Myeloproliferative neoplasms (MPNs) are characterized by excessive proliferation of one or more cell lines accompanied by the production of a large number of terminally differentiated blood cells. Classic MPNs can be divided into 3 categories: essential thrombocytosis (ET), polycythemia vera (PV), and primary myelofibrosis (PMF). The discovery of driver mutations in JAK2 (V617F or in exon 12), MPL (W515L), and CALR in MPN patients have greatly changed the diagnostic criteria of MPN. Patients who lack the JAK2, CALR, or MPL mutations are described as triple-negative patients and these patients always have a poor prognosis.[2,3]
Genes that are frequently mutated in other types of myeloid tumors, such as the epigenetic modification genes (eg, TET2, DNMT3A, ASXL1, EZH2), also were found to be mutated and to play important roles in MPN; their mutational frequency was 1% to 30%.[4,5] Moreover, patient with mutations in TP53 and TET2 were found to have a significant worse prognosis and higher risk of acute myeloid leukemia transformation than patients who did not have these mutations. MPNs were divided into 8 new genomic subtypes on the basis of a recent study that analyzed mutations in 2035 patients with MPN. In addition, a multivariate statistical model that incorporated 63 clinical and genomic variables was created to forecast disease outcomes, which might help to predict the best therapy on disease progression for newly diagnosed patients. This work was the first large-scale attempt to screen genomic variables and reclassify this disease. However, further confirmation is needed using other centers and prospective clinical trials before this approach can be used to make decisions about patients’ treatment. Some of the genomic variables have been incorporated into our sequencing panels.
Here, we customized a 50-gene target amplicon-based panel for next-generation sequencing to reveal the mutational profiles of 65 Chinese patients with a definite diagnosis of MPN by characterizing mutated genes that are already known to be involved in tumor pathogenesis, and to find novel mutations in triple-negative patients.
Subjects and methods
Sixty-five outpatients with MPN among the Chinese Han population were recruited from the Huashan Hospital and Changhai Hospital between 2015 and 2017 in this retrospective observational study. This study was conducted in accordance with the Declaration of Helsinki protocol and approved by the Institutional Review Board of Changhai Hospital, Naval Military Medical University, China. All the patients were re-diagnosed according to the 2018 World Health Organization criteria and were negative for the BCR–ABL1 fusion transcript. Data from some of these patient have been used in our previous work. Among the 65 patients, 38 were diagnosed with ET, 21 were diagnosed with PV, and 6 were diagnosed with PMF. The characteristics of the patients are given in Table 1. The samples from all 65 patients were sequenced using next-generation sequencing technologies. Informed consents were obtained from all individual participants included in the study.
Five bone marrow and 60 blood samples were collected from the patients for subsequent analysis. High quality DNA was extracted using a DNeasy Blood & Tissue kit (Qiagen, Germany) and quantified by NanoDrop 2000 (Thermo Fisher, San Diego, CA, USA).
Detection of mutations in JAK2, CALR, and MPL
The mutational status of these 3 driver genes was already known from the medical record system. However, JAK2, CALR, and MPL probes were included in the sequencing panel for use as internal positive controls.
Screening of target genes and customization of multiplex PCR primers
Fifty genes associated with the pathogenesis and evolution of MPN were selected by analyzing recent papers in PubMed (Table 2). These genes include mainly RNA splicing genes (SF3B1, SRSF2, U2AF1), chromatin histone modification genes (ASXL1, EZH2), DNA methylation genes (DNMT3A, IDH1, IDH2, TET2), and RAS signal transduction pathway genes (CBL, NRAS, KRAS). Multiplex polymerase chain reaction (PCR) primers were designed using the Ion AmpliSeq Designer (www.ampliseq.com).
The designed primers had a coverage rate of the targeted region of 99.25%. The multiplex PCR primers were divided into 2 pools; pool1 contained primers that amplified 193 amplicons and pool2 contained primers that amplified 182 amplicons. The lengths of the amplified products were 125 to 375 bp. The amount of original DNA used for the amplifications was 10 ng/tube.
