Myeloproliferative neoplasms (MPNs) are chronic, clonal hematopoietic disorders that include polycythemia vera (PV), essential thrombocytosis (ET), and primary myelofibrosis (PMF). A substantial proportion transforms into acute myeloid leukemia (AML) with a dismal prognosis.1 Mutually-exclusive hotspot driver mutations in JAK2, CALR, and MPL genes are present in 90% of PV, ET and PMF. The remaining 10% of cases are called triple negatives (TN). Mutations in other genes (e.g., ASXL1, DNMT3A, EZH2, IDH1/2, SH2B3, SRSF2, TET2) are also recurrent in MPNs.52 To better define the genetic events involved in MPN evolution and transformation,6532 we studied 57 MPNs in chronic phase and 38 post-MPN AMLs from the IPC/CRCM Tumor Bank (authorization #AC-2007-33). Patients were appropriately informed and asked to consent in writing, in compliance with French and European regulations.
MPN diagnosis was established according to the revised criteria of the World Health Organization.7 The 57 chronic cases were separated into two groups: 40 without acute evolution, arbitrarily called “steady” MPNs or sMPNs (median follow up was 17.9 years (11–28)); and 17 that transformed to AML, called tMPNs (median time between diagnosis of chronic and acute phases was 8.5 years (1–29)). The latter group had 17 matched post-MPN AML samples among the 38 cases of post-MPN AML. Genomic analysis was carried out for most patients in chronic phase during disease course, not at diagnosis. Median time between diagnosis and sample was 12.9 years (2–24 years) for sMPNs, and 5.7 years (0–25) for tMPNs (Online Supplementary Table S1). Post-MPN AML samples were obtained at diagnosis aside from 2 samples obtained during accelerated phase. History of the diseases is detailed in Online Supplementary Table S2.
DNAs from peripheral blood leukocytes (n=87) or bone marrow aspirates (n=10) were extracted and prepared for sequencing as described.98 For each sample, a library of all coding exons and intron-exon boundaries of 79 genes, selected on the basis of their known involvement in the pathogenesis of myeloid malignancies (Online Supplementary Table S3), was constructed with HaloPlex target enrichment system (Agilent technologies, Santa Clara, USA) and sequenced on an Illumina MiSeq (Online Supplementary Methods). Copy number aberrations (CNAs) were identified by using genome-wide, high-density 244K CGH Microarrays (Hu-244A, Agilent Technologies, Massy, France) as described.8 Fisher’s exact tests were used for qualitative variables with discrete categories, and Student’s t-test for continuous variables. P<0.05 was considered statistically significant.
In the 56 MPNs at chronic phase and 38 post-MPN AMLs analyzed by aCGH, we found CNAs in 40 samples. We classified aCGH profiles into three types. In type I profiles, called “normal-like”, no CNA was detected. Type II (intermediate profiles, between one and three CNAs) showed interstitial deletions/breaks/gains that affected few or single genes within a chromosomal region, such as deletions in 1p31 (NF1A), 4q24 (TET2), 8q21 (RAD21), 12p13 (ETV6, CDKN1B), Xq25 (STAG2), 17q11 (NF1, SUZ12), 20q12-q13 (L3MBTL1, PTPN1, PTPRT), 21q22 (RUNX1) or isolated gain or loss of one chromosome or chromosome arm (+8, −7, 1q+). Type III comprised complex profiles with more than three CNAs. In the 57 MPNs and 38 post-MPN AMLs analyzed by NGS, we retained 139 single-nucleotide variants (SNVs) and small insertions or deletions (indels) affecting 45 genes with at least one mutation in 92.8% of samples (Online Supplementary Tables S2 and S4).
