Cell lines and cell culture

HCC2814 cells were obtained from the Hamon Cancer Center Collection (University of Texas Southwestern Medical Center, Dallas, TX). H596 and KLN205 were purchased from ATCC. JH716 cells were provided by Chad Pecot (University of North Carolina, Chapel Hill, NC) [17]. To overexpress full-length ovalbumin (OVA), JH716 cells were transduced with lentivirus from pLVX-Hygro-IRES-OVA as previously described [18].

All tumor cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Corning #10-013-CV) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and streptomycin at 37 °C in a humidified atmosphere containing 5% CO2 and 95% air. Mycoplasma testing was performed on cultured cells every 6 months using the MycoStrip-Mycoplasma Detection kit (rep-mys-10, Invitrogen).

Generation of CRISPR- KO cell lines

Human and mouse knockout cell lines were generated as previously described [10]. In brief, sgRNAs targeting MLST8 were designed, amplified, and cloned into lentiCRISPR v2. Using standard protocols, lentiviruses were packaged in HEK293FT cells by transfecting cells with CRISPR or expression plasmids together with psPAX2 (lentiviral packaging) and pCMV-VSV-G (envelope) plasmids at 1:1:1 molar ratio using the Lipofectamine 2000 Reagent. Media was changed after 6 h of transfection and virus was collected after 48–72 h. Indicated cells were infected with virus in the presence of 8 μg/ml Polybrene for 24 h and selected with 1–2 μg/ml Puromycin or 20–40 μg/ml Blasticidin until non-transduced control cells died before being used for experiments.

The Pten and Mlst8 knockout KLN205 cell line was generated using the transfection of Cas9 ribonuleoprotein (RNP) complex technique. Cas9 protein (Integrated DNA Technologies, Inc.) was assembled with sgRNA for Pten or Mlst8 at 1:1 molar ratio in OptiMEM media. The Cas9/sgRNA RNP complex was reverse transfected into KLN205 cells with RNAiMAX (Thermofisher scientific). After 48 h post-transfection, transfected cells were seeded into a 96- well plate by limited dilution. Once cloned cells were expanded, double knockout cells (KLN205 Ptennull Mlst8null) were screened by western blot for PTEN and MLST8. sgRNA sequence were AATTAATGGGTGCGTTCACC for Mlst8 and AATTAATGGGTGCGTTCACC for Pten. Wild type Mlst8 or empty vector was re-expressed in the KLN205 Ptennull Mlst8null double knockout cell line to generate KLN205 Ptennull Mlst8null WT or Mlst8-KO.

From single clones obtained through limiting dilutions, JH716-OVA Mlst8 loss of function mutant cells (JH716-OVA-Mlst8-KO) and WT (JH716-OVA-WT) were generated as described above.

Mouse model

All mice were housed in a non-barrier animal facility. Rag1-deficient (C57BL/6, Rag1null) mice were obtained from Jackson Laboratories, and were maintained in sterile housing. DBA/2 mice were purchased from Envigo or The Jackson Laboratories. OT-I (C57BL/6) transgenic mice were purchased from The Jackson Laboratories. Mice used in experiments were between 8- and -12 weeks old.

Tumor models and treatment regimens

KLN205 Ptennull Mlst8null WT or Mlst8-KO cells (5 × 105) or H596 WT or MLST8-KO cells (5 × 106) were implanted subcutaneously into the dorsal flanks of recipient mice (DBA/2 or Rag1null, respectively. n = 10–13 mice per group). For HIF2α rescued experiment, 5 × 105 KLN205-Ptennull Mlst8-KO expressing empty vector (Mlst8-KO + EV) or constitutive active HIF2α cDNA (Mlst8-KO + HIF2α_CA) (Addgene) were implanted into DBA/2 mice subcutaneously (SQ) or via tail vein administration (n = 10-23 mice per group). All SQ tumors were measured using calipers, and tumor volume was calculated using the formula V = 0.5* (L x W2).