PCR amplification and Ampliseq library preparation
As described above, our multiplex PCR primers were divided into two pools, so the PCRs were performed in two 10-μL volume reactions, each reaction contains 5 μL master mix and 5 μL PCR primers. Master mix were prepared before the PCR reactions begin and is composed of Ion AmpliSeq HiFi Mix, targeted DNA, and Nuclease-free Water. Then, the 2 reaction systems were combined after target amplification. The PCR cycles were: 99°C for 2 minutes; followed by 20 cycles of 99°C for 15 seconds, and 60°C for 8 minutes; then 10°C hold overnight. After partial digestion, the mixed amplicons were barcoded with diluted Ion Xpress barcode 1 to 32 adapter mix (MAN0006735; Thermo Fisher) and purified. After the above reactions, these terminal barcoded amplicons were called libraries. The purified libraries were quantified on a 7500 HT Fast Real-Time PCR System (Thermo Fisher) using an Ion AmpliSeq Library kit 2.0 (Thermo Fisher) according to the manufacturer's protocol.
Template preparation and sequencing
The purified libraries were diluted to 90 pM, and 25 μL of each diluted library was pipetted into the bottom of the appropriate Ion Chef Library Sample Tube (MAN0014032; Thermo Fisher). Template preparation and subsequent chip loading were performed on an Ion Chef System (Thermo Fisher) using an Ion 530 chip (Thermo Fisher), then sequenced immediately on an Ion Torrent S5 XL instrument (Thermo Fisher).
Data processing and bioinformatics analysis
The raw data were analyzed using Ion Reporter software (www.ionreporter.thermofisher.com). The resultant BAM file were uploaded to the cloud of Ion Reporter and reformatted as BED files by Ion Ampliseq. All the called mutations were aligned using COSMIC (v85), dbSNP, and Clinvar. The minor allele and phenotype frequencies for the Chinese Han Beijing (CHB) population on the 1000 Genome Project website (www.1000genomes.org) were obtained and used as the reference healthy control.
Statistical analysis was performed using GraphPad Prism 6 (GraphPad Software, La Jolla, CA, USA). To compare the mutation frequencies of genes between the MPN subgroups or between patients and healthy controls, the Chi-square or Fisher exact tests were applied to evaluate statistical significance; P < 0.05 was considered as statistically significant.
Mutation status of JAK2, MPL, and CALR in the ET, PV and PMF samples
At least one of the three typical MPN-related genes (JAK2, CALR, MPL) was mutated in 82% (53/65) of the 65 patients with MPN in this study. The patients with ET, PV, and PMF carried 58% (22/38), 81% (17/21), and 67% (4/6) of the JAK2 mutations; 26% (10/38), 0% (0/21), and 17% (1/6) of the CALR mutations (type-1 and -2); and 3% (1/38), 0% (0/21), and 0% (0/6) of the MPL mutation (W515A), respectively. Similar to the current diagnostic criteria, JAK2 had the highest mutation frequency. Among the 65 patients, all mutations were JAK2 V617F (none were on exon 12), and no untypical mutations were found in CALR or MPL. Two female patients had both JAK2 and CALR mutations, which was rare. In patients with ET, those who had a single JAK2 mutation had a lower JAK2 V617F allele burden than those with double driver gene mutations (20 ± 9% vs 42 ± 7%, P = 0.3). Similarly, in patients with ET, those who had a single CALR mutation had a much lower CALR allele burden than those with double driver gene mutations (8 ± 4% vs 58 ± 9%, P = 0.02).
Mutational profiles of our study cohort
Mutational profiles of all 65 patients
We performed targeted sequencing on all 65 patients using our customized gene panel. The sequencing data are shown in Figure 1. We found that each patient harbored at least one pathogenic mutation and 51 (78%) patients harbored two or more mutations; the average was 1.2 mutated genes per patient. Besides the classical driver mutated genes (JAK2, CALR, MPL), 15 other pathogenic mutations were detected. Three of them, TET2 (n = 26), EZH2 (n = 19), and ASXL1 (n = 7), were mutated in >10% of the patients, and 12, MIR662 (n = 5), MLH1 (n = 5), MLH3 (n = 4), SF3B1 (n = 3), MSH6 (n = 3), BARD1 (n = 3), DNMT3A (n = 2), KIT (n = 2), MSH2 (n = 2), RUNX1 (n = 1), TP53 (n = 1), and NRAS (n = 1), were mutated in <10% of the patients (Fig. 2A). These 15 genes had different functions (Table 2), including DNA methylation, chromatin modification, RAS signal transduction pathway, and RNA splicing (Fig. 2B), but have not been routinely used to diagnose MPN.
