higher mutation rates than exons and introns. For intergenic DNA: MShef11 standard vs MShef11 low oxygen P < 0.001; For intronic DNA: MShef4 standard vs. MShef11 standard P = 0.004, MShef11 standard vs MShef11 low oxygen P = 0.017; For exonic DNA: MShef11 standard vs MShef11 low oxygen P < 0.0001. Asterisks indicate level of significance between groups assessed by independent pairwise two-tailed Mann-Whitney testing. ns: P > 0.05; *P ≤ 0.05; **P ≤ 0.01, ***P ≤ 0.001; ****P ≤ 0.0001. Source data are provided in Supplementary Data 1.
Fig. 3
Fig. 3. Pattern of base substitutions and mutation signatures.
a Bar and dot chart showing low-resolution mutational spectra and proportions of each of seven point mutation classes detected in MShef4 standard (N = 20), MShef11 standard (N = 19), MShef11 + Y27632 (N = 20), and MShef11 low oxygen (N = 21) subclones. Bars indicate the mean proportion of each class of mutation for each cell line/growth-condition group; dots indicate individual subclones in each group. C > A transversions and C > T transitions are most prevalent in the data. A two-tailed t-test showed MShef11 cultured in low oxygen to have a significant reduction in C > A transversions compared to standard conditions (P = 0.037) whilst Wilcoxon testing also showed a small increase in T > C transitions (P = 0.045). b High-resolution mutational spectra derived from the combined mutation data all subclones grown in each growth-condition. Each of the six possible point mutations is subdivided into 16 classes on the basis of the 5’ and 3’ nucleotides flanking the mutation, resulting in 96 possible substitution classes. C:G > A:T and C:G > T:A mutations are most prevalent in the data. These spectra can be correlated with 30 mutation signatures annotated in the Catalogue of Somatic Substitutions in Cancer (COSMIC) database to explore the aetiology of mutation. c Dendrogram showing the similarity of all growth-condition groups based on their mutational profiles. MShef4 and MShef11 cultured in standard conditions exhibited the most similar mutational profiles, whereas the low oxygen condition is the most dissimilar in mutation profile compared to all other groups. Source data are provided in Supplementary Data 2.
Fig. 4
Fig. 4. Mutation rate and genomic distribution of insertion and deletion mutations in different growth-conditions.
a Box and whisker plots showing haploid mutation rates all INDELs in subclones derived from clones grown in different conditions. MShef11 subclones derived from low oxygen (N = 21) showed a significantly reduced INDEL mutation rate compared to subclones from standard conditions (N = 19) (assessed by unpaired two-tailed t-test with Welch’s correction; P = < 0.02). Taken together, the mutation rates of INDELs were approximately 10-fold lower than the mutation rates of single-nucleotide variants. Boxes represent the 25th–75th percentiles of the data; whiskers show the min and max range of the data; horizontal lines indicate the median value. ns: P > 0.05; *P ≤ 0.05; **P ≤ 0.01, ***P ≤ 0.001; ****P ≤ 0.0001. b Box and whisker plots showing haploid mutation rates of insertion, deletion and complex (mixed insertion and deletion) mutations in cells grown in different conditions. MShef11 subclones derived from low oxygen (N = 21) showed a significantly reduced number of deletions compared to those from standard conditions (N = 19) (assessed by unpaired two-tailed t-tests with Welch’s correction; P = < 0.0008). The mutation rates of complex INDELs (mixed insertion and deletion) were 10-20X lower than the rates of insertions or deletions across all cell lines and growth conditions. Boxes represent the 25th–75th percentiles of the data; whiskers show the min and max range of the data; horizontal lines indicate the median value. ns: P > 0.05; *P ≤ 0.05; **P ≤ 0.01, ***P ≤ 0.001; ****P ≤ 0.0001. Source data are provided in Supplementary Data 3.
Fig. 5
Fig. 5. Relationship between gene size, expression, and mutation.
a Dot plot showing the mutational load of genes in relation to their size. Mutated genes from all growth-condition groups were binned by size into 20 kb bins. Dots represent individual bins containing genes of increasing size; the y-axis indicates the mean number of mutations per base-pair for each bin (natural log scale). Mutational load decreased with increasing gene size. b Density plot showing the distribution in gene size for the 10% of genes with highest mutation load (dashed grey line) compared to an equal number of the least-mutated genes (dashed black line) and to all mutated genes (solid black line). Genes with a high mutational burden were small compared to the overall size distribution and were significantly smaller than genes with a low mutational burden. c Dot plots showing the percentage of mutated genes per subclone with occurrences of mutation at increasing distances from transcription start sites (TSS). The x-axis denotes distance from the TSS (kb); the y-axis shows the percentage of genes within each bin that harbour mutation. The data show increasing mutation prevalence with increased distance from TSS. d Scatter plots showing of the relationship between mutational load (y-axis) and gene expression (x-axis) in MShef4 and MShef11 under all conditions (natural log scale). Plots indicate a weak positive correlation between expression and mutation. Solid coloured lines indicate the fitted linear regression models for each group; shaded regions around regression lines indicate 95% confidence intervals for the models. Spearman correlation test statistics and P-values are shown for each group (MShef4 standard:.rho = 0.24; MShef11 standard: rho = 0.22; MShef11 + Y27632: rho = 0.28; MShef11 low oxygen: rho = 0.2). Source data are provided in Supplementary Data 4 and 5.
Fig. 6
Fig. 6. Epigenetic changes.
a Beanplots showing global DNA methylation levels of MShef11 starting cultures, parental clones grown in standard or low oxygen conditions, and the subclones derived from those conditions. Methylation was assessed over tiled probes across the genome, each containing 100 CpGs. Both parental clones grown in low oxygen show a small decrease in methylation, particularly clone G8. This decrease in methylation was reversible, as subclones derived from low oxygen revert to normal methylation levels following expansion in standard conditions. Source data are provided in Supplementary Data 6. b DNA methylation of promoters in MShef4 subclones (N = 20). Promoter regions were defined as regions surrounding transcription start sites, with 1500 kb upstream and 500 kb downstream context. Beanplots (left panels) show DNA methylation levels of promoters with and without CpG islands, in a randomly selected subset of MShef4 J subclones. For both classes of promoter, the starting culture and the parental clone, B8, derived from that starting culture had comparable methylation levels, whereas each J subclone showed hypermethylation of a subset of CpG island-containing promoters compared to its parental clone, B8. Density scatter plots (right panels) show the distribution of promoter methylation levels between the MShef4 starting culture and its derived parental clone, B8, and between the parental clone B8 and an example subclone, J1. The density scatter plots show comparable levels of non-CpG island-promoter methylation between starting culture material and parental clone B8, but hypermethylation of CpG island-containing promoters in subclone J1 compared to parental clone B8. Source data are provided in Supplementary Data 9–12. c Top panel; heatmap showing the relative expression of de novo methyltransferase (DNMT) genes in MShef4 (N = 20) and MShef11 (N = 19) subclones from standard conditions. Lower panel; box and whisker plots showing the normalised expression counts (fragments per million reads) for DNMT genes in MShef4 and MShef11 standard subclones. Both DNMT3B and DNMT3L are more highly expressed in MShef4 subclones (P = 7.2−8 and P = 2−8). Boxes represent the 25th–75th percentiles of the data; whiskers show the min and max range of the data; horizontal lines indicate the median value. ns: P > 0.05; *P ≤ 0.05; **P ≤ 0.01, ***P ≤ 0.001; ****P ≤ 0.0001. Source data are provided in Supplementary Data 13.
Fig. 7
Fig. 7. Imprint methylation and expression.
a Heatmap showing the methylation levels of 26 imprint control regions (ICRs) in starting cultures, parental clones, and subclones, grouped by cell line and growth-condition. Individual ICRs show a wide range of methylation levels with some notable inter- and intra-group variation. Source data are provided in Supplementary Data 14. b Top panel; heatmap showing the relative expression of four selected imprinted genes: FAM50B, H19, ZIM2, and MIMT1 in MShef4 and MShef11 subclones. Both FAM50B and H19 are more highly expressed in MShef4 subclones; ZIM2 and MIMT1 are more variable in their expression. Lower panel; dot plots showing the correlation between ICR methylation and the expression of four imprinted genes, in a subset of MShef4 and MShef11 subclones with WGBS data of sufficient depth to permit such analysis (selected subclones indicated by asterisks in top panel). In all cases differences in gene expression correlate with a change in ICR methylation. Source data are provided in Supplementary Data 15.

