You are here

| Pluripotent Stem Cells

Genetic Instability in Induced Pluripotent Stem Cells: One Step Forward in Understanding, Two Steps Back from the Clinic?



Review by Stuart P. Atkinson

Recent studies in embryonic stem cell (ESC) and induced pluripotent stem cell (iPSC) biology have turned from comparative studies at the RNA and chromatin level to focus on what’s happening at the DNA level (See iPSC don´t Forget their Origins and Another Blow to the iPSC Field?). Genomic integrity is of course vital for the future use of such pluripotent cells in the clinic; changes at the genomic level can, at best, lead to failure of transplanted cell function, and in the worst case potentially lead to tumorigenesis. This review hopes to condense some of the recent high impact papers which have studied genomic alteration in iPSC in great detail.

Mayshar et al. first analysed the chromosomal integrity of human ESCs and iPSCs (hESCs and hiPSCs) through transcriptional means by identifying large regions where gene expression is co-ordinately increased or decreased, therefore identifying regions where duplication or insertion events have taken place. Using 104 unique gene expression profiles from 17 hESC lines (38 samples) and 46 iPSCs lines from 17 different fibroblastic origins (66 samples) they discovered that 12 of the 38 hESC lines had chromosomal abnormalities, with 8 of the observed aberrations involving entire chromosomes or large chromosomal regions. The most recurrent duplications occurred on chromosomes 12 and 17, previously identified as common aberrations in embryonic tumours and arising during ESC adaption (Baker et al. and Reuter). Studies in hiPSCs found abnormally high overexpression of genes on chromosome 12 in one clone at passage 14 with a further increase at passage 31, suggesting that the gain of chromosome 12 is positively selected for during extended passaging following a period of genomic mosaicism. Another clone showed evidence of gains in chromosome 3 and 12 at passage 56, whereas at passage 9 it was found to be normal. Chromosome 12 is of particular interest as it contains the pluripotency associated genes NANOG and GDF3, and the iPSCs gaining chromosome 12 showed the specific upregulation of these, but not other pluripotency associated genes. However, a significant enrichment of genes associated with a specific function was not observed for other genes present on chromosome 12 with upregulated gene expression. Overall, around 20% of the hiPSC lines studied contained genomic aberrations, with around 9% carrying at least one full trisomy.

This study was followed by another (Laurent et al.) which undertook a more specific and detailed analysis. High resolution single nucleotide polymorphism (SNPs) analysis was performed across a staggering range of cells; 69 hESC lines, 37 iPSC lines, 11 somatic cell lines, 41 primary cell lines and 20 tissue types, which showed that there are large areas of copy number variations (CNVs) in both hESC and hiPSCs including trisomy of chromosomes 3, 12 and 17 and the deletion of the long arm of chromosome 7. This is in agreement with the previous study which also discovered duplication events for chromosomes 3, 12 and 17. There also existed smaller regions of CNVs including both insertions and deletions, which were further studied to find recurrent duplications in pluripotent cells. Regions on chromosome 12 and 20 were identified as containing recurrent CNVs. Regions of chromosome 12 were duplicated in 9 of the 69 hESC lines, which included the NANOG pseudogene NANOGP1 and the glucose transporter SLC2A3 (GLUT3), previously linked to tumorigenesis (Macheda et al.). A region of chromosome 20 was duplicated in 7 of the 69 hESC lines and included the DNMT3B gene, which has been strongly linked to pluripotency. The duplication of a region near DNMT3B linked to carcinogenesis was also identified in 5 of 69 hESC lines, while NANOG itself lies close to the NANOGP1 pseudogene but outwith the region of duplication. This suggests that these areas, lying near to pluripotency-associated genes may be under positive selection in culture, or that upstream regulatory elements may lie in these areas. Further analysis found a large number of duplications in pseudogenes of pluripotency associated genes such as POU5F1 (OCT4) and NANOG which are known, alongside other genes active during embryogenesis, to have large numbers of pseudogenes (including GDF3 and DPPA3 (STELLA)). It has been suggested that pseudogenes may have an important role in regulating their “parent” gene, and this may be the selection pressure behind these duplications. They also may act as a buffer against differentiation-associated microRNA (miRNA) function.


