In-vivo-Analyse von Schlüsselgenen und Resistenzmechanismen in der Therapie der akuten myeloischen Leukämie
Final Report Abstract
The outcome of clinical AML treatment is highly heterogeneous and still less than 30% of patients can be cured with existing therapies. More than 100 disease alleles have been identified in AML and some have been associated with favourable or adverse treatment outcome. The heterogeneity and complexity of AML oncogenomes complicate the establishment of genotype-tailored regimens and augment the challenge of identifying functional drug targets. Genetically defined mouse models provide a tractable experimental system to dissect genotype-response patterns, evaluate rational drug targets and test new compounds in a physiological environment. As key resources for such studies I have (i) developed a panel of genetically defined AML mouse models involving 10 common human AML lesions in various combinations, (ii) integrated bioluminescent and fluorescent imaging and (iii) established a combined chemotherapy regimen that approximates to clinical induction therapy. In a first study I have demonstrated that murine AMLs harboring two common human genotypes, i.e. a combination of oncogenic Nras with either AML1/ETO or MLL/ENL fusion proteins, show remarkably diverse responses, to conventional therapy that precisely mirror clinical experience in patients. Next, we found that the resistance seen in MLL/ENL+Nras leukemia is caused by a dramatic attenuation of p53 drug-response programs. Furthermore we demonstrate that loss of p53 generates drug resistance in the responsive AML1/ETO+Nras context. These studies documented the importance of genetic information, in guiding the treatment of human AML, establish the p53 network as a central determinant of therapy response in AML, and demonstrate that genetically engineered mouse models of human cancer can accurately predict therapy response in patients. Beyond ongoing studies to further dissect genotype-response patterns for known AML mutations, we have started to use in-vivo RNAi to identify new tumor suppressor genes and drug targets in AML. To enable such studies, I initially focused on optimizing the RNAi technology for in-vivo screening approaches. Major improvements include the development of retroviral systems to codeliver ShRNAs with multiple transgenes, the generation of .powerful tet-regulatable RNAi expression systems and the establishment of new shRNA design algorithms yielding more potent shRNAs. In a collaborative effort with Prof. Gregory Hannon, I have pioneered a high-throughput approach to evaluate shRNA potency. After these technological improvements, our lab now has completed two in-vivo RNAi screens resulting in the identification of new tumor suppressor genes in lymphoma and liver cancer, which provide proof-of-principle evidence of for the applicability of this approach in two independent mosaic cancer models. Besides participating in these pioneering projects, I have initiated similar studies focusing on the identification of tumor suppressor genes in frequently deleted genomic regions in treatment resistant AML. In a collaborative effort with Prof. Jim Downing (St. Jude Children's Research Hospital) we have collected high-resolution array-CGH data for 200 AML patients, generated an shRNA library comprising 1600 shRNA targeting 450 commonly deleted genes and initiated several in-vivo RNAi screens in different genetic contexts, which are currently underway. In addition to positive selection screens focused on the discovery of tumor-suppressor and drug resistance promoting genes, I have pioneered strategies that apply in-vivo RNAi to identify and evaluate new drug targets in treatment resistant AML. To facilitate such negative selection RNAi studies in vivo I have further optimized tet-regulatable shRNA expression vectors and generated various chemoresistant tet-ON-competent luciferase+ AML lines. In proof-of-principle experiments I introduced shRNAs targeting essential genes (e.g. factors of the replication machinery) into chemoresistant AML lines and transplanted these into secondary recipients. Following leukemia onset and subsequent doxycycline treatment, we could demonstrate that expression of a single lethal shRNA can induce rapid regression and long-term remission of chemotherapy-resistant AML (e.g. in the MLL/ENL+Nras context). Based on these first proof-of-principle studies I have screened a panel of genotype-specific candidate drug-targets and identified shRNAs that rapidly kill AML cells, but leave other cell types (e.g. fibroblast, hepatoma cell lines) unaffected. Paralleling these focused candidate gene studies I have performed first large-scale lethal RNAi screens applying complex shRNA libraries and second-generation Solexa sequencing to readout shRNA representation. In result of these studies, I hope to provide a panel of shRNAs that are specifically lethal in certain genetic contexts of AML. Ultimately, such shRNAs will point at genes that are essential for leukemia maintenance, therefore define and evaluate new promising targets for small molecule inhibitors and/or RNAi therapeutics.
Publications
- (2007). Snapshot: genetic mouse models of cancer. Cell 129, 838
Zender, L., Zuber, J., and Lowe, S.W.
