Incidence and Prognostic Impact of SNPs Regulating PU.1 Gene Expression in AML

Background/Aim: The transcription factor PU.1 is crucial for hematopoiesis. The transcriptional control of PU.1 gene is mediated by a distal upstream regulatory element (URE). Mice lacking URE have a decreased PU.1 expression and develop AML. In human, 2 SNPs reported in the URE of PU.1 reduce PU.1 expression in myeloid progenitors and have been reported frequently associated with complex karyotype in AML patients (Steidl, JCI, 2007). The aim of this study was to investigate the frequency of PU.1/URE SNPs in different cytogenetical and molecular AML subgroups and to evaluate their prognosis impact. Methods: Patients with de novo AML homogeneously treated in the ALFA9802 protocol (15-50y, non PML) were screened by pyro-sequencing for both SNP1 and SNP2 status (+/+,+/- and -/-). Samples were available for 260 AML patients, also characterized for FLT3dup, FLT3d835, NPMm, CEBPAm and WT1m gene mutations. The SNP profile in AML patients was compared to 207 healthy controls. Results: SNP1(+) allele was detected at the same frequency in AML patients than in healthy controls (SNP1: 43% vs 43%, p=NS by the Pearson’s test).SNP2(+) allele was less frequent in AML patients (13% vs 18% p=.05). No difference was observed between cytogenetical subgroups (low-, int- and high- risk) or according to gene mutational status. In this cohort, SNPs allele frequencies were similar in patients with complex karyotype (>=5abn.). Homozygous SNP1(+/+) and SNP2(+/+) profiles were found in 53 (20.3%) and 6 (2.3%) AML patients respectively. Considering SNP+ patients presenting SNP1(+/+) or SNP2(+/+) genotype, a trend for a lower risk of relapse (5y-RR: 35% vs 52%, p=0.10) and a better EFS (5y-EFS: 54% vs 37%, p=.08) was observed in the global population. Considering patients with normal karyotype AML (NK-AML, n=106), SNP+ was predictive of a better outcome (5y-OS: 77% vs 44%;p=.03). This difference was due to a better CR rate (100% vs 77%, p=.2) but also to a lower relapse rate (5y-RR: 27% vs 52%,p=.11) with a global significant impact on EFS (5y-EFS: 70% vs 37%, p=.02). In patients with NK-AML, the other prognosis factors for OS were the WBC (p=.01), the NPMm+/FLT3dup- status (p=.03). In multivariate analysis considering WBC, FLT3dup, NPMm and SNP+ as covariates, factors influencing OS were NPMm (p=.03), WBC (p=.04) and SNP+ (p=.05; HR=.36[.13-.99]). Among the 66 patients with NPMm, no relapse was observed in SNP+ patients (5y-EFS: 80% vs 40%, p=.03). In patients with FLT3dup (n=36), a mutation reported to down-regulate PU.1 expression, no effect of SNP was observed (5y-EFS: 33% vs 20%, p=.40). Conclusion: In young adults with AML, PU.1/URE SNP2 allele is less frequent than in healthy donors. Furthermore, a homozygous genotype leading to PU.1 downregulation is found to be associated with a favorable outcome in patients with normal karyotype. These paradoxical effects of constitutive PU.1 downregulation in human deserve further functional and clinical studies to compare the impact of constitutional- versus oncogenic-related modulation of PU.1 expression in AML.