Genetic and Epigenetic Interactions in in vivo and in vitro Reprogramming

In cancer pathogenesis and induced pluripotent stem: iPS) cell production, an essential step for reprogramming is acquisition of self-renewal. In hematopoietic cells, HOX genes are partially responsible for self-renewal, and HOX gene dysregulation commonly occurs in acute myeloid leukemia: AML). HOX dysregulation is seen in AML with translocations involving HOX genes themselves: e.g. NUP98-HOXA9) and with other disease-initiating translocations: e.g. MLL translocations and inv(16)). However, HOX genes are also highly expressed in many AML samples without translocations; the mechanism that causes "dysregulation" in these cases is unknown. Whole genome sequencing of 45 de novo AML genomes showed that recurrent mutations in the HOX gene clusters are not responsible for the phenotype. Expression array data from 190 AML cases revealed that while translocations have unique HOX expression patterns, most AML cases predominantly express HOXA5, A9, A10, B2, B3, and MEIS1 in a canonical, highly coordinated pattern that is virtually identical to that found in normal human CD34 cells. HOX gene "dysregulation" in these cases may therefore represent the persistence of a normal, stem cell-specific HOX gene expression pattern that is probably required for self-renewal, and "captured" by mutations that initiate leukemia in hematopoietic stem cells. In vitro reprogramming induces self-renewal with overexpression of a cocktail of transcription factors, often in the form of integrating viruses, which have raised concerns about the genomic integrity of iPS lines. We performed whole genome sequencing of 10 murine iPS lines produced in 3 independent experiments. We found an average of 414 somatic nucleotide variants: SNVs) per iPS clone, with variant allele frequencies suggesting that the mutations occurred at or before reprogramming. In one experiment, four independent iPS clones contained 164 identical variants: 6 protein-coding SNVs, 157 non-coding SNVs and 1 structural variant) that were also found in rare parental cells, suggesting that these rare cells were extraordinarily "fit" for reprogramming. Our data suggest that most of the mutations detected in iPS cells occurred prior to reprogramming and are simply "captured" by cloning; however, some preexisting mutations provide an advantage for reprogramming, and may provide novel insights into the genetic underpinnings of this process.