Certain mutations and chromosome aberrations have been shown to cooperate in AML leukemogenesis e.g. t(15;17) + FLT3 mutations, t(8;21) + KITD816 mutations, TP53 + complex aberrant karyotype. Previously AML1/RUNX1 mutations were associated with activating mutations e.g. in FLT3 and NRAS. We now performed a detailed analysis focused on distinct cytogenetic subgroups. A total of 120 selected AML with normal karyotype or recurrent aberrations were analyzed: normal karyotype (NK) (n=43); monosomy 7 (n=32), trisomy 8 (n=10), trisomy 13 (n=14), trisomy 21 (n=9), inv(3)/t(3;3) AML (n=12). Of these 105 were de novo, 7 had t-AML after previous chemotherapy (1 with NK, 4 with −7, 1 with +8, 1 inv(3)) and 8 had sAML after MDS (1 NK, 6 with −7, 1 with inv(3)). RUNX1 mutations were detected in cases with NK: 5/43 (11.6%); −7: 10/32 (31%); +8: 2/10 (20%); +13: 14/14 (100%); +21: 5/9 (55.6%), and inv(3)/t(3;3): 2/12 (12%). Thus, the subgroup with the highest RUNX1 mutation rate was +13 followed by +21. All cases were also analysed for FLT3-length mutations (FLT3-LM), FLT3-TKD mutations, MLL-PTD, NRAS, NPM1 and 38 in addition for CEBPA. NPM1mut and CEBPAmut were found to be nearly mutually exclusive of RUNX1mut. Further analysis was done for subgroups. NK subgroup: In 2 of 5 (40%) RUNX1mut AML with NK a FLT3-LM, in 2 a MLL-PTD, and 1 an NRAS mutation was detected. Thus within the 5 RUNXmut NK-AML a cooperating mutation was detected in all cases. −7 subgroup: No additional mutation was detected in the 10 RUNX1mut cases with −7. In contrast in RUNX1 unmutated cases (RUNX1wt) with −7 1/21 had FLT3-LM, 2/21 (9.5%) CEBPAmut, and 7/20 (35%) an NRASmut. + 8 subgroup: In 2 RUNXmut cases with +8 no further mutation was detected. In contrast 2 of the RUNXwt with +8 had an FLT3-LM, 3 a NPM1mut and 2 a NRASmut. +13 subgroup: All cases with +13 were RUNX1mut and 2/14 (14.3%) had a FLT3-LM and 1/14 (7.1%) a MLL-PTD. In this specific subgroup a 4-fold-elevated FLT3 expression was suggesting to be a specific cooperating event. +21 subgroup: All 5 RUNXmut with +21 had an additional aberration, 4 (80%) had FLT3-LM and 1 NPM1mut. Inv(3) subgroup: In both RUNX1mut inv(3) cases no additional mutation was found. In contrast 3/9 RUNX1wt cases with inv(3)/t(3;3) were NRAS mutated. Overall, additional mutations in RUNX1 mutated AML are very frequent in subgroups with NK and +21 (100%), unfrequent in +13 (21%) and absent in others (−7, +8, inv(3)/t(3;3)). In total (except +13, where no RUNX1wt cases were available), the frequency of additional mutations was higher in the RUNX1wtcases (41.6% vs. 87.8%) This is in contrast to previous reports that suggested that additional activating mutation in RUNX1mut AML are frequent. In contrast we found a high incidence of NRASmut in inv(3) (33%), −7 (35%), +8 (29%) and normal karyotype (14.3%) in cases with RUNX1wt. Monosomy 7 + RUNX1mut + NRASmut previously has been highly correlated to therapy related AML and AML after MDS. In our cohort with predominantely de novo AML we found a high correlation of RUNX1mut with −7 and a high correlation of −7 with NRAS, but no association of RUNX1mut with NRASmut. In addition, FLT3-LM has been highly correlated to AML1mut. We found this correlation only in cases with +21. In conclusion, FLT3-LM and NRAS mutations were detected as frequent cooperating mutations in RUNX1mut AML with NK and +21. No mutation cooperating with RUNXmut was detected in −7, +8, +13, and inv(3)/t(3;3). Here alternative mechanisms may drive leukemogenesis e.g. overexpression of FLT3 in +13 or dosage effects due to monosomies or trisomies.

Disclosure: Ownership Interests: SS, WK, TH and CH own the MLL Munich Leukemia Laboratory. FD works for MLL. TH and WH run the MHP Munich Hematology Practice. SS and CH work for MHP.