Advertisement

DIA-based Proteomics Identifies IDH2 as a Targetable Regulator of Acquired Drug Resistance in Chronic Myeloid Leukemia

  • Wei Liu
    Affiliations
    Department of Clinical Pharmacology, College of Pharmacy, Dalian Medical University, Dalian, Liaoning, China

    Key Laboratory of Structural Biology of Zhejiang Province, School of Life Sciences, Westlake University, Hangzhou, Zhejiang, China

    Center for Infectious Disease Research, Westlake Laboratory of Life Sciences and Biomedicine, Hangzhou, Zhejiang, China

    Institute of Basic Medical Sciences, Westlake Institute for Advanced Study, Hangzhou, Zhejiang, China
    Search for articles by this author
  • Yaoting Sun
    Affiliations
    Key Laboratory of Structural Biology of Zhejiang Province, School of Life Sciences, Westlake University, Hangzhou, Zhejiang, China

    Center for Infectious Disease Research, Westlake Laboratory of Life Sciences and Biomedicine, Hangzhou, Zhejiang, China

    Institute of Basic Medical Sciences, Westlake Institute for Advanced Study, Hangzhou, Zhejiang, China
    Search for articles by this author
  • Weigang Ge
    Affiliations
    Westlake Omics (Hangzhou) Biotechnology Co., Ltd., Hangzhou 310024, China
    Search for articles by this author
  • Fangfei Zhang
    Affiliations
    Key Laboratory of Structural Biology of Zhejiang Province, School of Life Sciences, Westlake University, Hangzhou, Zhejiang, China

    Center for Infectious Disease Research, Westlake Laboratory of Life Sciences and Biomedicine, Hangzhou, Zhejiang, China

    Institute of Basic Medical Sciences, Westlake Institute for Advanced Study, Hangzhou, Zhejiang, China
    Search for articles by this author
  • Lin Gan
    Affiliations
    Key Laboratory of Structural Biology of Zhejiang Province, School of Life Sciences, Westlake University, Hangzhou, Zhejiang, China

    Center for Infectious Disease Research, Westlake Laboratory of Life Sciences and Biomedicine, Hangzhou, Zhejiang, China

    Institute of Basic Medical Sciences, Westlake Institute for Advanced Study, Hangzhou, Zhejiang, China
    Search for articles by this author
  • Yi Zhu
    Affiliations
    Key Laboratory of Structural Biology of Zhejiang Province, School of Life Sciences, Westlake University, Hangzhou, Zhejiang, China

    Center for Infectious Disease Research, Westlake Laboratory of Life Sciences and Biomedicine, Hangzhou, Zhejiang, China

    Institute of Basic Medical Sciences, Westlake Institute for Advanced Study, Hangzhou, Zhejiang, China
    Search for articles by this author
  • Tiannan Guo
    Correspondence
    Correspondence:
    Affiliations
    Key Laboratory of Structural Biology of Zhejiang Province, School of Life Sciences, Westlake University, Hangzhou, Zhejiang, China

    Center for Infectious Disease Research, Westlake Laboratory of Life Sciences and Biomedicine, Hangzhou, Zhejiang, China

    Institute of Basic Medical Sciences, Westlake Institute for Advanced Study, Hangzhou, Zhejiang, China
    Search for articles by this author
  • Kexin Liu
    Correspondence
    Correspondence:
    Affiliations
    Department of Clinical Pharmacology, College of Pharmacy, Dalian Medical University, Dalian, Liaoning, China
    Search for articles by this author
Open AccessPublished:December 15, 2021DOI:https://doi.org/10.1016/j.mcpro.2021.100187

      Highlights

      • 1.
        The dynamic and temporal changes of proteomes responsive to the imatinib or adriamycin-induced drug resistance were acquired.
      • 2.
        Four DIA software tools-based PulseDIA data analysis quantified 7082 proteotypic proteins from 98,232 peptides in K562 cells.
      • 3.
        Sirtuin Signaling Pathway was significantly enriched in resistant K562 cells.
      • 4.
        IDH2 was identified as a potential drug target correlated with the drug resistance phenotype.

      Abstract

      Drug resistance is a critical obstacle to effective treatment in patients with chronic myeloid leukemia (CML). To understand the underlying resistance mechanisms in response to imatinib (IMA) and adriamycin (ADR), the parental K562 cells were treated with low doses of IMA or ADR for two months to generate derivative cells with mild, intermediate and severe resistance to the drugs as defined by their increasing resistance index (RI). PulseDIA-based quantitative proteomics was then employed to reveal the proteome changes in these resistant cells. In total, 7082 proteotypic proteins from 98,232 peptides were identified and quantified from the dataset using four DIA software tools including OpenSWATH, Spectronaut, DIA-NN, and EncyclopeDIA. Sirtuin Signaling Pathway was found to be significantly enriched in both ADR- and IMA-resistant K562 cells. In particular, IDH2 was identified as a potential drug target correlated with the drug resistance phenotype, and its inhibition by the antagonist AGI-6780 reversed the acquired resistance in K562 cells to either ADR or IMA. Together, our study has implicated IDH2 as a potential target that can be therapeutically leveraged to alleviate the drug resistance in K562 cells when treated with IMA and ADR.