Similar articles

See all similar articles

References

    1. Mandai M, Kurimoto Y, Takahashi M. Autologous induced stem-cell-derived retinal cells for macular degeneration. N. Engl. J. Med. 2017;377:792–793. doi: 10.1056/NEJMc1706274. - DOI - PubMed
    1. Amps K, et al. Screening ethnically diverse human embryonic stem cells identifies a chromosome 20 minimal amplicon conferring growth advantage. Nat. Biotechnol. 2011;29:1132–1144. doi: 10.1038/nbt.2051. - DOI - PMC - PubMed
    1. Rouhani FJ, et al. Mutational history of a human cell lineage from somatic to induced pluripotent stem cells. PLoS Genet. 2016;12:e1005932. doi: 10.1371/journal.pgen.1005932. - DOI - PMC - PubMed
    1. Baker D, et al. Detecting genetic mosaicism in cultures of human pluripotent stem cells. Stem Cell Rep. 2016;7:998–1012. doi: 10.1016/j.stemcr.2016.10.003. - DOI - PMC - PubMed
    1. Andrews PW, et al. Assessing the safety of human pluripotent stem cells and their derivatives for clinical applications. Stem Cell Rep. 2017;9:1–4. doi: 10.1016/j.stemcr.2017.05.029. - DOI - PMC - PubMed
Feedback