Next-Generation Sequencing Mutational Landscape and Clinical Features of Chinese Adults with Myeloproliferative Neoplasms

Background: Myeloproliferative neoplasms (MPNs) include three classical subtypes: polycythemia vera (PV), essential thrombocythemia (ET), and primary myelobrosis (PMF). Since prebrotic primary myelobrosis (pre-PMF) was recognized as a separate entity in the 2016 revised classication of MPN, it has been a subject of debate among experts due to its indenite diagnosis. However, pre-PMF usually has a distinct outcome compared with either ET or overt PMF. In this study, we examined the clinical, haematologic, genetic, and prognostic differences among pre-PMF, ET, and overt PMF. Methods: We retrospectively reviewed the clinical parameters, haematologic information, and genetic mutations of patients who were diagnosed with pre-PMF, ET, and overt PMF according to the WHO 2016 criteria using next-generation sequencing (NGS). Results: Pre-PMF patients exhibited higher leukocyte counts, higher LDH values, a higher frequency of splenomegaly, and a higher incidence of hypertension than ET patients. On the other hand, pre-PMF patients had higher platelet counts and haemoglobin levels than overt PMF patients. Molecular analysis revealed that the frequency of EP300 mutations was signicantly increased in pre-PMF patients compared with ET and overt PMF patients. In terms of outcome, male sex, along with symptoms including MPN-10, anaemia, thrombocytopenia, and KMT2A and CUX1 mutations, indicated a poor prognosis for PMF patients. Conclusion: The results of this study indicated that comprehensive evaluation of BM features, clinical phenotypes, haematologic parameters, and molecular proles is needed for the accurate diagnosis and treatment of ET, pre-PMF, and overt PMF patients. NGS; next-generation sequencing; BM: bone marrow; OS: overall survival; HMR: high-molecular-risk; BMI: body mass index; VAF: variant allele frequency; ALL: acute lymphoblastic leukaemia; AML: acute myeloid leukaemia; LFS: leukaemia-free survival.


Patients
We conducted a retrospective study of patients with BCR-ABL-negative MPN from October 2014 to June 2020 in the Fourth A liated Hospital of Zhejiang University School of Medicine. Patients who were diagnosed with ET, pre-PMF or overt-MF according to the 2016 WHO Classi cation were included.
Patients with incomplete information from laboratory tests and medical records were excluded. This study was approved by the Ethics Committee of the Fourth A liated Hospital of Zhejiang University School of Medicine in line with the principles of the Declaration of Helsinki.

Data Collection
We retrieved patient characteristics, including sex, age, history of smoking, drinking, and thrombotic events, splenomegaly, cardiovascular risk factors, diabetes, hypertension, MPN-associated symptoms, body mass index (BMI), leukocyte counts, platelet counts, the level of LDH, and haemoglobin, and the variant allele frequency (VAF) of mutated genes. Full blood cell counts and LDH levels were taken as maximum values before diagnosis transformation, and other laboratory data were taken at diagnosis. OS was calculated as the interval between diagnosis and death or date of last follow-up.

Molecular analysis
Genomic DNA was puri ed from bone marrow with the Gentra Puregene Blood Kit (Qiagen, Hilden, Germany) according to the protocol. High-throughput gene sequencing was performed using ultrahigh multiple PCR exon enrichment technology with an average sequencing depth of 800×. Mutation analysis was performed using the Ion Reporter System and Variant Reporter Software.

Statistical analysis
Statistical analyses were performed using SPSS statistics 25.0 and GraphPad Prism 8.0 software. For comparisons among these groups, the Pearson chi-square test was used for categorical factors, single factor analysis of variance or t-test was used for normally distributed continuous variables, and the Kruskal-Wallis test (3 groups) or Mann-Whitney U test (2 groups) was used for continuous variables that did not conform to the normal distribution. The association between the clinical characteristics or molecular characteristics and OS was analysed using Kaplan-Meier curves and compared using the logrank test. P < 0.05 was considered statistically signi cant.