Patients who lack mutations in the 3 typical genes (JAK2, MPL, CALR) are referred to as triple-negative patients. Among the 65 patients recruited in this study, 12 were triple-negative; 7 with ET, 4 with PV, and 1 with PMF. Each of these patients harbored at least one pathogenic mutation and a total of 8 different mutations were detected; 4 (33%) patients had two mutations and one (8%) patient had three mutations (n = 1) (Fig. 3). The 8 mutated genes were TET2 (n = 5), EZH2 (n = 5), MIR662 (n = 2), MLH1 (n = 1), ASXL1 (n = 1), MSH6 (n = 1), KIT (n = 1), and MLH3 (n = 1). The average frequency of mutated non-typical genes in the triple-negative patients (1.5 non-typical mutations/patient) did not differ significantly from the average in the non-triple-negative patients (1.2 non-typical mutations/patient) (P = 0.82).
RNA splicing genes show different mutational patterns among the MPN subtypes
We found that 2 of the 6 patients with PMF carried mutations in the genes annotated as RNA splicing (SF3B1, SRSF2, U2AF1), whereas only one of the 38 patients with ET carried these mutations, and these mutations were not detected in the patients with PV. Thus, the RNA splicing genes were more frequently mutated in patients with PMF than in patients with ET (P = 0.006) and were not mutated in patients with PV (Fig. 2C), which is consistent with a previous study. Genes annotated to the other functional categories did not show such differences.
Mutational profile of a patient who progressed from PV to PMF
Studies have shown that mutations in RUNX1, NRAS, ASXL1, EZH2, and TET2 are associated with low survival rates, and patients with complex phenotypes have more mutations at the time of diagnosis and a worse prognosis than patients with simpler phenotypes. One patient (J316) in this study presented with severe anemia and reduced complete blood cells. Blood tests showed that hemoglobin was 32 g/L, mean cell hemoglobin was 23.9 pg, mean corpuscular hemoglobin was 291 g/L, red blood cell count was 1.34 × 1012/L, erythroblasts were 10/100 white blood cells, and immature cells appeared in the peripheral blood. This patient progressed from PV to post-PV PMF, which eventually developed to myelodysplastic syndrome. Seven pathogenic mutations were detected in this patient, which is much higher than the number of such mutations in the other patients with no disease progression. The mutated genes and their allele burden are shown in Table 3. However, no statistical significance could be drawn because of the limited source of samples.
Four single nucleotide polymorphisms that might confer susceptibility to MPNs
We detected single nucleotide polymorphisms (SNPs) in 4 genetic regions (LINC-PINT, THRB-RARB, HBS1L-MYB, GFI1B), namely rs4858647, rs9376092, rs58270997, and rs621940 (Fig. 2D). Genetic variations in 12 such regions are thought to increase the risk of developing MPNs, and are termed as the population attributable risk.
The mutation frequency of HBS1L-MYB was 66% in patients with ET and 52% in patients with PV or PMF. In the patients with ET, the HBS1L-MYB mutation seemed to be more frequent in JAK2-positive patients (68%) than in JAK2-negative patients (64%); the THRB-RARB mutation was more frequent in patients with PMF (67%) than in patients with PV or ET (59%); and the LINC-PINT and GFI1B mutations were more frequent in JAK2-negative patients than in JAK2-positive patients (100% vs 91%, 8% vs 7%, respectively). The mutation frequencies of these 4 SNPs were higher in triple-negative patients than in non-triple-negative patients. However, no statistically significant differences were detected between these indicators.