Interestingly, comparisons of CNVs in hESCs with hiPSCs and non-pluripotent cell types suggested that the average numbers of duplication and deletion events is higher in hiPSCs than in non-pluripotent cell types. It also showed that a small number of the hESC lines had a very large number of duplication events, implying that CNVs, a rare event in hESCs, may be more likely to arise in hiPSCs, perhaps due to the reprogramming process. The next logical step was to analyse the genomic aberrations that arise during iPSC generation and passaging. Comparison of parental fibroblasts with the iPSC generated from them allowed the identification of 11 deletions at the earliest passage tested (5-8) while 5 of 6 duplications arose during long-term culture. The authors do note that some deletions were ‘lost’ during culture, suggesting that while some deletions may be positively selected for during reprogramming, others may be under negative selection. Duplications also tended to arise in the coding sequences and/or promoter regions of coding genes that are positively associated with tumorigenicity or cell proliferation. Interestingly, the most rapidly arising genomic aberrations were identified in samples from directed hESC differentiation experiments, with some duplications arising between days 2 to 7 of differentiation. These duplications were linked to genes more highly expressed in pluripotent cells.

In the next study, Pasi et al. initially used the lentiviral transduction of mouse mammary progenitor cells, which lack stem cell properties, with a MycER construct, in order to induce mammary stem cells as a model system. Transduced cells had the ability to generate mammospheres, indicative of the attainment of a stem cell phenotype. Interestingly, when 4-hydroxytamoxifen (TAM) was added and nuclear MycER protein levels increased, mammosphere cultures were exhausted in one passage, and this was associated with increased phosphorylation and stabilisation of p53, expression of p53 target genes and apoptosis. This suggests that high levels of MycER resulted in a DNA damage response due to replication induced stress, and also that even at low levels (i.e. without TAM); MycER could induce some genomic instability. To address this point, mammary stem cells were induced and grown for 9 weeks without TAM, and then single cells were selected and clones grown for another 3-6 weeks followed by comparative genomic hybridization (cGH) analysis using high density arrays to around one quarter of the mouse genome. 4 of the 8 clones studied contained focal copy number changes (CNCs). Clone 1 had a 100kbp deletion in the Rora gene which maps to a common fragile site (CFS) in humans. CFSs are a major target for genomic rearrangements in oncogene-expressing and pre-neoplastic cells, and have been previously linked to replicative stress. Clone 2 had a 250kbp deletion within the Pde4d gene, Clone 3 had a 200kbp deletion in the Ptprg gene (which again maps to a human CFS) while Clone 4 had an amplifications of the Jarid2 and Dtnbp1 genes, Jarid2 having a known role in stem cell self-renewal (Landiera et al.). The authors suggest that the frequency of aberrations observed in their study, which covered only one quarter of the mouse genome in 8 clones, suggests that reacquisition of stemness in reprogramming may be in part due to genomic rearrangement.

These data suggested to the authors that genomic instability may be widespread in iPSC generated by standard protocols. Initial reprogramming of cells to pluripotency with three factors (without c-Myc) led to the generation of 6 clones, two of which gained signs of genomic instability (small insertions in the Odz3 and Cdh13 gene). Reprogramming with 4 factors (including c-Myc) led to the generation of 4 clones, two diploid and two aneuploid. Of the diploid clones, the first had a large amplification in chromosome 7 and a deletion affecting the Fto gene. Next the authors studied genomic lesions induced by replicative stress, as some of the sites of genomic instability observed within the study overlapped with CFSs (such as the Rora and Ptprg genes). This was established by the treatment of a mouse-human hybrid cell line (GM11713A, containing a single human chromosome 3) with aphidocolin for three weeks to induce a replicative stress response. This led to the discovery of deletions in the FHIT gene at a CFS in 4 of the 5 clones and deletions within the PTPRG gene at the same CFS in 2 of the 5 clones. Further deletions were also found, none of which mapped to CFSs. This suggests that CNCs found in induced stem cells are likely to be due to DNA replication stress and that genomic instability is an inherent characteristic of reprogramming across species.

Similar studies were published in Nature. Hussein et al. used a high-resolution single nucleotide polymorphism array to analyse CNVs of 22 human iPSC lines, 17 hESC lines and 3 fibroblast lines. Successful generation of iPSC had previously been demonstrated using retroviral or piggyback mediated gene delivery and initial analyses showed that these iPSC had roughly double the number of CNVs than hESC and fibroblasts, and that the majority of the CNVs in iPSC were unique. Interestingly, the passage number of iPSCs was correlated with a decrease in the number of CNVs, something not observed in hESC or fibroblasts, and was not linked to method of reprogramming, fibroblast source, viral integration or presence/absence of Myc. The authors suggest two reasons for this interesting correlation. A DNA damage and repair response could allow for the decrease in CNVs with increasing passage number, but the authors suggest that this method may not be efficient enough for the speed of CNV loss observed. The other reason, which the authors propose strongly, is an initial mosaicism of iPSCs in culture followed by selection for those cells containing little or no CNVs. Epigenetic differences and differentiation bias between different iPSC have also been observed to disappear over time in culture and with passage (Polo et al.) perhaps due to a similar phenomenon. To test this hypothesis, “new” iPSC were generated and tested at passage 2 or 3 for CNVs with fluorescent probes and, indeed, more aberrations were found in early passage iPSCs than in late passage iPSCs or fibroblasts. For evidence of selection they studied regions with homozygous deletions, which cannot be corrected, and discovered that these regions were selected against during passaging. This finding is slightly at odds with the other studies in which aberrations tended to be positively selected for during passage.

