The Number of Cooperating Mutations in AML Is Still Growing: A Study on 3789 Patients.

Abstract

Abstract According to a two hit model in AML at least two genetic alterations are required for the clinical manifestation of leukemia. Many recent studies have shown that one type comprizes genetic hits causing a stop in differentiation. In most cases these are loss of function mutations in transcription factors, e.g. through reciprocal translocations like AML1TO, CBFBMYH11, and PML-RARA or molecular mutations in CEBPA and AML1. Others are mutations in nuclear transport proteins like NPM1 or NUP214 (DEK). This type of mutations is reflected mainly in the morphologic phenotype of the AML and is called type II mutation. The other type of mutations (type I mutations) usually leads to enhanced proliferation and mostly comprizes activating mutations in genes coding for tyrosine kinases or molecules downstream of the respective pathways. These two kind of mutations often are referred to as cooperating mutations. Based on our molecular analyses of 3789 AML we found that the pattern in which these mutations occur is not random and that certain type I mutations prefer to occur together with certain type II mutations. This suggests that certain combinations may have optimal cooperative functions. Combinations of two type I mutations are very rare and type II mutations occur completely mutually exclusive. The following preferential combinations of type I mutations with chromosomal aberrations were found (significant associations are given in the table): FLT3-LM with t(15;17), t(6;9) and normal karyotype; FLT3-TKD with normal karyotype; NRAS with inv(16)/t(16;16) and inv(3)/t(3;3); KITD816 with t(8;21); KITexon11 with inv(16)(t(16;16). Newly detected was the association of JAK2V617F with trisomy 8, suggesting that on #8 a still unknown type I mutation may be located. In addition JAK2 was correlated with t-AML with t(8;21). Furtheron mutations in the TKD domain of VEGFR1 were associated with mutated AML1 and trisomy 13. In addition, some of the type I mutations were correlated to normal karyotypes and were associated with other molecular mutations: FLT3-LM is preferentially found in AML with MLL-PTD as well as in AML with NPM1-, CEBPA-, and AML1-mutations. For some combinations a prognostic relevance already was shown: FLT3-LM has unfavourable impact on EFS in normal karyotype (p=0.04) and on OS in t(15;17) (p=0.05), in CEBPA mutated AML (p=0.03) and even completely abrogates the favourable impact of NPM1 (p<0.001). KITD816 mutations were shown to confer an extreme bad impact on OS in t(8;21) (p<0.001). In contrast, NRAS seems associated with an improved outcome in cases in which all unfavourable molecular markers are negative (p=0.06). In conclusion this study supports the growing importance of molecular characterization of AML with respect to diagnosis and prognosis. Coincidience of type I and II mutations Type I (row), Type 2 (column) FLT3-LM (%) NRAS (%) KITD816 (%) KITexon8 (%) JAK2 (%) VEGFR1 (%) For clarity only significantly elevated frequencies are shown in comparison to overall frequencies in AML. Percentages of cases with type I alterations within the groups of type II are given. Prognostically favourable associations coded as #, unfavourable associations coded as * total 23.4 10.3 1.7 < 1 6.2 n.a. t(15;17) 35.3* t(8;21) 10.5* 9.5 t(rare;21) 75 inv(16)/t(16;16) 37.6# 10.8 inv(3)/t(3;3) 26.8 t(6;9) 75 MLL-PTD 35.2 CEBPA 36* AML1 18 14.2 NPM1 40* Trisomy 8; unknown gene 23.3