ALL cell lines, including Jurkat, CCRF-CEM, CEM/C1, Nalm6, Sup-B15, THP-1, HL-60, and Reh cells, and the human embryonic kidney cell line 293T, were purchased from the National Collection of Authenticated Cell Cultures (Shanghai, China). CCRF-CEM, CEM/C1, Nalm6, and Reh cells were cultured in RPMI 1640 medium (Gibco, Carlsbad, CA) containing 10% fetal bovine serum (FBS, Corning, New Zealand Sourced), and 1% penicillin–streptomycin (Gibco). Jurkat cells were cultured with RPMI 1640 containing 10% FBS, 1% GlutaMAX (Gibco), and 1% penicillin–streptomycin. THP-1 cells were cultured in RPMI 1640 containing 10% FBS, 0.05 mmol/l β-Mercaptoethanol (Gibco), and 1% penicillin–streptomycin. HL-60 and Sup-B15 cells were cultured in IMDM (Gibco) containing 20% FBS and 1% penicillin–streptomycin. 293T cells were cultured with Dulbecco’s modified Eagle’s medium (DMEM) (Gibco) containing 10% FBS and 1% penicillin–streptomycin. All cells were maintained at 37°C with 5% CO₂, and saturated humidity. Cells in the logarithmic growth phase were used for subsequent experiments. All cells were identified by short tandem repeat (STR) profiling and were mycoplasma negative.

The following antibodies were used at the indicated dilution for immunoblotting (IB) and immunoprecipitation (IP) analyses: VDAC3 (1:800 for IB, 3 µg for IP; Proteintech, Wuhan, China), FBXW7 (1:800 for IB; 3 µg for IP, Proteintech), Ubiquitin (1:800 for IB, Proteintech), and β-Actin (1:8,000 for IB, Proteintech). Horseradish peroxidase (HRP) labeled secondary antibody conjugates were purchased from Molecular Probes (Abbkine, Beijing, China). Erastin (APExBIO, Beijing, China), ferrostatin-1 (Fer-1), rapamycin (Rapa), MK-2206, chloroquine (CQ), 3-MA, bafilomycin A1 (BafA1), MG132, and imidazole ketone erastin (IKE) were obtained from Selleck Chemicals (Houston, TX).

Cell viability was detected by the MTS assay (Promega, Madison, WI). Briefly, cells were equally seeded into 96-well plates with various concentrations of erastin (0.01, 0.1, 1, 10, and 100 μmol/l). DMSO (Sigma-Aldrich, St. Louis, MO) was used as a negative control. Cell viability was measured using the MTS kit at 24 h. Absorbance was tested at 490 nm using a microplate analyzer (BioTek, Winooski, VT). Cell mortality was detected by a trypan blue assay (Beyotime Biotechnology, Shanghai, China). In brief, cells were stained with trypan blue and counted under an inverted microscope. The blue-stained cells were considered dead, and five random fields per insert were counted. Cell mortality was calculated according to the following formula: cell mortality (%) = number of dead cells/total number of cells × 100.

Total RNA was extracted using RNAiso Plus Reagent (TaKaRa, Dalian, China) and dissolved in 10 μl of RNase-free water. The purity and concentration were determined using the NanoPhotometer 50 (Implen, Munich, Germany). Next, 1 μg of total RNA was used for reverse transcription to synthesize cDNA using the PrimeScript RT Reagent Kit (TaKaRa). For quantitative real-time polymerase chain reaction (qRT-PCR), the amplification system consisted of 2 μl of cDNA, 1 μl of forward and reverse primers, 10 μl SYBR Premix Ex TaqII (2×), 0.4 µl of ROX Reference DyeⅡ (50×), and ddH₂O for a total volume of 20 μl. The amplification conditions were as follows: initial denaturation at 95°C for 30 s, followed by 40 cycles of 95°C for 5 s and 60°C for 30 s. Relative expression levels were quantified using the 2^−ΔΔCt method. The primer sequences were as follows (synthesized by GenScript, Nanjing, China): FBXW7, 5′-GGC CAA AAT GAT TCC CAG CAA-3′ (Forward), 5′-CCC TAC ATT GCA GAT GAG GCT C-3′ (Reverse); VDAC3: 5′-TTG TAC CGA ACA CAG GAA AGA AG-3′ (Forward); β-Actin, 5′-CAC CAT TGG CAA TGAGCGGTT C-3′ (Forward), 5′-AGG TCT TTG CGGATGTCCACG T-3′ (Reverse).