      Graphical abstract

      Keywords

      Introduction

      The treatment of chronic myeloid leukemia (CML) patients includes targeted therapy (tyrosine kinase inhibitors, TKIs), chemotherapy, biological therapy, hematopoietic cell transplant (HCT), and donor lymphocyte infusion (DLI) [
      • Radujkovic A.
      • et al.
      Donor Lymphocyte Infusions for Chronic Myeloid Leukemia Relapsing after Allogeneic Stem Cell Transplantation: May We Predict Graft-versus-Leukemia Without Graft-versus-Host Disease?.
      ,
      • Radich J.P.
      • et al.
      Chronic Myeloid Leukemia, Version 1.2019, NCCN Clinical Practice Guidelines in Oncology.
      ,
      • Pal S.K.
      • et al.
      Clinical Cancer Advances 2019: Annual Report on Progress Against Cancer From the American Society of Clinical Oncology.
      ]. Chemotherapy inhibits the rapidly proliferating tumor cells by interfering with cell replication. However, drug resistance leads to failed chemotherapy treatment in 90% patients [
      • Mansoori B.
      • et al.
      The Different Mechanisms of Cancer Drug Resistance: A Brief Review.
      ]. Adriamycin (ADR) is a traditional chemotherapeutic drug that disturbs the DNA replication process, which can be therapeutically leveraged against certain hematologic tumors. Using a K562 cell model, various mechanisms have been uncovered to explain ADR-induced drug resistance, such as transporter-mediated drug efflux [
      • Liu W.
      • et al.
      Targeting P-Glycoprotein: Nelfinavir Reverses Adriamycin Resistance in K562/ADR Cells.
      ,
      • Sun Y.
      • et al.
      Targeting P-glycoprotein and SORCIN: Dihydromyricetin strengthens anti-proliferative efficiency of adriamycin via MAPK/ERK and Ca(2+) -mediated apoptosis pathways in MCF-7/ADR and K562/ADR.
      ], altered mitochondrial function [
      • De Oliveira F.
      • et al.
      Effects of permeability transition inhibition and decrease in cytochrome c content on doxorubicin toxicity in K562 cells.
      ,
      • Li R.J.
      • et al.
      Down-regulation of mitochondrial ATPase by hypermethylation mechanism in chronic myeloid leukemia is associated with multidrug resistance.
      ], and changes in survival-related signaling pathways including EGFR,ERK, NF-κB, PTEN, and AKT pathways [
      • Zhao L.
      • et al.
      Functional screen analysis reveals miR-3142 as central regulator in chemoresistance and proliferation through activation of the PTEN-AKT pathway in CML.
      ,
      • Jiang L.
      • et al.
      Ivermectin reverses the drug resistance in cancer cells through EGFR/ERK/Akt/NF-κB pathway.
      ,
      • Dong J.
      • et al.
      Medicinal chemistry strategies to discover P-glycoprotein inhibitors: An update.
      ].
      As one of the most effective clinical regimens in the chronic phase, TKIs have dramatically improved survival rate in CML patients [
      • Holyoake T.L.
      • Helgason G.V.
      Do we need more drugs for chronic myeloid leukemia?.
      ]. Indeed, imatinib (IMA) is among the first generation of TKIs and the first-line drug for CML treatment, which targets BCR-ABL1 and inhibits tumor growth [
      • Buchdunger E.
      • O'Reilley T.
      • Wood J.
      Pharmacology of imatinib (STI571).
      ]. Notably, the five-year survival rate of CML patients with IMA treatment was increased to 89% [
      • Druker B.J.
      • et al.
      Five-year follow-up of patients receiving imatinib for chronic myeloid leukemia.
      ]. However, about 20-25% of CML patients showed a suboptimal response to IMA, who have likely developed drug resistance [
      • Milojkovic D.
      • Apperley J.
      Mechanisms of Resistance to Imatinib and Second-Generation Tyrosine Inhibitors in Chronic Myeloid Leukemia.
      ]. Several mechanisms have been proposed to explain the failed IMA treatment in CML patients, including for example altered conformation of the BCR-ABL1 kinase domain by mutations that reduces its binding affinity to IMA [
      • Holyoake T.L.
      • Helgason G.V.
      Do we need more drugs for chronic myeloid leukemia?.
      ]. Other resistance mechanisms independent of BCR-ABL1 have also been reported, including P-glycoprotein (P-gp) upregulation, activation of alternative PI3K/AKT, JAK-2, or MAPK-signaling [
      • Burchert A.
      Roots of imatinib resistance: a question of self-renewal?.
      ,
      • Weisberg E.
      • Griffin J.D.
      Resistance to imatinib (Glivec): update on clinical mechanisms.
      ], changes in the intracellular environment such as endoplasmic reticulum stress induced autophagy [
      • Bellodi C.
      • et al.
      Targeting autophagy potentiates tyrosine kinase inhibitor-induced cell death in Philadelphia chromosome-positive cells, including primary CML stem cells.
      ].
      Drug resistance remains a clinical hurdle to traditional chemotherapy and targeted therapy [
      • Cree I.A.
      • Charlton P.
      Molecular chess? Hallmarks of anti-cancer drug resistance.
      ]. Given its complex nature, both genetic mutations and non-genetic changes (such as epigenetics) may contribute to a drug-resistance phenotype [
      • Aleksakhina S.N.
      • Kashyap A.
      • Imyanitov E.N.
      Mechanisms of acquired tumor drug resistance.
      ]. Mass spectrometry (MS)-based proteomics could quantify thousands of proteins and provide unique insights into the pathways of dysregulation [
      • Aebersold R.
      • Mann M.
      Mass-spectrometric exploration of proteome structure and function.
      ], thus can be used to explore the mechanisms of drug resistance. A deep proteome profiling is essential for charactering relevant signaling proteins responsible for drug resistance. Some study of CML drug resistance reported relatively few protein numbers, i.e., 2059 proteins [
      • Monteleone F.
      • et al.
      SWATH-MS based quantitative proteomics analysis reveals that curcumin alters the metabolic enzyme profile of CML cells by affecting the activity of miR-22/IPO7/HIF-1α axis.
      ], 1344 proteins [
      • Xiong L.
      • et al.
      Global proteome quantification for discovering imatinib-induced perturbation of multiple biological pathways in K562 human chronic myeloid leukemia cells.
      ], and 477 proteins [
      • Corrêa S.
      • et al.
      A comparative proteomic study identified LRPPRC and MCM7 as putative actors in imatinib mesylate cross-resistance in Lucena cell line.
      ]. Some studies tried to characterize the changes of proteomes related to either drug-resistant or drug-sensitive phenotypes of CML cells, bone marrow extracellular fluids [
      • Corrêa S.
      • et al.
      A comparative proteomic study identified LRPPRC and MCM7 as putative actors in imatinib mesylate cross-resistance in Lucena cell line.
      ,
      • Ferrari G.
      • et al.
      Comparative proteomic analysis of chronic myelogenous leukemia cells: inside the mechanism of imatinib resistance.
      ,
      • Gjertsen B.T.
      • Wiig H.
      Investigation of therapy resistance mechanisms in myeloid leukemia by protein profiling of bone marrow extracellular fluid.
      ] or imatinib treatment CML cells in just 24 hours [
      • Xiong L.
      • et al.
      Global proteome quantification for discovering imatinib-induced perturbation of multiple biological pathways in K562 human chronic myeloid leukemia cells.
      ]. However, the difference in phenotype may only be partially attributed to imatinib. Furthermore, these are all proteomic profiling of CML cells under a certain physiological state [
      • Hrdinova T.
      • et al.
      Exosomes released by imatinib-resistant K562 cells contain specific membrane markers, IFITM3, CD146 and CD36 and increase the survival of imatinib-sensitive cells in the presence of imatinib.
      ].Here, we focused on the dynamic and temporal changes of proteomes responsive to the imatinib-induced drug resistance of CML cells at different stages.
      Data-independent acquisition (DIA) is an effective proteomic method with rigorous quantitative accuracy and reproducibility [
      • Gillet L.C.
      • et al.
      Targeted data extraction of the MS/MS spectra generated by data-independent acquisition: a new concept for consistent and accurate proteome analysis.
      ,
      • Collins B.C.
      • et al.
      Multi-laboratory assessment of reproducibility, qualitative and quantitative performance of SWATH-mass spectrometry.
      ]. Based on DIA-MS, our group has recently developed PulseDIA-MS as an improved approach that utilizes gas phase fractionation to achieve a greater proteome depth [
      • Cai X.
      • et al.
      PulseDIA: Data-Independent Acquisition Mass Spectrometry Using Multi-Injection Pulsed Gas-Phase Fractionation.
      ]. In this work, we employed the pressure cycling technology (PCT)-based peptide preparation [
      • Guo T.
      • et al.
      Rapid mass spectrometric conversion of tissue biopsy samples into permanent quantitative digital proteome maps.
      ,
      • Shao S.
      • et al.
      Minimal sample requirement for highly multiplexed protein quantification in cell lines and tissues by PCT-SWATH mass spectrometry.
      ], and the PulseDIA-MS strategy to quantify the dynamic proteome changes upon drug treatment, using time-series K562 drug-resistant cell line models treated by ADR and IMA, respectively. Due to the lack of correspondence information between precursor ions and fragment ions in DIA data, DIA data analysis has become a huge challenge. The different DIA tools have different scoring methods and core algorithm for peptides prediction, which may lead to certain technical deviations in the analytical results of the same DIA data [
      • Zhang F.
      • et al.
      Data-Independent Acquisition Mass Spectrometry-Based Proteomics and Software Tools: A Glimpse in 2020.
      ]. To reduce the technical deviations of the quantitative proteome by DIA software tools, we quantified the proteome with four commonly used independent tools, Spectronaut [
      • Bruderer R.
      • et al.
      Extending the Limits of Quantitative Proteome Profiling with Data-Independent Acquisition and Application to Acetaminophen-Treated Three-Dimensional Liver Microtissues.
      ], DIA-NN [
      • Demichev V.
      • et al.
      DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput.
      ], EncyclopeDIA [
      • Searle B.C.
      • et al.
      Chromatogram libraries improve peptide detection and quantification by data independent acquisition mass spectrometry.
      ] and OpenSWATH [
      • Rost H.L.
      • et al.
      OpenSWATH enables automated, targeted analysis of data-independent acquisition MS data.
      ]. With this cell model and quantitative approaches, we were able to pinpoint and characterize pathways that were significantly altered upon ADR or IMA treatment. Notably, we identified IDH2 a previously unknown potential target that can be therapeutically leveraged to reverse drug resistance in K562 cells.

      Experimental Procedures

       Experimental design and statistical rationale

      In this study, a total of 21 samples from seven different drug sensitivities K562 cell lines were subjected to mass spectrometric analysis. For each cell line, we harvested three individually cultured cell samples as biological replicates and performed sample preparation independently. We also selected five samples for duplicate injection in MS acquisition as technical replicates. All 26 samples were subject to four-part pulseDIA analysis as described previously [
      • Cai X.
      • et al.
      PulseDIA: Data-Independent Acquisition Mass Spectrometry Using Multi-Injection Pulsed Gas-Phase Fractionation.
      ], and generated a total of 104 DIA MS raw data. To ensure the stability and reliability of the MS data, a mouse liver digest was used for instrument performance evaluation, and we also run blank samples (buffer A) every 4 injections to minimize carry-over. We carried out library-based method analysis of the 106 pulseDIA MS data using four DIA tools. 37 data-dependent acquisition (DDA) acquired samples from 10 in-solution digestion files, 7 PCT-assisted digestion files, and 20 high-PH fractionations were used to generate the spectral library. Due to the presence of a gap in the MS2 window of each pulseDIA data, the standard peptides or proteins were not suitable for retention time correction [
      • Cai X.
      • et al.
      PulseDIA: Data-Independent Acquisition Mass Spectrometry Using Multi-Injection Pulsed Gas-Phase Fractionation.
      ]. Therefore, we used the software's internal default algorithm for retention time correction or the common index RT(CIRT) method [
      • Zhu T.
      • et al.
      DPHL: A DIA Pan-human Protein Mass Spectrometry Library for Robust Biomarker Discovery.
      ]. We obtained quantitative results of peptide levels by DIA software with peptide precursors 0.01 Q-value (FDR) cut-off. At the protein level, the linear regression of top 3 precursor intensity was used to peptide to protein inference [
      • Zhu T.
      • et al.
      ProteomeExpert: a Docker image-based web server for exploring, modeling, visualizing and mining quantitative proteomic datasets.
      ].

       Establishment of drug-resistance K562 cell models

      The parental sensitive K562 cells were purchased from Nanjing KeyGen Biotech Co., Ltd. (Nanjing, China) and authenticated via ShortTandem Repeat (STR) profiling by Shanghai Biowing Applied Biotechnology Co., Ltd (Shanghai, China) on Feb. 28, 2019. The parental K562 cells were cultured in RPMI medium (Cromwell, USA) with 10% fetal bovine serum (Waltham, USA) and 1% penicillin-streptomycin (HyClone, USA) at 37 °C with 5% CO2 and 95% humidity.
      The building of drug-resistance K562 cell models includes three phases (Fig. 1A). In the first phase, 1X105 K562 cells were treated with 0.1 μM ADR (A1) or IMA (I1). After one week, K562 cells were cultured to about 10 million cells. Then drugs were removed by centrifugation and cells were divided into four aliquots for cell cytotoxicity determination, subsequent drug-resistance induction, cell collection, and cell freezing. In the second phase, the concentration of ADR and IMA was increased to 0.4 μM (A2) and 0.8 μM (I2), respectively. The treatment lasted for two weeks. In the third phase, the concentration of ADR and IMA increased to 0.8 μM (A3) and 1.6 μM (I3), respectively. The treatment lasted for four weeks. In this way, we established two drug-resistance K562 models for ADR and IMA, separately.
      Figure thumbnail gr1
      Figure 1Establish derivitative K562 cells with mild, intermediate and severe resistance to ADR and IMA. (A) Overview of the drug resistance model. (B) Native K562 cells and IMA-resistant K562 cells from each of the three phases were treated with a series of cytotoxicity concentrations (4, 2, 1, 0.5……0 μM) IMA for 48 h. (C) Native K562 cells and ADR-resistant K562 cells from each of the three phases were treated with a series of concentrations (4, 2, 1, 0.5……0 μM) ADR for 48 h. The cell survival rate was calculated and plotted in each group. The downward shift of the survival curves by increasing treating concentration (I1, I2, I3 and A1, A2, A3) indicated suppressed proliferation.