Patient cohorts and characteristics
A total of 93 patients with classical BCR-ABL-negative MPN were included: 47 patients were diagnosed with ET, 13 with pre-PMF, and 33 with overt PMF. The median age was 53 (19-79) for ET patients, 67 (32-72) for pre-PMF patients, and 70 (29-88) for overt PMF patients. The male/female ratio was 1/1.76 for ET patients, 1/1.17 for pre-PMF patients, and 1/0.65 for overt PMF patients. The main clinical and haematologic characteristics are summarized in Table 1, and differences were noted. We found that the frequency of splenomegaly, hypertension, leukocyte counts, and LDH levels were signi cantly higher in pre-PMF patients than in ET patients. The haemoglobin and platelet counts were higher in pre-PMF patients than in overt PMF patients. Of note, history of thrombosis, smoking, drinking, diabetes and mutational frequencies of JAK2V617F, CALR and MPL were similar among the three groups ( Fig. 1,2).
The 84 genes were divided into 12 categories: signalling pathway, DNA methylation, histone methylation, chromatin modi ers, spliceosome and RNA metabolism, myeloid transcription factors, transcription factors, tumour suppressors, cohesin complex, cell adhesion, cell metabolism, and others. We found that spliceosome and RNA metabolism mutations and transcription mutations were signi cantly distinct in the three groups. Further analysis revealed that the genes belonging to the spliceosome and RNA metabolism were more frequently mutated in patients with overt PMF (6.8%) than in those with ET (1.0%, P = 0.033). In contrast, transcription factor genes were more frequently mutated in patients with ET (10.4%) than those with in overt PMF (1.6%, P = 0.008) ( Table 2). Abbreviations: ET, essential thrombocythemia; PMF, primary myelo brosis.
We analysed the 18 genes with mutation frequencies > 10%, and the results revealed that TET2, EP300, and TP53 mutations differed among the three disease entities. In detail, EP300 was more often mutated in patients with pre-PMF (60%) than in those with ET (10%; P = 0.012) or overt PMF (15.79%; P = 0.042).
Furthermore, TET2 and TP53 mutations were also signi cantly more frequent in patients with overt PMF than in those with ET (P = 0.004 and P = 0.014, respectively) ( Table 3).

The relationships between clinical and mutation presentations and OS in PMF
The median follow-up was 36 months for ET, 42 months for pre-PMF, and 53 months for overt PMF.
Overall survival between pre-PMF, overt PMF, and ET was signi cantly different (P < 0.001), as shown in Fig. 5. During the follow-up time, only one death of ET was registered, so we analysed the impact of clinical parameters and mutational status at diagnosis on outcome in PMF, including pre-PMF and overt PMF. We performed Kaplan-Meier curves to examine the relationships between the clinical parameters and patient survival. We found that male sex (P = 0.0107), MPN-10-related symptoms (P = 0.0354), anaemia (haemoglobin < 120g/L; P = 0.0239), and thrombocytopenia (platelet count < 100 ×10 9 /L, P = 0.0002) were signi cantly related to inferior OS (Fig. 6).
To evaluate whether the OS of PMF might be in uenced by mutation status, we also generated Kaplan-Meier curves for PMF, which revealed that mutations in KMT2A (P = 0.0386) and CUX1 (P = 0.0444) were signi cantly associated with worse OS in PMF. TP53 mutation, a widely accepted prognostic factor in PMF, was associated with poor survival in our study, but the difference was not signi cant (P = 0.0538). Moreover, no statistical signi cance was found in the other 15 genes for OS (Fig. 7).

