Recurrent SETD2 mutation in NPM1-mutated acute myeloid leukemia

Abstract

SETD2 is the only methyltransferase for H3K36me3, and our previous study has firstly demonstrated that it functioned as one tumor suppressor in hematopoiesis. Consistent with it, SETD2 mutation, which led to its loss of function, was identified in AML. However, the distribution and function of SETD2 mutation in AML remained largely unknown. Herein, we integrated SETD2-mutated AML cases from our center and literature reports, and found that NPM1 mutation was the most common concomitant genetic alteration with SETD2 mutation in AML, with its frequency even higher than MLL rearrangement and AML1-ETO. Though this result indicated the cooperation of SETD2 and NPM1 mutations in leukemogenesis, our functional study showed that SETD2 was required for the proliferation of NPM1-mutated AML cell line OCI-AML3, but not MLL-rearranged AML cell line THP-1, via maintaining its direct target NPM1 expression, which was just opposite to its role of tumor suppressor. Therefore, we speculated that SETD2 possibly had two different faces in distinct subtypes and stages of AML.

To the editor

SETD2 has been demonstrated as one tumor suppresser in hematopoiesis [1], and SETD2 mutation affected AML, in which its distribution remained not fully understood [2]. Herein, we analyzed the SETD2 mutation in NPM1-mutated AML.

One 36-year-old woman was committed due to abdominal pain and fever for 7 and 3 days, respectively. PB test showed WBC: 52.4 × 10⁹/L, Hb: 98 g/L, PLT: 48 × 10⁹/L, circulated blast: 80%. BM examination exhibited 67.5% myeblasts with the immunophenotype of CD11b-CD13^dim + CD14-CD15^dim + CD33 + CD34^partial + CD35-CD38^dim + CD45 + CD64-CD65^dim + CD71 + CD117 + CD123^dim + HLA-DR-. Though karyotype was normal and CBF or MLL rearrangements were negative, NPM1, SETD2, NRAS and ETV6 mutations were identified. Therefore, AML with mutated NPM1 was diagnosed. After receiving the operation for co-existed acute appendicitis, she accepted IA regimen as induction therapy, and CR1 was achieved. Subsequently, she received three cycle of medium-dose cytarabine regimen. However, AML relapsed at the 3 months after cessation of chemotherapy, and 72% myeloblasts re-emerged in BM. Due to the early recurrence, she accepted HAA and CLAG regimen successively, and achieved CR2. However, the leukemic clones were not eradicated reflected by persistent above mutations. Therefore, allogeneic semi-compatible HSCT was immediately conducted. As follow-up, CR was still maintained at the 15 months after HSCT (Fig. 1a).

Fig. 1

The distribution of SETD2 mutation in AML. a One SETD2-mutated AML case in our center. b SETD2^R2109X-mutated cancers in our report and COSMIC database. c SETD2-mutated AML cases from literature reports. d The common concomitant genetic alterations with SETD2 mutation in AML patients from literature reports. e and f The frequency of SETD2 mutation in NPM1^Mut or NPM1^WT AML (e) and it of NPM1 mutation in SETD2^Mut or SETD2^WT AML (f) were calculated, and all AML cases from our study and literature reports mentioned above were involved

In this patient, SETD2^R2109X was identified, and it was also found in other malignancies from COSMIC database (Fig. 1b), so SETD2^R2109X was one driver in cancer. However, SETD2 deficiency was not sufficient to generate AML, so additional hits were required [1, 3]. Therefore, we reviewed AML studies involving SETD2 mutation [2, 4,5,6,7,8,9], and found that NPM1 mutation rather than MLL rearrangement or AML1-ETO was the most common co-existed genetic alteration of SETD2 mutation in AML (Fig. 1c-d). To establish their association, we displayed subgroup analysis in above studies, then submitted it Pearson’s chi-square test, and calculated OR. Strikingly, SETD2 and NPM1 mutations were the concomitant mutation in AML (P = 0.031; OR = 3.28) (Fig. 1e-f). To address whether SETD2 mutation mediated drug resistance in AML, we analyzed their therapeutic response to standard chemotherapy. Among 22 SETD2-mutated AML patients, the data were available in 11 patients, while CR, PR, and NR was 72.7%, 9.09%, and 18.2%, respectively. Notably, the CR was comparable to it in total AML. Interestingly, all with NPM1-mutated AML achieved CR, and two with MLL-rearranged AML exhibited NR. Therefore, SETD2 mutation was possibly not one determinant in drug sensitivity for AML. Furthermore, we analyzed the OS between SETD2- mutated and wild-type groups with cBioPortal database [10, 11], but no significance between two groups was found (Additional file 1: Figure S1). Regretfully, the data about EFS were not available.

Loss of SETD2 function accelerated the progression of MLL-rearranged or AML1-ETO-positve AML, but whether it was the same in NPM1-mutated AML remained unknown. Herein, we displayed shRNA-mediated SETD2 knockdown, which simulated its loss of function caused by SETD2 frame-shift or nonsense mutation, in NPM1-mutated AML cell line OCI-AML3 and MLL-rearranged AML cell line THP-1. Interestingly, SETD2 knockdown impaired the proliferation of OCI-AML3 but not THP-1 cells (Fig. 2a-d). Furthermore, the proliferative defect of OCI-AML3 was caused by increased cell apoptosis (Fig. 2e) and cell cycle arrested at G1/G0 phase (Fig. 2f). It has been reported that the viability of OCI-AML3 relied on the function of NPM1 mutation [12], while NPM1 expression was regulated by the transcriptional activation mark, H3K36me3, which indicated by ChIP-Seq in the HSPCs of Mll-af9-positive leukemia (Fig. 2g) [13]. Consistently, we demonstrated that NPM1 and its direct targets MEIS, HOXA9 were significantly down-regulated in SETD2 knockdown OCI-AML3 cells (Fig. 2h-i). Therefore, our results indicated that SETD2 knockdown-mediated OCI-AML3 proliferation inhibition was possibly attributed to NPM1 down-regulation.

Fig. 2

SETD2 was required for the maintenance of NPM1-mutated AML cell line OCI-AML3. a and b The proliferation (a) and SETD2 expression (b) of scramble and SETD2 knockdown OCI-AML3 cells. c and d The proliferation (c) and SETD2 expression (d) of scramble and SETD2 knockdown THP-1 cells. e Annexin-V staining for detecting cell apoptosis in OCI-AML3 cells. f PI staining for cell cycle analysis in OCI-AML3 cells. g NPM1 has been demonstrated as one direct target of H3K36me3 in the literature report. h and i The expression of NPM1 (h) and its direct downstream targets, MESI and HOXA9 (i), was analyzed in scramble and SETD2 knockdown OCI-AML3 cells. ***, P < 0.001; **, P < 0.01; *, P < 0.05; T test was used for each graph

The detailed role of SETD2 mutation in NPM1-mutated AML remained mysterious. Theoretically, SETD2 and NPM1 mutations probably cooperated in leukemogenesis. However, our results showed that SETD2 was required for the maintenance of OCI-AML3. To our knowledge, two possibilities existed: firstly, SETD2 mutation played different roles in the initiation and maintenance of NPM1-mutated AML; secondly, additional genetic alteration influenced SETD2 function in NPM1-mutated AML. Therefore, further investigations were needed in the furture.