Furthermore, the phenotype frequency of these SNPs between the 65 patients in our study and the healthy controls (ie, the phenotype frequency for the CHB population on the 1000 Genomes Project (www.1000genomes.com) also did not differ significantly. These discrepancies may be explained by the limited number of samples in our study.
Multiple mutations in mismatch repair-related genes occurred in patients with MPN
By repairing errors in DNA replication, reducing chromatin rearrangements, and mediating cellular responses to specific types of DNA damage factors, DNA mismatch repair (MMR)-related genes help to maintain genomic stability. However, DNA MMR also can introduce mutations under specific conditions.[14–16] Thus, we suspected that the MMR pathway was involved in the pathogenesis of MPN, so we performed targeted sequencing of MMR-related genes in the 65 patients to obtain the mutation profile of these genes. The data showed that multiple types of mutations (SNPs, single nucleotide variations) were present in multiple loci of these genes (Table 4).
The frequency of such mutations in a healthy population was obtained using the CHB data from the 1000-Genome Project (www.1000genomes.com). The SNP mutation frequencies in the MMR-related genes were MLH3 (40%), MSH6 (34%), MSH3 (12%), MLH1 (8%), and MSH2 (7%), and the single nucleotide variation mutation frequencies were MSH6 (8%), MLH3 (5%), and MSH2 (3%). Overall, the mutation frequencies of the MMR-related genes in the 65 patients with MPN were MLH3 (43%), MSH6 (38%), MSH3 (12%), MSH2 (10%), and MLH1 (8%).
Among the various mutations, 11% have been confirmed as pathogenic (MLH1 c.655A > G, c.1151T > A, MSH2 c.23C > T), but the pathogenicity of the others is still controversial. For the 7 detected mutations with unclear clinical significance, we used PROVEAN (provean.jcvi.org) and SIFT to predict their pathogenicity. The prediction results indicated that 4 of them (MSH6 c.118G > C, c.1716G > T, c.2561A > T, and MLH3 c.3241G > C) may be pathogenic (Table 4).
Our customized 50-gene panel includes not only genes that have been confirmed to be associated with BCR–ABL1-negative patients with MPNs, but also genes that were frequently mutated in other types of myeloid tumors. To apply the panel, we performed next-generation sequencing of 65 BCR–ABL1-negative MPN patients. The sequencing data analysis showed that all known JAK2, MPL, and CALR mutations were successfully detected, demonstrating the reliability of the obtained sequencing data. In addition to the typical driver genes, we detected 15 genes (30%) from among the 50 genes on the panel that had atypical pathogenic mutations in at least one of the patients in this study, and 49 of the 65 (75%) patients harbored 2 or more pathogenic mutations. Mutations were detected in the three epigenetic regulator genes TET2 (DNA methylation), EZH2 (chromatin modification), and ASXL1 (chromatin modification genes) in more than 10% of the 65 patients with MPN. TET2 mutations have been found in 5% to 20% of patients with MPN and were predicted to affect gene regulation. Recently, TET2 mutations were shown to reduce the function of the intestinal barrier, which disturbed the expression of interleukin-6 by intestinal bacteria. An increase in interleukin-6 stimulates pre-leukemic transformation by switching the self-renewal hematopoietic stem cell differentiation into the myeloid lineage. Loss-of-function mutations in EZH2 have been detected in myeloid malignancies and are associated with poor survival; they occur most frequently in the myelodysplastic/myeloproliferative neoplasms category (10%–13%). TET2 was most frequently mutated together with EZH2 in patients with myelodysplastic/myeloproliferative neoplasms. We also found that patients with MPN had the highest number of EZH2 mutations, and that TET2 mutations frequently occurred together with EZH2 mutations in both non-triple-negative and triple-negative patients. ASXL1 (additional sex combs like 1) is located in the chromosomal region 20q11 and is affected in hematologic malignancies. The risk of progressing to acute myeloid leukemia in an individual carrying more than a 5% ASXL1 mutation load was doubled. Although in a previous report, none of the patients with MPN who had ASXL1-mutations had a JAK2 V617F mutation, we identified ASXL1 mutations in patients both with and without driver mutations. This difference might be explained by the higher number of patients with MPN in our study. A combined panel of ASXL1, EZH2, SRSF2, IDH1, and IDH2 has been used to predict the prognosis of patients with PMF. In our study, mutations in 2 of these genes (ASXL1, EZH2) were found not only in patients with PMF but also in patients with ET or PV, and the remaining 3 prognostic factors were not detected. This may be because there were only 6 patients with PMF in our study. This study, together with others, confirms that epigenetic modifications in tumor cells are correlated with mutations in specific genes. How many genes and at what stages they were involved in triggering a disease need further investigation.