       Cytotoxicity assay

      The cytotoxicity of Adriamycin (CAS: 25316-40-9, meilunbio, Dalian, China), imatinib mesylate (CAS: 220127-57-1, meilunbio, Dalian, China) or AGI-6780 (CAS: 1432660-47-3, MedChemExpress, New Jersey, USA) to native K562 cells and the model cells were detected by cell counting kit-8 (CCK-8) (Bimake, Houston, USA). Cells were planted into 96-well plates in a density of 5000 cells/100uL medium/well. After 24h, cells were treated with drugs for 48h. After 48h, 10μL CCK-8 was added to each well and incubated about 2h in dark. The absorbance value of each well was determined by a Synergy™ H1 (BioTek, State of Vermont, USA) at 450nm. The IC50 value was calculated by SPSS 22.0.

       PCT based peptide extraction

      The workflow of PCT based peptide extraction is described as Shao et al [
      • Shao S.
      • et al.
      Minimal sample requirement for highly multiplexed protein quantification in cell lines and tissues by PCT-SWATH mass spectrometry.
      ]. For each sample, 500,000 cells were harvested and cleaned by PBS three times to remove all traces of fetal bovine serum.
      Cells were transferred into PCT-Microube (Pressure Biosciences Inc.) with 30uL lysis buffer, 5 uL 1ug/uL DNAase (STEMCELL) and 15uL 0.1M ammonium bicarbonate (ABB) (GENERAL-REAGENT®). The lysis buffer includes 6M urea (SIGMA-ALDRICH®) and 2 M thiourea (SIGMA-ALDRICH®. Then the cells were lysis by Barocycler NEP2320-45k (PressureBioSciences Inc.) with 90 cycles containing 25 s of 45,000 p.s.i. high pressure plus 10 s at ambient pressure, at 30 °C The lysate solution was added 2.5 μL 800 mM tris (2-carboxyethyl) phosphine (ALDRICH®) and 5 μL 100 mM iodoacetamide (SIGMA®) to in PCT-Micro tube to dilute into a final concentration of 10 mM and 40 mM, followed by a 30 min incubation in the dark with gentle vortexing (800 rpm) at room temperature in shaker.
      After reduction and alkylation, the proteins solution were added with 57.5μL 0.1 M ABB and 25 μL 0.1 mg/mL Lys-C (Hualishi) and digested using Barocycler NEP2320-45k (Pressure Biosciences Inc., MA, USA) with 45 cycles containing 50 s of 20,000 p.s.i. high pressure plus 10 s at ambient pressure, at 30 °C.
      After Lys-C digestion, the solution was added with 10 μL 0.2 mg/mL trypsin (Hualishi) and were tryptic digested by Barocycler NEP2320-45k with 90 cycles containing 50 s of 20,000 p.s.i. high pressure plus 10 s at ambient pressure, at 30. Then, 15 μL 10% trifluoroacetic acid (Fisher Scientific®) was added to the lysate solution at a final concentration of 1% to stop digestion.
      Then the digested peptides were cleaned in micro spin columns (The Nest Group Inc.) and dried in CentriVap DNA Vacuum Concentrators (LABCONCO). Peptides were redissolved in ms buffer (0.1% formic acid and 2% acetonitrile in HPLC water). The peptide concentration was measured using ScanDrop2 (Analytik Jena).

       PulseDIA mass spectrometry

      The PulseDIA mass spectrometry method was performed as previously described [
      • Cai X.
      • et al.
      PulseDIA: Data-Independent Acquisition Mass Spectrometry Using Multi-Injection Pulsed Gas-Phase Fractionation.
      ]. The redissolved peptides of each sample were acquired and analyzed by EASY-nLC™ 1200 System (Thermo Fisher Scientific™) coupled to a QE HF-X mass spectrometer (Thermo Fisher Scientific™). The MS1 was acquired in an m/z range of 390 to 1210 with the resolution at 60,000, AGC target of 3e6, and the maximum ion injection time of 80 ms. The MS2 was performed with the resolution at 30,000, AGC target of 1e6, and the maximum ion injection time of 50 ms. Different from the conventional DIA method (the 400-1200 m/z mass range is divided into 24 windows) [
      • Gillet L.C.
      • et al.
      Targeted Data Extraction of the MS/MS Spectra Generated by Data-independent Acquisition: A New Concept for Consistent and Accurate Proteome Analysis.
      ], the PulseDIA symmetrically divided each window into four parts and acquired the data of each part independently. For one sample, we acquired four injections with different MS2 range. For each pulse acquisition, 0.2 μg of peptides was injected and separated across a linear 45 min LC gradient (from 8% to 40% buffer B) at a flowrate of 300 nL/min (precolumn, 3μm, 100 Å, 20mm*75μm i.d.; analytical column, 1.9um, 120 Å,150mm*75um i.d.). Buffer A was HPLC-grade water containing 0.1% FA, and buffer B was 80%ACN, 20%H2O containing 0.1%FA.

       Quality control samples

      Cells with mild, intermediate and severe resistance to ADR and IMA were analyzed in 3 duplicates for peptide extraction and acquisition as biological replicates. Five samples were repeatedly injected by PulseDIA as the technical replicates to evaluate the data quality.

       Generating Spectral library for DIA-MS

      To build the spectral library, we acquired 37 data-dependent acquisition (DDA) files including 10 in-solution digestion files, 7 PCT-assisted digestion files and 20 high-PH fraction files on a QE-HFX mass spectrometer in DDA mode. Library was built by Spectronaut (version: 13.5.190902.43655) for Spectronaut and DIA-NN analysis. Library was built by OpenSWATH (version 2.0) for OpenSWATH and EncyclopeDIA analysis. In the two libraries, data were searched against the SwissProt Human database (20, 269 entries). Trypsin and Lys-C were used to generate peptides in silico.
      For Spectronaut library building, carbamidomethyl (C) was set as the fixed modification, acetyl (Protein N-term) and oxidation (M) were set as the variable modification. Two missed trypsin cleavages were allowed. The precursor and fragment tolerance were set as dynamic. Two calibration searches were performed: based on the first-pass calibration (rough calibration), the ideal tolerance for the second-pass calibration was defined; based on the second-pass calibration (finer calibration), the ideal tolerance for the main search was defined (Spectronaut™ Manual, https://biognosys.com/resources/spectronaut-manual/). The FDR cutoff on precursor and protein was 0.01. Finally, a K562 library including identified 191,008 precursors, 133,025 peptides, 8226 protein groups and 8313 proteins was built.
      For OpenSWATH library building, the pFind [
      • Chi H.
      • et al.
      Comprehensive identification of peptides in tandem mass spectra using an efficient open search engine.
      ] (version 3.1.5) was used as a search engine with the parameters including carbamidomethyl (C) as a fixed modification and oxidation (O) as a variable modification. DDA library was built according to the workflow [
      • Schubert O.T.
      • et al.
      Building high-quality assay libraries for targeted analysis of SWATH MS data.
      ] (http://openswath.org/en/latest/ docs/pqp.html#id3). No missed trypsin cleavage was allowed by default. The precursor peptide mass tolerance was 20 ppm and fragment ion mass tolerance were 0.05 Da. The FDR cutoff on precursor and protein was 0.01 and other parameters were performed as default. Finally, a K562 library including identified 110,583 transition groups, 84,548 target peptides, 84,910 decoy peptides, 9511 target protein groups, 9575 decoy protein groups and 7935 target proteotypic proteins was built.
      Base on the analysis of K562 Spectral library, we finally identified a total of 8524 proteotypic proteins and up to 10,732 protein groups.

       PulseDIA data analysis

      Library-based PulseDIA data analysis was performed by Spectronaut, OpenSWATH, DIA-NN, and EncyclopeDIA.
      For Spectronaut (version: 13.5.190902.43655) analysis, the default setting of library-based DIA analysis was used for PulseDIA analysis. PulseDIA analysis was performed according to the standard workflow in Spectronaut (Spectronaut™ Manual, https://biognosys.com/resources/spectronaut-manual/). Retention time prediction type was set to dynamic iRT (correction factor for window 1). The MS Mass Tolerance was set as dynamic, which means Spectronaut calculated the ideal mass tolerances for data extraction and scoring based on its extensive mass calibration. At least three fragment ions used per peptide identification and major and minor group quantities were set to mean peptide and mean precursor quantity. The FDR was set to 1% at the peptide precursor level.
      For DIA-NN analysis, we used the same library with Spectronaut analysis. Library search was performed according to the DIA-NN manual (https://github.com/vdemichev/DiaNN/blob/ master/ DIA-NN%20GUI%20manual.pdf). For retention time prediction and extraction mass accuracy, we used the default parameter 0.0, which means DIA-NN performed automatic mass and retention time correction. Top six fragments (ranked by their library intensities) were used for peptide identification and quantification. The FDR was set to 1% at the peptide precursor level.
      For OpenSWATH (version: 2.4) analysis, the retention time extraction window was 300 seconds and m/z extraction for MS2 was performed with 30 ppm tolerance while m/z extraction for MS1 was performed with 20 ppm tolerance. Retention time was then calibrated using common index retention time standards (CiRT) peptides. The m/z extraction for CiRT peptides was performed with 50 ppm tolerance. Peak groups were used peptide identification if they contained peaks in three of five transitions, and the most intense peaks with most were prioritized. Peptide precursors that were identified by OpenSWATH and Pyprophet with FDR 0.01.
      For EncyclopeDIA (versions: 0.9.0) analysis, the library was converted from OpenSWATH library. The precursor, fragment, and library mass tolerance were set as 10 ppm for the PulseDIA data. The retention time model was generated from peptides detected at 1% FDR using a non-parametric kernel density estimation algorithm that follows the density mode over time. Peptide quantities were set to the sum of the top five transitions that pass EncyclopeDIA criteria, and peptides with at least three quantitative transitions were considered to be trustworthy. The search results were filtered at a 1% peptide-level FDR.
      The peptide matrixes from four DIA software tools were converted to protein matrixed by in-house R code (https://github.com/Allen188/PulseDIA/blob/master/Pulsedia_DIANNresult _combine.R) and ProteomeExpert platform [
      • Zhu T.
      • et al.
      DPHL: A DIA Pan-human Protein Mass Spectrometry Library for Robust Biomarker Discovery.
      ].

