NPM1 Mutant Mediated PML Delocalization and Stabilization Enhances Autophagy and Cell Survival in Leukemic Cells.

Accumulating evidence has defined nucleophosmin 1 (NPM1) mutation as a driver genetic event in acute myeloid leukemia (AML), whereas the pathogenesis of NPM1-mutated AML remains to be fully elucidated. In this study, we showed that mutant NPM1 elevated autophagic activity and autophagic activation contributed to leukemic cell survival in vitro. Meanwhile, we also found high expression of promyelocytic leukemia gene (PML) and its cytoplasmic dislocation in primary NPM1-mutated AML blasts and NPM1-mA positive OCI-AML3 cells. Mechanically, mutant NPM1 interacted with PML and mediated it delocalization as well as stabilization. Notably, NPM1-mA knockdown impaired autophagic activity, while induced expression of PML reversed this effect. Finally, we confirmed that PML modulated autophagic activity via AKT signal. These findings suggest that aberrant PML expression and autophagy are beneficial to the leukemic transformation driven by NPM1 mutations. This indicates an attractive therapeutic avenue for PML targeting and/or autophagy inhibition in the treatment of NPM1-mutated AML.

Introduction

Acute myeloid leukemia (AML) is a clinically, cytogenetically, and molecularly heterogeneous disease. Recurrent cytogenetic abnormalities help define subgroups with different prognosis, and identify patients who might benefit from targeted therapies 1. However, the clinical and molecular features of patients with cytogenetically normal AML (CN-AML) accounting for about 50% of all adult AML are still poorly understood 2. High-throughout DNA sequencing of thousands of primary AML patient samples have benefited us better knowledge of molecular landscape of CN-AML 3. Mutation in nucleophosmin 1 (NPM1) gene is one of the most frequently genetic lesions in about 30% of all AML and 50-60% of CN-AML 4. NPM1 mutation resulted in the loss of a nucleolar localization signal (NLS) and the acquisition of a novel nuclear export signal (NES), giving rise to skewed cytoplasmic accumulation of the NPM1 mutant protein (NPM1c+) 5. Because of distinct biological and clinical features, NPM1-mutated AML has been defined as a distinct molecular leukemia entity in the recent updated World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia 6-8. During the past decade, intensive efforts have been done to determine the contribution of NPM1 mutation in the pathogenesis of AML. Studies in vitro have evidenced that NPM1c+ could mediate tumor suppressors such as PTEN 9 and Fbw7γ 10 depletion from the nucleus, aiding to apoptosis resistance and proliferation induction. We previously identified the regulation role of NPM1 mutation in myeloid differentiation block and invasion promotion through upregulating miRNA-10b and matrix metalloprotease (MMPs), respectively 11, 12. Moreover, analysis derived from mouse models of NPM1-mutated AML has revealed the cooperation of NPM1 mutation with key molecular events to induce AML 13, 14. Although NPM1 mutation has been characteristically defined as a driver genetic event in AML, the precise mechanism of action of the NPM1 mutant in AML remains to be fully elucidated, and other contributing oncogenic events are still needed to be investigated.

Autophagy, literally 'self-eating', is a highly conserved and closely regulated homeostatic process and mediates the removal of dysfunctional or damaged organelles, which are digested and recycled for cellular metabolic need in all eukaryotic cells 15, 16. Mount of evidence has demonstrated an essential role of autophagy in the regulation of cancer progression including leukemia 17, 18. Preclinical investigations have documented that autophagic activation induced by histone deacetylase inhibitors (HDACi) may promote a selective advantage in t (8; 21) AML cell survival 19. Consequently, pharmacological inhibition of autophagy was showed to potentiate anticancer activity induced by chemotherapeutic drug in primary leukemia cells 20. In addition, the genetic ablation of autophagy was found to impede megakaryopoiesis and led to severely impaired platelet production and failed hemostasis 21. However, investigation conducted on the potential significance of autophagy in the pathogenesis of NPM1-mutated leukemia is limited.