Patients with MPN who carry multiple driver gene mutations may have a worse prognosis than patients who carry only one typical mutated gene. Such patients did not response well to traditional therapeutic drugs (hydroxyurea, tyrosine kinase inhibitors such as imatinib) and were intolerant to their side effects. A French intergroup of MPN study characterized the clinical and biological properties of patients with MPN who carried two driver mutations. They found that patients with double mutations were often older males who had lower hemoglobin and higher platelet counts than patients who had a single mutated gene. The JAK2 V617F allele burden in these patients also was much lower than in patients with a single-JAK2 mutation; the results were similar for CALR. The identification of the 2 female patients with double driver genes and the patient with 7 pathogenic mutations who progressed quickly from PV to post-PV MF, further confirms and enriches the previously reported data. However, the 2 female patients were only clinically assessed to harbor JAK2 V617F using the the step-wise algorithms for the detection of JAK2, CALR, and MPL mutations suggested by the international baseline workup for patients with Ph-negative MPN, and no further detection of CALR mutations was conducted. The patient with 7 pathogenic mutations also was only clinically assessed to harbor JAK2 V617F during the clinical diagnosis and treatment process. Therefore, simultaneous identification of many more of the important mutations at diagnosis will provide additional clinically useful prognostic information that can be used to select more appropriate treatment strategies, as has been suggested by other researchers.
The most important factor for triple-negative patients with MPN is the relatively poor prognosis, especially for patients with PMF. Sporadic mutations have been detected in epigenetic modification genes (ASXL1, TET2), splicing genes (SF3B1, SRSF2), and regulators of cytokine signaling genes (CBL, SH2B3) in triple-negative patients by applying targeted or whole-exome sequencing technologies. Other atypical mutations (non-JAK2 V617F, MPL W515L/K) also were detected in JAK2 and MPL exons, and these regions are not usually mutated in non-triple-negative patients. The sequencing of 12 triple-negative patients revealed that each patient carried at least one pathogenic mutation, indicating a clonal disease that is different from myeloid hyperplasia caused by reactive inflammatory factors. However, the frequency of these mutations was not very high in these 12 patients, indicating that these genes may not be important driver genes. The next-generation sequencing of the triple-negative patients is still valuable to some extent, because the obtained data can help clinicians confirm their diagnosis and choose treatment strategies. The prognosis of triple-negative patients is still unclear, so continuous monitoring of the mutated genes and their allele burden, combined with clinical data, will be helpful in predicting the prognosis.
Our 50-gene panel detected 4 SNPs that might confer individual susceptibility to patients with MPN. Before the discovery of JAK2 V617F, the germline susceptibility of MPN had been noted, and the most important reason for this is family clustering of MPN. The most well-known molecular variants related to germline susceptibility are the JAK2 46/1 haplotype and TERT (rs2736100), which can explain about 70% of population attributable risk, although germline variants in SH2B3, TET2, ATM, CHEK2, PINT, and GFI1B also are associated with JAK2 V617F-positive clonal hematopoiesis and MPN. In a recent study using 939 patients with MPN, HBS1L-MYB was associated with JAK2 V617F-positive ET, and THRB-RARB had a weak association with PMF.HBS1L-MYB resulted in a decrease in MYB-expressed proteins, causing ET-like phenotypes in mouse models, and differences in disease phenotypes were related to structural variance of MYB-expressed proteins. Although we found no significant differences among the sub-types of MPN, our results provide additional experimental proof of the previous findings. However, LINC-PINT and GFI1B, which have been reported to be associated with JAK2 V617F-positive MPNs, had more mutations in JAK2 V617F-negative patients than in JAK2 V617F-positive patients in our study. This inconsistency might be caused by the rare patients with CALR or MPL mutations in this study.
