EZH2 Mutations Are Related to Low Blast Percentage in Bone Marrow and -7/del(7q) in De Novo Acute Myeloid Leukemia

The purpose of the present work was to determine the incidence and clinical implications of somatic EZH2 mutations in 714 patients with de novo acute myelogenous leukemia by sequencing the entire coding region. EZH2 mutations were identified in 13/714 (1.8%) of AML patients were found to be more common in males (P = 0.033). The presence of EZH2 mutations was significantly associated with lower blast percentage (21–30%) in bone marrow (P<0.0001) and -7/del(7q) (P = 0.025). There were no differences in the incidence of mutation in 13 genes, ASXL1, CBL, c-KIT, DNMT3A, FLT3, IDH1, IDH2, MLL, NPM1, NRAS, RUNX1, TET2, and WT1, between patients with and without EZH2 mutations. No difference in complete remission, event-free survival, or overall survival was observed between patients with and without EZH2 mutation (P>0.05). Overall, these results showed EZH2 mutation in de novo acute myeloid leukemia as a recurrent genetic abnormality to be associated with lower blast percentage in BM and -7/del(7q).

EZH2, located in 7q36.1, is another important gene associated with epigenetic regulation of transcription. EZH2 encodes the catalytic component of the polycomb repressive complex 2 (PRC2), which is responsible for the methylation of lysine 27 on the N-terminal tail of histone H3 (H3K27), which influences stem cell renewal by epigenetic modification. [13][14] Macro-and microdeletions involving EZH2 have been detected in about 10% of MDS patients [15], and a few MDS patients have shown loss-ofheterozygosity caused by acquired uniparental disomy. [16][17] Somatic mutations of EZH2 were recently identified in lymphoma and myeloid neoplasmas. In lymphoma, somatic mutations affecting codon 641 of EZH2 were detected in 7% of follicular lymphomas and 22% of diffuse large cell B-cell lymphomas of germinal center origin, resulting in enhanced lymphomagenesis. [18] In myeloid neoplasms, mutations were found throughout the EZH2 and have been described in 10-13% of poor-prognosis myelodysplasia-myeloproliferative neoplasms (MDS/MPN), 13% of myelofibrosis (MF), and 6% of MDS. [16][17] However, the prevalence and clinical significance of somatic EZH2 mutations in patients with acute myelogenous leukemia (AML) remains largely unknown.
In this study, we investigated the prevalence and prognostic value of somatic EZH2 mutations in 714 patients with de novo AML by PCR amplification of the entire coding region of EZH2 followed by direct bidirectional DNA sequencing. Patients were also assessed for the presence of mutations in 13 genes, including ASXL1, CBL, c-KIT, DNMT3A, FLT3, IDH1, IDH2, MLL, NPM1, NRAS, RUNX1, TET2, and WT1.

Patients
From January 2005 to December 2010, a total of 714 patients who had been newly diagnosed with AML at Jiangsu Institute of Hematology (JIH) were enrolled in the present study. Patients with antecedent hematologic diseases, especially agnogenic hematocytopenia, and those experiencing MDS, MDS/MPN, or therapyrelated AML were excluded. In order to exclude AMLs attributable to MDS, all the samples from patients with karyotypes -7, 7q-, -5, 5q-, 20q-, or BM-blast less than 30% were examined retrospectively, and no evidence of morphologic myelodysplasia or atypical localization of immature progenitor (ALIP) was found. Diagnosis and classification of AML were defined according to the French-American-British classification (FAB) system and were revised using the World Health Organization (WHO 2008) classification system. The main characteristics of the patients in this study and the entire group are summarized in Table S1. This study was approved by the Ethics Committee of the First Affiliated Hospital of Soochow University. Written informed consent was provided for sample preservation and genetic analysis from every subject. For minors (56 patients, age 8-17 y) written informed consent of the parent or guardian was also obtained. Genomic DNA was extracted from frozen bone marrow mononuclear cells (BMMCs) after Ficoll gradient centrifugation using standard procedures.
For acute promyelocytic leukemia (APL) patients with t (15;17), all-trans-retinoic-acid-and arsenic-trioxide-based treatments were used in induction and consolidation therapy. Patients with non-APL AML were induced with standard first-line treatment, specifically the DA-like regimen, which consisted of daunorubicin (45 mg/m 2 , d1-3) and cytarabine (100-150 mg/m 2 , d1-7). With regard to the consolidation therapy, high-dose cytarabine based chemotherapy was performed on young patients. In addition, 27 patients received allogeneic hematopoietic stem cell transplantation (allo-HSCT) or autologous HSCT.

Statistical Analysis
Patient characteristics were analyzed by chi-square (x 2 ) or Fisher exact tests for univariate analysis. Kaplan-Meier analysis was used to evaluate patient survival. The log-rank test was used to compare survival difference. Binary logistic regression and COX model was used for the multivariate analysis of CR, EFS and OS, respectively. P-values less than 0.05 were deemed significant. All calculations were performed using the SPSS software package (version 13.0).