It has become apparent that both oncogenes and tumor suppressor genes help maintain the physiological autophagic homeostasis of the cell. Dysregulation of these genes often trigger an autophagic switch leading to autophagic cell death or survival 22-24. The tumor suppressor promyelocytic leukemia (PML) consisting of nine exons is located on chromosome 15 and was first discovered in acute promyelocytic leukemia (APL) 25. Being the central component of discrete sub-nuclear structures named PML-nuclear bodies (PML-NBs), PML plays multiple roles in a wide variety of biological processes such as apoptosis, proliferation and cell metabolism 26-28. In addition, PML was also defined as a member of TRIM (tripartite motif) family due to the presence of RING, B-boxes, coiled-coil (RBCC) domain, named as TRIM 19 29. Recently, analysis of small interfering RNA screen indicated that a large number of TRIM proteins may represent as a new class of autophagic regulators 30. Although there is no definitive evidence that PML is involved in autophagy in cancers, Huang et al. 31 have provided compelling evidence that PML-RARα fusion protein induced constitutive autophagic activation in myeloid leukemia U937 cells. Missiroli et al. 32 have provided further evidence that PML is indeed required for autophagy regulation and cancer development. Notably, aberrant localization of PML was observed in NPM1-mutated leukemic cells. Moreover, PML degradation induced by arsenic trioxide (ATO) has been implicated a potential therapeutic relevance in NPM1-mutated AML primary samples 33, 34. However, whether the aberrant PML expression participated in autophagic process and cell survival in NPM1-mutated AML remains unknown.

In light of these findings, we sought to evaluate the role of autophagy in NPM1-mutated leukemic cells and on the mechanistic consequences of aberrant PML expression. Here, we demonstrated that autophagic activation contributed to cell growth facilitated by NPM1 mutation type A (NPM1-mA). In addition, NPM1-mA interacted with PML and led to cytoplasmic delocalization and stabilization of PML protein. Importantly, enforced PML expression could rescue the effect of NPM1-mA knockdown on autophagy and cell proliferation. The study points to a potential therapeutic strategy for PML and/or autophagy targeting in NPM1-mutated AML therapy.

Methods

Patient Samples

The peripheral blood of 26 AML patients newly diagnosed through cytomorphology, cytogenetic and molecular genetic analysis, were obtained from Southwest Hospital of the Third Military Medical University and the First Affiliated Hospital of Chongqing Medical University. Samples were enriched for mononuclear cells by Ficoll gradient purification. The isolated mononuclear cells were used for LC3, p62 and PML relative expression analysis. Informed consent in accordance with the Declaration of Helsinki was obtained from the individuals examined, and the related study was approved by the Institutional Review Board of the Southwest Hospital of The Third Military Medical University and the First Affiliated Hospital of Chongqing Medical University. Details of the clinical characteristics of patients are provided in Table .

Cell culture

Human myeloid leukemia cell lines KG-1a, HL-60, THP-1, K562 and human embryonic kidney cells HEK293T were obtained from American Type Culture Collection (ATCC, USA). Human myeloid leukemia cell line OCI-AML3 (harboring NPM1 mutation type A, NPM1-mA) 35 was obtained from Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ, Germany). The myeloid leukemia cell lines were maintained in RPMI-1640 medium (Gibco, USA) supplemented with 10% fetal bovine serum (FBS, USA) and 1% penicillin-streptomycin (Sangon biotech, China). HEK293T cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin-streptomycin. Cultured cells were incubated in a humidity chamber (Thermo Fisher Scientific, USA) containing 5% CO2 at 37 ˚C.

Reverse transcription PCR and quantitative real-time PCR

Total RNA was isolated using TRIzol reagent (Takara, Japan) and transcribed into cDNA using PrimeScript™ RT reagent Kit (Takara, Japan). Quantitative real time-PCR (qRT-PCR) amplifications of LC3, p62, PML and NPM1-mA genes were performed using SYBR Green (KAPA Biosystems, USA) on a CFX ConnectTM real-time system (Bio-Rad, USA) with β-actin as a normalization control. Cycling conditions were 5 min at 94 °C for the initial denaturation, and amplification was performed with 40 cycles of 30 s at 94 °C, 30 s at 59 °C (for LC3), 56 °C (for p62 and NPM1-mA) or 56.4 °C (for PML), 50 s at 72 °C, and finally 10 min at 72 °C for extension. The melt curve was (65~95) °C at the rate of 0.5°C/0.05 s. Three independent experiments were performed. Details of the primer sequences used are shown in Table .

Western blot

Following the designated treatments, cells were harvested and lysed in ice-cold RIPA lysis buffer supplemented with protease inhibitor cocktail (Roche, Switzerland). Equal amount of proteins were separated by SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes. The specific primary antibodies used in this study were: rabbit monoclonal antibody LC3 І/II (Novus, USA, 1:2000), rabbit monoclonal antibody p62 (Abcam, USA, 1:2000), mouse monoclonal antibody Flag (sigma, USA, 1:1000), rabbit polyclonal mutant NPM1 antibody (Abcam, USA, 1:1000), rabbit polyclonal antibody PML (Abcam, USA, 1:1000), rabbit polyclonal antibody GFP (Abcam, USA, 1:1000), rabbit monoclonal antibody Bcl-2 and Bax (Santa, USA, 1:500), rabbit polyclonal antibody AKT and pAKT (CST, USA, 1:1000), rabbit polyclonal antibody β-actin (Proteintech, USA, 1:2000). Corresponding secondary antibodies were used, and visualization of the products was done using an ECL detection kit (Millipore, USA).

