Abstract
Background: In AML four types of acquired alterations of the RUNX1 gene have been described: 1. translocations involving RUNX1 leading to fusion genes such as RUNX1-RUNX1T1, 2. molecular mutations, 3. amplifications of RUNX1, 4. Partial or complete deletions of the RUNX1 gene.
Aim: To determine the frequency of different RUNX1 alterations and to characterize the spectrum of accompanying genetic abnormalities.
Patients and Methods: We screened 726 de novo AML patients (pts) for RUNX1 deletions (del) and translocations using a dual color break-apart probe covering the 5' and 3' part of RUNX1 (MetaSystems, Altlussheim, Germany) and in addition evaluated RUNX1 mutations (mut) by Sanger or next-generation amplicon deep-sequencing. Median age was 67 yrs (range: 18 to 100 yrs). For all patients cytogenetics was available and categorized according to MRC criteria (Grimwade et al. Blood 2010). Partial deletions of RUNX1 as detected by FISH were confirmed by array CGH (Agilent Technologies, Santa Clara, CA).
Results: In 89/726 pts (12.3%) abnormalities of the RUNX1 gene were detected by FISH: 10 pts (1.4%) showed a deletion encompassing the whole RUNX1 gene while additional 9 pts (1.2%) showed a partial loss of one RUNX1 copy. A gain of one RUNX1 copy was present in 45/726 (6.2%) pts. In 3 pts a gain of the 5' part of RUNX1 was accompanied by a loss of the 3' part while in 2 pts one copy of the 3' part was gained accompanied by a loss of the 5' part. A translocation affecting the RUNX1 gene was detected in 31 pts (4.3%). The partner gene was RUNX1T1 in 29 pts and located on 16q13 and 18p11 in one pt each. One pt with a RUNX1 translocation also showed a 5' RUNX1 deletion. In 110/726 pts (15.2%) a RUNX1mut was detected. Of these, 16 pts showed two and 5 pts three mutations in RUNX1. Thus, in total 136 mutations were detected in 110 pts: 58 (42.6%) were frameshift, 42 (30.9%) missense, 21 (15.4%) nonsense, 9 (6.6%) splice-site and 6 (4.4%) in-frame insertions/deletions. The RUNX1mut was homozygous in 15 pts, these were predominantly missense mutations (9/15; 60%). Within the subset of pts with RUNX1mut 2 harbored an additional RUNX1del and 9 pts a gain of a RUNX1 copy, while no RUNX1 translocation was present. In AML FAB type M0 both RUNX1mut and RUNX1del showed the highest frequencies (33.3% and 14.8%). 48.4% and 45.2% of cases with RUNX1 translocations were FAB type M1 and M2. While RUNX1mut were most frequent in the cytogenetic intermediate risk group (19.1%; favorable: 2.2%, adverse: 9.7%), RUNX1del were most frequent in pts with adverse risk cytogenetics (9.7%; favorable: 1.1%, intermediate: 0.8%). A comparable distribution was observed for a gain of RUNX1 copies (adverse: 19.4%, favorable: 4.5%, intermediate: 2.6%). With respect to additional molecular mutations all types of RUNX1 alterations were mutually exclusive of NPM1mut. Further, the frequency of DNMT3Amut and CEBPAmut was significantly lower in pts with RUNX1 alterations as compared to those without (14.3% vs. 34.3%; p<0.0001 and 6.4% vs. 13.4%; p=0.012). However, some striking differences between the different types of RUNX1 alterations were detected: ASXL1mut were significantly more frequent in pts with RUNX1mut (36.7%) but rather infrequent in pts with RUNX1del, gain and translocation (12.5%, 6.1%, and 6.7%). A comparable association was noticed for SF3B1mut which were frequent in RUNX1mut pts (23.8%) and rather infrequent in pts with RUNX1del, gain and translocations (0%, 10.5%, and 0%). In contrast, pts with either RUNX1del or RUNX1 gains showed a significantly higher TP53mut frequency (66.7% and 35.3%) as compared to RUNX1mut or RUNX1 translocated pts (7.1% and 4.8%). In the total cohort median overall survival (OS) was 18.7 months and differed significantly between the different types of RUNX1 alterations: for RUNX1 translocations, mutations, gains and deletions it was 35.5, 14.1, 12.4 and 4.3 months.
Conclusions: 1. The RUNX1 gene is altered in 26% of AML. 2. All types of RUNX1 alterations predominantly occur in AML M0, M1 and M2 and are rare in the remainder AML. 3. They are mutually exclusive of NPM1 mutations and show a negative association with DNMT3A mutations. 4. While RUNX1 mutations were most frequent in patients with intermediate risk cytogenetics, RUNX1 deletions and gains were most frequent in patients with adverse cytogenetics. 5. Outcome differs significantly and is best in patients with RUNX1 translocations and worst in cases with RUNX1 deletions.
Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Nadarajah:MLL Munich Leukemia Laboratory: Employment. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.
Author notes
Asterisk with author names denotes non-ASH members.