Both sMPNs and tMPNs had a majority of “normal-like” aCGH profiles (Figure 1 and Online Supplementary Table S2). Rare abnormalities such as del12p13, del20q12-q13 and +1q were observed; they may be associated with worse prognosis,11104 and should be taken into account at diagnosis. tMPNs patients were older than sMPNs at diagnosis (60.4 years-old vs. 46.6, P=0.0012) and more mutated than sMPNs (median number 3.18 vs. 1.38) (Online Supplementary Tables S1). tMPNs were more JAK2V617F-mutated (82.4% vs. 47.5%), less CALR-mutated (0% vs. 35%), and as often TN (11.8% vs. 10%). The frequency of alterations (mutations and CNVs) grouped according to the affected cellular function is depicted in Table 1. Genes involved in RNA splicing (SRSF2 P=0.001), chromatin modification (ASXL1 P=0.007), and signaling pathways were more affected in tMPNs (52.9%, 35.3% and 23.5% of the cases, respectively) than in sMPNs (2.5%, 7.5% and 2.5% of the cases, respectively) (Figure 1, Table 1).
Figure 1.Mutational spectrum and copy number aberrations in myeloproliferative neoplasms (MPNs) and post-MPN acute myeloid leukemias (AMLs). Driver mutations of MPNs are designated by squares shaded in grey and J (JAK2 V617F), C (CALR mutations), M (MPLW515); and white squares correspond to triple negatives. Genes studied are grouped according to the cellular function they are involved in. Colored squares indicate mutations. sMPNs: “steady” MPNs; tMPNs: MPNs that transformed to AML; ET: essential thrombocytosis; PV: polycythemia vera; MF: myelofibrosis; ND: not determined.
Table 1.Frequency of gene alterations (mutations and CNVs) grouped according to the affected cellular function: comparison between sMPNs and tMPNs and comparison between total MPNs and total post-MPN AML cohorts.
In the 38 post-MPN AMLs, the median time for acute transformation (AT) was 10.6 years, and the median age at diagnosis was 70.5 years (Online Supplementary Table S1). Fifteen of the 38 cases (39.5%) had a type III aCGH profile, 12 (31.6%) a type II, and 11 (28.9%) a type I; 57.9% were JAK2V617F-mutated, 13.2% were CALR-mutated, 1 was MPL-mutated, and 26.3% were TN. The median number of mutations in post-MPN AMLs was higher than in sMPNs (3.92 vs. 1.38, P=3.33*10) but the same as in tMPNs (3.92 vs. 3.18, P=0.26) (Figure 1, Table 1). The median number of mutations per case was different between JAK2-, CALR-, MPL-mutated, and TN cases (4.5 mutations, 2.8, 2 and 3.4, respectively). CALR-mutated cases had fewer other mutations than JAK2-mutated cases (P=0.04). Three of the 5 CALR-mutated post-MPN AMLs had a TP53 mutation and an aCGH type III profile (60%). In vivo cooperativity of TP53 alterations with JAK2V617F has already been reported,5 but not yet with CALR mutations. Half of the AMLs were mutated in genes involved in DNA methylation: TET2 (21.1%), DNMT3A (18.4%) and IDH1/2 (15.8%). Seventeen cases (44.7%) were mutated in RNA splicing genes: SRSF2 (15.8%), U2AF1 (13.2%), SF3B1 (10.5%), and PRPF8 (5.3%). TP53 mutations were found in 13 cases (34.2%) (Figure 2A, see Online Supplementary Figure S1 for TP53 mutations in our study and in others). Thirteen cases (34.2%) were mutated in genes involved in cell signaling: FLT3 (13.2%), NRAS (7.9%), and NF1 (7.9%). NF1-mutated cases also had NF1 deletions. Thirteen cases (34.2%) had mutations in transcription regulators with frequent damaging mutations of RUNX1 (13.2%). Twelve cases (31.6%) were mutated in chromatin regulators: ASXL1 (26.3%) and EZH2 (13.2%). (Figure 1, Table 1, Online Supplementary Table S2).
