Search

Abstract 670 The availability of multiple options for chronic myeloid leukemia (CML) treatment is not paralleled by the availability of biological predictors of outcome allowing to identify patients (pts) who are more likely to benefit from dasatinib or nilotinib rather than imatinib (IM). Pharmacogenetics has proven a potential source of biomarkers given the known influence of polymorphisms in key genes encoding drug transporters and metabolizing enzymes on drug delivery – hence effectiveness. In CML, only two studies had so far explored this field, but both were conducted in heterogeneous populations including pts at different stages of disease, not all receiving IM first-line. We thus aimed to investigate a panel of 20 single nucleotide polymorphisms (SNPs) in ABCB1, ABCG2, SLC22A1, OATP1A2, OCTN1, CYP3A4 and CYP3A5 genes that can be hypothesized to influence IM transport and metabolism in 189 newly diagnosed CML pts enrolled in the TOPS phase III trial (Cortes et al, J Clin Oncol 2010). Pts selection was exclusively based on availability of written informed consent and sufficient amount of archived material. Median age was 46 years; male to female ratio was 103 to 86; 156 (83%) pts were Caucasian and 23 (12%) were Asian; low, intermediate and high Sokal risk pts were 84 (44.4%), 65 (34.4%) and 40 (21.2%), respectively. Baseline demographic/clinical features did not differ significantly from those of the overall population. Treatment outcomes (complete cytogenetic response [CCyR]; major molecular response [MMR] and complete molecular response [CMR]) were compared according to i) each candidate genotype ii) summary measures based on combinations of SNPs in the same gene and iii) summary measures based on combinations of SNPs in functionally related genes (uptake; efflux). CC genotype in OCTN1 had a favorable impact on the achievement of MMR at 12 months (MMR@12m; P = 0.03). With respect to the summary measures, combination of SNPs in the SLC22A1 gene was significantly correlated with MMR@12m (P = 0.03). When considering summary measures of uptake and efflux, the former was found to be associated with both MMR@12m and CMR@12m (P = 0.003 and P = 0.01, respectively). A separate analysis limited to Caucasian pts (n=156) yielded similar results (Table 1). In addition, the analysis in the Caucasian subgroup evidenced a significant association between the CC genotype in ABCB1 rs60023214 and MMR@12m (P = 0.005) (Table 1). Cumulative incidence plots based on the Kaplan-Meier method were also analyzed in the overall population and in Caucasians, with comparable results. Representative plots are shown in Figure 1. There was evidence for difference among MMR cumulative incidence curves for 2 single SNPs and 2 score measures. Presence of the major allele in OCTN1 (CC) and of the minor allele in CYP3A4 rs2740574 (GG) were associated with increased MMR rate (P = 0.028 and P = 0.042, respectively, in the overall population and P = 0.027 and P = 0.038, respectively, in Caucasians). Similarly, an increase in the number of favorable alleles in the SLC22A1 gene was associated with increased MMR rate (P = 0.030 and P = 0.043 in the overall population and in Caucasians, respectively). In addition, the combination of favorable alleles in the genes involved in IM uptake was associated with increased rates of both MMR and CMR (P = 0.004 and P = 0.015, respectively, in the overall population and P = 0.005 and P = 0.009, respectively, in Caucasians). Our results suggest that SNP genotyping might be helpful in selecting pts who are more likely to benefit from first-line use of more potent inhibitors. Further assessment of the SNPs here identified in larger series of pts is warranted. Table 1 . Summary of significant associations Overall population (n=189) Caucasians (n=156) Gene, rs Outcome P Hazard ratio[95% CI] P Hazard ratio[95% CI] OCTN1 - rs1050152 MMR@12m 0.03 0.78 (0.63, 0.98) 0.03 0.77 (0.60, 0.97) ABCB1 (MDR1) - rs60023214 CMR@12m 0.06 0.71 (0.49, 1.02) 0.005 0.56 (0.37, 0.84) SLC22A1 (hOCT1) - SNP combination MMR@12m 0.03 1.24 (1.02, 1.50) 0.04 1.24 (1.00, 1.54) Uptake - SNP combination MMR@12m 0.003 1.21 (1.06, 1.37) 0.005 1.24 (1.06, 1.44) Uptake - SNP combination CMR@12m 0.01 1.30 (1.05, 1.61) 0.01 1.37 (1.08, 1.74) Download : Download high-res image (197KB) Download : Download full-size image Figure 1: . Cumulative incidence of pts achieving MMR and CMR according to the # of major alleles in uptake-related gene (SLC22A1, OATP1A2, OCTN1) SNPs Supported by Novartis Oncology, Clinical Development, TOPS Correlative Studies Network Disclosures: Hughes: Novartis: Honoraria, Research Funding, Speakers Bureau; Bristol-Myers Squibb: Honoraria, Research Funding; Ariad: Honoraria. White: Novartis: Honoraria, Research Funding; Bristol-Myers Squibb: Research Funding. Saglio: Novartis: Consultancy, Honoraria; Bristol-Myers Squibb: Consultancy, Honoraria. Rosti: Novartis: Consultancy, Honoraria; BMS: Consultancy, Honoraria. Hatfield: Novartis: Employment. Martinelli: Novartis: Consultancy, Honoraria; BMS: Honoraria; Pfizer: Consultancy.

Abstract 3283 Poster Board III-1 IM is the first-choice treatment in CML and allows to obtain excellent response rates. Some pts, however, experience suboptimal response or resistance, which highlights the need to find biological predictors of outcome in order to guide therapy optimization. Since blood and tissue concentrations of drugs may be influenced by variations between individuals (single nucleotide polymorphisms, SNPs) in genes encoding drug metabolizing enzymes and drug transporters, we have investigated if SNPs influencing the delivery if IM to target cells may account for differences in response. To test that hypothesis, we have genotyped a panel of SNPs known to affect the activity of the cytochrome p450 family isoforms and of the transporters that have been implicated in IM metabolism and transport in a subset of pts enrolled in the TOPS trial -; a randomized phase III study of IM 400 mg/d vs IM 800 mg/d in newly diagnosed CML (Cortes et al, ASH 2008). Included in this analysis were 18 SNPs in 7 genes: hOCT1 (ins/del, rs12208357, rs683369, rs4646277, rs4646278, rs2282143); ABCB1 (rs60023214, rs1128503, rs10245483, rs3213619, rs1128501), ABCG2 (rs2231137, rs2231142); OCTN1 (rs1050152); OATP1A2 (rs11568563); CYP3A5 (rs28365083, rs776746); CYP3A4 (rs28371759). Genotype distributions were tested for Hardy-Weinberg equilibrium (HWE). Additive models were used to assess association between individual SNPs and three response outcomes: major molecular response (MMR) at 12 months, optimal CgR and optimal response as defined by European LeukemiaNet (ELN) recommendations (Baccarani, Blood 2006). Summary measures based on SNPs from the same gene and those related functionally (uptake and efflux) were defined as the number of alleles, across the genes of interest, hypothesized to be associated with favorable response. Association of individual SNPs and summary variables with response was assessed with logistic regression models and likelihood ratio and exact tests. A total of 166 pts were genotyped, out of which 133 whites from Italy (N=21), the U.S.A. (N=66) and Australia (N=46); 22 Asian from Korea (N=17), Italy (N=1), and Australia (N=4); and 11 non-white and non-Asian from the U.S.A (N=6) and Australia (N=5). Selection of pts was based exclusively on availability of a written informed consent for correlative substudies and of a sufficient amount of archived DNA material. The response rates by 12 months among the subset of pts analyzed were 51% for MMR, 87% for optimal cytogenetic response according to ELN, and 85% for optimal response according to ELN. Our analyses have identified four monomorphic, non-informative SNP ( hOCT1: rs4646278, rs2282143; MDR1: rs1128501; and CYP3A4: rs28371759). The results have shown, when analyzing the patient population as a whole, that no association could be found between individual SNPs and any of the response variables considered. However, among whites, but not other pts, we have observed a trend of association between achievement of MMR and two SNPs in the hOCT1 and CYP3A5 genes. Namely, the wild type allele for hOCT1 rs683369 was associated with a higher response rate compared to heterozygote and homozygote mutants (60% vs 36%, p=0.057). Also, the mutant allele for CYP3A5 rs28365083 was associated with a higher response rate (p=0.058). In the subset of pts from the U.S.A, there was a evidence of association between achievement of MMR and three SNPs in the hOCT1 and MDR1 