Immunocytochemistry

OCI-AML3 and THP-1 cells were washed with PBS and cytospun onto coverslips at 500 g for 5 min, fixed with 4% paraformaldehyde for 20 min and permeabilized with 1% Triton for 15 min at room temperature. Following blocking with 1% bovine serum albumin in PBS for 30 min, cells were immunostained with mouse monoclonal PG-M3 antibody (Santa, USA, 1:200) overnight at 4 °C. The primary antibody was revealed using the immunoalkaline phosphatase APAAP technique. Cells were counterstained with hematoxylin and mounted in neutral gum and analyzed using a bright field microscope.

Immunofluorescence

Following the designated treatments, cells were washed and cytospun on slides and fixed for 20 min in PBS containing 4% parafomaldehyde, permeabilized in 1% Triton X-100 for 20 min, and incubated in blocking buffer (5% FBS in PBS) for 30 min. Cells were ringed with PBS and incubated overnight at 4 °C in dilution buffer containing primary antibodies. The cells were washed three times with PBS before being incubated with an appropriate fluorochrome-conjugated secondary antibody (ZSGB-Bio, China) for 1 h at 37 °C in the dark. After nuclear counterstaining with 4, 6-diamidino-2-phenylindole (DAPI, Beyotime, China), slides were coverslipped with mounting medium and visualized using Nikon microscope. For inhibition of nuclear exportation of NPM1-mA, OCI-AML3 cells were treated with 20 nM leptomycin B (LMB; Beyotime, China) for 0 h, 2 h, 6 h, followed by immunofluorescence analysis. Quantification of cytoplasmic versus nuclear PML in leukemic cells was analyzed by Image J software. The following antibodies were used: rabbit polyclonal mutant NPM1 antibody (Abcam, USA), mouse monoclonal Flag antibody (sigma, USA), mouse monoclonal PG-M3 antibody (Santa, USA), rabbit monoclonal LC3 І/II antibody (Novus, USA), rabbit polyclonal GFP antibody (Abcam, USA). Alexa Fluor fragment of goat anti-mouse IgG and goat anti-rabbit IgG (ZSGB-Bio, China), both at 1:200 dilutions in blocking medium.

Cell transfection and infection

Plasmids encoding pEGFPC1-NPM1-mA, pEGFPC1-NPM1-wt and empty vector pEGFPC1 were kindly provided by Dr. B. Falini (Institute of Hematology, University of Perugia, Perugia, Italy). Plasmids encoding Flag-NPM1-mA, Flag-NPM1-wt and empty vector were obtained from Dr. C.J. Sherr (Genetics and Tumor Cell Biology, St, Jude Children's Research Hospital, Memphis, USA). Plasmids encoding HA-AKT, Flag-PML-IV were obtained from Dr. H.K. Lin (Department of Molecular and Cellular Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA). All transfection experiments were conducted using Lipofectamine™ 2000 (Invitrogen, USA) according to the manufacturer's instruction. After 48 h of transfection, the cells were collected for qRT-PCR, western blot or immunofluorescence analyses. shRNA targeting PML (5'-CCCGCAAGACCAACAACAT-3') and scramble lentiviral vectors were purchased from Gene Pharma (Shanghai, China), respectively. OCI-AML3 cells and THP-1 were infected with shRNA lentivirus targeting NPM1 for 48 h in the presence of 5 μg/mL polybrene (Sigma, USA), following 2 μg/mL puromycin selection for 7 days (Sigma, USA), respectively. The puromycin-resistant cells were isolated and propagated for use in the experiments.

Immunoprecipitation

The cells were harvested after 48 h of transfection and washed twice with ice-cold PBS buffer. Cells were then sonicated in IP buffer [20 mM Tris-Cl, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 % (v/v) Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM β-glycerolphosphate, 1 mM Na3VO4, and protease inhibitor cocktail (Roche), pH 7.5] at 4 °C for 3 times per 5 s by Bio-ruptor UCD-200 (Diagenode, Belgium), followed by centrifuged at 14,000 g at 4 oC for 15 min to remove the cell debris. Expressing of the indicated proteins in the lysates was checked by western blot analysis using relevant antibodies to normalize total amounts of the inputs. After normalization, the supernatants were each incubated with specific antibodies or normal IgG (as control), and equal amounts of protein A/G beads (Santa, USA) overnight at 4 oC. The protein A/G beads with the bound antibodies and interacting proteins were pelleted and washed three times with IP buffer before boiled in 2× SDS-PAGE sample. The boiled samples were then resolved in SDS-PAGE and subject to western blot analysis.

