RUNX1 Mutations in Normal Karyotype MDS Cooperate with Class I Mutations in the Progression to S-AML Following MDS.

The mechanism that may cause progression of myelodysplastic syndrom (MDS) to acute myeloid leukemia (AML) is genetically poorly defined and several different scenarios might account for this phenomenon. Previously, RUNX1 mutations (RUNX1+) but also cytogenetic aberrations and mutations of the RAS signal transduction pathway have been implicated in this process. Therefore, we analyzed a cohort of MDS (n = 179) and s-AML following MDS (n = 93). The MDS cohort consisted of RARS (n=2), RCMD (n=15), 5q- syndrome (n=2), CMML-1/-2 (n=13), RAEB-1 (n=48), RAEB-2 (n=58) and MDS not further classified (n=41). The entire coding region was screened for RUNX1+ by denaturing high performance liquid chromatography (DHPLC) and mutations were called by direct DNA sequencing. Furthermore, FLT3 was screened for length mutations (FLT3- LM, FLT3-ITD), NRAS for mutations in codons 12/13 and 61, MLL for partial tandem duplications (MLL-PTD) and NPM1 for exon-12 gene mutations. Cytogenetic analysis was done by chromosome banding and if needed for further clarification with FISH analysis. RUNX1+ were detected in 26 (14.5%) of 179 MDS cases and in 32 (34.4%) of 93 sAML. Thus the incidence of RUNX1+ in the MDS cohort was significantly lower compared to the incidence of RUNX1+ in s-AML following MDS (p<0.001, Fisher’s exact test). Next, we categorized the total cohort in cases with normal vs aberrant karyotype and with or without a mutation in one of the following genes: FLT3, NRAS, MLL and NPM1. Whereas the proportion of patients (pts) with aberrant karyotype (AK) was not significantly different between MDS (39%) and s-AML following MDS (42.7%) (p=.595), the percent of patient associated with additional gene mutations was significantly increased in the latter group (MDS=11.2% vs s-AML following MDS=32.3%, p<0.001). As the frequency of RUNX1+ was significantly increased in the s-AML vs the MDS pts, we concentrated on events associated with RUNX1+. Therefore, we separated our MDS cohort into pts with normal karyotype (NK) without RUNX1 mutations (RUNX1-) (n=90), NK/RUNX1+ (n=13), aberrant karyotype (AK)/RUNX1- (n=53) and AK/RUNX1+ (n=13) and compared these groups with the respective s-AML groups (n=33, n=18, n=26 and n=12, respectively) in order to analyze the acquisition of cooperating gene mutations. Whereas no significant increases in the frequencies of cooperating gene mutations comparing MDS and s-AML stages of AK/RUNX1- (20.8% vs. 34.6%, p=0.27) and AK/RUNX1+ (15.4% vs. 25.0%, p=0.645) were detected, significantly increased frequencies were seen in NK/RUNX1- (5.6% vs. 21.2%, p=0.016) as well as in NK/RUNX1+ (15.4% vs. 55.6%, p=0.032). The prognostic impact of RUNX1+ was analyzed with Cox regression analysis on the overall survival of s-AML pts with RUNX1- (n=37) vs. RUNX1+ (n=18) or NK s-AML following MDS with RUNX1- (n=24) vs. RUNX1+ (n=10). RUNX1+ had no significant prognostic impact in cytogenetically unselected s-AML following MDS, whereas RUNX1+ attained borderline prognostic significance in s-AML with NK (p=0.065, HR=2.819). No significant prognostic impact with the same model was attained for MDS with or without RUNX1 mutation. In conclusion these data implicate that RUNX1 mutations in MDS with normal karyotype cooperate with mutations in receptor tyrosine kinases and the RAS signaling pathways, but also with MLL-PTD to promote the progression to AML. Therefore, RUNX1 mutations in MDS might indicate a high risk for progression into AML.