Previous studies have suggested that MPN and some other hematopoietic malignancies may be the consequence of an inherent genomic instability and inadequate response to DNA damage. Impaired MMR might cause microsatellite instability in therapy-related myelodysplastic syndromes and in patients with acute myeloid leukemia. However, few studies have elucidated the role of MMR-genes in MPN. Here, we detected MMR-genes associated with Lynch syndrome to determine whether mutations in these genes contributed to genomic instability and potentially played some role in the development of MPN. Not surprisingly, different mutations were identified in MLH3, MSH6, MSH3, MSH2, and MLH1, 11% of which were known to be pathological. Importantly, 4 other potential pathogenic mutations were newly identified in our cohort, as predicted by PROVEAN and SIFT. Compared with a recent large-scale study, the mutation types detected in our study are relatively small (single nucleotide variations or indels) because of the relatively small sample size and limitations of the platforms used, so other mutation types (eg, gain/loss of heterozygosity, splicing) were not detected. However, our results suggest that a wide range of mutations have occurred in MMR-genes, which might predisposition individuals to MPN.
In summary, using a customized 50-gene sequencing panel, we performed next-generation sequencing on a relative large number of Chinese patients with MPN to further describe the mutation landscape of classic MPNs. In addition to the driver mutations in JAK2, CALR, or MPL, many patients had additional abnormal mutations that spanned a wide range of cancer genes, with individual variations in the genetic landscape. Among the abnormal mutated genes, for the first time, we demonstrated potential roles of MMR abnormality in the development of MPN. Further studies using larger cohorts are needed to reveal how the mutational landscape affects the disease phenotype and response to therapy. Integrating these important genes into clinical diagnosis and treatment might greatly improve the clinical prognosis and survival of patients with MPN. The continual challenge is how to incorporate the increasing genetic abnormal discoveries into the evolving MPN molecular diagnostic and prognosis algorithms.
MG and GT designed the study. XH, JW, and GT wrote the paper. JW, XD, XX, XZ, and WM performed the study. TH contributed analytic tools. JY, MG, and GT analyzed the data. All authors approved the final version of the paper.
This study was funded by the National Natural Science Foundation of China (No. 81672105) and Naval Military Medical University Special Application Research on Precision Medicine Transformation, China (No. 2017JZ18).
Institutional review board statement
This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Changhai Hospital, Naval Military Medical University, China.
Declaration of participant consent
The authors certify that they have obtained the patient consent forms. In the forms, patients have given their consent for their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity.
Conflicts of interest
The authors declare that they have no conflicts of interest.
. Nangalia J, Massie CE, Baxter EJ, et al Somatic CALR mutations in myeloproliferative neoplasms with nonmutated JAK2. N Engl J Med 2013;369:2391–2405.
. Langabeer SE. Chasing down the triple-negative myeloproliferative neoplasms: implications for molecular diagnostics. JAKSTAT 2016;5:e1248011.
. Chang YC, Lin HC, Chiang YH, et al Targeted next-generation sequencing
identified novel mutations in triple-negative myeloproliferative neoplasms. Med Oncol 2017;34:83.
. Tenedini E, Bernardis I, Artusi V, et al Targeted cancer exome sequencing reveals recurrent mutations in myeloproliferative neoplasms. Leukemia 2014;28:1052–1059.
. Reinig E, Yang F, Traer E, et al Targeted next-generation sequencing
in myelodysplastic syndrome and chronic myelomonocytic leukemia aids diagnosis in challenging cases and identifies frequent spliceosome mutations in transformed acute myeloid leukemia. Am J Clin Pathol 2016;145:497–506.
. Lundberg P, Karow A, Nienhold R, et al Clonal evolution and clinical correlates of somatic mutations in myeloproliferative neoplasms. Blood 2014;123:2220–2228.
. Agarwal MB, Malhotra H, Chakrabarti P, et al Myeloproliferative neoplasms working group consensus recommendations for diagnosis and management of primary myelofibrosis, polycythemia vera, and essential thrombocythemia. Indian J Med Paediatr Oncol 2015;36:3–16.