For drug studies, KLN205-Ptennull WT or Mlst8-KO (5 × 105) were seeded into the lung via tail vein injection (n = 7–16 mice per group) and the mice were randomized and treated with IgG control (10 mg/kg, BioXCell #BE0091; Lebanon, NH, USA) or anti-VISTA (10 mg/kg, BioXCell, #BE0310; Lebanon, NH, USA) starting on day 4. The mice were treated every 3–4 days for 7 doses via intraperitoneal injection. Animals were euthanized on day 25, and their lungs were flushed at least twice with PBS via the left ventricle of the heart. The weight of the lungs was measured, and detailed images were captured using an Olympus stereo microscope.

Flow cytometry

Flow cytometry experiment was performed as previously described [19]. Briefly, tumors were dissociated in RPMI-1640 media (Corning #MT10040CV, Corning, NY, USA) supplemented with 5% FBS, 1 mg/ml collagenase IA (Sigma-Aldrich #C9891, St Louis, MO, USA), and 0.25 mg/ml DNase I (Sigma-Aldrich #DN25) for 30 min at 37 °C followed by filtering the digested tissue through a 70-µm strainer. Red blood cells were lysed using ACK Lysis Buffer (KD Medical #RGF-3015, Columbia, MD, USA) and samples washed with PBS before staining with Ghost Dye Violet V510 (Tonbo Biosciences #13-0870, San Diego, CA USA) to exclude dead cells. After washing with buffer (0.5% BSA, 2 mM EDTA in PBS), samples were blocked in anti-CD16/32 mouse Fc block (Tonbo Biosciences #70-0161) and cell surface proteins were analyzed using antibodies against: CD45, TCRβ, CD4, CD8a, CD25, CD162, PD-1, and/or CD107a. Intracellular staining for GZMB was accomplished using a Cytofix/Cytoperm solution kit (BD 554714, Franklin Lakes, NJ, USA) on paraformaldehyde-fixed cells, according to the manufacturer’s protocol. Fluorescence minus one (FMO) or IgG controls were included in the gating controls when needed using splenocytes or tumor cell suspensions. Flow cytometry data was obtained on a BD Fortessa FACS Diva software and analyzed by blinding using FlowJo software (v10). Antibodies used in flow panels are detailed in Supplementary Table 1.

RNA sequencing

RNA sequencing was performed as previously described [19]. In brief, RNA was extracted from HCC2814 sgControl (WT, n = 3) or sgMLST8 (MLST8-KO, n = 3) using the Direct-zol RNA Miniprep Kit (Zymo Research, Irvine, CA, USA) according to manufacturer’s instructions. RNA sequencing was performed by BGI Americas (Cambridge, MA, USA) using the DNBSEQ platform. Raw data was filtered after sequencing to remove reads with high rates of unknown bases, low quality reads, and reads of adapter sequences. Clean reads were aligned to the reference genome (Homo Sapiens, version GCF_000001405.38_GRCh38.p12) using HISAT and aligned to reference genes using BowTie2. Using the Dr. Tom platform (BGI), differentially expressed genes (DEG) were identified using DESeq2 (q value < 0.05). Pathway enrichment analysis was performed against Molecular Signatures Database (MSigDB) Hallmark gene sets using ShinyGO Software (0.77) [20].

Gene alteration analysis

In total, 110 input gene identifiers from the human REACTOME_PI3K_AKT_SIGNALING_IN_CANCER gene set obtained from MSigDB [21,22,23] were used to examine their alterations from 5 TCGA Pan-cancer data sets [NSCLC: TCGA PanCancer Atlas (n = 1144); LUSC: CPTAC (n = 80), TCGA Firehose Legacy (n = 511); LUAD: CPTAC (n = 110), TCGA Firehose Legacy (n = 586)] using tools developed by cBioPortal (https://www.cbioportal.org/). The percentage of total alteration frequency was calculated from one or more datasets, as indicated. Genes in the mTORC2 pathway were further evaluated on cBioPortal to illustrate the genetic alterations in the PI3K-mTORC2-AKT pathway.