Cycloheximide assay

For cycloheximide assays, HEK293T cells were co-transfected with pEGFPC1-NPM1-mA and Flag-PML-IV. In addition, OCI-AML3 cells were infected with either shNPM1 lentivirus or vector. Next, the HEK293T and OCI-AML3 cells were treated with 100 μg/ml cycloheximide (CHX; Sigma, USA) for the indicated times, respectively. Following designated treatment, whole-cell lysates were prepared using RIPA buffer with protease inhibitors. Equal amounts of protein for each time point were analyzed by western blot.

Cell viability assay

Cell viability was determined using Cell Counting Kit-8 (Dojindo Laboratories, Japan) according to the manufacturer's instruction. In Brief, cells were seed at 2×103 cells per well in a 96-well plate for 12 h and subsequently treated with rapamycin or 3-methylademine (3-MA) reagents for indicated times. At the indicated time point, cells were treated with CCK-8 at 10 μl/well at 37°C for 3 h, and the numbers of cells per well were determined by measuring absorbance at 450 nm.

Colony formation assay

Methylcellulose clonogenic assays was carried out to determine cell colony formation ability by planting 1×103 cells per well in triplicate in 24 well-plate, and maintained in RPMI 1640 medium containing 20% FBS at 37°C in incubator. Colony numbers were scored 10 days later. The colony forming units (CFU), defined as cell clusters consisting of more than 5 cells, were counted using an inverted microscope.

The cancer genome atlas analysis

Expression levels of gene and clinical information of 200 AML patients were retrieved from The Cancer Genome Atlas (TCGA). Total of 173 samples had IlluminaGA RNASeq data for gene expression. The values (normalized RNA-Seq data) represented the expression and mutation status in the gene expression matrix and the gene mutation matrix, respectively. Median value of the measured PML expression levels based on the range of expression of PML in AML samples was used to dichotomize patients into high and low cohorts. NPM1 mutation occurs frequently in AMLs and age is one of the most important risk factors among patients with AML, therefore, 33 NPM1-mutatd cases older than 40 years of age were included in our study. Finally, comparison of the overall survival (OS) according to the levels of PML in NPM1-mutated AML cases was done by long-rank test. The OS curves were plotted according to the Kaplan-Meier methods.

Statistical analysis

All data were derived from three independent experiments. P values for comparisons of gene expression between groups were obtained using unpaired t-tests or one-way analyses of variance (ANOVA). The Kaplan-Meier estimation and the log-rank test were used to compare the survival difference. The SPSS (Version 13.0) software and GraphPad (Prism 5) was used for statistical analyses. P < 0.05 was considered statistically significant.

Results

Autophagy activation facilitated by NPM1-mA contributes to leukemic cell survival

To investigate the involvement of autophagy in NPM1-mutated leukemic cell growth, we firstly assessed the levels of autophagy marker in the NPM1-mutated cell lines and primary blasts. The results showed that mutant NPM1-expressing OCI-AML3 cell line had higher LC3 I/II levels and lower p62 levels, as compared to the wild-type NPM1-expressing THP-1 and KG-1a cell lines (Figure Similar results were obtained by Immunofluorescence analysis, as indicated by the accumulated LC3 puncta in OCI-AML3 cells (Figure . In addition, higher LC3 I/II and lower p62 mRNA levels were also observed in primary NPM1-mutated AML blasts, as compared to primary NPM1-unmutated AML blasts (Figure .

Next, we assessed the autophagic activity upon NPM1-mA expression. THP-1 cells derived from human monocytic leukemia were transiently transfected with Flag-NPM1-mA, Flag-NPM1-wt, and empty vector, and designated as vector group, NPM1-wt group and NPM1-mA group, respectively. The data showed that NPM1-mA overexpression increased endogenous LC3-II and decreased p62 levels (Figure . We also found that overexpression of NPM1-mA resulted in increased accumulation of LC3 puncta (Figure . To validate the effect of NPM1-mA on autophagy flux, 3-methyladenine (3-MA), an autophagy inhibitor, was added in NPM1-mA-enforced THP-1 cells. Consistent with previous observation revealing the capacity of 3-MA to prevent autophagy in leukemic cells 31, 3-MA treatment suppressed LC3-II accumulation and p62 downregulation induced by NPM1-mA overexpression (Figure . Furthermore, we observed that enforced NPM1-mA expression promoted cell proliferation in vitro, whereas exposure to 3-MA abrogated the enhancement of NPM1-mA-induced autophagic activation and growth advantage (Figure . Similar results were obtained from another wild-type NPM1-expressing KG-1a cell line (Figure .