       Calculation of IC50 values

      The IC50 values were calculated by a non-linear least squares regression model to fit the data log (inhibitor) vs. normalized response in GraphPad Prism 6.

       Bliss independence model analysis

      Bliss independence model analysis was performed using Combenefit software Combenefit software (version:2.021) [
      • Di Veroli G.Y.
      • et al.
      Combenefit: an interactive platform for the analysis and visualization of drug combinations.
      ]. With the cytotoxicity assay results as input, we calculated the cell survival rate after jointly treated with AGI-6780 and imatinib or adriamycin.

       Statistical analysis

      The statistical significance of protein intensity in different types of resistant cells and parental K562 cells was determined by one-way analysis of variance (ANOVA), and p values were adjusted using Benjamini & Hochberg correction. Proteins with p values less than 0.05 were considered as statistically significant. Soft clustering analysis of statistically significant proteins was performed by the R/Bioconductor package Mfuzz [
      • Kumar L.
      • Mfuzz E.F.M.
      a software package for soft clustering of microarray data.
      ]. The average protein intensity of the parental K562 cells and each type of resistant cells as the input data for clustering. The time-series were separated according to the cell sensitivity to ADR or IMA, with the initial being the parental K562 cells. Ingenuity Pathway Analysis (IPA, QIAGEN) was performed to outline the significant canonical pathways [
      • Jimenez-Marin A.
      • et al.
      Biological pathway analysis by ArrayUnlock and Ingenuity Pathway Analysis.
      ]. In IPA, the p value was calculated using the right-tailed Fisher's exact test and a p value less than 0.05 is considered significant [
      • Kramer A.
      • et al.
      Causal analysis approaches in Ingenuity Pathway Analysis.
      ].

      Results

       Establishment of drug-resistance cell models

      Drug-resistant K562 cell models were developed as shown in the workflow (Fig. 1A). The parental K562 cells were sensitive to ADR and IMA treatment, and then treated with increasing concentrations of each drug for one, two, and four weeks to obtain derivative cell lines with differential drug sensitivities. Each model contains three time points during the drug resistance acquisition, with each time point containing cells showing a different degree of drug sensitivity (Fig. 1A, method).
      At this point, we have generated derivative K562 cells with different degrees of IMA resistance: mild, intermediate and severe, which were defined by a resistant index (RI) at 2.76, 6.41 and 7.06 (Fig. 1B), respectively, as shown in Table 1. Similarly, we have mild, intermediate, and severe ADR-resistant K562 cells with a RI at 2.00, 3.49 and 11.59, respectively (Fig. 1C; Table 1). These K562 cells with different degrees of IMA and ADR resistance were all collected for PulseDIA-based proteomic analysis (Fig. 2A).
      Table 1IC50 values and resistance index of the derivative K562 cells.
      ModelIC50 (μM, 95%CI)
      IC50 values represent the best-fit values and 95% CI of three independent experiments performed.
      Resistance Index
      Resistance Index was calculated by dividing the best-fit IC50 values of native K562 cells in response to ADR and IMA to the best-fit IC50 values of model cells.
      IMAADR
      Native K562 cells0.37(0.26 to 0.54)0.29 (0.24 to 0.35)1
      Model IMA phase 10.74 (0.61 to 0.90)/2
      Model IMA phase 21.29 (1.06 to 1.58)/3.49
      Model IMA phase 34.29 (3.03 to 6.36)/11.59
      Model ADR phase 1/0.80 (0.51 to 1.24)2.76
      Model ADR phase 2/1.86 (1.5 to 2.30)6.41
      Model ADR phase 3/2.05 (1.32 to 3.32)7.06
      a IC50 values represent the best-fit values and 95% CI of three independent experiments performed.
      b Resistance Index was calculated by dividing the best-fit IC50 values of native K562 cells in response to ADR and IMA to the best-fit IC50 values of model cells.
      Figure thumbnail gr2
      Figure 2PulseDIA raw data acquisition and analysis. (A) The workflow of peptide extraction from samples and analysis of mass spectrometry raw data. (B) The numbers of identified peptides by four DIA software tools, and their Venn diagram (C). (D)The numbers of identified proteins by four DIA software, and their Venn diagram (E).

       Peptide and protein identification

      PCT-assisted peptide preparation and PulseDIA-MS were then carried out to analyze the parental K562 cells and the derivative resistant cells in biological triplicates (Fig. 2A). In total, 98,232 peptides, 8630 protein groups,7082 proteotypic proteins were quantified from 26 independent PulseDIA-MS runs (Fig. 2B, D; Table. S1). 30,289 peptides and 4493 proteotypic proteins were consistently quantified by four DIA software (Fig. 2C, E). We used these commonly identified proteins for the subsequent quantitative analysis of proteome changes during the development of drug resistance in K562 cells.

       Quality control (QC) of PulseDIA proteome dataset

      We examined the reproducibility of PulseDIA data by calculating the Pearson correlation coefficient between technical duplicates and the coefficient of variation (CV) of the protein intensity among the biological triplicates. Using four different DIA analysis tools as described above, the five pairs of technical duplicates showed a strong correlation (r>0.9) (Fig. 3A), and the median CV of protein intensity among biological triplicates was around 20% (Fig. 3B). These results confirmed the high reproducibility of MS data acquired by the PulseDIA method. To compare the quantified result of the same peptides and proteins in the four DIA software, we calculated the Spearman correlation of the overlapped 30,289 peptides and 4,493 proteins in four DIA software quantified. At the peptide and protein levels, DIA-NN and Spectronaut exhibited the strongest correlation (r >0.9), with the correlation of any two DIA software as no less than 0.85 (Fig. 3C, D). To check whether the quantified proteome thus acquired classify different drug resistance model, we performed also principal component analysis (PCA) analysis between the cells in ADR- and IMA-resistant cells, and parental K562 cells. The result showed that the two resistant cell lines and parental K562 cells were separated into three clusters as analyzed by the four software tools (Fig. 3E-H).
      Figure thumbnail gr3
      Figure3PulseDIA proteome data quality control (QC) analysis. (A)The box plot showed the Pearson correlation coefficient of 5 technical replicates in four DIA software quantification. (B) The violin plot showed the distribution of the CV of each protein's quantitative values in three biological replicates. Three lines of the black box inside the violin represented lower quartile, median, higher quartile. (C) Spearman correlation coefficient of overlapped 30,289 peptides quantitative values in four DIA software. (D) Spearman correlation coefficient of overlapped 4493 proteotypic proteins quantitative values in four DIA software. (E, F, G, H) PCA plot shows the distribution of the sample in the first two principal component levels. The red, blue, green dots mean samples from model ADR, model IMA, native K562 cells, respectively.

       Dynamic proteomic changes during acquisition of drug resistance

      To minimize the statistical variation for each analytic step, these commonly identified 4,493 proteins were selected and those with less than 25% missing ratio in each derivative resistant cell line were subject to the analysis of variance (ANOVA). Proteins with adjusted p value less than 0.05 were selected for further downstream analysis.
      By fuzzy c-means clustering [
      • Kumar L.
      • Mfuzz E.F.M.
      a software package for soft clustering of microarray data.
      ,
      • Futschik M.E.
      • Carlisle B.
      Noise-robust soft clustering of gene expression time-course data.
      ], we identified four clusters related to the resistance to ADR and IMA (Fig. S1; Fig. 4), amongst which only those that were continuously upregulated and downregulated were further selected and studied (Fig. 4). In addition, we selected 1035 (by Spectronaut), 1273 (by EncyclopeDIA), 1088 (by OpenSWATH), and 2161 (by DIA-NN) proteins related to resistance to ADR, (Table S2; Fig. 4A), and 1662 (EncyclopeDIA), 747 (OpenSWATH), 950 (Spectronaut) and 1027 (DIA-NN) proteins related to resistance to IMA (Table S3; Fig. 4B).
      Figure thumbnail gr4
      Figure 4Protein cluster analysis. The cluster for proteins that were continuously upregulated and downregulated from K562 cells resistant to ADR (A) and IMA (B).The horizontal axis represented the progress of the model (K-native K562 cells, A1/I1‐first phase of model ADR/IMA, A2/I2‐second phase of model ADR/IMA, A3/I3‐third phase of model ADR/IMA). The vertical axis represents proteins expression changes in each cluster.

       Activated sirtuin signaling pathway and abnormal mitochondrial function in drug resistant K562 cell

      The selected proteins from the four DIA analytic tools as described above in resistance to ADR and IMA were further analyzed by IPA (Table S4, S5), with the top three pathways (sorted by p value) as listed in Figure 5A and 5B. Notably, EIF2 signaling and sirtuin signaling pathway were enriched from as least three DIA tools related to ADR resistance. In contrast, sirtuin signaling pathway and oxidative phosphorylation were frequently enriched related to IMA resitance (Fig. 5A, B). These data indicate that the sirtuin signaling pathway was significantly enriched in both types of drug resistance with a positive activation mode (Z-score>0, Table 2).
      Figure thumbnail gr5
      Figure 5Enriched proteins that exhibit continuous changes as identified by IPA in ADR- and IMA-resistant cells. (A) In ADR-resistant cells, the most significantly changed 3 pathways were enriched by IPA from four DIA software. (B) In IMA-resistant cells, the most significantly changed 3 pathways were enriched by IPA from four DIA software. (C) The heatmap showed the proteins (adjusted p value<0.05 calculated by ANOVA) involved in Sirtuin signaling pathway from four DIA software expression in IMA- and ADR-resistant cells, with the overlapped quantified proteins being highlighted in red. Each row indicated a protein, and each column indicated a sample. The protein intensity matrix was normalized by Z-score and colored in the heatmap. (D) Venn diagram showed the 16 overlapped proteins involved in Sirtuin Signaling Pathway in ADR-resistant cells. (E) Venn diagram showed the 5 overlapped proteins involved in Sirtuin Signaling Pathway in IMA-resistant cells.
      Table 2The enrichment statistics of Sirtuin Signaling Pathway in four DIA software.
      SoftwareP value
      P-value is the result of Fisher’s exact test.
      Ratio
      Ratio is the number of proteins from different DIA software that maps to the pathway divided by the total number of proteins that map to the same pathway.
      Z-score
      Z-score is calculated by the IPA z-score algorithm to predict the direction of change for the function. An absolute z-score of ≥ 2 is considered significant.
      Model
      OpenSWATH4.480.08251.155Model IMA
      Spectronaut13.70.1552.414Model IMA
      DIA-NN8.970.1342.449Model IMA
      EncyclopeDIA20.60.2513.202Model IMA
      OpenSWATH9.870.1442.4Model ADR
      Spectronaut6.80.121Model ADR
      DIA-NN21.80.2961.336Model ADR
      EncyclopeDIA9.310.1553.138Model ADR
      a P-value is the result of Fisher’s exact test.
      b Ratio is the number of proteins from different DIA software that maps to the pathway divided by the total number of proteins that map to the same pathway.
      c Z-score is calculated by the IPA z-score algorithm to predict the direction of change for the function. An absolute z-score of ≥ 2 is considered significant.
      We then focused on the proteins involved in the sirtuin signaling pathway, whose expression levels were shown in a heatmap (Fig. 5C). As a result, 16 proteins, which were involeved in sirtuin signaling pathway, were quantified between all four DIA analytic tools in model ADR (Fig. 5D), and 5 proteins in model IMA (Fig. 5E). The overlapped proteins between all four DIA analytic tools showed the same expression pattern in the analysis results of four DIA software (Fig. 5C). Among the overlapped proteins, 13 out 16 related to ADR resistance were involved in maintaining mitochondrial function, and three out of five related to IMA resistance participated in the formation of the mitochondrial respiratory chain (Fig. 6A).
      Figure thumbnail gr6
      Figure 6Proteins involevd in mitochondrial function. (A) Schemathic diagram of TCA cycle and electron respiratory chain in mitochondria. (B) Normalized protein expression in ADR- and IMA-resistant cells from four DIA software.
      NDUFB1, NDUFB6, NDUFA4, and NDUFA6 are the subunits of NADH-ubiquinone oxidoreductase (Complex I) of the mitochondrial respiratory chain (Fig. 6A) [

      Brandt and Ulrich, Energy Converting NADH: Quinone Oxidoreductase (Complex I). Annual Review of Biochemistry. 75(1): p. 69-92.

      ]. UQCRFS1 is a subunit of ubiquinol-cytochrome c oxidoreductase (complex III) (Fig. 6A). The main physiological function of Complex I and Complex III is to oxidize NADH, transfer electrons to ubiquinone and from ubiquinone to cytochrome C, and carry out the next electron transfer in the mitochondrial respiratory chain (Fig. 5A) [
      • Sarewicz M.
      • Osyczka A.
      Electronic connection between the quinone and cytochrome C redox pools and its role in regulation of mitochondrial electron transport and redox signaling.
      ]. In addition, Complex 1and Complex III also regulate mitochondria production of reactive oxygen species (ROS) [
      • Sharma L.K.
      • Lu J.
      • Bai Y.
      Mitochondrial respiratory complex I: structure, function and implication in human diseases.
      ]. Our results showed decreased expression levels of these Complex I and III components (Fig. 6A, B), suggesting low oxidative stress levels in drug-resistant cells.
      We also identified upregulation of succinate dehydrogenase complex iron sulfur subunit B (SDHB), which participates in the electron transfer process of succinate dehydrogenase (Complex II) from succinate to ubiquinone (Fig. 6A, B), and when mutated, is closely related to pheochromocytoma [
      • Astuti D.
      • et al.
      Gene mutations in the succinate dehydrogenase subunit SDHB cause susceptibility to familial pheochromocytoma and to familial paraganglioma.
      ].
      Voltage dependent anion channel proteins (VDACs) are the most abundant protein [
      • Camara A.K.S.
      • et al.
      Mitochondrial VDAC1: A Key Gatekeeper as Potential Therapeutic Target.
      ] on the outer mitochondrial membrane (Fig. 6A), and function to maintain free diffusion of small molecules inside and outside the mitochondrial membrane [
      • Mathupala S.P.
      • Pedersen P.L.
      Voltage dependent anion channel-1 (VDAC-1) as an anti-cancer target.
      ]. In tumor cells, the interaction between VDAC and hexokinase inhibits apoptosis [
      • Pastorino J.G.
      • Hoek J.B.
      Regulation of hexokinase binding to VDAC.
      ], and therefore targeting both molecules could potentially offer improved anti-tumor benefits. Our results showed that the three VDAC isoforms, namely VDAC1, VDAC2, and VDAC3, were significantly downregulated upon acquision of ADR resistance (Fig. 6C), suggesting that targeting VDAC may not be an effective strategy in drug-resistant tumor cells.
      The translocase of the outer mitochondrial membrane complex proteins (TOMMs) regulates entry of mitochondrial protein precursors into the mitochondria cytoplasm (Fig. 6A) [
      • Sokol A.M.
      • et al.
      Mitochondrial protein translocases for survival and wellbeing.
      ]. Our data showed that the subunits of TOMMs, TOMM40, TOMM5, TOMM6, TOMM22, and TOMM20, were significantly downregulated in cells resistant to ADR or IMA (Fig. 5C; Fig. 6A, D). In contrast, we found increased expression of TOMM34 upon acquision of drug resistance (Fig. 5D). Located in the cytoplasm, TOMM34 is known to interact with HSP70 and HSP90, and regulate the activity of ATPase [
      • Durech M.
      • et al.
      Novel Entropically Driven Conformation-specific Interactions with Tomm34 Protein Modulate Hsp70 Protein Folding and ATPase Activities.
      ]. High expression of TOMM34 has been found in colon cancer[
      • Shimokawa T.
      • et al.
      Identification of TOMM34, which shows elevated expression in the majority of human colon cancers, as a novel drug target.
      ], breast cancer[
      • Muller P.
      • et al.
      Tomm34 is commonly expressed in epithelial ovarian cancer and associates with tumour type and high FIGO stage.
      ], ovarian cancer [
      • Aleskandarany M.A.
      • et al.
      TOMM34 expression in early invasive breast cancer: a biomarker associated with poor outcome.
      ].
      Glutaminase (GLS) can convert glutamine into glutamic acid, which constitutes the major source for α-ketoglutarate (α-KG) production(Fig. 6A) [
      • Matés J.M.
      • et al.
      Glutaminase isoenzymes as key regulators in metabolic and oxidative stress against cancer.
      ,
      • Wu N.
      • et al.
      Alpha-Ketoglutarate: Physiological Functions and Applications.
      ], which promotes cell differentiation through dioxygenases [
      • Dang L.
      • Yen K.
      • Attar E.C.
      IDH mutations in cancer and progress toward development of targeted therapeutics.
      ]. Our results showed significantly down-regulated expression of GLS when K562 cells became resistant to ADR (Fig. 6E, Right panel).
      Isocitrate Dehydrogenase (NADP(+)) 2 (IDH2) catalyzes the oxidative decarboxylation of isocitrate into α-ketoglutarate (α-KG ) in tricarboxylic acid cycle (TCA cycle), and NADPH was synchronously produced at the biochemical process [
      • Reitman Z.J.
      • Yan H.
      Isocitrate dehydrogenase 1 and 2 mutations in cancer: alterations at a crossroads of cellular metabolism.
      ] (Fig. 6A). NADPH is essential in protecting cells from oxidative damage[
      • Biaglow J.E.
      • Miller R.A.
      The thioredoxin reductase/thioredoxin system: novel redox targets for cancer therapy.
      ]. We found the IDH2 abundance increased upon resistance to both ADR and IMA (Fig. 6E, Left panel), indicating that IDH2 overexpression may promote cellular resistance against high doses of both therapeutic drugs.
      Taken together, our results showed significantly changed abundance of proteins involved in mitochondria function that likely mediate survival of drug-resistant cells, which may be therapeutically targeted to enhance drug sensitivity and response in tumor cells.

