Purpose: Amplification of cyclin E1 (CCNE1) is associated with poor outcome in breast, lung, and other solid cancers, and is the most prominent structural variant associated with primary treatment failure in highgrade serous ovarian cancer (HGSC). We have previously shown that CCNE1-amplified tumors show amplicon-dependent sensitivity to CCNE1 suppression. Here, we explore targeting CDK2 as a novel therapeutic strategy in CCNE1-amplified cancers and mechanisms of resistance.Experimental Design: We examined the effect of CDK2 suppression using RNA interference and smallmolecule inhibitors in SK-OV-3, OVCAR-4, and OVCAR-3 ovarian cancer cell lines. To identify mechanisms of resistance, we derived multiple, independent resistant sublines of OVCAR-3 to CDK2 inhibitors. Resistant cells were extensively characterized by gene expression and copy number analysis, fluorescence-activated cell sorting profiling and conventional karyotyping. In addition, we explored the relationship between CCNE1 amplification and polyploidy using data from primary tumors.Results: We validate CDK2 as a therapeutic target in CCNE1-amplified cells by showing selective sensitivity to suppression, either by gene knockdown or using small-molecule inhibitors. In addition, we identified two resistance mechanisms, one involving upregulation of CDK2 and another novel mechanism involving selection of polyploid cells from the pretreatment tumor population. Our analysis of genomic data shows that polyploidy is a feature of cancer genomes with CCNE1 amplification.Conclusions: These findings suggest that cyclinE1/CDK2 is an important therapeutic target in HGSC, but that resistance to CDK2 inhibitors may emerge due to upregulation of CDK2 target protein and through preexisting cellular polyploidy.

To identify patients who developed secondary clonal cytogenetic aberrations (CCA) following therapy for acute promyelocytic leukemia (APL), we retrospectively analyzed cytogenetic results from 123 patients diagnosed with APL between 1995 and 2007, who had ongoing cytogenetic analysis undertaken in our laboratory. During follow-up for APL we identified 12 patients (9.8%) who developed CCA, not detected at diagnosis of APL and unrelated to their original APL karyotype. All patients had received all-trans retinoic acid (ATRA) and chemotherapy and were in complete remission for APL when secondary CCA were identified. The median latency period between diagnosis of APL and emergence of secondary CCA was 27.5 months (range: 2-54 months). To date, four patients with CCA have been diagnosed with therapy-related myelodysplastic syndrome (t-MDS)/acute myeloid leukemia (t-AML), giving a median t-MDS/AML free survival of 78 months, with follow-up ranging between 20 and 136 months from APL diagnosis. Three patients have died: two patients died of t-AML and another developed relapsed APL with persistence of his secondary clone but no diagnosis of t-MDS/AML and died from transplant-related complications. Two patients are alive with t-MDS. Seven patients with CCA are alive with no morphological evidence of MDS at the time of their last known follow-up; thus median survival has not been reached. The appearance of these abnormalities in the absence of morphological evidence of MDS in the majority of patients is unusual, and highlights the importance of continued cytogenetic follow-up in these patients. Am. J. Hematol., 2009. (c) 2009 Wiley-Liss, Inc.