In complementary loss-of-function studies, we established the NPM1-mA-silenced OCI-AML3 cells, which has been previously characterized and used for this type of analysis 11. The results showed that depletion of NPM1-mA decreased LC3-II levels and increased p62 levels, along with the reduced accumulation of LC3 puncta numbers in OCI-AML3 cells (Figure . Additionally, treatment with autophagy activator rapamycin reversed the changes in autophagic activity and cell proliferation caused by NPM1-mA knockdown (Figure . Collectively, these results demonstrated that autophagic activity was enhanced upon NPM1-mA expression and autophagic activation provided a survival benefit to leukemic cells in vitro.

PML is aberrantly expressed in NPM1-mutated leukemic cells

Recent studies have implicated the critical role of PML in the regulation of autophagy in cancer cells 32. We hypothesized that aberrant PML might participate in autophagy and cell survival in NPM1-mutated AMLs. To test this notion, we first evaluated PML expression and intracellular distribution in myeloid leukemia cell lines. Most PML isoforms were detected at various levels in five myeloid leukemia cell lines, and relative high expression levels of PML were observed in OCI-AML3 cells harboring NPM1-mA (Figure . Consistent with above data, we also found that PML expression were higher in NPM1-mutated AML samples compared with that in NPM1-unmutated AMLs (Figure . In addition, analysis from immunocytochemistry staining showed that, unlike the mainly restrained to the nuclear fractions in THP-1 cells, PML exhibited nucleoplasm and/or cytoplasmic expression patterns in OCI-AML3 cells by the presence of the red precipitate particles (Figure . Similar results were obtained by immunofluorescence analysis (Figure . Importantly, we detected the endogenous co-localization of PML and mutant NPM1 protein in OCI-AML3 cells (Figure . These results confirmed the aberrant expression of PML in OCI-AML3 cells and indicated the potential relationship between PML and mutant NPM1 protein.

PML delocalization and stabilization is mediated by mutant NPM1

Based on the above data and the critical role of NPM1 in proper nucleolar localization and stability of some nuclear proteins 36, 37, we proposed that NPM1-mA may be involve in regulation of PML delocalization and stabilization. Firstly, we examined the effect on PML localization upon NPM1-mA expression. THP-1 cells were transfected with pEGFPC1-NPM1-mA plasmids, followed by immunofluorescence analysis of PML localization. The results showed that PML was delocalized to the cytoplasm of the transfected cells in NPM1-mA group while restrained to the nuclear in the two control groups (Figure . Since normal nucleo-cytoplasmic traffic of NPM1 occurs in a CRM1-dependent manner, we then tested the effects of the specific CRM1 inhibitor leptomycin (LMB) treatment on NPM1 mutant as well as PML localization. Our data showed that treatment with LMB relocated cytoplasmic NPM1 mutant to the nucleus in OCI-AML3 cells, accompanied by restoration of PML nuclear localization (Figure . Then, we determined whether NPM1 mutant protein physically interacts with PML. Co-immunoprecipitation (Co-IP) analysis revealed that immunoprecipitation with mutant NPM1 antibody pulled down endogenous PML along with NPM1-mA in OCI-AML3 cells (Figure . Similarly, immunoprecipitation with PML antibody pulled down endogenous NPM1-mA along with PML (Figure . In support of this result, HEK293T cells were transfected with pEGFPC1-NPM1-mA and Flag-PML-IV (the best studied PML isoform), and the interaction between exogenous of NPM1-mA and PML was confirmed (Figure . These data indicated that mutant NPM1 interacted with PML, giving rise to cytoplasmic delocalization of PML protein in OCI-AML3 cells.

Next, we investigated whether NPM1-mA plays a role in the turnover of PML protein. pEGFPC1-NPM1-mA was transfected into THP-1 cells and PML protein levels were monitored. Our results revealed that enforced NPM1-mA expression (Figure led to upregulation of PML protein (Figure . In contrast, shRNA-mediated NPM1-mA downregulation decreased PML protein levels in OCI-AML3 cells (Figure , whereas the levels of PML remained constant in NPM1-silenced THP-1 cells (Figure . Next, we undertook to identify the effect of NPM1-mA on half-life of PML protein. HEK293T cells were co-transfected with pEGFPC1-NPM1-mA and Flag-PML-IV, followed by cycloheximide (CHX, 100 μg /mL) to block de novo protein synthesis. Our data showed that NPM1-mA overexpression in HEK293T cells alleviated the degradation of exogenous PML protein caused by CHX treatment in a time-dependent manner (Figure . In contrast, NPM1-mA knockdown in OCI-AML3 cells sped up the degradation of endogenous PML protein (Figure . Next, we sought to determine the potential mechanism of which mutant NPM1 regulated PML stability. Consistent with previous report that PML is subject to proteasome-dependent proteolysis 38, our data showed that treatment with CHX resulted in PML protein levels decreasing in a time-dependent manner, whereas addition of proteasome inhibitor MG132 (10 μM) reversed the changes in PML protein levels caused by CHX treatment (Figure Further experiments demonstrated that the addition of MG132 could alleviate the downregulation of PML mediated by NPM1-mA knockdown in OCI-AML3 cells (Figure . These data supported that mutant NPM1 mediated PML stabilization through inhibiting proteolysis. Additionally, we also determined the changes in PML mRNA levels upon NPM1-mA expression (Figure , indicating other potential mechanisms underlying aberrant PML expression in NPM1-mutated AML cells.

