This study was approved by the Ethics Committee of the Sun Yat-Sen Memorial Hospital of Sun Yat-Sen University. Written informed consent was obtained from all the participants or their parents.

Bone marrow samples were collected from patients aged from 0 to 14 years who suffered from childhood acute lymphoblastic leukaemia (ALL) or a non-malignant blood disease. In total, 30 subjects with B-cell ALL (12 patients in the primary phase, 12 in the complete remission phase who were still undergoing treatment, 3 patients in the complete remission phase who had completed their treatment course and 3 in the relapse or non-remission phase) and 8 subjects with non-malignant blood disease were enrolled. The diagnoses of ALL were based on morphologic and flow cytometric analyses of the immunophenotype, and the stratification of ALL was based on the IC-BFM 2002 criteria for ALL. Bone marrow mononuclear cells (BMMCs) were isolated via Ficoll (MP Biomedicals, Santa Ana, CA, USA) density gradient centrifugation (14).

Total RNA was isolated from BMMCs using TRIzol reagent (Takara Bio Inc., Otsu, Shiga, Japan) and then reverse-transcribed into cDNA. The sequences of primers used were as follows: GAPDH: forward, 5′-GGTCGGAGTCAACGGATTTGGTCG-3′ and reverse, 5′-CCTCCGACGCCTGCTTCACCAC-3′; for HMGB1: forward, 5′-TTTCAAACAAAGATGCCACA-3′ and reverse, 5′-GTTCCCTAAACTCCTAAGCAGATA-3′; for Beclin1: forward, 5′-ATCCTGGACCGTGTCACCATCCAGG-3′ and reverse, 5′-GTTGAGCTGAGTGTCCAGCTGG-3′; for FIP200: forward, 5′-TCAGTAAGTTGCGCAGTAGCA-3′ and reverse, 5′-TTCAGGTGCACAAGCTCCAT-3′. Reactions were carried out in a Roche Light Cycler 480 system (Roche Diagnostics GmbH, Mannheim, Germany) with a SYBR Premix ExTaq kit (Takara Bio Inc., Otsu, Shiga, Japan). With the 2^−ΔΔCt method, Data were normalized to GAPDH expression. The control group was set as 1.

The human leukaemia cell line, Jurkat and Reh, were purchased from the Type Culture Collection of the Chinese Academy of Sciences, Shanghai, China and cultured in RPMI-1640 medium (Gibco, Grand Island, NY, USA) with 10% FBS (Hyclone, Logan, UT, USA) in a humidified incubator with 5% CO₂ and 95% air (15).

The antibodies to FIP200, ULK1, mAtg13, PI3KC3, Phospho-ULK1 (Ser555) and Rabbit (DA1E) mAb IgG XPTM Isotype Control were obtained from Cell Signaling Technology (Danvers, MA, USA). The antibodies to LC3 were obtained from Sigma-Aldrich (St. Louis, MO, USA). The antibodies to HMGB1, Beclin1, Bcl-2 and P62 were obtained from Abcam (Cambridge, UK). The antibodies to actin, GAPDH and tubulin were obtained from EarthOx Life Science (Millbrae, CA, USA). All secondary antibodies were purchased from Bioss (Beijing Biosynthesis Biotechnology Co., Ltd.). Vincristine (VCR), daunorubicin (DNR), 3-methyladenine (3-MA), chloroquine (CQ) and ethyl pyruvate (EP) were purchased from Sigma-Aldrich.

Lentivirus with shRNA was purchased from GenePharma (Shanghai, China). Transfection was conducted with lentivirus using 5 μg/ml polybrene and selected with 1 μg/ml l puromycin under the instruction of GenePharma Recombinant Lentivirus Operation Manual (16). The interference efficiency was evaluated by RT-qPCR and western blotting. We selected the best sequence which processed the highest interference efficiency out from 3 different sequences. The optimal target sequence for FIP200 was 5′-GCCGTCAAACTATTGCCAAAC-3′, for HMGB1 was 5′-CCCGTTATGAAAGAGAAATGA-3′ and for control was 5′-TTCTCCGAACGTGTCACGTTTC-3′.

Cells were plated in 96-well plate at a density of ~6×10⁴ cells/well. Cell viability was measured after medication treatments by CCK-8 test (Dojindo Molecular Technologies Inc., Kumamoto, Kyushu, Japan) according to the manufacturer’s instructions.