genes. Those with the wild type allele for hOCT1 ins/del had a lower MMR rate compared to those carrying at least one minor allele (48% vs 81%, p=0.08). The wild type for hOCT1 rs683369 was associated with a higher response rate compared to heterozygote and homozygote mutants (65% vs 36%, p=0.05). MDR1 rs60023214 (p=0.08) and the MDR1 gene summary measure of total number of favorable alleles among all MDR1 SNPs (p=0.08) were correlated with MMR, with improved response in the wild type group. Our study suggests for the first time a role of ethnicity in determining correlations between SNP-based molecular signature and response to treatment. If confirmed by larger studies, this suggests that ethnicity should be taken in consideration to better interpret pharmacogenetic analyses in CML and could explain the differences in results of the studies published so far. Supported by Novartis Oncology, Clinical Development, TOPS Correlative Studies Network Disclosures: White: Novartis and Britol-Myers Squibb: Research Funding. Baccarani:Novartis: Research Funding, Speakers Bureau; Bristol-Myers Squibb: Research Funding, Speakers Bureau.

Abstract 3252 Poster Board III-1 A better understanding of the mechanisms which regulate drug resistance is critical for improving therapy for patients at risk of poor response. Even for CML patients the long-term benefits of the treatment are limited by the emergence of resistance, although the therapy has been dramatically improved by Imatinib Mesylate, which inhibits Bcr-Abl activity. One of the most important mechanisms of resistance is the disregulation of various members of the highly conserved family of transmembrane proteins characterized by an ATP-binding cassette domain, called ABC superfamily of transporters (Dean et al., Genome Res 2001). In CML cells an aberrant expression of ABC transporter genes has been described to be mediated by specific transcription factors, which are in turn affected by an aberrant activity of Bcr-Abl. It has been shown that Bcr-Abl can indirectly activate c-Myc via the JAK2 pathway, which increases a translation of c-Myc mRNA. In addition, in a subgroup of CML patients a chromosome 8 trysomy, has been observed, which can be associated with c-Myc amplification. In this study we have investigated whether c-Myc can regulate transcription of ABC tranporter genes in CML. Initially, ABC transporter gene expression has been monitored in HL60, a human promyelocytic cell line in which c-MYC is overexpressed. We have examined the expression level of all 48 human ABCs transporters as a function of c-MYC silencing. Our results have demonstrated that c-Myc regulates the transcription of several ABC genes, such as ABCA2, ABCB9, ABCB10, ABCC1, ABCC4, ABCE1, ABCF1, ABCF2, a majority of which has been found implicated in drug resistance. Furthermore, by performing chromatin immunoprecipitation (ChIP) we have shown a direct binding of c-Myc to the promoters of those ABC transporter genes in HL60 cell line. In addition, by ChIP we have demonstrated that c-MYC is physically bound to the promoter of tested ABC genes in CML cell lines KG-1a and K562. Based on those findings we have investigated the expression level of c-Myc in CD34+ progenitor cells derived from CML patients. Also, we have evaluated the effects of highly expressed c-Myc on ABC transporters. Our results have shown in a group of 20 newly diagnosed chronic phase (CP)-CML patients an increased expression of c-Myc in CD34+ cell fraction, when compared to the expression level of c-Myc in the entire population of mononuclear cells from which CD34+ cell fraction has been purified. In those cells we have identified in association with an increased level of c-Myc an increased expression of the same subset of ABC genes, which we have observed in cell lines. Furthermore, we have evaluated whether a differential expression of c-Myc and ABC transporter genes is associated with low Sokal risk and high Sokal risk patients. This analysis has been performed on CD34+ cells purified from 19 CML patients of whom 12 were scored as low Sokal risk and 7 as high Sokal risk. Our results have shown that expression of c-Myc and ABCC4 is different in those two patients population (p Disclosures: No relevant conflicts of interest to declare.