       IDH2 is a potential target for reversing drug resistance in K562 cells

      As discussed above, among the dysregulated proteins, IDH2 was upregulated in K562 cells resistant to both ADR and IMA. To validate its biological function, we utilized a selective inhibitor of IDH2 [
      • Quek L.
      • et al.
      Clonal heterogeneity of acute myeloid leukemia treated with the IDH2 inhibitor enasidenib.
      ], AGI-6780, and treated sensitive or resistant K562 cells with ADR and IMA alone or combined with AGI-6780, followed by monitoring cell survival with a cytotoxicity assay. Our results showed that, compared with ADR and IMA treatment alone, combination with AGI-6780 did not affect the survival in the parental sensitive K562 cells (Fig. 7A, B; Table 3). However, in resistant K562 cells, the sensitivity to ADR and IMA was significantly increased (IMA+AGI-6780, IC50=0.53μM; ADR+AGI-6780, IC50=0.29μM), when combined with AGI-6780. We further show in IMA-resistant cells the reversal index of 2 μM and 4 μM AGI-6780 was 1.92 and 2.94, respectively (Fig. 7C; Table 3), whereas in ADR-resistant cells, the reversal index of 2 μM and 4 μM AGI-6780 was 1.60 and 2.74, respectively (Fig. 7D; Table 3). To further confirm the synergistic sensitization effect of AGI-6780 in combination with imatinib or adriamycin on drug resistance models, we performed Bliss independence model using Combenefit software (version:2.021) [
      • Di Veroli G.Y.
      • et al.
      Combenefit: an interactive platform for the analysis and visualization of drug combinations.
      ]. The surface plots showed the dose-response results of AGI-6780 combination with imatinib or adriamycin (Fig. 7E-H). As shown in the surface plot, in the parental K562 cell line, the combination of AGI-6780 with imatinib (Fig. 7E) or adriamycin (Fig. 7F) did not produce a significant synergistic effect. In the imatinib resistance model, the combination of 0.03-0.25 μM imatinib and 4 μM AGI-6780 produced a significant synergistic effect (Fig. 7G). In the adriamycin resistance model (Fig. 7H), the combination of 0.03-0.125 μM adriamycin and 2-4 μM AGI-6780 produced a synergistic effect, although this synergy is relatively weak.
      Figure thumbnail gr7
      Figure 7IDH2 enhanced the sensitivity to ADR and IMA in K562 cells. (A) and (C) The parental sensitive and IMA-resisant K562 cells were treated with various concentrations (4, 2, 1, 0.5……0 μM) IMA alone or combined with 4 μM or 2 μM AGI-6780 for 48h. (B) and (D) The parental sensitive and ADR-resisant K562 cells were treated with various concentrations (4, 2, 1, 0.5……0 μM) ADR alone or combined with 4 μM or 2 μM AGI-6780 for 48h. The cell survival rate was calculated and plotted in each group. The downward shift of the survival curves indicated suppressed proliferation. Columns are expressed as mean ± SD. (E-H) Surface plots show the synergistic or antagonistic effects from the combination of two drugs. The parental K562 treated with AGI-6780 (0, 2, 4 μM) combined with imatinib (0-4 μM) (E) or adriamycin (0-4 μM) (F). (G) The Imatinib resistance model cell treated with AGI-6780 (0, 2, 4 μM) combined with imatinib (0-4 μM). (H) The Adriamycin resistance model cell treated with AGI-6780 (0, 2, 4 μM) combined with adriamycin (0-4 μM). Each point represents the mean of two independent cytotoxicity assay results. Plots were generated using Combenefit by applying the Bliss independence model.
      Table 3IC50 values and reverse index of the derivative resistant K562 cells
      DrugsIC50 (μM, 95%CI)
      IC50 values represent the best-fit values and 95% CI of three independent experiments were performed.
      Resistance Index
      Resistance Index was calculated by dividing the best-fit IC50 values of parental K562 cells to ADR or IMA by the best-fit IC50 values of the model cells.
      Reversal Index
      Reversal Index was calculated by dividing the Resistance Index in the absence of ACI-6780 by the Resistance Index in the presence of AGI-6780.
      Native K562Model IMAModel ADR
      IMA0.26 (0.19 to 0.36)1.55 (1.12 to 2.19)/5.961
      IMA+AGI-6780(2 μM)0.35 (0.26 to 0.48)1.09 (0.80 to 1.49)/3.111.92
      IMA+AGI-6780(4 μM)0.26 (0.20 to 0.33)0.53 (0.27 to 1.16)/2.032.94
      adramycin0.37 (0.31 to 0.45)/0.67 (0.49 to 0.95)1.811
      adramycin+AGI-6780(2 μM)0.38 (0.31 to 0.46)/0.43 (0.27 to 0.68)1.131.60
      adramycin+AGI-6780(4 μM)0.44 (0.31 to 0.61)/0.29 (0.18 to 0.47)0.662.74
      a IC50 values represent the best-fit values and 95% CI of three independent experiments were performed.
      b Resistance Index was calculated by dividing the best-fit IC50 values of parental K562 cells to ADR or IMA by the best-fit IC50 values of the model cells.
      c Reversal Index was calculated by dividing the Resistance Index in the absence of ACI-6780 by the Resistance Index in the presence of AGI-6780.
      Taken together, the cytotoxicity assay confirmed the ability of AGI-6780 to reverse drug resistance in vitro, implicating IDH2 as a potential target to improve drug sensitibity and clinical responses in patients.

      Discussion

      Two major paradigms have been widely proposed to elucidate the underlying mechanism of drug resistance. One is based on the principle of Darwinian selection, which posits that a fraction of tumor cells is inherently more tolerant and thus become enriched after drug treatment, thereby gradually developing, and exhibiting a drug-resistant phenotype. The other mechanism, which is called Lamarckian induction, proposes that drug treatment induces a resistance phenotype in tumor cells [
      • Pisco A.O.
      • Huang S.
      Non-genetic cancer cell plasticity and therapy-induced stemness in tumour relapse: 'What does not kill me strengthens me'.
      ]. In our resistant cell line model, our focused was on the proteins that showed continuous and consistent changes upon acquisition drug resistance. Here, we uncovered potential association between multiple proteins and drug responsiveness in K562 cells. These proteins serve as potential therapeutic targets for combating drug resistance.
      While establishing the resistance cell lines, we also constantly monitored drug sensitivity in the parental sensitive K562 cells (Fig. 1B, C). Different from other similar studies [
      • Piskareva O.
      • et al.
      The development of cisplatin resistance in neuroblastoma is accompanied by epithelial to mesenchymal transition in vitro.
      ], we did not consider morphology changes as a parameter, primarily because K562 cells are suspension cells in the medium in a spherical shape while the drug resistant phenotype was established. After K562 cells with mild, intermediate and severe resistance to ADR and IMA were developed and collected, we performed PCT-based peptides extraction. The efficient and prompt sample preparation under high pressure (lysis in 45K psi; digest in 20K psi) by PCT ensures that the process of peptides extraction is reproducible and stable [
      • Shao S.
      • et al.
      Minimal sample requirement for highly multiplexed protein quantification in cell lines and tissues by PCT-SWATH mass spectrometry.
      ].
      The resistant model K562 cells and parental K526 cells were acquired by PulseDIA [
      • Cai X.
      • et al.
      PulseDIA: Data-Independent Acquisition Mass Spectrometry Using Multi-Injection Pulsed Gas-Phase Fractionation.
      ]. Due to the inherent complexity of DIA data, multiple computational software tools have been developed to analyze DIA data, among which Spectronaut [
      • Bruderer R.
      • et al.
      Extending the Limits of Quantitative Proteome Profiling with Data-Independent Acquisition and Application to Acetaminophen-Treated Three-Dimensional Liver Microtissues.
      ], DIA-NN [
      • Demichev V.
      • et al.
      DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput.
      ], EncyclopeDIA [
      • Searle B.C.
      • et al.
      Chromatogram libraries improve peptide detection and quantification by data independent acquisition mass spectrometry.
      ], and OpenSWATH [
      • Rost H.L.
      • et al.
      OpenSWATH enables automated, targeted analysis of data-independent acquisition MS data.
      ] are all broadly used. These software tools identify and quantify peptides and proteins by various different algorithms with some commonality and difference. The exact algorithm of commercial software Spectronaut is not transparent to users. However, all software tools meet the rigorous statistical criteria of false discovery rate. Pedro Navarro et al. [
      • Navarro P.
      • et al.
      A multicenter study benchmarks software tools for label-free proteome quantification.
      ] reported the quantified proteins of several DIA analysis software (OpenSWATH, SWATH2.0, Skyline, Spectronaut, DIA Umpire) for the same SWATH-MS proteomics data, with the number of proteins quantified in the library-based parsing algorithm ranging from 3673 to 4692, and shared 3064 proteins in common. In our study, the number of proteins quantified from the four DIA tools varied from 4926 to 6935 for the same set of proteomics raw data, with 4493 shared proteins (Fig. 2D). Therefore, we focused on the 4493 shared proteins for a more unbiased discovery of potential candidates. This approach is a useful way to further narrow down the candidates for functional validation. While most DIA studies used only one software tool for data interpretation and prioritization of proteins of interest, our study presents a more rigorous approach for protein selection. Nevertheless, this approach may miss certain useful candidate, and may not be the best option for all.
      Our results showed that the coefficient of variation of protein intensity in three biological replicates was around 20% (Fig. 3B). For PulseDIA based mass spectrometry raw data acquisition, we identified and quantified 83.1% (7082/8524) proteotypic proteins of the DDA library from 26 samples by the four DIA datasets, and 63.4% (4493/7082) proteotypic proteins were detected by all four tools (Table. S1; Fig. 2D, E). The PulseDIA data stability was assessed by the Pearson correlation of two technical replicates, which was no less than 0.9 (Fig. 3A). The PCA result demonstrated that the resistant K562 cells can be discriminated from the parental cells at the whole proteome level (Fig. 2E-H). These high-quality mass spectrometry data provides a key basis for our subsequent data mining analysis.
      We independently analyzed 4493 proteins identified by all four tools. Based on cluster analysis and ANOVA (p≤0.05), we selected proteins that exhibited continuous changes during development of IMA and ADR resistance (Table S2, S3; Fig. 4), followed by IPA analysis. Pathway analysis revealed common characteristics associated with drug resistance. Significant changed stress signaling pathways including Oxidative Phosphorylation and the Sirtuin Signaling Pathway (Fig. 5A, B) are known to enhance the resistance and plasticity of tumor cells[
      • Vellinga T.T.
      • et al.
      SIRT1/PGC1α-Dependent Increase in Oxidative Phosphorylation Supports Chemotherapy Resistance of Colon Cancer.
      ,
      • Verdin E.
      • et al.
      Sirtuin regulation of mitochondria: energy production, apoptosis, and signaling.
      ,
      • Alexa-Stratulat T.
      • et al.
      What sustains the multidrug resistance phenotype beyond ABC efflux transporters? Looking beyond the tip of the iceberg.
      ]. We then decided to focus on the identified proteins involved in the Sirtuin Signaling Pathway as identified by all four DIA tools, including IDH2, NDUFB1, NDUFB6, NDUFA4, NDUFA6, SHDB, and GLS, which are involved in mitochondrial ATP generation (Fig. 5D, E; Fig. 6A)[
      • Brandt U.
      Energy converting NADH:quinone oxidoreductase (complex I).
      ,
      • Parker S.J.
      • Metallo C.M.
      Metabolic consequences of oncogenic IDH mutations.
      ]. These results suggest that inhibiting ATP production and blocking P-gp energy supply could potentially enable reversal of drug resistance [
      • Liu W.
      • et al.
      Targeting P-Glycoprotein: Nelfinavir Reverses Adriamycin Resistance in K562/ADR Cells.
      ,
      • Hedley D.
      • et al.
      A novel energy dependent mechanism reducing daunorubicin accumulation in acute myeloid leukemia.
      ].
      IDH2 mutation catalyzes D-2-hydroxyglutarate (2-HG) production, which leads to competitive inhibition of α-KG-dependent DNA demethylases and consequently promotes tumorigenesis [
      • Dang L.
      • Yen K.
      • Attar E.C.
      IDH mutations in cancer and progress toward development of targeted therapeutics.
      ]. Preclinical experiments have shown that IDH2 inhibitors promote leukemic cell differentiation in 40% of the patients with relapsed/refractory AML[
      • Quek L.
      • et al.
      Clonal heterogeneity of acute myeloid leukemia treated with the IDH2 inhibitor enasidenib.
      ]. Here, we used the selective IDH2 inhibitor AGI-6780, and demonstrated IDH2 as a potential target for reversing drug resistance in tumor cells. Recent research has shown that 5 μM AGI-6780 selectively impaired wild type IDH2 enzymatic activity in multiple myeloma cells in vitro[
      • Matés J.M.
      • et al.
      Glutaminase isoenzymes as key regulators in metabolic and oxidative stress against cancer.
      ]. In contrast, our data showed no effects on cell proliferation even 48 h after 4 μM AGI-6780 treatment in either sensitive or resistant K562 cells (Fig. S2). As a potential therapeutic avenue for reversing drug resistance, IDH2 inhibitors are specifically efficacious in resistant tumors cells.
      Derivative cell lines that have developed drug resistance could provide an important and informative model to help elucidate the underlying mechanism that can be strategically leveraged to manipulate and improve the sensitivity of tumor cells to therapeutic treatment. Single-cell proteomics studies of melanoma-derived cells with different levels of drug sensitivity revealed changes in intracellular signals before drug resistance has developed [
      • Su Y.
      • et al.
      Single-cell analysis resolves the cell state transition and signaling dynamics associated with melanoma drug-induced resistance.
      ]. By constructing a cisplatin-resistant neuroblastoma cell line, Olga et al. characterized the epithelial to mesenchymal transition during the development of drug resistance [
      • Piskareva O.
      • et al.
      The development of cisplatin resistance in neuroblastoma is accompanied by epithelial to mesenchymal transition in vitro.
      ]. Herein, we have developed multiple K562 cell lines with different degrees of resistance to ADR and IMA treatment, which could be used for preclinical investigation of the molecular and cellular mechanisms that underlie the therapeutic resistance in CML patients. We further identified and characterized IDH2 as a potentially useful target that can be utilized to enhance tumor cell sensitivity to targeted treatment.

