Cell lines culturing and treatment

Human PATC cell lines (PATC53, PATC66, PATC108, PATC118, PATC124, and PATC148) were generated from pancreatic duct adenocarcinoma by patient-derived xenograft (PDX) models as previously described9. Cells were maintained and treated in DMEM/low glucose (Cytiva HyClone) medium supplemented with 10% heat-inactivated fetal bovine serum (FBS, Gibco) and 1% Penicillin/Streptomycin (Cytiva HyClone) antibiotics. All PATC cell lines were cultured in standard incubation conditions (5% CO 2 , 37 °C) and treated with compounds at early passages (within 20 passages).

Complex I inhibitor IACS-010759 was synthesized and kindly provided by the Institute for Applied Cancer Science (IACS), the University of Texas MD Anderson Cancer Center (MDACC)13. Metformin hydrochloride (#PHR1084), phenformin hydrochloride (#PHR1573), oligomycin A (#75351), bovine serum albumin fraction V (#03117057001), oleic acid-albumin from bovine serum (O3008), and stearic acid (FA 18:0, #85679) were purchased from Sigma-Aldrich. Antimycin A (#ab141904) was ordered from Abcam. MitoPQ (#18808), MitoQ (#89950), oleic acid (FA 18:1, #24659), α-linolenic acid (FA 18:3, #21910), arachidonic acid (FA 20:4, #90010) were obtained from Cayman Chemical Company. Ether phospholipids 1-O-hexadecyl-2-oleoyl-sn-glycero-3-phosphocholine (O-C16-18:1 PC, #878112), 1-O-hexadecyl-2-(8Z,11Z,14Z-eicosatrienoyl)-sn-glycero-3-phosphocholine (O-C16-20:3 PC, #878122), and 1-O-hexadecyl-2-butyryl-sn-glycero-3-phosphocholine (O-C16-04:0 PC, #878115) were provided by Avanti Polar Lipids.

Generation of CRISPR/Cas9 knockout cell lines

The sgRNAs targeting SOD2 / GNPAT / AGPS / FAR1/ UQCC3 were designed using CRISPick website software (https://portals.broadinstitute.org/gppx/crispick/public) supported by BROAD Institute. The sequences are SOD2, ACAAACCTCAGCCCTAACGG; GNPAT, ATGGCTAAAAGGCTTAACCC; AGPS, TTTGTTCAAATACGGTCAGT; FAR1, AGCACTAATCCTTTCCACTG; UQCC3, #1 CTCCCGGGGTCACGATAACG, #2 GCGGAAGCAGGAAATGCTAA. The sgRNAs were individually inserted into LentiCRISPR v2 one vector system (Addgene, Plasmid #52961) and transfected in HEK-293 cells (CRL-1573TM) for 48 hr. The culturing medium was isolated and filtered (0.22 μm pore size, MilliporeSigma™ SCGP00525). After ultracentrifuge at 25,000 g for 2 hr, virus pellets were suspended by PBS and infected into PATC lines. The stable CRISPR/Cas9 knockout PATC lines targeting indicated genes were generated after puromycin (1 mg/mL, Invivogen, # ant-pr-1) selection. Western blotting was detected for the loss of proteins.

Cell sensitivity studies

Cell death assay

A total of 3 × 105 cells per well were seeded in 6-well plates and treated with indicated concentrations of compounds for specific durations. Cells were harvested, stained with propidium iodide (PI, BD-Pharminogen#51-66211E), and detected within 1 hr of staining by using a Gallios flow cytometer (Beckman Coulter, Inc. Brea, CA) and following the manufacturer’s instructions. Data were analyzed by Kaluza and Flowjo (Version 10.6.2) software.

Crystal violet stain

A total of 1 × 105 cells were plated and treated with 10 nM IACS-010759 or with vehicle for 5 days. Cells were fixed and stained with a 1% crystal violet solution that included 20% methanol. After 1 hr of staining at room temperature, cells were thoroughly washed with PBS, dried, and scanned with Epson Perfection V700 Photo.

Bright field images

A total of 3 × 105 cells per well were cultured in 6-well plates and treated with indicated concentrations of compounds for specific durations. Cells were captured and analyzed by Olympus microscopy.