Figure 2.TP53, ASXL1 and DNMT3A mutations in chronic MPNs and/or post-MPN AMLs. (A) TP53 mutations were more frequent in post-MPN AMLs than in MPNs; (B) TP53 mutations were associated with aCGH type III complex profiles; (C) High TP53 VAF was associated with aCGH type III; (D) TP53 mutations were more frequent in post-ET/PV/sMF AML, whereas ASXL1 mutations were associated with post-PMF AML; (E) DNMT3A mutations were associated with post-ET/PV AML; (F) TP53, ASXL1 and DNMT3A genes are not often co-mutated in post-MPN AML. ET, essential thrombocytemia; PV, polycythemia vera; PMF, primitive myelofibrosis; sMF, secondary myelofibrosis.
When we compared the mutational status of the 57 MPNs to that of the 38 post-MPN AMLs, we found that mutations in genes involved in DNA methylation, tumor suppressors, signaling, RNA splicing and transcription factors were more frequent in post-MPN AMLs. Also, whilst chronic phases were associated with “normal-like” aCGH profiles (45/56, 80.4%), the 38 AMLs had more type III profiles (15/38, 39.5%) (Table 1).
Seventeen matched chronic phases were available for the 38 post-MPN AML patients. For these pairs, median time between diagnosis of chronic and acute phases was 8.5 years (Online Supplementary Table S1). The acute phases had an average gain of one mutation (4.24 vs. 3.18), and a worse aCGH profile (35.3% type I, 47.1% type II, 17.6% type III) than the chronic phases (75% type I, 25% type II) (P=0.037). VAF evolution between chronic and acute phase for each patient is represented in Online Supplementary Figure S3. For most mutations in chronic phase, the median VAF was around 40%, and there was no increase of VAF in acute phases (VAF variation <20%), suggesting that these mutations were present in the founding clone (as previously demonstrated6) (Online Supplementary Figure S2). In some cases, an increase of VAF was supported by a loss of heterozygosity (EZH2: AML43; TP53: AML48, 51; TET2: AML42). We also observed the apparition of TP53 mutations with an average VAF of 40.2% in three cases (AML44, 49, 51) (Online Supplementary Figure S3A). Considering all cases, when the TP53 VAF was >10%, the patients generally displayed a complex aCGH profile and complex karyotype, suggesting that mutated TP53 confers genomic instability. Conversely, when TP53 VAF was <10%, the cytogenetic analysis could not detect complex abnormalities (Figure 2B and 2C).
These results suggest that genome alterations and mutations in TP53, DNA methylation, and transcription factors are simultaneous with (and potentially responsible for) evolution to AML, whereas alterations in RNA splicing, chromatin modification, and signaling pathways, which are found in chronic phases before AT, may increase AT risk (Online Supplementary Figure S3). These results are in agreement with recent studies of myelodysplastic syndromes12 and AML patients.13 In two recent NGS studies of 197 MPNs in chronic phase6 and 33 post-MPN AMLs5, the distribution of mutations according to the cellular function affected was similar to that of our study. Here, we also compared the mutational status of chronic phase MPNs with stable course to chronic phase MPNs that transformed to AML, which showed a different profile (Table 1).
Post-PV/ET or post-secondary MF AMLs (13/31) had more TP53 mutations than post-PMF AMLs (0/7). Conversely, they had less ASXL1 mutations (4/31 vs. 6/7), and this was independent of their JAK2/CALR/MPL mutational status (Figure 2D). DNMT3A mutations were more frequent in post-ET/PV AMLs than in post-MF (secondary or PMF) AMLs (7/20 vs. 0/18) (Figure 2E). In total, 25 out of the 38 post-MPN AMLs were ASXL1 and/or TP53, ASXL1 and/or DNMT3A, TP53 and/or DNMT3A, with very rare co-mutated cases (Figure 2F). These results suggest the existence of three mechanisms of AT depending on the MPN subtype, with TP53 or DNMT3A mutation in post-ET/PV AML and ASXL1 mutations in PMF.
