Clinical Relevance of Alternatively Spliced Isoform AML1-ETO9a in Adult Patients with t(8;21)(q22;q22.1) Acute Myeloid Leukemia (AML): A Study of the German-Austrian AML Study Group (AMLSG)

Background: AML with t(8;21)(q22;q22.1) defines a distinct disease entity within the WHO category 'AML with recurrent genetic abnormalities' and results in the formation of the AML1-ETO (AE) fusion transcript. Recent data showed that alternative splicing generates a truncated variant of AE, namely AML1-ETO9a (AE9a), containing an additional exon 9a. Co-expression of AE and AE9a was shown to induce a more aggressive leukemic phenotype with a rapid onset of AML in a retrovirally transduced mouse model (Yan et al., Nat Med 2006). However, the clinical relevance of the alternatively spliced AE9a isoform in adult patients (pts) with t(8;21) AML is not well determined yet.

Aim: We sought to assess the incidence and prognostic significance of AE9a co-expression in the context of clinical and genetic factors in a large clinically well-annotated cohort of pts with t(8;21) AML.

Methods: 129 pts were included based on the availability of a diagnostic bone marrow (BM) or peripheral blood (PB) sample; all pts were either enrolled on an AMLSG clinical trial protocol (n=93) or received standard intensive chemotherapy regimens. AE9a mRNA transcript levels (TL) were determined by quantitative reverse transcription PCR (qRT-PCR) using TaqMan technology. Co-expression of AE9a as a fraction of the full-length AE transcript was reported as AE9a / AE ratio (%). ABL gene expression was used as housekeeping control. Additional gene mutation data were available for KIT, FLT3 -ITD/TKD, NRA S, and ASXL2 .

Results: The median age of the pts was 49 years (yrs; range, 18-75 yrs); the alternatively spliced isoform AE9a was identified in all pts; the AE9a / AE ratio (range 3%-77%, median 32%) did not significantly differ between BM (n=116, range 8%-77%, median 31%) and PB (n=13, range 3%-66%, median 52%). Median AE9a / AE ratio was not correlated with clinical features (sex, age, WBC, platelets, BM blasts) or gene mutations affecting KIT (p=0.10), FLT3 -ITD/TKD (p=0.77) or ASXL2 (p=1.00); pts with an additional NRAS mutation showed higher AE9a / AE ratios as compared to wild-type pts (median, 44% vs 31%) without reaching statistical significance (p=0.06). Using Cox regression analysis, AE9a / AE ratios were not associated with the clinical endpoints overall survival (OS), event-free survival (EFS) and cumulative incidence of relapse (CIR). The same remained true when performing univariate analyses comparing AE9a / AE ratios dichotomized at the median (AE9a / AE^>median vs AE9a / AE^≤median): AE9a / AE had no impact on 4-yr OS (71% vs 79%; p=0.37), 4-yr EFS (59% vs 67%; p=0.83), and 4-yr CIR (32% vs 33%; p=0.35). Correlation of AE9a / AE ratios according to quartiles of distribution did also not affect OS (p=0.14), EFS (p=0.67), and CIR (p=0.70). Finally, we evaluated a possible correlation between AE9a / AE ratio and NRAS or KIT mutations which frequently co-occur in this AML subtype. In a subgroup analysis according to KIT or NRAS mutation status, no significant differences with regard to the clinical endpoints OS, EFS and CIR were found between AE9a / AE^>median and AE9a / AE^≤median.

Conclusion: Alternative splicing of the AML1-ETO fusion transcript is a common feature in pts with t(8;21) AML. This finding is also supported by a recent collaborative transcriptome study where novel high-throughput sequencing (HTS) technologies were applied (Faber et al., Nat Genet 2016). Using different explorative approaches, AE9a has no impact on clinical outcome which precludes its potential as a novel independent prognostic marker in this AML subtype.

MA was supported by the Else Kröner-Forschungskolleg.