Mitochondrial oxygen consumption rate (OCR) detection

A total of 5 × 103 cells were seeded with 80 µL Dulbecco’s Modified Eagle Medium (DMEM) in 96-well Seahorse XFe96 V3 PS Cell culture microplates (#101085-004) and treated with indicated concentrations of IACS-010759 or vehicle for 24 hr. Cells were incubated in Seahorse XF base medium supplemented with 5.5 mM D-glucose, 2 mM glutamine, and 1 mM pyruvate. Oxygen consumption rate (OCR) was then measured by Seahorse XFe96 Analyzer (Agilent). Seahorse OCR analysis for PATC cell lines were performed according to Seahorse Biosciences instructions from Seahorse XF Cell Mito Stress Test Kit (#103015-100). Oligomycin (0.5 μM), FCCP (optimized based on cell lines, 0.125-0.25 μM), and rotenone/antimycin A (0.5 μM) were injected into ports and measured by XFe96 sensor cartridges (#102416-100) consecutively. Data were normalized by Hoechst stain-positive cell numbers obtained by the Operetta CLS High-Content Analysis System (PerkinElmer).

Metabolic stress assays

Mitochondrial reactive oxygen species (ROS) and lipid peroxidation detection

A total of 3 × 105 PATC cells were plated and treated with indicated compounds for specified durations. Cells were harvested, washed with cold PBS, and stained with 5 μM MitoSOX (Life Technologies, #M36008) or 1 μM BODIPY 581/591 C11 (Thermo Fisher Scientific #D3861) for mitochondrial ROS detection or lipid ROS detection, respectively, for 20 min at 37 °C. Cells were then washed with cold PBS twice and filtered into single-cells suspensions. Cells stained for MitoSOX were then detected by a PE-Texas red filter and separated by flow cytometry. The PE-Texas Red filter was also used for detecting reduced BODIPY-C11, and FITC was used for detecting oxidized BODIPY-C11. Approximately 10,000–20,000 cells were collected and analyzed with Kaluza and Flowjo (Version 10.6.2) software. Peak shifts detected by MitoSOX-based flow cytometry indicated the medium fluorescence intensity (MFI) of mitochondrial ROS production. Further, the ratio oxidized/reduced BODIPY-C11 indicated levels of lipid peroxidation in each group.

Expression of H 2 O 2 probes and indication in subcellular scale

Ultrasensitive H 2 O 2 probes pCS2+MLS-HyPer7 and pCS2 + HyPer7-NES were gifted by Dr. Vsevolod Belousov (Addgene plasmids, RRID: Addgene_136470; RRID: Addgene_136467). For plasmid transfection, 3 × 105 PATC66 cells were seeded in 6-well plates for 24 hr. Cells were transfected with pCS2+ plasmids using Lipofectamine 3000 reagent (Life Technologies, #L3000015) according to the manufacturer’s protocol. After 24 hr, transfected cells were reseeded onto Nunc TM lab-Tek TM II 8-well chambered cover glass (Thermo Scientific, #155360) and treated with IACS-010759, MitoPQ, or Paraquat (PQ) with specified durations. Cells were then fixed with 4% paraformaldehyde (Boston Bioproducts Inc. BM155500ML) and treated with blocking/permeabilization buffer containing 0.2% Triton X-100 and 1% BSA for 10 min. Afterwards, each well was treated with anti-TOMM20 primary antibody (Proteintech, #11802-1-AP, 1:2000 dilution in PBS) and incubated overnight at 4 °C. Cells were then washed three times with PBST (0.05% Triton X-100) and incubated with a goat polyclonal secondary antibody (Goat anti-Rabbit Alexa Fluor 555, Fisher Scientific, #PIA32732) and Hoechst 33342 nucleic acid stain (Life Technologies, #H1399) for 40 min at room temperature. After washing with cold PBS, samples were covered with mounting medium (VECTOR LABS, #H-1200). Cell imaging was performed by a Leica SP8 Laser scanning confocal microscope and wide-field Nikon Eclipse-Ni microscope with a Hamamatsu C11440 digital camera.

Automatic image segmentation and quantification were performed using a MATLAB 2020b (The MathWorks, Inc.). Otsu’s thresholding method and marker-controlled watershed algorithm were employed to segment and specify DAPI-positive nuclear regions of individual cells. Expression levels of NES and MLS in the individual segmented area were then quantified as pixel intensity of green fluorescence. At least 50 cells with positive green fluorescence were then randomly selected for further statistical analysis. The quantified data were plotted in a violin graph using the fluorescent intensity level. Statistical significance was assessed using the Kruskal-Wallis test and the Dunn pairwise test.