Abstract 854 It has been proposed that most protein-encoding genes may be regulated by small multifunctional RNAs which can control transcript turnover and/or protein translation (siRNA and miRNA). Specifically, miRNAs have been shown to act predominantly at the level of translation by blocking the access or sliding of ribosomes to mRNAs (3'UTR). Several studies have shown that multiple miRNAs can be disregulated in tumours as compared to normal tissues. For example, the miR-17-92 cluster is up regulated in CML, suggesting their involvement in leukemiagenesis (Venturini et al., Blood 2007). However, a role of miRNA in preventing tumor development and progression has also been suggested. Bueno et al have shown that hsa-miR-203 can specifically target BCR-ABL and reduce its expression in CML derived cell lines (Bueno et al., Cancer Cell 2008). Since multiple miRNAs seem to act in a combinatorial fashion to regulate mRNA translation, we hypothesised that other miRNAs in addition to hsa-miR-203 might be involved in BCR-ABL expression control. To test this hypothesis we conducted a search for miRNAs specifically targeting Bcr-Abl, using the miRBASE program to scan human genome (http://microrna.sanger.ac.uk/). This search identified 15 miRNAs potentially able to target the BCR-ABL 3' UTR. Further investigation showed that only hsa-miR-451, hsa-miR-515-3p and hsa-miR-760 had a sufficiently high score to predict genuine interactions with the 3'UTR of BCR-ABL and therefore only these miRNAs were utilized for further analyses. Initially, the three miRNAs were transfected in K562 Ph+ cells, together with hsa-miR-203 as a positive control and a scrambled miRNA used as a negative control. BCR-ABL expression was monitored at both mRNA ( qRT-PCR) and protein level (Western Blotting). Our results demonstrated that BCR-ABL mRNA expression was unaffected by miRNAs, whereas protein expression was significantly reduced. Considering the combinatorial function of miRNAs on their target mRNAs, we also tested whether a pool of the four identified miRNAs could be effective in suppressing BCR-ABL expression. In those experiments, the molar concentration of each miRNA was a quarter of that used for a single miRNA transfection. The results have shown that the pool of miRNAs worked efficiently in suppressing Bcr-Abl expression, which suggested a cooperative function of the four miRNAs in controlling both the expression and translation of Bcr-Abl. To further confirm that miRNAs directly targeted BCR-ABL, the 3'UTR of the gene was cloned downstream of a reporter renilla luciferase gene. The reporter was co-transfected in HEK293 cells together with single miRNAs or a pool of them, and luciferase activity was quantified. Those results show that the presence of each miRNAs significantly reduced the luciferase activity as compared to that obtained by transfecting cells with a scrambled miRNA. These experiments therefore confirmed a direct effect of the four miRNAs on the 3'UTR of BCR-ABL. Again the “pool” of miRNAs showed the strongest effect on the luc reporter expression. Through an additional deletion analysis we mapped the regions of 3'UTR of BCR-ABL targeted by the four miRNAs, in order to confirm that the predicted binding regions of miRNAs are critical to mediate repression of BCR-ABL. In conclusion, our study identified three new human miRNAs having a potential to specifically target BCR-ABL and suppress its translation and expression in CML cells. BCR-ABL, which plays a critical role in CML, is effectively suppressed by TK inhibitors, exemplified by Gleevec and our findings provide a rationale to exploit miRNA as an alternative therapeutic approache which could further improve CML treatment, or to complement TK inhibitors in an effort to eradicate minimal residual disease. Supported by: Novartis Oncology, Clinical Development, TOPS Clinical Correlative Studies Network Disclosures: No relevant conflicts of interest to declare.