Abstract
Despite the major progress obtained in prognosis with the use of tyrosine kinase inhibitors (TKI), the great majority of patients with CML remain on long-term therapy and progression occurs in patients with either primary or secondary resistance. The mechanisms of progression towards accelerated phase (AP) and blast crisis (BC) have been studied so far only in primary patient samples in BC. Currently, there is no in vitro model to study sequentially the molecular events leading from CP to BC as only some primary sequential samples are amenable to analysis. Using induced Pluripotent Stem Cell (iPSC) technology, it is now possible to reprogram the primary leukemic cells to pluripotency and generate a major source of stem cells. To determine the feasibility of studying progression of CML towards AP and BC, we have used a patient-specific iPSC line that we have generated from the primary leukemic cells of a patient who later showed a TKI-resistance requiring an allogeneic stem cell transplant. These iPSC expressed BCR-ABL, exhibited all pluripotency markers and after injection in NSG mice, generated teratoma with differentiation into three germ layers. In hematopoietic differentiation assays using day 19-embryoid bodies (EB), increased numbers of hematopoietic progenitors were found as compared to control iPSC (5-fold increase n= 3). We have then treated leukemic iPSC with the mutagenic agent N-ethyl-N-nitrosourea (ENU) during regular medium changes. After 61 days in ENU cultures, day-19 derived embryoid bodies generated hematopoietic cells (>90% CD45+, CD43+) which proliferated in liquid cultures with myeloid and some blast cell morphology. Cytogenetic analyses of iPSC revealed chromosomal abnormalities such as loss of Y and loss of der q9+, both alterations known to occur in CML during progression. They exhibited increased numbers of micronuclei (MN) as compared to leukemic iPSC without ENU (X 3 Fold increase) suggesting acquisition of a progressive chromosomal instability. CGH array analyses were performed using ENU-treated iPSC-derived hematopoietic cells in two different timepoints as compared to leukemic iPSC cultured without ENU. Genomic aberrations were analyzed by Agilent Cytogenomics software with Mosaicism workflow on HG19 genome. 249 gene loci alterations were detected after polymorphism filtration on European population. These analyses showed DNA losses and DNA gains in genes implicated in mesoderm development and hematopoietic lineage as well as genes implicated in DNA damage response. Several loci of transcription factors were found to be involved such as IKZF1 described in imatinib-refractory chronic myeloid leukemia (Bolton-Gillespie et al. 2013). The aberrations included SIRT1and BLM which is implicated in DNA repair. Several cancer genes were found to be involved, some known to be altered in leukemia (BLM, IKZF1, NCOA2, ALK, EP300, ERG, MKL1, PHF6 and TET1). Remarkably, transcriptome geodataset GSE4170 (Radich et al. 2006) allowed us to associate 125 of 249 of the aberrations that we detected in CML iPSC, with the CML progression genes already described during progression from chronic and AP to BC (p-value =9.43E-32, after ANOVA with 1000 permutations). 38 most predictive aberrations allowed perfect reclassification of BC and chronic phase samples by unsupervised classification. Among these candidates, eleven of them have been described in CML physiopathology and connected to TKI resistance and genomic instability. Majority of them ( 7/11) are connected to chronic phase (FAS, ACTB, TRIM21, ANPEP, MLK1, CSF2RA, and MAGEC2) whereas a minority of them (4/11) are connected to BC (ACP1, SH3YL1, FHL1, IL3RA). Interestingly, these experiments also allowed us to discover the connection of a new multidrug resistance molecule associated to BC and having the ability to modify interferon pathway connected to the TKI sensitivity. Thus, genomic instability pattern that we have generated using a single leukemic iPSC allowed duplication of genomic abnormalities described in CML progression and allowed identification of some novel genes. Overall, these results demonstrated that we have generated for the first time to our knowledge, an in vitro BC model, reproducing genomic events described in patients with BC. This "blast crisis in a dish" tool using patient-derived iPSC will be of major interest to study CML progression and eventually to test novel therapies.
No relevant conflicts of interest to declare.