Immunoblotting assays

Western blotting and antibodies

After treatment of compounds or knocked-out genes, PATC cells were harvested by digestion with 0.25% trypsin (Corning, #MT25053CI) and washed twice with ice-cold PBS. Cells were then lysed with RIPA lysis buffer (Fisher Scientific, #NC9193720) and a protease and phosphatase inhibitor cocktail (Life Technologies, #78447) on ice for 30 min, and centrifuged at 20,000 g for 15 min. Supernatants were then transferred into new EP tubes, and the total protein concentration was quantified by a BCA protein assay kit (Thermo Scientific, #23227). Samples were then diluted 6× by using Laemmli loading buffer (Fisher Scientific, #NC9140746) and heated to 95 °C for 5 min. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) was performed in 4-15% precast polyacrylamide gels (Bio-rad, #4561085). Samples were transferred onto a nitrocellulose membrane (Bio-rad, #1704158). After blocking with 5% BSA or milk in PBST (including 0.1% Tween 20), membranes were incubated with indicated primary antibodies overnight at 4 °C. Samples were then washed with PBST and incubated with indicated secondary antibodies for 1 hr at room temperature and washed with PBST again. Immunoblotting images were developed using a Pierce ECL western blotting substrate kit (VWR, PI32209). The following antibodies were used: anti-Cleaved Caspase 3 (Cell Signaling Technology, #9664 L,1:1000 dilution), anti-Ki67 (Life Technologies, #MA514520,1:200 dilution), anti-SOD2/MnSOD (Abcam, #ab13533,1:2000 dilution), anti-GNPAT (Proteintech, #14931-1-AP, 1:1000 dilution), anti-AGPS (Sigma-Aldrich, #HPA030211, 1:1000 dilution), anti-FAR1 (Novus Biologicals, #NBP1-89847, 1:1000 dilution), total OXPHOS human WB antibody cocktail (Abcam, #ab110411, 1:1000 dilution), anti-ATP5A (Abcam, #ab14748, 1:2000 dilution), anti-vinculin (EMD Millipore, #05-386, clone V284, 1:3000 dilution).

Mitochondria isolation and Blue Native western blotting

Mitochondria were isolated using a Mitochondria/Cytosol Fractionation Kit (BioVision, #K256) according to the manufacturer’s instruments. Briefly, 1–5 × 107 PATC cells were harvested and washed with ice-cold PBS. Cells were resuspended by 1 mL 1× cytosol extraction buffer (with DTT and protease inhibitors) and lysed on ice for 10 min. Cells were homogenized on ice (40-50 strokes) and centrifuged at 700 g for 10 min at 4 °C. Supernatant was transferred to a 1.5 mL tube and centrifuged at 10,000 g for 30 min at 4 °C. Mitochondria fraction pellets were resuspended with 100 μL ice-cold PBS. Mitochondria quantification was performed by the BCA protein assay kit (ThermoFisher, #23227) according to the manufacturer’s instruments. Fifty micrograms of mitochondria pellets were isolated by centrifuge at 10,500 g for 15 min at 4 °C. Mitochondria samples were then immediately stored at −80 °C, or were immediately used to perform Blue Native polyacrylamide gel electrophoresis (BN-PAGE) for detecting mitochondrial supercomplexes. We follow the BN-PAGE and western blotting protocol described by Jha et al.54. Briefly, pellets were mixed with 20 μL NativePAGE sample buffer cocktail (Thermo Scientific #BN2003 for 4×NativePAGE sample buffer, BioVision #2082-1 for EZSolution Digitonin, digitonin/protein ratio is 8 g/g) on ice for 20 min and centrifuged at 20,000 g 10 min at 4 °C. Fifteen microlitres of supernatant was mixed with 2 μL Coomassie G-250 sample additive (Thermo Scientific, #BN2004) in a new PCR tube. Afterwards, 15 μL of the resulting mixture was reserved for BN-PAGE and while the remaining 2 μL was reserved for detecting APT5A levels (control protein) using SDS-PAGE. We then used an XCell SureLock Mini-Cell electrophoresis system to run the 15 μL samples on NativePAGE 3%-12% gradient Bis-Tris gels (Thermo Scientific, #BN1003BOX) in a 4 °C cold room. Specifically, NativePAGE anode buffer (Thermo Scientific, #BN2001) and Dark blue cathode buffer (0.2 g/L Coomassie blue G-250 in anode buffer) were used to run the gel at 150 V for 30 min. We then changed the dark blue cathode buffer to the light blue cathode buffer (0.02 g/L Coomassie blue G-250), and ran the gel at 250 V for 90 min. The iBlot gel transfer device was then prepared, and samples were transferred to a PVDF membrane with 25 V for 10 min. Subsequently, the PVDF membrane was fixed with 8% acetic acid for 5 min and then washed with H 2 O three times. The fixed PVDF membrane was then incubated in methanol for 5-10 min to remove Coomassie blue and then washed with H 2 O three times. The membrane was blocked with blocking buffer (5% milk in TBS + 0.1% Tween 20) for 30 min at room temperature and then incubated with 10 mL of the primary antibody (total OXPHOS human WB antibody cocktail) overnight with gentle shaking at 4 °C. The PVDF membrane was then washed with TBST three times and incubated with 10 mL of the secondary anti-mouse HRP-linked antibody (Cell Signaling Technologies, 1:2000, #7076 V) at room temperature for 60 min. After washing the membrane with TBST three times, membrane signaling was developed using horseradish peroxidase chemiluminescent substrate (Thermo Scientific, #34096). All western blotting data were replicated three times. Full scans of all X-films were obtained.