PML knockdown inhibits cell growth through suppressing autophagy in OCI-AML3 cells

To gain evidence that the effect of aberrant PML expression mediated by NPM1 mutant is biologically relevant, OCI-AML3 cells were stably infected with shRNA lentivirus targeting PML and cell growth was observed. As expected, PML mRNA and protein levels were downregulated caused by shRNA mediated PML depletion in OCI-AML3 (Figure . Importantly, results from CCK-8 analysis revealed that loss of PML resulted in cell proliferation inhibition (Figure . In line with the data, colony formation assay showed decreased number of cell clusters morphologically and lower proportion of colony forming units in the PML-silenced OCI-AML3 cells as compared to scramble group (Figure . Additionally, we measured the changes in apoptosis-associated protein Bcl-2 and Bax levels and found that silenced PML decreased Bcl-2 levels and increased Bax protein levels (Figure . In support of these results, we accessed to the RNA-Seq data and clinical information of 33 NPM1-mutated AMLs from TCGA database to analyze the possible correlation between PML expression and prognosis. The results revealed that NPM1-mutated AML patients expressing high PML levels had a shorter survival compared with those expressing low PML expression (Figure . Collectively, these results revealed that knockdown of PML could suppress cell growth in OCI-AML3 cells.

Next, we evaluated the effect of PML knockdown on autophagic activity in OCI-AML3 cells. Results from western blot analysis revealed that PML depletion decreased the levels of endogenous LC3-II and increased p62 levels (Figure . Consistent with these observations, we found that loss of PML resulted in reduced LC3 puncta distribution (Figure . In addition, rapamycin treatment could reverse the changes in LC3-II conversion and p62 levels mediated by PML depletion (Figure . Furthermore, autophagic activation by rapamycin rescued the inhibiting effect of PML depletion-mediated cell proliferation (Figure . Finally, we sought to explore the potential mechanism by which PML modulated autophagy. As a key regulator of cell proliferation and survival, the phosphoinositide 3-kinase (PI3-kinase)-AKT signaling pathway has been implicated an important role in the regulation of autophagy 39. Thus, we assessed whether AKT was involved in autophagy regulation mediated by PML. Results from western blot showed that the levels of total AKT and phosphorylated AKT were decreased in PML-silenced OCI-AML3 cells (Figure . Notably, enforced expression of AKT could alleviate the impairment of PML depletion-mediated autophagy in OCI-AML3 cells, as indicated by the increased LC3-II and decreased p62 levels (Figure . These results indicated that PML was involved in autophagy regulation at least partially due to AKT signal in OCI-AML3 cells.

PML expression is critical for NPM1-mA-mediated autophagy and cell survival

On the basis of above results, we further evaluated the pivotal role of PML in NPM1-mA mediated autophagy and cell growth. The NPM1-mA-silenced OCI-AML3 cells were transiently transfected with Flag-PML-IV plasmid. We found that introduction of exogenous PML resulted in LC3-II accumulation and p62 downregulation (Figure . In addition, we also observed the increased accumulation of LC3 puncta in PML-transfected cells as compared to vector group (Figure . Next, we examined the effect of PML overexpression on cell growth in NPM1-mA-silenced OCI-AML3 cells. Results from CCK-8 analysis confirmed that NPM1-mA knockdown resulted in cell growth inhibition, whereas overexpression of PML bypassed the effect of NPM1-mA depletion on cell proliferation (Figure . Additionally, we measured the changes in apoptosis-associated protein and found that PML overexpression increased Bcl-2 levels and decreased Bax levels in NPM1-mA-silenced OCI-AML3 cells (Figure Collectively, our data indicated that PML was involved in NPM1-mA-mediated autophagy and cell proliferation in OCI-AML3 cells.

