RUNX1 Mutations Play a Major Role in the Progression of MDS to S-AML Following MDS: A Genetic and Cytogenetic Analysis of Sequential Samples

The genetic and cytogenetic events underlying the development of myelodysplastic syndrome (MDS) as well as the mechanisms leading to progression of MDS into acute myeloid leukemia (AML) are largely unknown. Activating mutations in receptor tyrosine kinases like FLT3 or members of the RAS family have been implicated in this process. Recently, mutations of RUNX1 (AML1) have been detected in MDS as well as in AML. As the development of AML is considered as a multistep process, in which mutations of genes with proliferative potential (class I) cooperate with gene mutations in hematopoietic transcription factors (class II), we hypothesized that similar mechanisms might be responsible for the progression of MDS into s-AML. Therefore, we investigated a cohort of 21 MDS patients (pts) (14 male, 7 female; median age at diagnosis of MDS=67.1/range 49.1–76.5), with WHO categories: MDS RARS (n=1), MDS RCMD-RS (n=1), MDS RAEB-1 (n=8), MDS RAEB-2 (n=3), MDS/AML (n=1), CMML-1 (n=1), MDS/MPS overlap (n=1), MDS not further classifiable (n=4) and one patient with reactive condition in morphology only, but del(5q) in chromosome banding analysis. In all cases specimens at the time of progression into s-AML following MDS were available and all specimens were characterized for RUNX1 mutations (entire coding region), FLT3-ITD, MLL partial tandem duplications (MLL-PTD), NPM1-exon 12 mutations and NRAS (codons12/13 and 61) mutations. Furthermore, in all cases chromosome banding analyses (CBA) was performed. At diagnosis of MDS, a normal karyotype (NK) was present in 11/21 cases, whereas cytogenetic aberrations were detected in the remaining 10 cases (del(5q) (n=2), del(20q) (n=1), +8 (n=2), complex (n=2), −Y (n=1), t(1;3)(p36;q21) (n=1) and t(1;14) (p34;q32), −7 (n=1)). The following molecular aberrations were detected at diagnosis of MDS: RUNX1 (n=5), MLL-PTD (n=1), NRAS (n=1), NPM1 (n=1), but never FLT3-ITD. Only one pt had two of these molecular aberrations in combination at the MDS stage (RUNX1/NPM1). After progression to s-AML, 3 of the 11 pts with NK at the MDS stage had developed cytogenetic aberrations (+8 (n=1), +11 (n=1), del(5q), del(7q) (n=1)). The other 10 already at the MDS stage cytogenetically aberrant cases remained unchanged. Thus at the s-AML stage there were 8 pts with NK and 13 with cytogenetic aberrations. At this stage the cohort was again characterized for molecular aberrations and following gains were detected: RUNX1 (n=3), MLL-PTD (n=2), NRAS (n=1), NPM1 (n=1) and FLT3-LM (n=2). The median time between diagnosis of MDS and s-AML was 207 days (d) (range 27–960 d). However, no significant difference could be detected for pts with RUNX1 mutations at the MDS stage (n=5, mean=210 d) vs. the remaining cohort (n=16, mean=266 d) (p=0.536, t-test) and pts with NK at the MDS stage (n=11, mean=200 d) vs. pts with cytogenetic aberrations (n=10, mean=324 d) (p=0.190). The pattern of acquisition of the different aberrations was complex and therefore the cohort was divided in three different subgroups based on the cytogenetic profile:

1. cytogenetically stable, NK: this group included pts without molecular aberration at all stages (n=5), pts with RUNX1 mutation at the MDS stage (n=2), where one acquired secondary a MLL-PTD and the other a FLT3-ITD and one pt, who acquired a RUNX1 mutation at progression.

2. cytogenetically stable, AK: 3 pts without molecular aberrations, 4 pts with molecular aberrations (RUNX1, RUNX1/NPM1, MLL-PTD, NRAS), where the pt with RUNX1/NPM1 acquired FLT3-ITD at progression and 3 pts, who acquired molecular aberrations at progression (RUNX1, NPM1, NRAS).

3. cytogenetically unstable: 2 pts without aberrations at MDS, who progressed with an AK with or without RUNX1 mutation and one pt with a RUNX1 mutation, who progressed with an additional AK plus MLL-PTD.

In conclusion, RUNX1 mutations, when present at the MDS stage, seem to play a role in the progression to s-AML following MDS. Cooperating molecular events for progression to s-AML are the acquisition of FLT3-ITD (n=2) and MLL-PTD (n=2). FLT3-ITD was never detected at the MDS stage in our cohort and seems to be a strong indicator for the progression to s-AML, whereas MLL-PTD can be already detected at earlier stages. Finally, the combination of RUNX1 mutation with FLT3-ITD or MLL-PTD was always associated with progression.