Knowledge of mutations, including early detection of small abnormal clones, could modify therapeutic choices. In particular, the detection of mutations at diagnosis (e.g., ASXL1, DNMT3A, splice, TP53) and during disease course, may help identify an MPN with risk of AT and help adapt the treatment.
References
- Cervantes F, Tassies D, Salgado C, Rovira M, Pereira A, Rozman C. Acute transformation in nonleukemic chronic myeloproliferative disorders: actuarial probability and main characteristics in a series of 218 patients. Acta Haematol. 1991; 85(3):124-127. PubMedhttps://doi.org/10.1159/000204873Google Scholar
- Abdel-Wahab O, Pardanani A, Patel J. Concomitant analysis of EZH2 and ASXL1 mutations in myelofibrosis, chronic myelomonocytic leukemia and blast-phase myeloproliferative neoplasms. Leukemia. 2011; 25(7):1200-1202. PubMedhttps://doi.org/10.1038/leu.2011.58Google Scholar
- Beer PA, Delhommeau F, LeCouedic JP. Two routes to leukemic transformation after a JAK2 mutation-positive myeloproliferative neoplasm. Blood. 2010; 115(14):2891-2900. PubMedhttps://doi.org/10.1182/blood-2009-08-236596Google Scholar
- Brecqueville M, Rey J, Devillier R. Array comparative genomic hybridization and sequencing of 23 genes in 80 patients with myelofibrosis at chronic or acute phase. Haematologica. 2014; 99(1):37-45. PubMedhttps://doi.org/10.3324/haematol.2013.091454Google Scholar
- Rampal R, Ahn J, Abdel-Wahab O. Genomic and functional analysis of leukemic transformation of myeloproliferative neoplasms. Proc Natl Acad Sci U S A. 2014; 111(50):E5401-E5410. PubMedhttps://doi.org/10.1073/pnas.1407792111Google Scholar
- Lundberg P, Karow A, Nienhold R. Clonal evolution and clinical correlates of somatic mutations in myeloproliferative neoplasms. Blood. 2014; 123(14):2220-2228. PubMedhttps://doi.org/10.1182/blood-2013-11-537167Google Scholar
- Arber DA, Orazi A, Hasserjian R. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood. 2016; 127(20):2391-2405. PubMedhttps://doi.org/10.1182/blood-2016-03-643544Google Scholar
- Bertucci F, Finetti P, Guille A. Comparative genomic analysis of primary tumors and metastases in breast cancer. Oncotarget. 2016; 7(19):27208-27219. Google Scholar
- Brecqueville M, Rey J, Bertucci F. Mutation analysis of ASXL1, CBL, DNMT3A, IDH1, IDH2, JAK2, MPL, NF1, SF3B1, SUZ12, and TET2 in myeloproliferative neoplasms. Genes Chromosomes Cancer. 2012; 51(8):743-755. PubMedhttps://doi.org/10.1002/gcc.21960Google Scholar
- Harutyunyan A, Klampfl T, Cazzola M, Kralovics R. p53 lesions in leukemic transformation. N Engl J Med. 2011; 364(5):488-490. PubMedhttps://doi.org/10.1056/NEJMc1012718Google Scholar
- Klampfl T, Harutyunyan A, Berg T. Genome integrity of myeloproliferative neoplasms in chronic phase and during disease progression. Blood. 2011; 118(1):167-176. PubMedhttps://doi.org/10.1182/blood-2011-01-331678Google Scholar
- Pellagatti A, Roy S, Di GC. Targeted resequencing analysis of 31 genes commonly mutated in myeloid disorders in serial samples from myelodysplastic syndrome patients showing disease progression. Leukemia. 2016; 30(1):248-250. Google Scholar
- Lindsley RC, Mar BG, Mazzola E. Acute myeloid leukemia ontogeny is defined by distinct somatic mutations. Blood. 2015; 125(9):1367-1376. PubMedhttps://doi.org/10.1182/blood-2014-11-610543Google Scholar