For the immunoblotting analysis, cells were washed with phosphate-buffered saline (PBS) and lysed in radioimmunoprecipitation assay (RIPA) lysis buffer (Solarbio, Beijing, China) containing 1% PMSF (Solarbio) and 1% protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO). The cells were lysed on ice for 30 min, vibrated and mixed for 10 s every 10 min, and then centrifuged at 14,000 rpm for 30 min at 4°C. The bicinchoninic acid (BCA) method was applied to detect the protein concentration in the supernatant. Next, 30 µg of total protein samples were separated by 10% SDS-PAGE and transferred to polyvinylidene difluoride membranes; the membranes were incubated with primary antibodies (anti-VDAC3, anti-FBXW7, or Ubiquitin antibody) overnight at 4°C. Subsequently, the membranes were blocked with 5% skim milk and then incubated with HRP-conjugated goat anti-rabbit or goat anti-mouse IgG (H+L) as secondary antibodies for 2 h at room temperature. Immunoreactive protein bands were visualized using the ECL Detection Kit (Thermo Fisher Scientific, Waltham, MA).

Cells were harvested and lysed with IP lysis buffer containing 1% PMSF and a 1% protease inhibitor cocktail. Approximately 1 mg of protein extract was used for each sample. First, the cell extract was combined with 3 μg of IP antibody (anti-FBXW7/anti-VDAC3, normal IgG) per sample on a rotating wheel at 4°C overnight. Next, the mixture was co-precipitated with 25 µl of protein A/G beads (MCE, Shanghai, China) on a rotating wheel at 4°C for 1 h. The beads were washed four times with IP lysis buffer, followed by the addition of 20 μl of 2× loading buffer at 100°C for 5 min at 100°C. The supernatants in the 2× loading buffer were prepared for subsequent analyses.

For ubiquitylation assays, cells were treated with 10 μmol/l MG132 for 6 h before harvesting. Cell extracts were incubated with the VDAC3 antibody and protein A/G beads on a rotating wheel for 2 h at 4°C. Subsequently, the beads bound proteins were eluted by boiling them in loading buffer and subjected to immunoprecipitation.

The FBXW7 overexpression lentiviral vector (pRRLSIN-cPPT-SFFV-MCS-3FLAG-E2A-EGFP-SV40-puromycin) and FBXW7 RNAi lentiviral vectors (pRRLSIN-cPPT-U6-shRNA-SFFV-EGFP-SV40-puromycin) were purchased from GeneChem (Shanghai, China). 293T cells, Jurkat cells, and Reh cells were infected following the manufacturer’s protocol. The cells were subjected to puromycin selection for 1 week, and those surviving cells were used for subsequent experiments.

NOD/SCID mice were purchased from HFK Bioscience (Beijing, China). The mice (males, 6–7 weeks old, and 19–21 g weight) were bred under pathogen-free conditions. The mice were provided free access to food and water in a 12 h light/12 h dark cycle animal facility. CCRF-CEM cells (5 × 10⁶/200 μl) were injected into the tail veins of mice. Flow cytometry (FCM) of peripheral blood (PB) was performed to detect huCD45⁺mCD45^– cells every week. PB from the orbital venous plexus was sampled and flushed in PBS solution. CCRF-CEM cells from PB were labeled with PerCP-conjugated anti-huCD45 (BioLegend, San Diego, CA) and FITC-conjugated anti-mCD45 (BioLegend) to determine the fraction of human ALL cells (huCD45⁺mCD45^–) by FCM. The mice were treated after the disease burden was sufficient (>1% huCD45⁺/mCD45^– blasts detected in PB) (Teachey et al., 2006; Maude et al., 2012). The mice were randomly divided into four groups (8 mice per cage) and treated with rapamycin (2 mg/kg), IKE (20 mg/kg), rapamycin (2 mg/kg) + IKE (20 mg/kg), and vehicle (2% DMSO + 30% PEG300 + 68% normal saline) once every other day by i.p. injection for 2–3 weeks. The survival time for each group was recorded. All animal experiments were reviewed and approved by the Shengjing Hospital of China Medical University Institutional Animal Care and Use Committee (No: 2019PS162K).