In vivo study

Xenograft models

We used 6–8 weeks nude mice (Strain #002019, The Jackson Laboratory) for the development of subcutaneous xenograft models. All mice were housed in caging under HEPA filtered air supply and exhaust and sterile equipment conditions with a photoperiod of 6:00 a.m. to 6:00 p.m. at 72 ± 2 F and 50% ± 10% humidity. For each mouse, 3 × 105 early-passages PATCs PDX cells were suspended in 50 μL 1×HBSS (Corning, #21022CV) and mixed with 50 μL matrigel (Westnet, #356254) in a 10 mL tube on ice. The 100 μL cell suspension was then subcutaneously injected into nude mice. At 4–5 weeks after implantation, all tumors were measured with digital caliper. Mice harboring 150–200 mm3 tumors were randomly divided into groups (n > 3), maintained on a fasting/feeding cycle protocol (described below), and treated with indicated compounds. Specifically, mice were treated with vehicle control (0.5% methylcellulose solution), IACS-010759 (prepared in 0.5% methylcellulose solution, 5 mg/kg), or A939572 (7.5 mg/kg,) every two days when fasting via oral gavage. Body weight and tumors were measured 2 times every week with an electronic scale and a digital caliper, respectively. Tumor volumes were calculated with the formula (Length × width × width)/2. For all mice experiments, maximum tumor volumes were not exceeded 2000 mm3 to meet the standards of IACUC. All mice were operated and treated under protocols approved by M.D. Anderson’s Institutional Animal Care and Use Committee (IACUC).

Fasting/feeding cycle mice model

We followed the fasting/feeding cycle protocols described by Elgendy et al.18. Briefly, mice were treated with a 24 hr fasting/feeding cycle in which they were allowed food for 24 hr (5 pm–5 pm) and then fasted for 24 hr. During each fasting cycle, indicated compounds were administered at 8 am. To verify the fasting effect, murine blood glucose was measured by a OneTouch UltraMini meter and test strips. Mice were weighed and observed carefully twice every week to ensure that mice were maintained weight and appeared healthy.

Immunohistochemistry (IHC) analysis

Harvested tumor samples were cut in half, immediately fixed with 4% formaldehyde for 24 hr, and then incubated with 70% ethanol. Samples were embedded in paraffin blocks (Leica ASP300S), cut into 4 μm sections, and mounted on slides to dry. Samples were deparaffinization and a sodium citrate solution was used for antigen retrieval. After boiling, endogenous peroxidase was blocked by 3% hydrogen peroxide, and non-specific sites were blocked by Rodent Block M solution (Biocare Medical, #RBM961H). Samples were incubated with indicated primary antibody diluted in incubation buffer (0.1% Triton X-100 in TBS + 3% BSA) overnight at 4 °C. Afterwards, samples were washed with TBST (0.05% Triton X-100 in TBS) three times, and incubated with a secondary antibody (ImmPRESS HRP Goat anti-Rabbit IgG Polymer Detection Kit, Peroxidase; Vector Biolabs, #MP-7451-50) for 40 min. Afterwards, ImmPRESS NovaRED substrate (Fisher Scientific, #NC9925963) was added to detect the IHC signal, and Harris Hematoxylin 100 sec was added for staining nuclei. To detect hypoxic areas using an Hypoxyprobe Omni Kit (Fisher Scientific, #NC0132724), 60 mg/kg pimonidazole HCl was administered through an intraperitoneal injection at 4 hr before tumor harvest. Following our IHC protocol, anti-pimonidazole rabbit antisera were used as the primary antibody, and Goat anti-rabbit IgG ImmPRESS secondary antibody [HRP polymer] (Vector Biolabs MP-7451-50) was incubated as the secondary antibody.

