Clonal Architecture of TET2 Mutations Influences the Clinical Phenotype of Patients with Myeloid Neoplasms

Somatic TET2 mutations (TET2^MT) are frequently found in myeloid neoplasia (MN) with a frequency that increases with age, consistent with the presence of subclinical TET2^MT clones in asymptomatic older controls. Despite the prominent pathophysiologic role they play based on their prevalence; phenotype-genotype and prognostic associations have not been reliably established. We hypothesized that the diversity of TET2^MT with regard to their configuration, genetic topography, and sub-clonal context, result in differences in biological, morphological, and clinical features.

We studied blood and bone marrow samples from 4930 patients with MN using deep sequencing for common myeloid somatic mutations, including TET2 . This analysis identified 1205 TET2^MT patients with 1781 distinct TET2^MT, as many had more than one TET2^MT. The most common TET2^MT were frame shift (47%) and nonsense (34%) mutations, both gene truncating events. Truncations were subdivided into those occurring upstream enough to effectively delete the whole gene (46%) and those that only led to deletion of the catalytic domain. TET2^MT were found in 17% of patients with MDS, 46% with MDS/myeloproliferative neoplasm (MDS/MPN), 19% with MPN, 21% with pAML and 24% with sAML. Across all MN's, single mutations were more common than biallelic or homozygous, except in MDS/MPN, where there were similar proportions of single and multiple mutant cases (22% single vs . 24% biallelic/homozygous TET2).

TET2^MT co-occurred most commonly with another TET2^MT (43%), ASXL1 (21%), SRSF2 (18%), and NPM1 (13%). A similar distribution was seen in MDS, but with SF3B1 replacing NPM1 ; and in MDS/MPN, but with SF3B1 replaced by RUNX1 and CBL . In sAML the distribution of sub clonal lesions was similar to pAML, though with more mutations in NPM1 and FLT3^ITD/TKD in pAML. Genes with the most significant differences in mutation rates between TET^MTvs . TET2^WT MDS/MPN patients were increased SRSF2 (34% vs . 15%, p<.001) and KRAS (8% vs . 3%, p<.001).

The rank of TET2^MT within the clonal hierarchy is heterogeneous. TET2^MT were found to be ancestral hits in 40% of mutant cases. A second TET2^MT was the most common secondary hit following an ancestral TET2 hit. Other common 2^nd hits included mutations in SRSF2, ASXL1, DNMT3A, and SF3B1 . TET2^MTirrespective of ancestral or secondary configurations had no impact on survival outcome. However, when comparing TET2^MT clonal size, TET2 allelic burden negatively affected survival (p=.016) for all patients with MNs.

Having ascertained that the majority of TET2 lesions are ancestral hits, we set to determine their impact on progression, extent of dysplasia, and proliferation (monocytosis, increase megakaryocytes fibrosis). We used logistic regression to assess odds ratios (OR) of a gene being associated with MDS vs. MDS/MPN. We then determined univariate Cox regression hazard ratios for each gene and used these to construct 2-dimensional regression plots (Figure 1). This plot shows that TET2^MT as a sole ancestral hit implies a greater likelihood of dysplasia and less of progression to high risk (HR) disease as compared to low risk (LR). Subsequent secondary hits change the features of the disease as shown in Figure 1. A second (biallelic) hit in TET2 increased the odds for the disease being myeloproliferative (OR=1.17), a secondary U2AF1 hit increased the odds of progression (OR=2.64), and a sub clonal SRSF2 mutation increased both (OR=6.61, 3.89). RAS family mutations had a moderate effect on progression (OR=2.66, 1.71), but induced proliferative features (OR=8.77, 7.80), whereas SF3B1 mutations associated with dysplasia and LR (OR=0.26, 0.32). In comparison, all combined TET2 hits (irrespective of the secondary context) were plotted as a circle (Figure 1).

In sum, our analyses demonstrate the genotypic/phenotypic impact of primary TET2 hits and how each subsequent mutation can affect the clinical phenotype of the disease.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal