Bi-allelic TET2 alterations are frequently found in NPM1 mutated AML and constitute a distinct subgroup with unfavorable prognosis

Abstract Background: Mutations (mut) in TET2 and NPM1 are common in myeloid neoplasms, with NPM1mut defining a distinct genetic entity in acute myeloid leukemia (AML). However, NPM1mut does not always constitute the initiating event and AML-NPM1 represents a heterogeneous group with considerable phenotypic and genetic variability. Generally, AML-NPM1 is associated with a favorable prognosis, especially in the absence of FLT3-ITD. TET2mut, also prevalent in AML, often occurs as an early event in AML development, and double hit (dh) TET2 alterations (alt) are frequently detected. TET2dh is associated with a poor prognosis but, so far, is not part of any genetic AML (sub-) classification. Around 20% of NPM1mut cases also have TET2mut, however, the clonal evolution and cellular origin of these cases remains to be investigated. Aim: Analyzing biallelic TET2alt and concurrent NPM1mut to determine the sequence of genetic events and characterization of these cases. Patients and methods: We characterized 479 patients (pts) with TET2dh and concurrent NPM1mut. Clinical data was available for 290 pts. Cryopreserved cells of 10 pts were processed with the Single-Cell DNA + Protein Sequencing Protocol (MissionBio). 65 pts were analyzed by whole transcriptome sequencing (WTS). 141 AML-NPM1 without TET2dh were used as control cohort (n=51 for overall survival (OS) analysis). Results: Of 3,276 pts diagnosed with AML-NPM1, 479 were found to harbor a TET2dh, accounting for 15% of all AML-NPM1. The median age was 74 years [30-93 years], with 51% being female. Immunophenotypic data available for 285 pts showed aberrant CD56 expression in 74% of pts. 392/479 (82%) showed ≥2 TET2mut (mut+mut), 63 (13%) a TET2mut with an accompanying CN-LOH (mut+LOH), and 24 (5%) had a TET2mut and a TET2del (mut+del). In 41% of TET2dh AML-NPM1 pts, the corrected TET2mut variant allele frequency (VAF) was >15% higher than that of NPM1, suggesting that NPM1mut often occurred as secondary event. This contrasts with AML NPM1+ cases with single TET2 hit (sh, n=255 patients), where only 21% showed a similar pattern (p < 0.0001), with no significant differences regarding age and sex (TET2sh: median age 70 years, 59% female) between the two groups. The secondary nature of NPM1mut in TET2dh cases was confirmed by proteogenomic single-cell analysis, which further revealed that in TET2 mut+mut and TET2 mut+del, all cell types (including B and T cells) carried TET2mut, whereas in TET2 mut+LOH, only progenitor and myeloid cell populations harbored TET2. Additionally, TET2dh AML-NPM1 cases exhibited significantly reduced OS compared to AML-NPM1 cases with NPM1 as initiating event (median 14 months (m) vs. 44m, p=0.031). This effect was observed in particular for TET2 mut+LOH (8m vs. 44m, p=0.004) and TET2 mut+del (4m vs. 44m, p=0.006). The negative impact of TET2dh on OS was also evident when considering FLT3-ITD status. Mutational profile analysis showed TET2dh AML-NPM1 has a distinct co-mutation landscape from AML-NPM1 without TET2dh: TET2dh pts showed a lower frequency of FLT3-ITD (TET2dh vs. without TET2dh, 31% vs. 49%), DNMT3A (27% vs. 54%), PTPN11 (7% vs. 14%), IDH1 (1% vs. 12%) and IDH2 (2% vs. 19%) mut, while SRSF2 (26% vs. 9%)and ASXL1 (10% vs. 1%)mut were more frequent. In line with this, TET2dh AML-NPM1 pts with available longitudinal diagnostic data showed a high frequency of CMML, MDS or MPN diagnoses prior to occurrence of NPM1mut. WTS analysis revealed that 5/65 samples with low NPM1 VAF clustered with CMMLs. Median time of progression from MDS/MPN to AML-NPM1 (defined by occurrence of NPM1mut) was 36 months.Conclusions: (1) TET2dh AML-NPM1 constitutes a significant proportion (15%) of all AML-NPM1 and has been identified and characterized as a distinct subgroup. TET2dh AML-NPM1 pts differ from other AML-NPM1 cases in several ways: (i) they exhibit a unique co-mutation pattern with increase in MDS-related mut and a decrease in AML-associated mut; (ii) NPM1mut predominantly occur as secondary event in a TET2 background; (iii) they show a worse OS than AML-NPM1 cases with NPM1mut as the initiating event; (iv) they are characterized by a high frequency of aberrant CD56 expression, which in AML can indicate a poor prognosis. (2) Data from pts prior to occurrence of NPM1mut and WTS analyses suggest that some of these cases may originate from CMML. (3) Due to the unique features of this subgroup, we suggest that TET2dh AML-NPM1 should be considered as a separate entity.