Discussion

The NPM1 mutation is one of the most frequent genetic alterations thus far identified in AML. Despite progress in the clinical and biological characterization of NPM1-mutated AML, the role played by the NPM1 mutations in AML development has not been fully elucidated. Herein, our data showed that NPM1 mutant mediated tumor suppressor PML delocalization and stabilization, further enhanced autophagic activity and cell survival in AML (Figure .

In recent years, autophagy being suggested as an important mechanism in leukemia pathogenesis has aroused much attention 40. Notably, increased autophagy-associated transcripts levels were reported in NPM1-mutated AML during analysis of the potential target of autophagy pathway in human cancers with recurrent molecular alteration 41. In this study, we determined the autophagic activity in leukemia cell lines as well as primary blasts and found that enforced NPM1-mA expression could promote autophagy in leukemia cells. It has been reported that NPM1 was essential for autophagy when the nucleolus was disrupted by inhibiting RNA polymerase I (Pol I) transcription 42. Indeed, NPM1 is found mainly in the nucleolus. Consequent to the mutations, the NPM1 leukemic mutants export from nucleus to cytoplasm, leading wild-type NPM1 to cytoplasm dislocation 43. These findings indicated that NPM1 mutant may be involved in autophagy in leukemia. Additionally, our data revealed that NPM1 mutation promoted leukemic cell growth in vitro. Evidence from transgenic mouse model showing the critical role of NPM1 mutant in myeloproliferation induction further confirmed our notion 44. In addition, our data showed that autophagy inhibitor 3-MA treatment could abrogate the enhancement of NPM1-mA-mediated autophagic activation on cell survival. In line with our data, disruption of autophagy by 3-MA was shown a synergism antineoplastic effect with HDAC inhibitors in chronic myeloid leukemia Ba/F3 cells 45. These findings and our data were consistent with the conception that activated autophagy enhance cell survival in the established tumors 40.

In addition to the well-documented role of autophagy-related gene (ATG) in autophagy, the function of tumor suppressor genes such as p53 linking autophagy in tumor development has been proposed in recent years 46. PML, a pleiotropic tumor suppressor controlling a variety of biological processes, has been recently defined a regulation role in autophagy 32. Notably, aberrant PML expression was reported in NPM1-mutated AMLs 33. In the present study, we confirmed the aberrant localization of PML protein in mutant NPM1-expressing OCI-AML3 cells. It has been acknowledged that alternative splicing of C-terminal exon 6 encoding the nuclear localization signal can give rise to cytoplasmic localization of PML protein 47. More recent research by Buczek et al. 48 has evidenced that the cytoplasmic localization of PML may also be mediated by its nuclear export in a chromosomal maintenance 1 (CRM1)-dependent manner. Here, we provided the first evidence that the cytoplasmic localization of PML could be mediated by interacting with mutant NPM1. Indeed, NPM1 mutant protein has been previously reported to delocalize some nuclear protein such as ARF to cytoplasm 49. The exact domain required for interaction between mutant NPM1 and PML needs to be further determined.

In addition to the aberrant localization of PML, we also observed higher expression of PML in NPM1-mutated AML blasts and cultured cells. Furthermore, loss- and gain-of function assays revealed that PML was subject to similar-mutant NPM1-dependent regulation. Mechanically, we verified that NPM1-mA could stabilize PML through inhibiting proteasome-mediated degradation. Aberrant PML post-translational modifications are the most characterized mechanism accounting for PML loss in cancer. Recent studies have demonstrated the pivotal roles of ubiquitination in PML stability mediated by an E3 ligase UHRF1 or deubiquitinase USP11 in cancers 50, 51. Considering that NPM1 mutant regulated PETN stabilization through blocking HAUSP-mediated deubiquitination in AML 9, we aim to identify whether the interaction of mutant NPM1 with PML may protect PML protein from ubiquitination-mediated degradation in the future study. In addition, we also observed the change in PML mRNA levels upon NPM1-mA expression. Indeed, the transcriptional up-regulation of PML was previously reported to be accompanied by elevated protein levels 28, 52. Certainly, other potential multiple mechanisms underlying aberrant PML expression in NPM1-mutated AML cells are worthy to be further investigated.

The activity of PML has been intensively investigated in the context of suppression in tumorigenesis. Here, we uncovered an unexpected function of PML in NPM1-mutated leukemic cells. Preliminarily in vitro experiments showed that PML knockdown led to cell proliferation suppression and colony formation ability impairment, as well as anti-apoptotic Bcl-2 downregulation and pro-apoptotic Bax upregulation. Further analysis of TCGA AML dataset revealed a negative correlation between PML expression and overall survival in NPM1-mutated AMLs, which is consistent with the observation that high PML expression correlated to poor prognosis in chronic myeloid leukemia 53. In fact, evidence of the function of PML beyond tumor suppression in certain cellular settings is quickly accumulating 28, 54. More recent study has showed PML upregulation in breast cancer cells and pharmacological inhibition of PML by arsenic trioxide reduced the tumor formation capacity in vivo 55. Interestingly, a growing body of clinical evidence has also implicated that arsenic trioxide-induced PML degradation has a potential therapeutic role in NPM1-mutated AMLs 33, 34.

