Search for a command to run...
Mutations associated with chronic myeloproliferative neoplasms (MPNs), polycythemia vera (PV), essential thrombocythemia (ET) and primary myelofibrosis (PMF), include MPN phenotype-driver genes, JAK2, MPL and CALR, plus “myeloid neoplasm associated” mutations (MNAMs).1 The MNAMs affect genes with functions in epigenetics, spliceosome, oncogenesis and signal transduction. They are found also in myelodysplastic syndromes and acute myeloid leukemias, as well, usually at low variant allele frequency (VAF), in subjects with “age-related clonal hematopoiesis” (ARCH). Most frequently affected genes are TET2, DNMT3A, ASXL1, SF3B1, SRSF2, IDH1, IDH2, TP53. The MNAMs may be diagnostically useful to establish clonal hematopoiesis in ET and PMF patients lacking driver mutations (triple negative cases); furthermore, selected MNAMs (ASXL1, SRSF2, EZH2, IDH1/2 and U2AF1) are prognostically informative, and are included in the molecularly-enhanced MIPSS-70 score for patients with PMF.2 Mutations in the nuclear factor erythroid-2 gene (NFE2) were reported recently in MPNs3; these mutations originate premature stop codons that affect protien's DNA binding function. Mice expressing NFE2 mutations develop erythrocytosis and thrombocytosis, due to expansion of progenitor and stem cell compartment, and in the long term may develop leukemias with concomitant and/or isolated myelosarcoma.4 In the original report on 456 MPN patients, NFE2 mutations were discovered in 3 PV and 5 PMF patients, accounting for 1.7% of the total, and 2.1% and 2.6% of PV and PMF cases.3 In the largest study available, that provided sequence information on 2040 MPN patients, NFE2 mutations were found in 1.2%, 4.5% and 2.9% of ET, PV and MF patients, suggesting some enrichment in PV patients.1 The NFE2 frame-shift and missense mutations were discovered in four of six patients with isolated myelosarcoma.5 Whether NFE2 mutations associate with unique clinical features, or outcome, remains debated. To this end, we analyzed all unselected cases of MPNs in our data base that had the entire NFE2 coding sequence available (N = 631). The study was approved by IRB, in accordance with the Helsinki Declaration; patients provided informed, written consent. Mutation analysis was performed on granulocytes with a 24-gene panel, as described.2, 6 A threshold of ≥5% VAF was chosen to include any variant previously reported in the literature or considered damaging by in silico analysis. Survival was calculated from diagnosis to death or last follow-up, or censored at stem cell transplantation. The cumulative probability of overall survival (OS) and leukemia-free survival (LFS) was calculated by the Kaplan-Meier method; differences were estimated by the log-rank test. Continuous variables were analyzed with the Mann-Whitney U test (two groups) or Kruskal-Wallis test with Dunn's method for multiple comparisons. A P < .05 was considered statistically significant. This series included 257 patients with ET, 130 PV, 244 PMF. Patients were diagnosed, or the original diagnosis was reviewed, according to the 2016 WHO classification; their clinical characteristics are reported in Table S1. We found 22 unique variants in 35 patients with NFE2, accounting for 5.5% of the entire series, 3.9% of patients with ET, 8.7% PV and 4.1% PMF. Such preferential association of NFE2 mutations with PV is in agreement with a previous report.1 Most patients (n = 16; 46%) had mutations located between the NH2-terminal transactivation and the DNA binding domain (Figure 1); our figure is lower than reported by Jutzi et al (six of 8 patients with mutations; 75%)3 but comparable to Grinfeld et al. (17 of 35; 48%).1 Considering together two previous studies and the current cohort, a total of 35 codons were targets of mutation, of which only nine were shared in ≥2 of the three cohorts, indicating great mutation heterogeneity; of interest, no mutation described in MPN patients was listed in myelosarcoma patients.5 We consider the 261st codon as a hotspot, accounting for 20% of all mutated cases in our series, and 43% of all NFE2 mutations located in the domain between amminoacid 206 and 268, similar to Grinfeld et al.1 (26% and 53%, respectively). The majority of patients had frameshift mutations (n = 17, 49%), while point mutations leading to amino acid changes were found in 34% (n = 12), and deletions in 17% (n = 6). All mutations were heterozygous; the variant allele frequency (VAF) was 33.2% ± 14.4%, range 5% to 52%, resulting no different according to diseases: 36.1% ± 14.7%, 34.5% ± 14.7 and 32.0 ± 14.3% in ET, PV and PMF. In ET, NFE2 mutations were preferentially associated with MPL (30%) and CALR (40%) mutations (30%; P = .008). While there was no direct correlation between NFE2 and JAK2V617F VAF in different diseases, the frequency of NFE2 mutations was double in patients with >50% VAF (9.7% vs 4.1%, P = .04). Regarding the MNAMs, we found co-mutation patterns of NFE2 with mutations of ASXL1, EZH2, TET2, KIT and SH2B3 (Figure S1). Correlative analysis of NFE2 mutation with hematologic and clinical phenotype of the individual diseases (Table S1) disclosed only marginal differences depending on the mutation status. Among patients with ET and PMF, NFE2 mutation showed a trend of association with younger age, while in PMF they were associated with lower platelet count. In ET, NFE2 mutated patients had higher rate of myelofibrotic transformation (70%) compared to wild type (36%; P = .04). However, we cannot exclude the concurrent role of CALR and MPL mutation that were prevalent among NFE2 mutated patients; due to small numbers a multivariable analysis was not feasible. There was no impact of NFE2 mutations on clinical phenotypes, including rate of cardiovascular events or bleeding. As regarded overall survival, presence of NFE2 mutation showed only a trend for improved survival across the three diseases; HR for survival was 0.40 (95% CI, 0.1-1.8; P = .23) in ET, 0.71 (95% CI, 0.2-3.1; P = .66) in PV and 0.46 (95%CI, 0.1-1.9; P = .27) in PMF. In summary, our data confirm that NFE2 mutations are rare, heterogenous mutations with a prevalence of frameshift and deletion, and are slightly enriched in PV. With the limitation of a small number of patients, they do not appear to have meaningful hematologic and clinical correlates, nor contribute prognostic information. The correlation with progression to myelofibrosis in ET patients with would suggest some role for NFE2 in myelofibrosis development, and merits further studies. This work was supported by AIRC 5x1000 call “Metastatic disease: the key unmet need in oncology” to MYNERVA project, #21267 (MYeloid NEoplasms Research Venture AIRC). A detailed description of the MYNERVA project is available at http://www.progettoagimm.it. All authors declare no conflict of interest. P.G., A.M.V. designed the research, analyzed the data, and wrote the manuscript. A.P., E.C., G.R., N.B., S.F. performed molecular analysis, and contributed to data analysis; G.C., F.M., L.M., B.S., G.G.L., C.P. provided patient samples and clinical information; All authors have read and approved the final version. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.