All statistical analyses were performed using SPSS 25.0 software (Chicago, IL) and GraphPad Prism 8.0 software (San Diego, CA). Data are expressed as means ± SD. Quantitative data were analyzed using Student’s t-test for comparisons between two groups and one-way analysis of variance (ANOVA) with Bonferroni correction for comparisons among multiple groups. The survival curve was drawn by the Kaplan–Meier method and evaluated with the log-rank test. All experiments were performed at least three times, and a p-value < 0.05 was considered statistically significant.

To determine the cytotoxic effect of erastin on ALL cell lines, we treated Reh, Jurkat, CCRF-CEM, CEM/E1, Nalm6, THP1, and HL-60 cells with various concentrations of erastin (0–100 μmol/l) for 24 h and assessed cell viability using MTS assays. As shown in Figures 1A,B, cell viability following erastin treatment varied among ALL cell lines. Reh cells exhibited the most sensitive response, while others displayed the resistance to erastin-induced ferroptosis.

To verify whether autophagy could regulate ferroptosis in ALL cells, the cell mortality of Jurkat, CCRF-CEM, and Reh cells were evaluated after treatment with erastin and Rapa or CQ. As shown in Figure 1C, the mortality rates of Jurkat and CCRF-CEM cells were significantly higher for the combination of Rapa with erastin than for Rapa or erastin alone (p < 0.001 for both). Jurkat and CCRF-CEM cells were further treated with the ferroptosis inhibitor ferrostatin-1 (Fer-1). Fer-1 attenuated the combined effect of Rapa and erastin in cell death, indicating that Rapa participated in the ferroptosis. In addition, Rapa promoted erastin-induced ferroptosis in other ALL cell lines (Nalm6 and Sup-B15 cells, in Supplementary Figure S1). These results indicated that Rapa promotes erastin-induced ferroptosis in ALL cells. Moreover, Jurkat, CCRF-CEM, Nalm6, and Sup-B15 cells were treated with another autophagy activator, MK-2206, and similar results were obtained (Figure 1D). These data indicated that autophagy activation could promote the erastin-induced ferroptosis in ALL cells. In contrast, autophagy inhibitors such as CQ, BafA1, and 3-MA reduced Reh cell mortality induced by erastin (p < 0.001 for all, Figure 1E). These findings demonstrated that activation of autophagy promoted the ferroptosis of ALL cells induced by erastin, and blockage of autophagy protected ALL cells from erastin-induced ferroptosis.

VDAC3 is a direct target of erastin. The expression of VDAC3 was explored following Rapa administration in ALL cells. As shown in Figures 2A,B, in Jurkat and CCRF-CEM cells, VDAC3 expression increased substantially after Rapa treatment, and decreased significantly in Reh cells following CQ treatment (p < 0.05). VDAC3 expression in Reh cells was also higher after Rapa treatment while CQ treatment downregulated in Jurkat and CCRF-CEM cells (both p < 0.05, Supplementary Figure S2). Notably, despite substantial alterations in VDAC3 at the protein level, Rapa had no effect on mRNA expression (Figure 2C). The alteration of autophagic activity (by Rapa or CQ) did not affect the VDAC3 transcription, suggesting that autophagy contributed to the post-translational regulation of VDAC3. Protein degradation in eukaryotic cells occurs mainly via the autophagy lysosome pathway and UPS. Since the blockage of autophagy lysosome degradation did not preserve VDAC3 expression (Figure 2A), VDAC3 protein degradation was predicted mainly to be regulated by UPS. When cells were treated with autophagy regulators (CQ or Rapa) and MG132, as depicted in Figures 2D–F, MG132 inhibited the Rapa-induced VDAC3 expression in Jurkat and CCRF-CEM cells. VDAC3 expression in Reh cells increased significantly after combined treatment with CQ and MG132. These results supported the prediction that autophagy activation promotes the VDAC3 expression post-transcriptionally, mainly via the UPS. Next, we examined the specific mechanism underlying VDAC3 degradation via the UPS.