The next stage of the study was to identify novel CNVs in early-passage iPSC. When compared against what were deemed to be neutral CNVs from 270 healthy individuals from two combined studies in the HapMap project (McCarroll et al. and Conrad et al.), it was discovered that 37% of the CNVs in iPSCs lay outside this neutral set, while this was reduced to 25% in hESC and 15% in fibroblasts. The study went on to show that these CNVs are likely to arise due to the reprogramming process and that early-passage hiPSC undergo a strong selective pressure to rid of them. The obvious next question is what functional consequences these CNVs may have. Several deletions observed at early passage were observed at regions known to be important for the maintenance of the undifferentiated state (EGFR, FGF2, CTNNB1 and the miRNAs MIRLET7C and MIR125b) and also at Polycomb-bound regions. Such regions are known to be vitally important to the control of differentiation-associated genes and are likely to be bivalent with regards to their chromatin structure. Genomic aberrations at these sites could lead to negative or positive expression of such genes and therefore provide the positive or negative selective pressure. Further, novel CNVs in iPSCs were again linked to replicative stress as a high proportion of novel iPSC CNV sites overlapped with known CFSs which, as mentioned before, have been previously linked to replicative stress. Telomeric regions are also known for the relative fragility and selection against large deletions in these areas was also observed in iPSCs.

The other Study published in Nature (Gore et al.) began by sequencing the protein coding exons (exomes) of 22 hiPSC lines that were reprogrammed in 7 different laboratories using three integrative and two non-integrative methods, and nine matched fibroblasts, allowing for the discovery of single base changes, small insertions/deletions and alternative splicing variants. This led to the discovery of 124 mutations in the iPSC lines, but no small insertions or deletions, with the majority of the mutations being mis-sense (83 of 124), nonsense (5 of 124) or splice-variants (4 of 124). 53 of the mis-sense mutations have been predicted to alter protein function and 50 of these genes have been found to be mutated in some cancers (e.g. ATM, NTRK1 and NTRK3), while 14 of the 22 lines studied contained mutations with known roles in human Mendelian disorders. Through detailed cell and sequence analysis, the authors go on to propose that reprogramming-associated mutations can arise through several mechanisms and suggest while it is likely that some mutations are present in the starting fibroblasts at low frequency, most mutations occur during reprogramming and subsequent culture. Interestingly, in this study 7 of the mutations observed at low passage (in this case P9) remained fixed at passage 40, with an additional 4 mutations fixed at P40.

Overall, these studies paint a clear picture. While some aberrations arise in cultured hESCs and some are inherent in fibroblasts, many are a result of the replicative stress which arises during the reprogramming process of fibroblasts to iPSCs. This suggests that iPSCs are more likely to be tumorigenic and cells differentiated from these iPSCs may have a reduced functional capability. Alongside gene expression and chromatin studies, these genomic studies provide more evidence to suggest that iPSCs may not be equivalent to ESCs after all. But we also know that individual ESC lines vary from each other, so how similar do iPSCs need to be in order to be deemed useful for the clinic? Perhaps the next step in such comparative studies should be the start of wide-scale, long-term functional tests of cells differentiated from ESC and iPSC. Maybe this will tell us whether iPSCs, as we know them today, are destined for use in the clinic. Or perhaps not. Perhaps these studies suggest that we need to step back from proving the immediate clinical value of iPSCs and, for the time being, concentrate on the basic molecular mechanisms behind the reprogramming process and iPSC cell biology. This may allow safer, more reproducible and efficient reprogramming of somatic cells and their informed progression to the clinic for patient-specific treatment.



Identification and classification of chromosomal aberrations in human induced pluripotent stem cells.
Mayshar Y, Ben-David U, Lavon N, Biancotti JC, Yakir B, Clark AT, Plath K, Lowry WE, Benvenisty N.
Cell Stem Cell. 2010 Oct 8;7(4):521-31.

Adaptation to culture of human embryonic stem cells and oncogenesis in vivo.
Baker DE, Harrison NJ, Maltby E, Smith K, Moore HD, Shaw PJ, Heath PR, Holden H, Andrews PW.
Nat Biotechnol. 2007 Feb;25(2):207-15

Origins and molecular biology of testicular germ cell tumors.
Reuter VE.
Mod Pathol. 2005 Feb;18 Suppl 2:S51-60.