Recent studies have evidenced that aberrant PML promoted breast cancer cell survival and hematopoietic stem cell maintenance through regulation of peroxisome proliferator-activated receptor (PPAR) signaling and fatty acid oxidation metabolism 28, 56. Here, we showed that the inhibiting effect of PML knockdown on cell proliferation might be associated with impaired autophagic activity, which is consistent with recent observation that autophagy is a survival mechanism of AML precursors during dual mTORC2/mTORC1 targeting 57. It is well established that a variety of kinases was involved in autophagy-signaling network. AKT, also known as protein kinase B, has profound effects on autophagy via the activation of kinase target of rapamycin (TOR) protein, which antagonizes autophagy 58. Unexpectedly, our data showed that PML knockdown downregulated AKT levels. Moreover, ectopic expression of AKT could rescue the autophagy inhibition mediated by PML depletion in NPM1-mutated leukemic cells. These data are in agreement with the observation recently reported by Khan et al. 59 that sucrose induced autophagy in chondrocytes dependent on the activation of AKT/mTOR/P70S6K signaling pathways. In addition, a recent paper has revealed that pharmacological inhibition of AKT reduced autophagic activity and enhanced radiosensitivity in prostate cancer cells 60. Thus, defining the precise roles of AKT in autophagy depends on the specific cellular context. Finally, we performed a rescue assay and further highlighted the critical role of PML in NPM1-mA-mediated autophagy and cell proliferation.

In the present study, OCI-AML3 cell line that is accessible and preferable mutant NPM1-expressing cell line was selected for research. Our observation warrant further research on the effects of NPM1 mutations on autophagy including clinical studies and experiments with animal models. Further studies are needed to elucidate the role of PML in cell proliferation and autophagic activity as well as pro- versus anti-apoptotic cell markers using another mutant NPM1-expressing IMS-M2 cell line and MEF PML-/- versus MEF PML+/+ in the context of NPM1 mutation. In addition, the use of primary blasts from wild type or mutant NPM1- expressing AML patients will be tried in the continuous study to better translate the biology of the disease.

In summary, our data suggest for the first time that autophagy is beneficial to leukemogenesis driven by NPM1 mutation. Additionally, a mechanistic link between NPM1-mA expression and PML delocalization and stabilization illustrates an anticipated role of PML in NPM1-mutated AML. Collectively, these results conclusively indicate that pharmacological inhibitors of PML and/or autophagy are potential therapeutics for NPM1-mutated AML therapy.

Table 1

Patient characteristics

Characteristics	Median(range)	No. of cases
Sex
Female		14
Male		12
Total		26
Median age, y	53.8y (26-79) y
Younger than 40 y		6
40-60 y		12
Older than 60 y		8
Median WBC, ×109/L	44 (0.3-295)
Median platelets, ×109/L	57.3 ( 3.0 - 655.0)
FAB classification
M1		2
M2		4
M3		4
M4		6
M5		9
unclassified		1
Karyotype
Normal		14
t(8;21)		2
t(15;17)		3
inv(16)		5
Unknown		2
Gene mutations
NPM1		14
FLT3/ITD		8
WT1		9
CBFβ-MYH11		2

Abbreviations: AML, acute myeloid leukemia; y, year old; WBC, white blood cell; FAB classification, French-American-British classification, a classification of acute leukemia produced by three-nation joint collaboration.

Table 2

Primer sequences used in this study

Genes	Sequence (5'- 3')
LC3	F: 5'-GACCGCTGTAAGGAGGTGC-3'
	R: 5'-CTTGACCAACTCGCTCATGTTA-3'
p62	F: 5'-GGGGACTTGGTTGCCTTTT-3'
	R: 5'-CAGCCATCGCAGATCACATT-3'
PML	F: 5'-ACATCTTCTGCTCCAACCC-3'
	R: 5'-CGCCAAAGGCACTATCC-3'
NPM1-mA	F: 5'-TGGAGGTGGTAGCAAGGTTC-3'
	R: 5'-CTTCCTCCACTGC CAGACAGA-3'
β-actin	F: 5'-TAGTTGCGTTACACCCTTTCTTG-3'
	R: 5'-TGCTGTCACCTTCACCGTTC-3'

Abbreviations: F stands for forward; R stands for reverse