For survival analyses, relapse-free survival (RFS) and overall survival (OS) based on mean mTORC2-related gene (PIK3CA, PTEN (inverted), AKT1, AKT2, AKT3 and RICTOR) expression in stage II-IV lung squamous cell carcinoma were downloaded from KM plot (http://www.kmplot.com) [24]. Automatic cutoff was used to define “high” versus “low” expression from RNA-seq data, with cutoff values defined as RFS = 1422.17 (range = 536–11,116) and OS = 1793.67 (range = 536–11,116). Similar survival analyses were performed using 24 mTORC1 gene identifiers obtained from the human REACTOME_MTORC1_MEDIATED_ SIGNALLING and seven mTORC1-related gene (RPTOR, RPS6, RPS6KA1, AKT1S1, ELF4EBP1, ELF4EBP2, ELF4EBP3) expression in stage II-IV lung squamous cell carcinoma were downloaded from KM plot. Same as above, automatic cutoff was used to define “high” versus “low” expression from RNA-seq data, with cutoff values defined as RFS = 4068.73 (range = 2774–11,354) and OS = 4679.68 (range = 2774–11,354).

Western blot and co-immunoprecipitation (Co-IP)

Western blot and co-IP assays were performed as previously described [10]. Cultured tumor cells were lysed with CHAPS lysis buffer (40 mM Tris, pH 7.5, 120 mM NaCl, 1 mM EDTA, 0.3% CHAPS) and RIPA buffer supplemented with protease inhibitors and phosphatase inhibitors (Complete Mini and PhosStop inhibitor cocktail, Roche). Protein concentration of each sample was determined by Pierce BCA Protein Assay kit. Equal amounts of protein extracts were mixed with 4x Laemmli sample buffer, separated by electrophoresis on a 10% or a gradient 4–20% SDS-PAGE gel, and transferred onto nitrocellulose membranes. Membranes were blocked with 5% skim milk in TBS/T buffer and incubated with corresponding primary antibodies and IRdye-conjugated secondary antibodies. Immunoreactivity was detected using the Odyssey scanner (Li-cor Biosciences, Lincoln, NE, USA). To perform immunoprecipitation, equal amounts of input lysates (2–4 mg) were incubated with the primary antibodies (1–2 μg) for 2 h to overnight at 4 °C. Protein G Dynabeads were added, and lysates were incubated for 1 hr and washed four times with CHAPS lysis buffer.

Immunohistochemistry

Immunohistochemical staining of tumor lung tissue was performed on formalin-fixed, paraffin-embedded blocks as previously described [19]. Briefly, serial sections were deparaffinized in xylene and rehydrated through a graded series of ethanol concentrations. Antigen retrieval was achieved by microwave-boiling the sections for 10 min in citrate buffer. Sections were allowed to cool down for 30 min followed by PBS washes. Intrinsic peroxidase activity was blocked with 3% hydrogen peroxide in methanol for 10 min. Sections were blocked using goat serum (2.5%; Sigma) solution, and the optimally diluted primary antibodies (CD8a, 1:100, # 14-0808-82, eBioscience, San Diego, CA, USA; Cleaved Caspase-3, 1:100, #9664 L, CST, Danvers, MA, USA) were applied to cover the specimen and incubated at 4 °C overnight. After three washes in 1x PBS for 5 min each, slides were incubated with biotinylated secondary antibody at room temperature for 1 h. Following three additional washes in PBS, slides were incubated with streptavidin peroxidase reagent # SA-5704-100 (Vector Laboratories, Newark, CA, USA) for 45 min followed by additional PBS washes. Sections were developed with DAB substrate (#550524, BD Pharmingen) solution per manufacturer’s protocol and counter stain lightly with hematoxylin. Slides were dehydrated and mount with Cytoseal (#48212-196, VWR, Radnor, PA, USA). Images of at least five fields of view were obtained using an Olympus inverted fluorescence microscope [40x]. The intensity of the staining as well as the percentage of positive cells was recorded in a blinded manner.

Seahorse analysis

Cellular and mitochondrial bioenergetics of human and murine LUSCs were measured using the Seahorse XFe/XF-24 Extracellular Flux Analyzer (Seahorse Bioscience, North Billerica, MA, USA). Briefly, human (HCC2814 WT or MLST8-KO cells; H596 WT or MLST8-KO) and murine (KLN205-PTENnull WT or Mlst8-KO) cells were plated in XF-24 cell culture plates (1 × 105 cells/well) in 100 μL DMEM medium as mentioned above for 1 h before adding 400 µL complete medium and incubated overnight. Glycolysis stress and Mito stress tests were performed as per the manufacturer’s instructions (103020-100 and 103015-100, Agilent Technologies). For glycolysis stress test, tumor cells were washed in XF Assay media supplemented with 2 mM glutamine and 1 mM pyruvate. Glycolytic stress test drugs were made in prepared XF media at the following final concentrations: glucose (10 mM), oligomycin (0.25 uM), and 2-Deoxy-D-glucose (2-DG) (10 mM) To perform the mitochondrial stress test, tumor cells were washed in XF Assay Media supplemented with 10 mM glucose, 2 mM glutamine and 1 mM pyruvate. Mitochondrial stress test drugs were prepared in XF media at the following final concentrations: Oligomycin (1.5 μM), Carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone [(FCCP), 2.0 μM], and Rotenone/Antimycin A [(Rot/AA), 0.5 μM].