EZH2 Mutations in de novo AML
In this study, a total of 14 EZH2 mutations were documented in 13 patients with de novo AML. These involved exons 2, 9, 13, 17, and 18 (one case each) and exons 4, 12, 14, and 20 (2, 2, 3, and 2 cases, respectively). In this way, the frequency of somatic EZH2 mutations was 13/714 (1.8%) in de novo AML patients analyzed. The EZH2 mutations detected in the present study are listed in Table 1 and depicted graphically in Figure S1. Nonsense, frameshift, and missense mutations accounted for 7.1% (1/14), 28.6% (4/14), and 64.3% (9/14) of all of EZH2 mutations, respectively, showing a heterozygous pattern. Mutation of EZH2 Y641 was not found in this group of patients. Four out of fourteen mutations were located in the conserved catalytic SET domain (amino acids 618-731), which is essential to the methyltransferase activity of EZH2. We analyzed the genomic DNA of 3 EZH2mutated cases upon diagnosis and again during remission. Genomic DNA confirmed the somatic origin of EZH2 mutations (frame-shift mutations: Tyr297fs, Glu731fs, Cys534fs; missense mutation: His516Asn).

Correlation of EZH2 Mutations with Clinical and Cytogenetic Features
A comparison of clinical characteristics of patients with and without EZH2 mutations is shown in Table 2. Analysis of gender distribution in EZH2-mutated AML patients showed that EZH2 mutations in male patients (11/396; 2.8%) were 4.6 times more common (11/396; 2.8%) than in female patients (2/318; 0.6%) with AML (P = 0.033). Patients with EZH2 mutations showed lower BM blast counts (P,0.0001), than patients without EZH2 mutations. Among de novo AML patients with 20-30% of blast in BM, 14.3% (4/28) of them harbored EZH2 mutations. However, EZH2 mutations were detected in only 1.3% (9/686) of de novo AML patients with .30% of blast in BM. No difference in frequency of EZH2 mutations was observed between adult and pediatric patients (2.4% vs. 1.8%, P.0.05).

Association of EZH2 Mutations with Other Molecular Abnormalities
To investigate the interactions of gene mutations in the leukemogenesis of de novo AML, a comprehensive mutational screening of EZH2 and 13 other genes was performed. After    (3)

Impact of EZH2 Mutations on Response to Therapy and Clinical Outcome
For correlation with clinical outcome, patients with APL were excluded. EZH2 mutations had no influence on achieving complete remission (CR) (Figure 1). The CR rates of cases with and without EZH2 mutations were not significantly different (61.5% vs. 75.0%, P = 0.329). No significant difference in eventfree survival (EFS) (3 year rates, 33.1% vs. 45.2%; P = 0.2283) or overall survival (OS) (3 year rates, 41.0% vs. 49.1%; P = 0.5001) with a median follow-up of 46.6 months (range, 1-79 months) were observed.
Collectively, the results of the present study demonstrate that EZH2 mutations could be detected in a substantial proportion of patients with de novo AML and that these mutations occurred more frequently in male patients than in female patients. No difference in CR rates, EFS, or OS was observed between AML patients with and without EZH2 mutation. The presence of EZH2 mutations in AML was found to be closely associated with lower BM blast percentage (20-30%) and -7/del(7q). In the FAB classification of MDS, patients with 21-30% of blasts in BM were defined as having refractory anemia with excess blasts in transformation (RAEB-t). [23] The WHO classification criteria lowered the threshold to 20% for the number of blasts required for the diagnosis of AML. [1] This arbitrary threshold in blast percentage eliminated RAEB-t in MDS and patients in this category were considered to have AML. However, several reports have argued that the biology of RAEB-t is distinct from that of AML and should be considered a subtype of MDS. [24][25] In the present study, results showed the frequency of EZH2 mutations in patients with 21-30% of BM blasts to be much higher than that in patients with .30% of blasts in BM (14.3% versus 1.3%, P,0.0001). In addition, EZH2 is located on the distal part of chromosome 7q. EZH2 mutations were observed more frequently in cases with -7/del(7q) than in cases without -7/del(7q) (7.7% versus 1.6%, P = 0.025). It was concluded that somatic mutations of EZH2 might play an important role in pathogenesis of de novo AML patients with -7/del(7q) or with 21-30% of blasts in BM, which is classified as RAEB-t in FAB classification. Because of rarity of EZH2 mutations in de novo AML, the prognostic impact of EZH2 mutations in AML is still uncertain, and will need to be assessed in larger cohorts of patients collected on multi-center cooperative studies, though there no significant difference in EFS or OS was observed between EZH2 mutated patients and wild-type in the present study.