Cells with different treatments were collected and lysed. The cell lysate was separated by 8% (10, 12 and 15%) sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and electrophoretically transferred onto polyvinylidene fluoride membrane (Millpore, Billerica, MA, USA). The membrane was blocked with 5% non-fat milk, and then incubated with the primary antibodies at a dilution of 1:1,000 overnight at 4°C. After rinsing, the membranes were incubated for 1 h at room temperature with secondary antibodies at a dilution of 1:1,000 and then washed with TBST (15,17,18). The membranes were detected with enhanced chemiluminescence reagent (Millpore) by G: BOX XT4 (Syngene, Cambridge, UK). The expressions of the target proteins were quantified by Quantity One version 4.6.2.

Immunoprecipitation (IP) analysis was performed by Pierce Classic IP kit (Pierce Biotechnology, Rockford, IL, USA) according to the manufacturer’s protocol. In brief, cells were collected and lysed by ice cold IP lysis buffer. Then the protein was pre-cleared with the control agarose resin at 4°C for 1 h. The pre-cleared cell lysate was incubated with the Rabbit (DA1E) mAb IgG XP™ Isotype Control or specific antibodies overnight at 4°C to form the immune complex and then with protein A/G plus agarose for 1 h at 4°C to capture the immune complex. After washed the resin with wash buffer, the immune complex was eluted at 100°C for 10 min. The eluate was collected and subjected to SDS-PAGE for immunoblot analysis (15,18). The Clean-Blot IP Detection reagent (Pierce Biotechnology) was used as secondary antibody at a dilution of 1:1,000 for detecting the target proteins without interference from denatured IgG in western blot analysis.

Cells were collected and fixed in 4% formaldehyde for 20 min, permeabilized in 0.3% Triton X-100 for 5 min and incubated in blocking buffer for 1 h. Then cells were incubated overnight at 4°C with primary antibodies. After rinsing, cells were incubated with FITC- or Cy3-conjugated secondary antibodies for 1 h at room temperature in the dark and then with DAPI for 3 min. After rinsing, cells were resuspended. Several drops of cell suspension were added in a laser confocal Petri dish (12,19). The images were captured by a Carl Zeiss LSM 710 (Carl Zeiss Inc., Oberkochen, Germany).

Cells were fixed with 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 mol/l phosphate buffer (pH 7.4), followed by 1% OsO₄. After dehydration, thin sections were stained with uranyl acetate and lead citrate for observation by Tecnai G2 Spirit Twin electron microscope (FEI, Hillsboro, OR, USA). Transmission electron microscopic (TEM) assessment of autophagosome-like structures was carried out as previously described (19,20).

Data are shown as mean ± standard deviation (SD). The difference among groups was examined using the Student’s t-test or ANOVA in measurement data and Fisher’s exact probabilities in small size enumeration data. The Pearson product-moment correlation method was used to analyze the linear correlation between two groups. A value of p<0.05 was regarded as statistically significant. All statistical analyses were conducted by SPSS version 18.0.

In total, 30 children with B-cell ALL and 8 control subjects with a non-malignant blood disease were included in the present study.

As shown in Fig. 1A, HMGB1 was upregulated in the BMMCs obtained from the patients with primary, relapsed or non-remission leukaemia, whereas its expression level in the BMMCs from the ALL patients who exhibited complete remission was as low as that from those with a non-malignant blood disease. The expression level of Beclin1 was nearly identical between the groups (Fig. 1B). The expression levels of HMGB1 and Beclin1 appeared to be irrelevant to risk stratification (p>0.05, data not shown).

DNR and VCR are basic drugs used for leukaemia therapy. As shown in Fig. 2A, both drugs caused striking damage to Jurkat and Reh cells in a dose-dependent manner, and this trend was particularly apparent in the Reh cells. Interestingly, in the presence of 3-methyladenine (3-MA), an inhibitor of early-phase autophagy, the cytotoxic effect was much stronger and cell viability was more sharply reduced (Fig. 2B). This result suggested that the inhibition of autophagy may augment the lethality of anticancer agents.

To explore the potential role of FIP200, a key protein involved in autophagy, a target-specific shRNA against FIP200 was transfected into Jurkat and Reh cells. This treatment resulted in a significant decrease in FIP200 at both the mRNA and protein levels (Fig. 2C). Cell viability analysis showed that the silencing of FIP200 mimicked the inhibitory effect of 3-MA and rendered the leukaemia cells more sensitive to DNR and VCR (Fig. 2A).

Autophagy is a highly dynamic, multi-step process. To assess the entire autophagic process, we measured autophagosomes containing LC3 puncta using immunofluorescence staining and analysed the LC3-II/I ratio via western blot analysis. Additionally, we observed ultrastructural changes via electron microscopy and assessed autophagic flux based on lysosomal turnover and the P62 level via western blot analysis (19,21,22).

