Abstract
Cytarabine is the backbone of modern AML chemotherapy. However, extensive inter-patient variation in treatment response, development of resistance, and severe toxicity remain as major hurdles to effective cytarabine chemotherapy. Although the genes involved in cytarabine’s activation/inactivation and transport are well defined, we still lack in understanding of the PD genes of relevance. So far several efforts have been made to identify cytarabine response genes have some of these utilized transcriptional profiling to identify gene expression signatures that differentiate sensitive and resistant cell lines; others have utilized gene expression profiles from diagnostic specimens, to identify transcripts predictive of therapeutic outcome. To identify genes that show a biologically meaningful pattern of association with multiple pharmacologic and clinical variables, we have recently developed and applied the innovative PROMISE (PRojection Onto the Most Interesting Statistical Evidence) statistical analysis procedure. In our preliminary study using PROMISE, we identified transcriptional signatures (consisting of 60 genes) that were predictive of therapeutically beneficial (n=46 genes) or detrimental (n=14 genes) patterns of association with multiple clinical parameters. We then used this information, combined with genes reported in literature as being important to ara-C pharmacology and/or identified in relevant genome-wide screens, to define a high-priority list of 300 candidate genes, which we subjected to a multi-phased, high throughput siRNA/ cytarabine modifier screening campaign in human AML cells. The first screening phase used 3 individual siRNAs to target each of the 300 selected candidate genes in THP1 cells, tested alongside standard transfection controls in 384 well plates. Post siRNA transfection, cells were treated with different concentrations of cytarabine (0µM; 0.1µM-IC10; 0.8µM-IC50; and 10µM-IC90) followed by multi-parametric nuclear morphometry assays using automated microscopy to document the individual and combined phenotypic effects of siRNA gene silencing and cytarabine on cell growth and proliferation. Genes were classified as suppressors if their siRNA knockdown inhibited the drug response (i.e increased resistance) and enhancers if the knockdown enhanced the drug response (i.e increased drug sensitivity). This analysis yielded 72 candidate cytarabine modifier genes, which were then subjected to further technical and biological validation tests to document experimental reproducibility, siRNA targeting specificity and cell line specificity of observed phenotypes in THP1 (0, 0.8 and 10µM) and Kasumi cells (0, 0.05 and 1µM for IC50 and IC90 concentrations for Kasumi). Targeting specificity was assessed by comparing microscopy phenotypes to target knock-down as measured by RT-qPCR for candidate hit siRNAs and matched non-cleaving “C911” control siRNAs. Our results yielded clear validation of ara-C modifier effects for several genes with known functions in ara-C-relevant pathways, including DNA damage repair response factors and deoxynucleotide metabolism/catabolism enzymes. As such, these results confirm the patho-physiological relevance of our screening campaign in strengthening the predictive value of markers previously identified by transcriptional profiling analyses from AML patient samples. They also confirm our screen design’s potential for identifying novel modulators of cytarabine -induced phenotypes. Indeed, even from this relatively small screen of only 300 high-priority genes, our validation data supported ara-C suppressor effects from knocking down APOBEC3G, ANXA5, (DCK: involved in activation of cytarabine), KPNA2, NCF1, PLEKHM1, REPIN1 and XRCC1, as well as enhancer effects from knocking down CFLAR, CHEK1, DERA, EIF4F2, EXO1, GNB5, GRPEL1, IER3IP1, IQGAP1, MAPK11, NFKB2, NKX2, NUBP1, RPL31. These genes represent valuable biomarker candidates whose expression levels in AML patients may predict ara-C responsiveness. The identified enhancer genes also represent novel therapeutic target candidates for developing more effective ara-C combination treatments for AML.
Disclosures
Echeverri: Cenix BioScience Inc: Employment, Equity Ownership. Korn:Cenix BioScience Inc: Employment. Kochar:Cenix BioScience Inc: Employment.
Pheno-RNA, a method to identify genes linked to a specific phenotype: application to cellular transformation