Based on the results indicating that VDAC3 was regulated post-transcriptionally by autophagy via the UPS, the UbiBrowser database was used to predicted the E3 ligase of VDAC3 (Li et al., 2017). The top 10 E3 ligases were shown in Figures 3A,B. NEDD4 was reported to degrade VDAC2/3 via the UPS to inhibit the sensitivity of melanoma cells to erastin (Yang et al., 2020). We failed to show the interaction between ITCH/MDM2 and VDAC3 by co-immunoprecipitation (Co-IP) and IB assays. We then investigated whether FBXW7 could interact with VDAC3. As shown in Figures 3C,D, FBXW7 was enriched when VDAC3 was targeted by IP, while VDAC3 was enriched when FBXW7 was the target protein. This confirmed the interaction between VDAC3 and FBXW7. Moreover, FBXW7 expression decreased in Jurkat and CCRF-CEM cells after Rapa treatment, whereas FBXW7 expression in Reh cells was significantly lower than that in the control group after CQ treatment (p < 0.05 for all, Figures 3E,F). These findings revealed the interaction between FBXW7 and VDAC3, and FBXW7 protein expression could be regulated by autophagic activity.

To clarify how FBXW7 affects VDAC3 degradation, FBXW7 was knocked-down or overexpressed in 293T cells. The transfection was effective as confirmed by IB and qRT-PCR (Figures 4A,B). As depicted in Figures 4C,D, VDAC3 protein levels increased significantly after FBXW7 was silenced (p < 0.001) and decreased significantly when FBXW7 was overexpressed (p < 0.001). These results confirmed that VDAC3 protein expression was related to the level of FBXW7. Additionally, the levels of VDAC3 ubiquitination were detected by IP and IB assays after FBXW7 expression was altered. As shown in Figures 4E,F, the ubiquitination of VDAC3 decreased after FBXW7 silencing and increased after FBXW7 overexpression, suggesting that FBXW7 degraded VDAC3 via the UPS. Then, FBXW7 was knocked-down in Jurkat cells and overexpressed in Reh cells. The efficiency of transfection was determined by IB (Figure 4G). As shown in Figure 4H, mortality rates were significantly higher for Jurkat cells treated with sh-FBXW7 and erastin than in cells treated with sh-FBXW7 or erastin alone (both p < 0.001). Similar results were obtained for Nalm6 and Sup-B15 cells (Supplementary Figure S3). In contrast, oe-FBXW7 reduced the sensitivity of Reh cells to erastin (p < 0.05). These results indicated that the silencing of FBXW7 increased the sensitivity of ALL cells to erastin.

To confirm the therapeutic effect of the combination of Rapa with erastin in vivo, we used a xenograft mouse model. The erastin-resistant cell line CCRF-CEM was transplanted into NOD/SCID mice (Figure 5A) and huCD45⁺mCD45^– cells were evaluated by FCM every week. At week 2, huCD45⁺mCD45^– leukemic cells in PB were detected (Figure 5B). After huCD45⁺mCD45^– cells reached >1%, Rapa, IKE (erastin for in vivo administration), Rapa+IKE, and vehicle were administered once every other day by i.p. injection for 2–3 weeks. After transplantation, animals manifested signs of advanced disease, such as dull hair, arched back, rear limb weakness, or paralysis (Figure 5C). Compared to normal mice, the femur of recipient mice showed a more apparent white appearance and the spleen and liver were significantly enlarged (Figure 5D). These findings indicated that the leukemia engrafted successfully. The frequency of huCD45⁺mCD45^– cells in PB of the combination treatment group was significantly lower than that in the Rapa or IKE alone treatment groups (both p < 0.05, Figure 5F). Rapa+IKE extended the survival mice with advanced disease compared to the single treatment groups, as shown by Kaplan–Meier curves (Figure 5E). These data suggested that the administration of an autophagy activator could enhance the therapeutic effect of erastin.