      Data availability

      All data are available in the manuscript or the supplementary material. The proteomics data are deposited in ProteomeXchange Consortium (https://www.iprox.cn/page/SSV024.html;url=1631617793214PbAJ, password: 2ztF). All the data will be publicly released upon publication. The project data analysis codes are deposited in GitHub (https://github.com/guomics-lab/KMDR).

      Declaration of interests

      T.G. and Y.Z. are shareholders of Westlake Omics Inc. W.G. is an employee of Westlake Omics Inc. The remaining authors declare no competing interests.

      Acknowledgments

      This work is supported by grants from National Key R&D Program of China (No. 2020YFE0202200), the National Natural Science Foundation of China (81874324, 81972492, 21904107, 81473280, U1608283), Zhejiang Provincial Natural Science Foundation for Distinguished Young Scholars (LR19C050001), Hangzhou Agriculture and Society Advancement Program (20190101A04), and Westlake Education Foundation. We thank Westlake University Supercomputer Center for assistance in data storage and computation.

      Supplementary data

      References

        • Radujkovic A.
        • et al.
        Donor Lymphocyte Infusions for Chronic Myeloid Leukemia Relapsing after Allogeneic Stem Cell Transplantation: May We Predict Graft-versus-Leukemia Without Graft-versus-Host Disease?.
        Biol Blood Marrow Transplant. 2015; 21: 1230-1236
        • Radich J.P.
        • et al.
        Chronic Myeloid Leukemia, Version 1.2019, NCCN Clinical Practice Guidelines in Oncology.
        J Natl Compr Canc Netw. 2018; 16: 1108-1135
        • Pal S.K.
        • et al.
        Clinical Cancer Advances 2019: Annual Report on Progress Against Cancer From the American Society of Clinical Oncology.
        Journal of Clinical Oncology. 2019; 37: 834-849
        • Mansoori B.
        • et al.
        The Different Mechanisms of Cancer Drug Resistance: A Brief Review.
        Advanced pharmaceutical bulletin. 2017; 7: 339-348
        • Liu W.
        • et al.
        Targeting P-Glycoprotein: Nelfinavir Reverses Adriamycin Resistance in K562/ADR Cells.
        Cell Physiol Biochem. 2018; 51: 1616-1631
        • Sun Y.
        • et al.
        Targeting P-glycoprotein and SORCIN: Dihydromyricetin strengthens anti-proliferative efficiency of adriamycin via MAPK/ERK and Ca(2+) -mediated apoptosis pathways in MCF-7/ADR and K562/ADR.
        J Cell Physiol. 2018; 233: 3066-3079
        • De Oliveira F.
        • et al.
        Effects of permeability transition inhibition and decrease in cytochrome c content on doxorubicin toxicity in K562 cells.
        Oncogene. 2006; 25: 2646-2655
        • Li R.J.
        • et al.
        Down-regulation of mitochondrial ATPase by hypermethylation mechanism in chronic myeloid leukemia is associated with multidrug resistance.
        Ann Oncol. 2010; 21: 1506-1514
        • Zhao L.
        • et al.
        Functional screen analysis reveals miR-3142 as central regulator in chemoresistance and proliferation through activation of the PTEN-AKT pathway in CML.
        Cell Death Dis. 2017; 8: e2830
        • Jiang L.
        • et al.
        Ivermectin reverses the drug resistance in cancer cells through EGFR/ERK/Akt/NF-κB pathway.
        J Exp Clin Cancer Res. 2019; 38: 265
        • Dong J.
        • et al.
        Medicinal chemistry strategies to discover P-glycoprotein inhibitors: An update.
        Drug Resist Updat. 2020; 49: 100681
        • Holyoake T.L.
        • Helgason G.V.
        Do we need more drugs for chronic myeloid leukemia?.
        Immunol Rev. 2015; 263: 106-123
        • Buchdunger E.
        • O'Reilley T.
        • Wood J.
        Pharmacology of imatinib (STI571).
        European journal of cancer. 2002; 38: S28-S36
        • Druker B.J.
        • et al.
        Five-year follow-up of patients receiving imatinib for chronic myeloid leukemia.
        New England Journal of Medicine. 2006; 355: 2408-2417
        • Milojkovic D.
        • Apperley J.
        Mechanisms of Resistance to Imatinib and Second-Generation Tyrosine Inhibitors in Chronic Myeloid Leukemia.
        Clin Cancer Res. 2009; 15: 7519-7527
        • Burchert A.
        Roots of imatinib resistance: a question of self-renewal?.
        Drug Resist Updat. 2007; 10: 152-161
        • Weisberg E.
        • Griffin J.D.
        Resistance to imatinib (Glivec): update on clinical mechanisms.
        Drug Resist Updat. 2003; 6: 231-238
        • Bellodi C.
        • et al.
        Targeting autophagy potentiates tyrosine kinase inhibitor-induced cell death in Philadelphia chromosome-positive cells, including primary CML stem cells.
        J Clin Invest. 2009; 119: 1109-1123
        • Cree I.A.
        • Charlton P.
        Molecular chess? Hallmarks of anti-cancer drug resistance.
        BMC Cancer. 2017; 17: 10
        • Aleksakhina S.N.
        • Kashyap A.
        • Imyanitov E.N.
        Mechanisms of acquired tumor drug resistance.
        Biochim Biophys Acta Rev Cancer. 2019; 1872: 188310
        • Aebersold R.
        • Mann M.
        Mass-spectrometric exploration of proteome structure and function.
        Nature. 2016; 537: 347-355
        • Monteleone F.
        • et al.
        SWATH-MS based quantitative proteomics analysis reveals that curcumin alters the metabolic enzyme profile of CML cells by affecting the activity of miR-22/IPO7/HIF-1α axis.
        J Exp Clin Cancer Res. 2018; 37: 170
        • Xiong L.
        • et al.
        Global proteome quantification for discovering imatinib-induced perturbation of multiple biological pathways in K562 human chronic myeloid leukemia cells.
        J Proteome Res. 2010; 9: 6007-6015
        • Corrêa S.
        • et al.
        A comparative proteomic study identified LRPPRC and MCM7 as putative actors in imatinib mesylate cross-resistance in Lucena cell line.
        Proteome Sci. 2012; 10: 23
        • Ferrari G.
        • et al.
        Comparative proteomic analysis of chronic myelogenous leukemia cells: inside the mechanism of imatinib resistance.
        J Proteome Res. 2007; 6: 367-375
        • Gjertsen B.T.
        • Wiig H.
        Investigation of therapy resistance mechanisms in myeloid leukemia by protein profiling of bone marrow extracellular fluid.
        Expert Rev Proteomics. 2012; 9: 595-598
        • Hrdinova T.
        • et al.
        Exosomes released by imatinib-resistant K562 cells contain specific membrane markers, IFITM3, CD146 and CD36 and increase the survival of imatinib-sensitive cells in the presence of imatinib.
        Int J Oncol. 2021; 58: 238-250
        • Gillet L.C.
        • et al.
        Targeted data extraction of the MS/MS spectra generated by data-independent acquisition: a new concept for consistent and accurate proteome analysis.
        Mol Cell Proteomics. 2012; 11 (O111 016717)
        • Collins B.C.
        • et al.
        Multi-laboratory assessment of reproducibility, qualitative and quantitative performance of SWATH-mass spectrometry.
        Nature communications. 2017; 8: 1-12
        • Cai X.
        • et al.
        PulseDIA: Data-Independent Acquisition Mass Spectrometry Using Multi-Injection Pulsed Gas-Phase Fractionation.
        Journal of Proteome Research. 2020;
        • Guo T.
        • et al.
        Rapid mass spectrometric conversion of tissue biopsy samples into permanent quantitative digital proteome maps.
        Nat Med. 2015; 21: 407-413
        • Shao S.
        • et al.
        Minimal sample requirement for highly multiplexed protein quantification in cell lines and tissues by PCT-SWATH mass spectrometry.
        Proteomics. 2015; 15: 3711-3721
        • Zhang F.
        • et al.
        Data-Independent Acquisition Mass Spectrometry-Based Proteomics and Software Tools: A Glimpse in 2020.
        PROTEOMICS. 2020; 20: 1900276
        • Bruderer R.
        • et al.
        Extending the Limits of Quantitative Proteome Profiling with Data-Independent Acquisition and Application to Acetaminophen-Treated Three-Dimensional Liver Microtissues.
        Molecular & Cellular Proteomics. 2015; 14: 1400-1410
        • Demichev V.
        • et al.
        DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput.
        Nat Methods. 2020; 17: 41-44
        • Searle B.C.
        • et al.
        Chromatogram libraries improve peptide detection and quantification by data independent acquisition mass spectrometry.
        Nat Commun. 2018; 9: 5128
        • Rost H.L.
        • et al.
        OpenSWATH enables automated, targeted analysis of data-independent acquisition MS data.
        Nat Biotechnol. 2014; 32: 219-223
        • Cai X.
        • et al.
        PulseDIA: Data-Independent Acquisition Mass Spectrometry Using Multi-Injection Pulsed Gas-Phase Fractionation.
        Journal of Proteome Research. 2021; 20: 279-288
        • Zhu T.
        • et al.
        DPHL: A DIA Pan-human Protein Mass Spectrometry Library for Robust Biomarker Discovery.
        Genomics, Proteomics & Bioinformatics. 2020; 18: 104-119
        • Zhu T.
        • et al.
        ProteomeExpert: a Docker image-based web server for exploring, modeling, visualizing and mining quantitative proteomic datasets.
        Bioinformatics. 2021; 37: 273-275
        • Gillet L.C.
        • et al.
        Targeted Data Extraction of the MS/MS Spectra Generated by Data-independent Acquisition: A New Concept for Consistent and Accurate Proteome Analysis.
        Molecular & Cellular Proteomics. 2012; 11 (O111.016717)
        • Chi H.
        • et al.
        Comprehensive identification of peptides in tandem mass spectra using an efficient open search engine.
        Nature Biotechnology. 2018; 36: 1059-1061
        • Schubert O.T.
        • et al.
        Building high-quality assay libraries for targeted analysis of SWATH MS data.
        Nat Protoc. 2015; 10: 426-441
        • Di Veroli G.Y.
        • et al.
        Combenefit: an interactive platform for the analysis and visualization of drug combinations.
        Bioinformatics. 2016; 32: 2866-2868
        • Kumar L.
        • Mfuzz E.F.M.
        a software package for soft clustering of microarray data.
        Bioinformation. 2007; 2: 5-7
        • Jimenez-Marin A.
        • et al.
        Biological pathway analysis by ArrayUnlock and Ingenuity Pathway Analysis.
        BMC Proc. 2009; 3: S6
        • Kramer A.
        • et al.
        Causal analysis approaches in Ingenuity Pathway Analysis.
        Bioinformatics. 2014; 30: 523-530
        • Futschik M.E.
        • Carlisle B.
        Noise-robust soft clustering of gene expression time-course data.
        J Bioinform Comput Biol. 2005; 3: 965-988
      1. Brandt and Ulrich, Energy Converting NADH: Quinone Oxidoreductase (Complex I). Annual Review of Biochemistry. 75(1): p. 69-92.

