Mutations of NPM1, FLT3 and RAS are common in AML occurring in ∼ 30%, 25% and 10% of cases. The incidence and prognostic implications of the combinations of wildtype (WT) and mutant (MU) NPM1 and FLT3 are well described and have different prognostic implications. How these mutations effect the cellular biology of leukemic blasts at the protein level is unknown as is the effect of a RAS mutation on these combinations.
To understand the biology underlying these changes we utilized a Reverse Phase Protein Array (RPPA) to measure the protein expression of 195 different antibodies in cases of newly diagnosed AML with different combinations of NPM1, FLT3 and RAS mutations. We also examined the MDACC experience with these events in newly diagnosed AML (non APL) patients that were treated at MDACC.
NPM1-MU was present in 20% (4/461) including 36% of diploid (67/185) cases and for all or diploid cases was associated with slightly better overall survival (OS, p= 0.07 and 0.15) and significantly better Event Free Survival (EFS, P=0.002 and 0.001) and longer remission duration (RemDur, p=0.0009, 0.001). FLT3-ITD or D835 mutations were present in 22% (99/450) of cases and 35% (65/183) of diploid cases, in which it was associated with inferior OS (p= 0.05 for all and diploid). RAS-MU were found in 16% (64/403) of cases and was not predictive of OS (p=.49) or RemDur (p=0.62). Mutation in both NPM1 and FLT occurred in 9.3% of cases, more commonly than expected by chance (O/E =2.14, Χ2= 39, P <0.00001). As observed by others, WT FLT3&NPM1 had superior OS (Median 212 weeks, vs. 50–54 weeks for the other combinations, p=0.03), EFS (Median 74 vs. 22–24 weeks, P = 0.002), but RemDur was similar for NPM1-MU cases regardless of FLT3 status (157 &d 132 weeks ) and was superior to NPM1-WT cases (42 & 48 weeks, P=0.002). Similar prognostic implications were found in diploid cases. Combinations of NPM-MU & RAS-MU occurred at expected frequencies (p=.63). Although only 11 patients had both NPM1 and RAS mutations (Median age 47 yrs, 2 favorable, 6 diploid and 3 unfavorable cytogenetics) these patients had markedly superior outcomes (median OS, EFS and RemDur, not reached) with 9 of 11 alive. All other combinations of NPM1 and RAS had similar outcomes. The observed frequency of combinations of all 3 (NPM/FLT3/RAS) were less than expected for MU-MU-MU, MU-WT-WT and WT-MU-WT (O/E.4, 0.61 and 0.66 respectively) and more than expected for MU-MU-WT (O/E 2.59). Only one patient had all three mutations. Patients with MU-WT-MU had superior OS, RemDur and EFS. For OS, NPM1-MU and FLT3 and RAS MU cases fared next best. While those with WT-WT-WT and MU-MU-WT formed a third group, finally those with only FLT3MU or only RAS MU fared worst. The RemDur of those with only NPM1-MU or NPM1 and FLT3 MU lasted longer than those with only RAS-MU or FLT3-MU while those with both FLT-MU and RAS-MU fared worst. In NPM1-MU of the 195 proteins assessed by RPPA, 14 (AKTp473, BAK, CyclinD3, CyclinE, ELKp383,FLI1, HSP90, JAB1, JNK2, MTOR, Nurr77, RAFB, SHIP2, TSC2) were significantly higher (P < bonferonni corrected p <0.00026) and 4 were significantly lower ( FAK, FOXO1.3A, MDM2, YAPp). Cases with both NPM1&FLT3 MU typically showed an accentuated version of this protein expression pattern, typically showing higher expression than either NPM1-MU or FLT3-MU individually. In addition, higher levels of Caspase 8, GAB2p, TNK1 ZNF342, and lower levels of FOXO3A and SHIP were observed in NPM1-FLT3 dual mutants. In contrast, the addition of a RAS mutation blunted many NPM1-MU associated changes ( BAX, Cyclin D3 and E, ELKp383, FAK, Fibronectin, IntegrinB3, NFKBP65, PDK1p, PKCdeltap507 and p664 PTENp, RAFB, VASP), amplified others (DJ1, FOXO3A, P38p180, SSBP, ZNF342) and led to new increases (EIF4E, EGR123, JAB1 SHP2 SMAD1) or decreases (ARC, GATA3, P70S6K, S6RPp235NPM1).
We confirmed the prognostic impact of NPM1 and FLT3-ITD combinations as reported by others groups. We identified a new highly favorable prognostic group of cases (2.5%) with simultaneous NPM1 and RAS mutations and WT FLT3. The proteomic data identified that NPM1 mutation is strongly associated with activation of the RAF-AKT-MTOR axis suggesting that combinations of sorafenib and serolimus might be employed to interrupt this axis and increase chemosensitivity in patients with NPM1 mutations.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.