. Grinfeld J, Nangalia J, Baxter EJ, et al Classification and personalized prognosis in myeloproliferative neoplasms. N Engl J Med 2018;379:1416–1430.
. Huang X, Deng X, Xu X, et al Mutational profile of myeloproliferative neoplasms detected by a customized NGS-based gene panel. Zhonghua Jianyan Yixue Zazhi 2019;42:427–434.
. Tate JG, Bamford S, Jubb HC, et al COSMIC: the catalogue of somatic mutations in cancer. Nucleic Acids Res 2019;47:D941–D947.
. Wei CH, Phan L, Feltz J, et al tmVar 2.0: integrating genomic variant information from literature with dbSNP and ClinVar for precision medicine. Bioinformatics 2018;34:80–87.
. Delic S, Rose D, Kern W, et al Application of an NGS-based 28-gene panel in myeloproliferative neoplasms reveals distinct mutation
patterns in essential thrombocythaemia, primary myelofibrosis and polycythaemia vera. Br J Haematol 2016;175:419–426.
. Hinds DA, Barnholt KE, Mesa RA, et al Germ line variants predispose to both JAK2 V617F clonal hematopoiesis and myeloproliferative neoplasms. Blood 2016;128:1121–1128.
. Iyer RR, Pluciennik A, Napierala M, et al DNA triplet repeat expansion and mismatch repair. Annu Rev Biochem 2015;84:199–226.
. Scott LM, Rebel VI. JAK2 and genomic instability in the myeloproliferative neoplasms: a case of the chicken or the egg? Am J Hematol 2012;87:1028–1036.
. Zhou T, Chen P, Gu J, et al Potential relationship between inadequate response to DNA damage and development of myelodysplastic syndrome. Int J Mol Sci 2015;16:966–989.
. Villa F, Iacca C, Molinari AF, et al Inhalation versus endovenous sedation in subarachnoid hemorrhage patients: Effects on regional cerebral blood flow. Crit Care Med 2012;40:2797–2804.
. Meisel M, Hinterleitner R, Pacis A, et al Microbial signals drive pre-leukaemic myeloproliferation in a Tet2-deficient host. Nature 2018;557:580–584.
. Rinke J, Müller JP, Blaess MF, et al Molecular characterization of EZH2 mutant patients with myelodysplastic/myeloproliferative neoplasms. Leukemia 2017;31:1936–1943.
. Carbuccia N, Murati A, Trouplin V, et al Mutations of ASXL1 gene in myeloproliferative neoplasms. Leukemia 2009;23:2183–2186.
. Abelson S, Collord G, Ng SWK, et al Prediction of acute myeloid leukaemia risk in healthy individuals. Nature 2018;559:400–404.
. Vannucchi AM, Lasho TL, Guglielmelli P, et al Mutations and prognosis in primary myelofibrosis. Leukemia 2013;27:1861–1869.
. De Roeck L, Michaux L, Debackere K, et al Coexisting driver mutations in MPN: clinical and molecular characteristics of a series of 11 patients. Hematology 2018;23:785–792.
. Mansier O, Luque Paz D, Ianotto JC, et al Clinical and biological characterization of MPN patients harboring two driver mutations, a French intergroup of myeloproliferative neoplasms (FIM) study. Am J Hematol 2018;93:E84–E86.
. Langabeer SE. Double-mutant myeloproliferative neoplasms. Med Oncol 2018;35:137.
. Tashi T, Swierczek S, Prchal JT. Familial MPN predisposition. Curr Hematol Malig Rep 2017;12:442–447.
. Trifa AP, Bănescu C, Bojan AS, et al MECOM, HBS1L-MYB, THRB-RARB, JAK2, and TERT polymorphisms defining the genetic predisposition to myeloproliferative neoplasms: a study on 939 patients. Am J Hematol 2018;93:100–106.
. Tapper W, Jones AV, Kralovics R, et al Genetic variation at MECOM, TERT, JAK2 and HBS1L-MYB predisposes to myeloproliferative neoplasms. Nat Commun 2015;6:6691.