        • Sarewicz M.
        • Osyczka A.
        Electronic connection between the quinone and cytochrome C redox pools and its role in regulation of mitochondrial electron transport and redox signaling.
        Physiological reviews. 2015; 95: 219-243
        • Sharma L.K.
        • Lu J.
        • Bai Y.
        Mitochondrial respiratory complex I: structure, function and implication in human diseases.
        Curr Med Chem. 2009; 16: 1266-1277
        • Astuti D.
        • et al.
        Gene mutations in the succinate dehydrogenase subunit SDHB cause susceptibility to familial pheochromocytoma and to familial paraganglioma.
        Am J Hum Genet. 2001; 69: 49-54
        • Camara A.K.S.
        • et al.
        Mitochondrial VDAC1: A Key Gatekeeper as Potential Therapeutic Target.
        Frontiers in Physiology. 2017; 8
        • Mathupala S.P.
        • Pedersen P.L.
        Voltage dependent anion channel-1 (VDAC-1) as an anti-cancer target.
        Cancer biology & therapy. 2010; 9: 1053-1056
        • Pastorino J.G.
        • Hoek J.B.
        Regulation of hexokinase binding to VDAC.
        Journal of bioenergetics and biomembranes. 2008; 40: 171-182
        • Sokol A.M.
        • et al.
        Mitochondrial protein translocases for survival and wellbeing.
        FEBS Letters. 2014; 588: 2484-2495
        • Durech M.
        • et al.
        Novel Entropically Driven Conformation-specific Interactions with Tomm34 Protein Modulate Hsp70 Protein Folding and ATPase Activities.
        Molecular & Cellular Proteomics. 2016; 15: 1710-1727
        • Shimokawa T.
        • et al.
        Identification of TOMM34, which shows elevated expression in the majority of human colon cancers, as a novel drug target.
        Int J Oncol. 2006; 29: 381-386
        • Muller P.
        • et al.
        Tomm34 is commonly expressed in epithelial ovarian cancer and associates with tumour type and high FIGO stage.
        J Ovarian Res. 2019; 12: 30
        • Aleskandarany M.A.
        • et al.
        TOMM34 expression in early invasive breast cancer: a biomarker associated with poor outcome.
        Breast Cancer Research and Treatment. 2012; 136: 419-427
        • Matés J.M.
        • et al.
        Glutaminase isoenzymes as key regulators in metabolic and oxidative stress against cancer.
        Curr Mol Med. 2013; 13: 514-534
        • Wu N.
        • et al.
        Alpha-Ketoglutarate: Physiological Functions and Applications.
        Biomol Ther (Seoul). 2016; 24: 1-8
        • Dang L.
        • Yen K.
        • Attar E.C.
        IDH mutations in cancer and progress toward development of targeted therapeutics.
        Ann Oncol. 2016; 27: 599-608
        • Reitman Z.J.
        • Yan H.
        Isocitrate dehydrogenase 1 and 2 mutations in cancer: alterations at a crossroads of cellular metabolism.
        J Natl Cancer Inst. 2010; 102: 932-941
        • Biaglow J.E.
        • Miller R.A.
        The thioredoxin reductase/thioredoxin system: novel redox targets for cancer therapy.
        Cancer Biol Ther. 2005; 4: 6-13
        • Quek L.
        • et al.
        Clonal heterogeneity of acute myeloid leukemia treated with the IDH2 inhibitor enasidenib.
        Nature medicine. 2018; 24: 1167
        • Pisco A.O.
        • Huang S.
        Non-genetic cancer cell plasticity and therapy-induced stemness in tumour relapse: 'What does not kill me strengthens me'.
        Br J Cancer. 2015; 112: 1725-1732
        • Piskareva O.
        • et al.
        The development of cisplatin resistance in neuroblastoma is accompanied by epithelial to mesenchymal transition in vitro.
        Cancer Lett. 2015; 364: 142-155
        • Navarro P.
        • et al.
        A multicenter study benchmarks software tools for label-free proteome quantification.
        Nat Biotechnol. 2016; 34: 1130-1136
        • Vellinga T.T.
        • et al.
        SIRT1/PGC1α-Dependent Increase in Oxidative Phosphorylation Supports Chemotherapy Resistance of Colon Cancer.
        Clinical Cancer Research. 2015; 21: 2870-2879
        • Verdin E.
        • et al.
        Sirtuin regulation of mitochondria: energy production, apoptosis, and signaling.
        Trends in biochemical sciences. 2010; 35: 669-675
        • Alexa-Stratulat T.
        • et al.
        What sustains the multidrug resistance phenotype beyond ABC efflux transporters? Looking beyond the tip of the iceberg.
        Drug Resist Updat. 2019; 46: 100643
        • Brandt U.
        Energy converting NADH:quinone oxidoreductase (complex I).
        Annu Rev Biochem. 2006; 75: 69-92
        • Parker S.J.
        • Metallo C.M.
        Metabolic consequences of oncogenic IDH mutations.
        Pharmacology & therapeutics. 2015; 152: 54-62
        • Hedley D.
        • et al.
        A novel energy dependent mechanism reducing daunorubicin accumulation in acute myeloid leukemia.
        Leukemia. 1997; 11: 48-53
        • Su Y.
        • et al.
        Single-cell analysis resolves the cell state transition and signaling dynamics associated with melanoma drug-induced resistance.
        Proc Natl Acad Sci U S A. 2017; 114: 13679-13684