Metabolomics and lipidomics analysis

Metabolomics analysis

A total of 8 × 106 PATC cells were seeded in dishes (90% confluency) and changed medium 2 hr before harvested. All media were aspirated and immediately 4 mL MeOH: H 2 O 80:20 (v/v) was added to dishes. All dishes were kept at −80 °C for 10 min. Next, everything was scraped off and pipette into a 15 mL conical tube. Centrifuge the mix and aliquot volumes equivalent to the protein concentration (from parallel dishes) into 1.5 EP tubes. Keep the samples at −80 °C until shipped to Metabolon. Inc for analysis. All methods utilized followed manufacturers’ instructions for using the Waters ACQUITY UPLC System for a Waters ACQUITY ultra-performance liquid chromatography (UPLC), and a Thermo Scientific Q-Exactive high resolution/accurate mass spectrometer (Thermo Scientific) interfaced with a heated electrospray ionization (HESI-II) source, and an Orbitrap mass analyzer operated at 35,000 mass resolution.

Mitochondrial lipidomics analysis

For the lipidomic assay, mitochondria were purified (as described above) from 5 × 107 PATC cells. For each mitochondrial pellet suspended in 100 μL ice-cold PBS solution, 5 µL of Avanti SPLASH® LIPIDOMIX® Mass Spec Standard (330707) in methanol, 3 µL of 10 mM butylated hydroxytoluene in methanol, and 242 µL of −80 °C ethanol were added. The tubes were vortexed for 5 min and then centrifuged at 17,000 g at 4 °C for 10 min. Supernatants were then transferred to a Phenomenex Impact Protein Precipitation Plate (CE0-7565) and filtered through by using a vacuum manifold. The pellets obtained from centrifugation were re-extracted with 300 µL of ethanol, and the resulting supernatants were once again passed through the protein precipitation plate. Plate wells were then rinsed with 200 µL ethanol to elute residual lipids. The lipid extracts were transferred to new Simport tubes and pre-washed with methanol. 200 µL ethanol was then used to wash the collection wells of the plate. The wash solutions and extracts were combined, and samples were dried with a centrifugal vacuum concentrator. Dried samples were reconstituted in 100 µL ethanol. Samples were analyzed for lipidomics analysis according to manufacturer’s instructions for performing chromatography with a Thermo Fisher Scientific Accucore C30 column (2.6 µm, 150 × 2.1 mm).

Data analysis was performed with Thermo Scientific LipidSearch software (version 4.2.23). Statistical analyses were conducted using R (version 3.6.0). A two-factor mixed model ANOVA was used to compare biochemical levels between sensitivity, with the cell line treated as a random effect. All P values were two-tailed. P values obtained from multiple tests were adjusted using the false discovery rate (FDR). P ≤ 0.05 was considered statistically significant, unless otherwise specified.

13C 18 -Oleic acid Isotype flux tracing

PATC cells were incubated in DMEM medium containing 10% dialyzed FBS with 13C 18 -oleic acid for 0 or 24 hr. Ethanol was added to extract ePLs from whole-cell or mitochondria samples and vortexed for 5 min, put on ice for 10 min, and then centrifuged at 18,000 g for 10 min at 4 °C. The supernatant was transferred to an autosampler vial and 10 µL was injected into an HPLC-HRMS system. Mobile phase A (MPA) was 40:60 acetonitrile:water with 0.1 % formic acid and 10 mM ammonium formate. Mobile phase B (MPB) was 90:9:1 isopropanol:acetonitrile:water with 0.1 % formic acid and 10 mM ammonium formate. The chromatographic method included an Accucore C30 column (2.6 µm, 150 × 2.1 mm) maintained at 40 °C, autosampler tray chilling at 8 °C, a mobile phase flow rate of 0.200 mL/min, and a gradient elution program as follows: 0–3 min, 30% MPB; 3–13 min, 30–43% MPB; 13.1–33 min, 50–70% MPB; 33–48 min, 70–99% MPB; 40–55 min, 99% MPB; 55.1–60 min, 30% MPB. A Thermo Orbitrap Fusion Lumos Tribrid mass spectrometer with heated electrospray ionization source was operated in positive ionization mode, with scan ranges of 600–900 m/z and a resolution of 240,000 (FWHM). A spray voltage of 3600 V was used. Vaporizer and ion transfer tube temperatures were set at 275 and 300 °C, respectively. The sheath, auxiliary, and sweep gas pressures were 35, 10, and 0 (arbitrary units), respectively. Data were analyzed using Thermo Scientific TraceFinder software (version 5.1).

Statistical analysis

Each experiment was repeated independently at least three times unless otherwise indicated, and all data are presented as mean ± standard deviation (S.D.). Data analysis was performed using GraphPad Prism 9 and ImageJ. Two-tailed unpaired Student’s t-test was used for two-group comparison and one-way ANOVA analysis was used for multiple groups comparisons. P value <0.05 was considered statistically significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.