Based on western blot analysis, we found that silencing FIP200 suppressed the chemotherapy-induced high expression of LC3-II and the corresponding degradation of P62 (Fig. 3A and B). The accumulation of LC3-II in the mock-treated and control cells was enhanced in the presence of the lysosomal inhibitor chloroquine (CQ) in the presence or absence of stimuli, but this effect was not observed in the FIP200 shRNA-treated cells (Fig. 3A). These changes demonstrated that the accumulation of LC3-II was not due to its decreased degradation but rather increased autophagic flux. However, the changes of LC3-II and P62 were not observed in the presence of 3-MA (Fig. 3B). Based on immunofluorescence analysis, the change in endogenous LC3 puncta was consistent with the results of western blot analysis in both cell lines (Fig. 3C). Moreover, ultrastructural analysis revealed that the FIP200 shRNA-treated cells exhibited fewer autophagosomes during chemotherapy compared with the control shRNA-treated cells (Fig. 3D).

The data shown in Figs. 2 and 3 demonstrate that in leukaemia, the anticancer agent-induced cytotoxic effects were due to the autophagy-related drug resistance, and these effects were ameliorated by the depletion of FIP200 or the application of an autophagy inhibitor.

We transfected Jurkat cells with HMGB1 shRNA, which resulted in a prominent decrease in the mRNA and protein expression of HMGB1 (Fig. 4A). As shown in Fig. 4B, the phosphorylation of ULK1 at Ser555, which plays a critical role in responding to autophagic stimuli, was significantly increased after anticancer agent treatment. Immunoprecipitation analysis indicated that silencing HMGB1 did not affect ULK1-mAtg13-FIP200 complex assembly or ULK1 phosphorylation following DNR treatment.

Conversely, the absence of FIP200 did inhibit the assembly of the HMGB1-Beclin1 complex or the subsequent formation of the PI3KC3-Beclin1 complex after chemotherapy stimulation (Fig. 4C). In the control shRNA-treated Jurkat cells, Beclin1 separated from Bcl-2 after DNR treatment and subsequently complexed with HMGB1 and PI3KC3. In contrast, Beclin1-Bcl-2 was not separated after DNR treatment in the absence of FIP200 or HMGB1. Thus, the HMGB1-Beclin1 and PI3KC3-Beclin1 complexes failed to coalesce. Furthermore, we observed that the depletion of HMGB1 or FIP200 reduced the phosphorylation of Bcl2 (Fig. 4D), which may result in the stable binding of Bcl2 to Beclin1.

Although primarily localised to the nucleus, HMGB1 can translocate to the cytoplasm and the extracellular space during cell activation and cell death. To investigate the role of HMGB1 in autophagy, we focused on its localization. By detecting the expression of HMGB1 in the nucleus and the cytosol separately, we found that the chemotherapy promoted the translocation of HMGB1 from the nucleus to the cytosol in the cultured Jurkat and Reh cells (Fig. 5A). This translocation was also detected using immunofluorescence techniques (Fig. 5B). Remarkably, HMGB1 did not translocate in the absence of FIP200 (Fig. 5A and B). When combining the data of Figs. 4B and C and 5A and B, we concluded that the activation of the ULK1-mAtg13-FIP200 complex induced the translocation of HMGB1 and the subsequent dissociation of Beclin1-Bcl-2 and formation of the HMGB1-Beclin1 complex, ultimately influencing autophagy-related chemoresistance.

As shown in Fig. 5B and C, the Jurkat and Reh cells were treated with DNR in the presence or absence of EP, an inhibitor of HMGB1 translocation. We found that blocking HMGB1 translocation mimicked the knockdown of FIP200 and suppressed autophagy. At the ultrastructural level (Fig. 5D), EP reduced the formation of autophagosomes and autophagolysosomes to a great extent, and this effect was similar to that of HMGB1 or FIP200 depletion (Figs. 3D and 5D). Moreover, in the absence of HMGB1 translocation, autophagy was suppressed, and the leukaemia cells became more sensitive to the chemotherapy (Fig. 5E).

The above results suggested that HMGB1 is involved in the positive regulation and maintenance of autophagy in stressed leukaemia cells and that HMGB1 translocation is a necessary component of the autophagic process. Blocking HMGB1 translocation by interrupting the upstream signal or using a translocation inhibitor may represent a useful strategy for reducing autophagy-induced drug resistance.