Discussion
In the present study, we reviewed the clinical, haematologic, and molecular parameters of patients who were diagnosed with ET, pre-PMF, and overt PMF according to the 2016 revised WHO criteria. A body of evidence has been produced by several groups indicating that pre-PMF presents more aggressive features than ET but milder features than overt PMF (3,7,8,(10)(11)(12)(13)(14). We observed that pre-PMF patients had higher leukocytes, LDH values, and splenomegaly frequency than ET patients and higher platelet counts and haemoglobin levels than overt PMF patients, consistent with historical literature, verifying the value of clinical and haematologic parameters in the precise diagnosis of pre-PMF. Moreover, this analysis reinforced the signi cance of applying the 2016 revised WHO criteria in clinical practice.
In addition, this study revealed that pre-PMF patients exhibited a higher incidence of hypertension than ET patients. This phenomenon might be due to the presence of leukocytosis in pre-PMF patients, as an elevated leukocyte count was associated with increased blood pressure value and hypertension (15,16).
The results of this study also indicated that EP300 mutations were more frequent in pre-PMF patients, and the frequencies were 10%, 60%, 15.79% in ET, pre-PMF, and overt PMF, respectively. EP300 is a transcriptional coactivator and major lysine acetyltransferase that connects sequence-speci c DNAbinding factors with the basal transcriptional machinery to regulate transcription and acetylates transcription factors and histones to facilitate transcription (17). EP300 mutation, which occurs frequently in follicular lymphoma (FL) and diffuse large B-cell lymphoma (DLBCL), impairs the tness of germinal centre B cells and abrogates germinal centre formation to promote the incidence of DLBCL and alters tumour-associated macrophage polarization via the FBXW7-NOTCH-CCL2/CSF1 axis to accelerate the progression of DLBCL (18, 19). However, there have been few studies pertaining to EP300 mutations in MPN. In EP300 mutant mice, disordered megakaryocytopoiesis, peripheral blood thrombocytosis independent of thrombopoietin signalling, erythroid hypoplasia, and anaemia that met the characteristics of MPN implied that this mutation presumably plays an important role in the incidence of MPN (20). A published study assayed EP300 mutations in 56 MPN patients (polycythemia vera = 24, ET = 19, PMF = 6, MPN-unclassi able = 7) and revealed only a single base change in EP300 exon 29 of polycythemia vera patients (21). The distinction between it and our study might contribute to the small number of PMF patients and testing methods of EP300 mutation. Additional investigations of the relationship between EP300 mutation and pre-PMF are warranted.
Furthermore, male sex, MPN-10-related symptoms, anaemia, thrombocytopenia, and KMT2A and CUX1 mutations indicated poor outcomes in PMF patients. Therefore, PMF patients with the aforementioned factor should be monitored more intensively in the clinical course. Unlike HMR mutations, the prognostic effect of KMT2A and CUX1 mutations in PMF patients has rarely been reported in published studies. KMT2A (previously known as MLL) mutation is recognized as an important prognostic factor of acute lymphoblastic leukaemia (ALL) and acute myeloid leukaemia (AML) (22,23). KMT2A mutation was relevant to the incidence of ALL and AML (23,24), which might contribute to inferior OS in PMF patients.
In addition, the CUX1 mutation was identi ed as a mutation that may be involved in MPN development and phenotype modi cation (25). A recent study revealed that the CUX1 mutation was associated with inferior OS and leukaemia-free survival (LFS) (26) consistent with our research. The negative prognostic implication of CUX1 mutation may be related to the risk of progression to leukaemia (27)(28)(29). Our ndings call for further studies to validate the prognostic impacts of KMT2A and CUX1 mutations in PMF patients.

Conclusion
In conclusion, we identi ed differences in the clinical, haematologic, and molecular presentations of ET, pre-PMF, and overt PMF patients, indicating that comprehensive evaluation of not only BM features but also clinical, haematologic, and molecular pro les is needed for accurate diagnosis and treatment of these three disease entities. The molecular analysis revealed that pre-PMF might be relevant to EP300 mutation, demonstrating the value of molecular examination.  Differences in quantitative parameters in ET, pre-PMF, and overt PMF. (A) Overt PMF patients were older than ET patients; (B) overt PMF patients had lower body mass index (BMI) than ET patients; (C) patients with pre-PMF and overt PMF patients had higher LDH value than ET patients, but no statistically signi cant difference was found between pre-PMF and overt PMF; (D) the white blood cell (WBC) counts were higher in pre-PMF and overt PMF patients; (E) the haemoglobin (HB) level was lower in overt PMF patients compared with ET and pre-PMF patients; and (F) overt PMF patients had lower platelet (PLT) counts than ET and pre-PMF patients.

Figure 2
The differences in qualitative parameters in ET, pre-PMF, and overt PMF. (A) ET patients had a lower frequency of splenomegaly than pre-PMF and overt PMF patients; (B) pre-PMF and overt PMF patients were more likely to have hypertension than ET patients; (C) overt PMF patients were more likely to have a higher frequency of cardiovascular risk factors than ET patients; (D) patients with overt PMF were more frequently symptomatic than patients with ET; (E) the percentage of smokers was similar among the three groups; (F) the percentage of drinkers was similar among the three groups; (G) the history of thrombosis was similar among the three groups; (H) the sex distribution was different between ET patients and overt PMF patients; (I) the history of diabetes was similar among the three groups; (J) the JAK2V617F mutation rate was similar among the three groups; (K) the CALR mutation was distributed evenly among the three groups; and (L) the MPL mutation was distributed evenly among the three groups.