Dynamic changes in the copy number of pluripotency and cell proliferation genes in human ESCs and iPSCs during reprogramming and time in culture.
Laurent LC, Ulitsky I, Slavin I, Tran H, Schork A, Morey R, Lynch C, Harness JV, Lee S, Barrero MJ, Ku S, Martynova M, Semechkin R, Galat V, Gottesfeld J, Izpisua Belmonte JC, Murry C, Keirstead HS, Park HS, Schmidt U, Laslett AL, Muller FJ, Nievergelt CM, Shamir R, Loring JF.
Cell Stem Cell. 2011 Jan 7;8(1):106-18.

Molecular and cellular regulation of glucose transporter (GLUT) proteins in cancer.
Macheda ML, Rogers S, Best JD.
J Cell Physiol. 2005 Mar;202(3):654-62. Review

Genomic instability in induced stem cells.
Pasi CE, Dereli-Öz A, Negrini S, Friedli M, Fragola G, Lombardo A, Van Houwe G, Naldini L, Casola S, Testa G, Trono D, Pelicci PG, Halazonetis TD.
Cell Death Differ. 2011 Feb 11.

Inactive yet indispensable: the tale of Jarid2.
Landeira D, Fisher AG.
Trends Cell Biol. 2011 Feb;21(2):74-80

Copy number variation and selection during reprogramming to pluripotency.
Hussein SM, Batada NN, Vuoristo S, Ching RW, Autio R, Närvä E, Ng S, Sourour M, Hämäläinen R, Olsson C, Lundin K, Mikkola M, Trokovic R, Peitz M, Brüstle O, Bazett-Jones DP, Alitalo K, Lahesmaa R, Nagy A, Otonkoski T.
Nature. 2011 Mar 3;471(7336):58-62.

Cell type of origin influences the molecular and functional properties of mouse induced pluripotent stem cells.
Polo JM, Liu S, Figueroa ME, Kulalert W, Eminli S, Tan KY, Apostolou E, Stadtfeld M, Li Y, Shioda T, Natesan S, Wagers AJ, Melnick A, Evans T, Hochedlinger K.
Nat Biotechnol. 2010 Aug;28(8):848-55.

Integrated detection and population-genetic analysis of SNPs and copy number variation.
McCarroll SA, Kuruvilla FG, Korn JM, Cawley S, Nemesh J, Wysoker A, Shapero MH, de Bakker PI, Maller JB, Kirby A, Elliott AL, Parkin M, Hubbell E, Webster T, Mei R, Veitch J, Collins PJ, Handsaker R, Lincoln S, Nizzari M, Blume J, Jones KW, Rava R, Daly MJ, Gabriel SB, Altshuler D.
Nat Genet. 2008 Oct;40(10):1166-74.

Origins and functional impact of copy number variation in the human genome.
Conrad DF, Pinto D, Redon R, Feuk L, Gokcumen O, Zhang Y, Aerts J, Andrews TD, Barnes C, Campbell P, Fitzgerald T, Hu M, Ihm CH, Kristiansson K, Macarthur DG, Macdonald JR, Onyiah I, Pang AW, Robson S, Stirrups K, Valsesia A, Walter K, Wei J; Wellcome Trust Case Control Consortium, Tyler-Smith C, Carter NP, Lee C, Scherer SW, Hurles ME.
Nature. 2010 Apr 1;464(7289):704-12.

Somatic coding mutations in human induced pluripotent stem cells.
Gore A, Li Z, Fung HL, Young JE, Agarwal S, Antosiewicz-Bourget J, Canto I, Giorgetti A, Israel MA, Kiskinis E, Lee JH, Loh YH, Manos PD, Montserrat N, Panopoulos AD, Ruiz S, Wilbert ML, Yu J, Kirkness EF, Izpisua Belmonte JC, Rossi DJ, Thomson JA, Eggan K, Daley GQ, Goldstein LS, Zhang K.
Nature. 2011 Mar 3;471(7336):63-7.


Also See

iPS cell aberrations
Natalie de Souza
Nature Methods 7, 948–949 (2010)

Stem cells: Out for the count
Nicola McCarthy
Nature Reviews Cancer 10, 742-743 (November 2010)

Genomic instability in iPS: time for a break.
Blasco MA, Serrano M, Fernandez-Capetillo O.
EMBO J. 2011 Mar 16;30(6):991-3.

Stem cells: The dark side of induced pluripotency.
Pera MF.
Nature. 2011 Mar 3;471(7336):46-7. No abstract available.