Quantitative real-time polymerase chain reaction

Total RNA was isolated and reversely transcribed using the RNeasy kit (Qiagen, #74104, Hilden, Germany) and iScript cDNA synthesis kit (Bio-Rad, # 1708891, Hercules, CA, USA) according to the manufacturers’ directions. qPCR was performed using the SYBR Green PCR Master Mix (#4368706, Thermo Fisher Scientific, Walthman, MA USA) on StepOne system (Applied Biosystems, Foster City, CA, USA). Each condition was assayed in triplicate. The primer sequences used for each gene are presented on Supplementary Table 2. Quantitation was performed using the ΔΔCt method.

Tumor interstitial fluid metabolic assays

Tumor interstitial fluid was collected from tumors as previously described [18] with some modifications. Tumors were harvested, weighed, and then washed with PBS to remove excess blood from the tumor surface. Excess PBS was carefully removed by blotting. Tumors were placed in a 20 μm pluriStrainer (pluriSelect, #43-10020-40, Leipzig, Germany) mounted in a microcentrifuge tube, then centrifuged at 100 g at 4 °C for 30 min. The cleared interstitium was deproteinated and lactate was detected using the Lactate Assay Kit (Sigma-Aldrich, MAK064) according to the manufacturer’s instructions. Known lactate standards were used to calculate concentrations.

Cytotoxicity assay

In vitro cytotoxicity assay was described previously [18] with modifications. Briefly, splenocytes were harvested from OT-1 mice and activated with SIINFEKL peptide (1 μg/mL; Invivogen #vac-sin) for 48 h. 2 × 105 JH716-OVA-sgControl or sgMlst8-KO cells were cultured in a 12-well plate for 48 h to collect conditioned media. For the co-culture, 1 × 104 JH716-OVA-sgControl or sgMlst8-KO cells were allowed to attach in a 96-well plate overnight. CD8+ T cells were isolated using mouse CD8a microbeads (Miltenyi Biotec, 130-045-201) according to the manufacturer’s instructions. Purified CD8+ T cells were co-cultured with tumor cells at a 1:1 ratio in the appropriate conditioned media. CyQuant LDH cytotoxicity assay (Invitrogen, #C20300) was performed in triplicate as described by the manufacturer, and cytotoxicity was calculated, accounting for non-T cell mediated tumor cell death. To evaluate cytotoxicity in response to VISTA or PSGL-1 neutralization, 2 × 105 JH716-OVA-sgControl cells were co-cultured with activated CD8 + OT-1 T cells as described above in the presence of anti-VISTA (100 μg/mL), anti-PSGL-1 (100 μg/mL), or an IgG control (100 μg/mL) for 48 h.

Statistics

GraphPad Prism software (10.0.3) was used to generate graphs and performed statistical analyses. Summary data are expressed as the mean ± SEM. For in vivo tumor experiments, data are reported from 2 to 3 independent experiments, where each dot represents a mouse. Statistical comparisons between two groups were performed using a two-tailed Student’s T test. For multiple comparisons, one- or two-way analysis of variance (ANOVA) was performed, and individual comparisons were evaluated using either Tukey’s, Dunnett’s, or Sidak’s post hoc analysis as indicated. Outliers were excluded using the ROUT method (Q = 5%). Differences with a p value less than 0.05 were considered statistically significant. For Kaplan-Meier survival curve analysis, log-rank analysis was performed between groups using the Mantel-Cox method. Hazard ratios (HR) and 95% confidence intervals (CI) were determined using the log-rank test. Pearson correlation (r) was performed on TCGA datasets.