The nucleoside analogue cytarabine (AraC) has served as the backbone of acute myeloid leukemia (AML) treatment for nearly forty years. About one-third of expressed genes are abnormally spliced in AML yet alternative exon usage (AEU) plays a role in the plasticity of tumor cells and may influence the response to treatment. Here the exon expression profiles of the erythroleukemia K562 cell line were compared to that of its AraC-resistant variant K562/AraC through Affymetrix HTA2 exon arrays. 5140 exon events harbored by 2583 genes distinguished the 2 cell lines. Among these, the skipping of TET2 exon 2 was identified in K562 cells sensitive to AraC whereas TET2 gene expression remained unchanged at the whole transcript level. The results were confirmed by exon-specific RTPCR (ESPCR). Microarray analysis did not evidenced any significant change in mRNA splicing for the 10 remaining exons of the TET2 gene.

TET2 is a dioxygenase that catalyzes the conversion of 5-methylcytosine (5-mC) to 5-hydroxymethylcytosine (5-hmC) and promote DNA demethylation. TET2 somatic mutations occur in about 25% AML, distributing across the whole coding sequence without obvious hot spots. These mutations decrease TET2 enzymatic activity by truncating the protein or affecting its catalytic activity. TET2 exon 2 is spliced in a mutually exclusive manner with exon 1 yet it is used as an alternative promoter ( However, TET2 exon 2 is not translated into protein and its role in TET2 regulation is still unknown. Having found that AraC sensitive cells harbor the spliced TET2 isoform, we investigated whether or not skipping of TET2 exon 2 correlate with disease outcome in AML patients treated with AraC-based intensive chemotherapy (AraC-IC).

The discovery cohort included 106 consecutive AML patients treated with AraC-IC (median age 57.91, 64 males). RNA was extracted from bone marrow MNCs and assayed for TET2 exon 2 skipping through real-time quantitative ESRTPCR (qESRTPCR) amplification of E1E3 (spliced) and E2E3 (unspliced) TET2 isoforms. TET2 exon 2 skipping was quantified by calculating the ratio E1E3/E2E3. For statistical analysis, the ratio E1E3/E2E3 was dichotomized using median value as the cutoff value. Skipping of TET2 exon 2 was associated with a significantly lower response rate: 65% vs 92%, p = 0.001 but with a significantly lower relapse rate: 39% vs 85%: p<10-4. Univariate analysis (27 months median follow-up) showed that in addition to age and cytogenetic risk, TET2 exon 2 skipping significantly influenced DFS; the higher the level of E1E3 isoform expression, the better the outcome [HR=0.27 (95% confidence interval (CI):0.13-0.53), p<10-4]. The favorable effect of TET2 exon 2 skipping on DFS remained in the 49 patients with normal karyotype: HR=0.31 (CI: 0.12-0.84), p=0.022. Multivariate analysis showed that age [HR=1.83 (CI:1.03-3.24), p=0.039], cytogenetics [HR=3.03 (CI:1.55-5.96), p=0.001], and TET2 exon 2 skipping [HR=0.38 (CI:0.19-0.77), p=0.007] were independent prognostic factors for DFS. The validation cohort included 103 patients with normal karyotype treated with AraC-IC (median age 61.12, 63 males). Skipping of TET2 exon 2 was associated with a lower response rate: 87.5% vs 91.1% (NS) and a significantly lower relapse rate: 3% vs 66%: p<10-4. Univariate analysis showed that in addition to age, FLT3 internal tandem duplication (FLT3-ITD), nucleophosmin (NPM1) exon-12 gene mutation, TET2 exon 2 skipping significantly influenced DFS, EFS, and OS. In multivariate analysis, TET2 exon 2 skipping was the sole prognostic factor for DFS [HR=0.07 (CI: 0.02-0.3), p<10-4], EFS [HR=0.32 (CI: 0.14-0.74), p=0.008], and OS [HR=0.42 (CI: 0.19-0.90), p=0.025]. As an additional control, the prognostic effect of TET2 exon 2 skipping was evaluated in a series of 36 AML unfit for IC and treated with hydroxyurea, azacitidine, decitabine, low dose AraC, or supportive care only. TET2 exon 2 skipping possessed no significant impact on the response rate, the relapse rate, OS, DFS, and EFS in this population.

In conclusion TET2 exon 2 skipping is associated with a low relapse rate and possesses a strong favorable prognostic impact in AML treated with AraC-CTI. The mechanism linking this alternative exon usage with resistances to AraC are currently investigated while our results suggest that the determination of TET2 exon 2 splicing status might assist risk stratification.


Nicolini:Novartis: Consultancy.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution