Materials availability

Information and materials can be obtained from the lead contact, Daniel S Peeper, on reasonable request.

Data and code availability

The RNA sequencing data presented in this article can be accessed via theNational Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) database using accession numbers GSE161211 and GSE252956. Due to the extent to which patient consent was given, only the respective expression values, and not the full sequencing data, were deposited for the TIL cohort (online supplemental table 4). The mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD026984. General code can be found on the Peeper laboratory github (https://github.com/PeeperLab/Rtoolbox); specific code can be requested from the lead contact.

Experimental model and subject detailsCell lines, primary cultures and human participants

The B16F10 cell line (male) was obtained from American Type Culture Collection (ATCC) and transduced with a lentiviral construct encoding for the full-length OVA protein.102 The Platinum-E cell line (female) was obtained from the internal Peeper laboratory stock. Both cell lines were routinely confirmed to be Mycoplasma-negative by PCR.103 They were maintained in Dulbecco’s Modified Eagle Medium (DMEM), containing 9% fetal bovine serum (FBS; Sigma), penicillin (100 U/mL, Gibco) and streptomycin (100 µg/mL, Gibco). Human CD8 T cells were isolated from healthy donor buffycoats (Sanquin) of both male and female anonymous donors, who provided written informed consent approved by the Dutch Ministry of Health. These cells were maintained in Roswell Park Memorial Institute Medium (RPMI) with 9% FBS, penicillin (100 U/mL), streptomycin (100 µg/mL) and 100 U/mL human IL-2 (Proleukin, Novartis). Murine CD8 T cells were isolated from the spleens of both male and female OT-I/Cas9 mice. These cells were maintained in RPMI with 9% FBS, penicillin (100 U/mL), streptomycin (100 µg/mL), 2-Mercaptoethanol (50 µM, Merck), murine IL-2 (10 ng/mL, ImmunoTools), murine IL-7 (0.5 ng/mL, ImmunoTools) and murine IL-15 (1 ng/mL, ImmunoTools). For TIL analysis, RNA of infused TIL products from both male and female, patients with stage IV melanoma was used87 (NCT00287131). These patients signed an informed consent approved by the Israeli Ministry of Health (Helsinki approval no. 3518/2004). Due to the extent to which patient consent was given, only the respective expression values, and not the full sequencing data, were deposited. For PD-L1 induction on B16F10-OVA cells, 1.2E5 tumor cells were seeded in 12-well plates and treated with 25 ng/mL IFN-γ for 24 hours (PeproTech). Cells were harvested by trypsinization (Thermo Fisher) prior to fluorescent staining and flow cytometric analysis as per below.

In vivo animal studies

All animal studies were approved by the animal ethics committee of the Netherlands Cancer Institute (NKI) and performed under approved NKI CCD14 (Centrale Commissie Dierproeven) according to the ethical and procedural guidelines established by the NKI and Dutch legislation. Mice were housed in single-use standard cages at controlled filtered air humidity (55%), temperature (21°C) and light cycle. All housing material, food and water were autoclaved or irradiated before use. Female, C57BL/6 (Janvier) mice were used between the ages of 8–12 weeks for animal experiments. For ACT experiments, female, Rag2–/– (Janvier) animals were used. To generate OT-I/Cas9 mice, OT-I mice (The Jackson Laboratory) were crossed with Cas9-EGFP mice (The Jackson Laboratory) and subsequently backcrossed for at least 10 generations. Genotypes for both OT-I and Cas9 were confirmed homozygous by PCR as specified by the supplier. Spleens used for CD8 T cell isolation were harvested from OT-I/Cas9 mice with a maximum age of 12 months.

Method detailsIsolation, maintenance and treatment of splenic murine CD8 T cells

Spleens were harvested from OT-I/Cas9 mice, mashed subsequently through 100 µm and 70 µm cell strainers (Corning) and washed (addition of buffer, followed by centrifugation at 1,000× g for 5 min) in phosphate-buffered saline (PBS) (Gibco) containing 0.1% bovine serum albumin (BSA, Sigma; isolation buffer) before being resuspended and incubated in red blood cell lysis buffer (155 mM NH4Cl, 10 mM NaHCO3, 0.1 mM EDTA in distilled water; all Sigma) for 5 min. Splenocytes were then washed twice (once in PBS, once in isolation buffer) and resuspended in 1 mL isolation buffer. CD8 T cells were then isolated from this cell suspension using the Dynabeads Untouched Mouse CD8 Cells Kit (Thermo Fisher Scientific) following manufacturer’s instructions. Briefly, 100 µL antibody mix and 100 µL FBS was added to the cell suspension and incubated for 20 min on a rotator at 4°C. The cells were then washed in isolation buffer, resuspended in 2 mL isolation buffer and 1 mL pre-washed Dynabeads was added and the suspension was incubated for 15 min on a rotator at room temperature (21°C). This cell suspension was then passed over a magnet, and the non-bound fraction was harvested. The CD8 T cells that were obtained were then resuspended in medium and activated on non-tissue culture-treated 24-well plates (Corning) that were pre-coated with anti-CD3 (0.25 µg per well, Thermo Fisher Scientific) and anti-CD28 antibodies (2.5 µg per well, Thermo Fisher Scientific). After 48 hours, cells were resuspended at a concentration of 1 million cells per milliliter in medium and either retrovirally transduced, or maintained at this concentration by daily replacement of culture medium for at least 10 days until further analysis. In most experiments, cells were either left resting or were activated. To do so, after 10 days cells were grown on 24-well non-tissue culture-treated plates that were pre-coated with anti-CD3 antibody (for activated cells, 1.25 µg per well) or that were left untreated (for resting cells). In select experiments, cells were treated with BrefA for 24 hours (200 ng/mL, Invivogen). In select experiments, cells were treated with BafA for 24 hours (100 nM, Invivogen). In select experiments, cells were treated with AGN192403 (also known as BRD4780) for 24 hours (100 µM, Tocris)

Retroviral sgRNA vector and retroviral library construction

For the generation of a retroviral sgRNA vector, a geneblock (Integrated DNA Technologies) with the sgRNA cassette of lentiCRISPR V.2 (Addgene #52961) where the BsmBI sites were replaced by BbsI sites was cloned into the pMSCVpuro backbone (Clontech) by standard molecular techniques. To insert single sgRNA sequences into this backbone, gene-specific Cas9 target sequences were predicted using CHOPCHOP104 and cloned into the pMSCVpuro-sgRNA backbone by Golden-Gate cloning.105 For retroviral library construction, the sgRNA cassette of the Brie library (Addgene) was amplified by PCR and cloned into the pMSCVpuro backbone by standard molecular techniques. The integrity of the retroviral library was confirmed by deep sequencing (online supplemental table 1).

Retrovirus production

For retrovirus production, 3 million Platinum-E cells were seeded in a 10 cm dish (Greiner). After 24 hours, these cells were transfected by polyethyleneimine (45 µg/10 µg DNA, Sigma) with 5 µg of pCL-ECO (Addgene) plasmid and 5 µg of the transfer vector. After another 24 hours, the medium was replaced by Opti-MEM (Thermo Fisher Scientific) containing 2% FBS, penicillin (100 U/mL) and streptomycin (100 µg/mL). After a further 24 hours, the supernatant containing retrovirus was harvested, filtered through a 0.45 µm filter and stored at 4°C. Fresh medium was added to Platinum-E cells. The next day, the supernatant was again harvested and filtered, combined with the supernatant of the first harvest and concentrated 10 times by spin-filter centrifugation (100 kDa pore size, Merck). The concentrated supernatant was snap-frozen and stored at −80°C until use.

CD8 T cell transduction and selection

One million preactivated OT-I/Cas9 CD8 T cells were mixed with 1 mL concentrated retroviral supernatant in a non-tissue culture-treated 24-well plate well (Corning) that was pre-coated with RetroNectin (25 µg/well, Takara). The plate was then centrifuged at 3,000× g in a centrifuge with minimum acceleration and no brake. After centrifugation, the plate was placed in the incubator. The next day, cells were resuspended at a concentration of 1 million cells per milliliter in the medium. 24 hours later, the cells were put on puromycin selection (4 µg/mL, Sigma) and maintained for at least 8 days until analysis.

Flow cytometry

0.3 million cells per sample were spun down in 96-well V bottom plates (Brand) and washed in a 0.1% BSA in PBS solution (FACS buffer). Antibodies against markers of interest were then diluted in FACS buffer following manufacturer’s instructions, for a staining volume of 50 µL per well, or 100 µL for in vivo tumor samples. The cells were stained on ice for 30 min, protected from light. After incubation, cells were washed twice in FACS buffer and resuspended in FACS buffer before being analyzed on either an LSRFortessa Flow Cytometer or an LSR II Flow Cytometer (both BD). Dead cells were identified through the use of DAPI (BD) or the LIVE/DEAD Fixable Near-IR stain (Thermo Fisher Scientific). Antibodies against murine and human PD-1 (both PE-conjugated), murine and human CD137 (both APC-conjugated), murine CD8a (FITC-conjugated), murine CD39 (PE-Vio770-conjugated), murine CD45 (APC-Vio770-conjugated; all Miltenyi Biotec), murine CTLA-4 (PE-Cy7-conjugated, BioLegend), murine PD-L1 (BV711-conjugated, BD) human CD8A (BB515-conjugated, BD) and human IgG Fc (PE-conjugated, BioLegend). Isotype antibodies for Rat IgG2a (PE-conjugated) and Hamster IgG (APC-conjugated, both BioLegend) were used. In select experiments, a fusion protein consisting of the extracellular domain of PD-L1 and a human antibody Fc domain was used to label the cells (1 µg/well, BioLegend) for 1 hour on ice, before continuing with regular cell-surface staining.

Whole genome screen and analysis

300 million OT-I/Cas9 CD8 T cells were isolated, anti-CD3-activated, transduced with the retroviral Brie library and selected with puromycin for 7 days. After 2 days of puromycin selection, a library reference sample was taken (of 1,000× coverage). After 7 days of puromycin selection, a bulk sample was taken (of 1,000× coverage). On the seventh day, 2×109 cells were re-activated. The next day, cells were harvested and stained with DAPI and antibodies targeting CD137 and PD-1. These cells were then sorted on FACSAria Fusion Cell Sorters (BD): DAPI–CD137+ cells were selected, and from that population, the top 10% PD-1 expressors and bottom 10% of PD-1 expressors were selected and sorted. At least 6×107 cells were sorted for each arm of the screen. The screen was performed in duplicate, where each duplicate was performed with OT-I/Cas9 CD8 T cells from independent spleens. DNA was isolated from all populations by the Blood and Cell Culture DNA Maxi Kit (Qiagen) as per the manufacturer’s instructions. The sgRNA sequences present in the isolated DNA were amplified using the NEBNext High-Fidelity 2× PCR Master Mix (New England BioLabs) following manufacturer’s instructions, using the Brie_Fw and Brie_Rv primers (online supplemental table 5). The stretch of N nucleotides in the Brie_Fw primer denotes a unique DNA barcode used for each sample in the PCR. The amplicons generated were then analyzed on an Illumina HiSeq 2500 Sequencing system (Illumina). The identified sgRNA sequences were aligned to the Brie library, where reads with any mismatches were excluded from analysis. The resulting read count table (online supplemental table 1) was used as input for MAGeCK analysis (V.0.5.6),63 using the non-targeting sgRNA sequences as controls. To assess the relative depletion of essential genes, the library reference and bulk samples were used, and compared with known non-essential and core essential genes.62

Western blot

For western blot, cells were harvested, washed twice in PBS and lysed in RIPA buffer (50 mM TRIS pH 8.0, 150 mM NaCl, 1% Nonidet P40, 0.5% sodium deoxycholate, 0.1% SDS with Halt Protease and Phosphatase Inhibitor (Thermo Fisher Scientific)) for 30 min on ice. After incubation, the lysate was centrifuged for 10 min at 17,000× g and supernatant was harvested. Protein concentration was then measured by Bradford assay (Bio-Rad) and normalized. NuPAGE LDS Sample Buffer (Thermo Fisher Scientific) and 2-Mercaptoethanol (final concentration 2.5% v/v) was added and samples were boiled at 95°C in a heating block. Samples were then run on NuPAGE 4–12% Bis Tris gels (Thermo Fisher Scientific) using Precision Plus Protein Dual Color Standard (Bio-Rad) as a size indicator at 150 V for approximately 1 hour on ice. After electrophoresis, the protein was transferred to nitrocellulose membranes using the iBlot system (Thermo Fisher Scientific) following manufacturer’s instructions. After protein transfer, the membranes were incubated for 1 hour in blocking buffer (5% BSA, 0.2% Tween-20 in PBS). After blocking, membranes were incubated with primary antibodies, diluted in a blocking buffer, for 24 hours. At that point, membranes were washed three times in washing buffer (0.2% Tween-20 in PBS) for 5 min, after which the membranes were incubated with secondary antibodies for 1 hour. After this incubation, membranes were washed three times in washing buffer for 5 min before being developed using SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific), with images being captured on a ChemiDoc MP (Bio-Rad). Fiji (V.2.0.0) was used to quantify protein bands. Primary antibodies against mouse PD-1, human PD-1, Myc-tag, Vinculin (all Cell Signaling Technology), TMED2, TMED10 (both Santa Cruz), TMED9 and Tubulin (both Thermo Fisher Scientific), NFkB p65 (Santa Cruz Biotechnology), phospho-NFkB p65 (CST) and Vinculin (Sigma) and Cyclophilin B (CST) were used. Horseradish peroxidase (HRP)-conjugated, secondary antibodies against mouse and rabbit primary antibodies (both Thermo Fisher Scientific) were used to detect the primary antibodies. Ponceau S (Merck) was used to assess relative protein loading. Uncropped Western blot images can be found in online supplemental file 1.

Quantitative PCR

For qPCR, cells were harvested, washed twice in PBS and total RNA was isolated using the Isolate II RNA Mini Kit (Bioline) following manufacturer’s instructions. After isolation, 1 µg of RNA was transcribed into complementary DNA (cDNA) using the Maxima First Strand cDNA Kit (Thermo Fisher Scientific) according to manufacturer’s instructions. To perform qPCR, 0.5% of this cDNA preparation was used per reaction. Primers against Pdcd1 and Actb (0.4 µM each; online supplemental table 5) and SensiFAST SYBR Hi-ROX reaction mix (Bioline) were added and the reaction was carried out by, and read out by, a StepOnePlus Real-Time PCR system (Thermo Fisher Scientific). Relative gene expression was determined using the ∆∆CT method.106

Cell-surface protein biotinylation and purification

Cell surface proteins were biotinylated using the EZ-Link Sulfo-NHS-SS-Biotin reagent (Thermo Fisher Scientific) following manufacturer’s instructions. Briefly, cells were washed three times in PBS containing MgCl2 and CaCl2 and were then resuspended at 1 million per milliliter, and Sulfo-NHS-SS-Biotin was added at a final concentration of 0.1 mg/mL and incubated for 30 min on ice. After incubation, cells were washed three times in quenching buffer (PBS containing MgCl2 and CaCl2 and 50 mM glycine) before being lysed in IP lysis buffer (30 mM Tris-HCl pH 7.4, 120 mM NaCl, 2 mM EDTA, 2 mM KCl, 1% Triton X-100, supplemented with Halt Protease and Phosphatase Inhibitor), incubated for 30 min and centrifuged at 17,000× g for 10 min. The supernatant, containing protein, was then harvested and labeled protein were immunoprecipitated using Pierce streptavidin beads (Thermo Fisher Scientific) for 1 hour. Proteins were retrieved by boiling samples at 95°C in sample buffer for 5 min. Instead, for mass spectrometry, samples were washed three times in PBS. Western blot was then performed as described above.

Immunoprecipitation

48 million CD8 T cells per condition were activated overnight as described above. The next day, cells were harvested, washed twice with PBS and lysed in PD-1-IP lysis buffer (0.5% NP-40 and 2 mM DTT in PBS, pH 7.4) for 30 min. The lysate was then centrifuged at 17,000× g for 10 min. The protein-containing supernatant was harvested, quantified and 8 mg of protein was incubated on a rotator with anti-PD-1 antibody (CST) or isotype control (Thermo Fisher Scientific) for 1 hour at 4°C. After incubation, pre-washed protein A beads (Bio-Rad) were added and incubated for another hour. After IP, the beads were washed three times in lysis buffer and once in PBS, after which bound proteins were retrieved by boiling samples at 95°C in sample buffer for 5 min. Immunoblot analysis of the IPs was performed as per above.

RNA sequencing

The total RNA was isolated using the RNeasy Mini Kit (Qiagen), including an on-column DNase digestion according to the manufacturer’s instructions. Quality and quantity of the total RNA were assessed by the 2100 Bioanalyzer using a Nano chip (Agilent). Total RNA samples with RNA Integrity Number (RIN) >8 were subjected to library generation. Strand-specific libraries were generated using the TruSeq Stranded mRNA sample preparation kit (Illumina) according to the manufacturer’s instructions (Illumina). Briefly, polyadenylated RNA from intact total RNA was purified using oligo-dT beads. Following purification, the RNA was fragmented, random primed and reverse transcribed using SuperScript II Reverse Transcriptase (Invitrogen) with the addition of actinomycin D. Second strand synthesis was performed using polymerase I and RNase H with replacement of deoxythymidine triphosphate (dTTP) for deoxyuridine triphosphate (dUTP). The generated cDNA fragments were 3’ end adenylated and ligated to Illumina paired-end sequencing adapters and subsequently amplified by 12 cycles of PCR. The libraries were analyzed on a 2100 Bioanalyzer using a 7500 chip (Agilent, Santa Clara, California, USA), diluted and pooled equimolar into a multiplex sequencing pool. The libraries were sequenced with 65 base single reads on an HiSeq 2500 using V4 chemistry (Illumina).

Bioinformatic analyses

Raw read counts were aligned to mouse reference genome GRCm38 Ensembl V.69 using STAR (V.2.7.1a) using two-pass mode and default settings. Differentially expressed genes were identified by DESeq2 (V.1.24). Gene ontology term enrichment of biological process gene sets was performed on significantly differentially expressed genes (false discovery rate (FDR)<0.05) using Panther (V.16.0).107 This list was then used as input for REVIGO.80 Representative GO terms were then analyzed for directionality by GSEA using GSEA software (V.4.1.0).

For scRNA sequencing analyses, data was downloaded from the TISCH database108 (online supplemental table 4). For the correlation between TMED complex expression and T cell dysfunction, we used a 30-gene T cell dysfunction signature computed using MetaCell18 (online supplemental table 4). For the analysis of TMED complex expression in exhausted versus other CD8 T cells, populations were defined using canonical markers109–111 (online supplemental table 4). For each cohort, a mean TMED complex expression across CD8 and exhausted CD8 T cells available in the cohort was computed. The response status of the melanoma cohort was classified according to Response Evaluation Criteria in Solid Tumors (RECIST) for each patient.12

For TIL analysis, RNA was extracted from infused TIL products using Tri Reagent (Sigma-Aldrich) according to the manufacturer’s protocol. RNA sequencing libraries were prepared with Illumina’s Ribo Zero Gold and TruSeq stranded library prep kits and sequenced on the Illumina HiSeq 2500 platform using paired-end sequencing with read length of 2×125–150 bps. Reads were aligned to the human genome reference build hg38 using STAR aligner112 and were quantified with FeatureCounts.113 After filtration of lowly expressed genes (counts below 10 in more than 90% of samples), raw counts were normalized in the R environment according to the LIMMA pipeline.114

For RNA sequencing analyses for the ACT experiment, analyses were performed for each isolation condition separately (input, tumor digest and CD8-enriched fractions). Genes with summed counts lower than 10 within one isolation condition were excluded from analysis.

Protein motif searches were performed using MEME.115

Cytokine release assay

7.5×104 B16F10-OVA cells were seeded in 12-well plate wells and CD8 T cells were added in a 1:8 ratio, in the presence or absence of anti-PD-1 antibody (10 µg/mL; Bio X Cell). After 48 hours, 20 µL supernatant was removed from the well, centrifuged at 1,000× g, and the supernatant was analyzed for the concentration of IFN-γ, TNF and IL-2 by cytometric bead array (BD) following manufacturer’s instructions. For experiments with monocultures, 2 million CD8 T cells were activated in 24-well non-tissue culture-treated plates that were pre-coated with anti-CD3 antibody (for activated cells, 1.25 µg per well) or that were left untreated (for resting cells), instead of activation by tumor cells.

Proteomics and analysis

T cell pellets were lysed in heated guanidine lysis buffer as described previously.116 Protein concentrations were determined with a Bradford assay (Pierce), after which lysates were diluted to 2 M GuHCl and equal aliquots were taken for a 4-hour trypsin digestion (Sigma-Aldrich, enzyme:protein 1:50) at 37°C, followed by an additional trypsin digestion (1:50) overnight. Digestion was stopped by the addition of 5% formic acid, after which digests were desalted on Sep-Pak C18 cartridges (Waters, Massachusetts, USA). Eluates were dried in a SpeedVac concentrator and stored at −80°C until liquid chromatography tandem mass spectrometry (LC-MS/MS). After reconstitution in 2% formic acid, peptide mixtures were analyzed by nano LC-MS/MS on a Q Exactive HF-X Hybrid Quadrupole-Orbitrap Mass Spectrometer coupled to an EASY-NLC 1,200 system (Thermo Scientific). Samples were loaded directly onto the analytical column (ReproSil-Pur 120 C18-AQ, 1.9 µm, 75 µm×500 mm, packed in-house). Solvent A was 0.1% formic acid/water and solvent B was 0.1% formic acid/80% acetonitrile. For AGN192403-treated and vehicle-treated samples, peptides were eluted from the analytical column at a constant flow of 250 nL/min in a 210 min gradient containing a 195 min linear increase from 6% to 26% solvent B, followed by a 15 min wash. For Tmed10 KO and WT samples, peptides were eluted with a constant flow of 250 nL/min in a non-linear 210 min gradient containing the following percentages of solvent B: 10% at 5 min; 24% at 130 min; 35% at 170 min; 60% at 190 min and a final wash at 100%. Raw data files were analyzed with label-free quantitation (LFQ) in MaxQuant (V.1.6.17.0)117 using standard settings. MS/MS data were searched against the murine Swissprot database (release 2021_04, 17,073 entries) complemented with a list of common contaminants and concatenated with the reversed version of all sequences. The maximum allowed mass tolerance was 4.5 ppm in the main search and 20 ppm for fragment ion masses. FDR for peptide and protein identification were set to 1%. Trypsin/P was chosen as cleavage specificity allowing two missed cleavages. Carbamidomethylation (C) was set as a fixed modification, whereas oxidation (M) and protein N-terminal acetylation were used as variable modifications. LFQ intensities were log2-transformed in Perseus (V.1.6.14.0),118 after which protein abundance values were filtered for at least two valid values (out of three total) in at least one condition for the comparison of AGN192403-treated versus vehicle-treated cells, whereas in the case of Tmed10 KO versus WT cells, protein abundance values were filtered for at least five valid values (out of six total) in at least one condition. Missing values were then replaced by imputation based a normal distribution, using a width of 0.3 and a downshift of 1.8. Differentially regulated proteins were determined using a t-test (thresholds: p<0.05 and a 1.5-fold change in expression).

For the mass spectrometry of cell surface protein-enriched fractions, IP beads were heated for 7 min. at 95°C in 1× S-Trap Lysis buffer (5% SDS, 50 mM TEAB pH 8.5), after which proteins were reduced, alkylated and digested overnight with trypsin (Sigma-Aldrich; 2 µg per sample) on S-Trap Micro spin columns following the manufacturer’s instructions (ProtiFi, New York, USA). Peptides were eluted, vacuum dried and stored at −80°C until LC-MS/MS analysis.

Samples were analyzed by LC-MS/MS on an Exploris 480 mass spectrometer (Thermo Scientific) with a 135 min gradient, the LC set-up being the same as described above. RAW files were analyzed with LFQ in Proteome Discoverer (V.2.5.4.0, Thermo Scientific) using standard settings. MS/MS data were searched against the same database as described above using SEQUEST HT, with the same search parameters and CAMthiopropanoyl (K and N-terminus) as additional variable modifications. The maximum allowed precursor mass tolerance was 50 ppm and 0.06 Da for fragment ion masses, the remaining search parameters being the same as for the T cell proteomes. After filtering for Peptide Spectrum Match (PSM) Xcorr>1, the Proteome Discoverer output file containing protein LFQ abundances was loaded into Perseus and processed as described for the proteome analysis.

Human CD8 T cell isolation

Human CD8 T cells were isolated from healthy donor buffycoats as described previously.102 Briefly, peripheral blood mononuclear cells (PBMCs) were isolated through Ficoll gradient separation and CD8 T cells were purified by magnetic bead isolation using the Dynabeads CD8 Positive Isolation Kit (Thermo Fisher Scientific) following manufacturer’s instructions. Isolated CD8 T cells were activated for 48 hours on non-tissue culture-treated 24-well plate wells (Corning) that were pre-coated with anti-CD3 (5 µg/well, eBioscience) and anti-CD28 antibodies (5 µg/well, eBioscience). After activation, CD8 T cells were maintained for at least 7 days, refreshing every third day, in T cell medium at a concentration of 1×106 cells/mL before being used for experiments.

Murine in vivo ACT experiment

0.5×106 B16F10-OVA cells were injected in both flanks of C57BL/6 mice and allowed to establish for 4 days. After 4 days, mice were randomized and administered PBS (mock), or 2 million sgCtrl or sgTmed10 CD8 T cells (which were generated as per above) by intravenous tail injection. Mice received 1E5 units of human IL-2 (Clinigen) two times a day for 3 days after ACT by intraperitoneal injection. Tumor growth was followed by measuring tumor volume three times weekly by calipers, where tumor volume is calculated with the formula: . Seven days after ACT, sentinel tumors were harvested. The rest of the mice were followed until the total tumor volume exceeded 1,500 mm3. Mice were given ad libitum access to drinking water, chow and a nutritionally fortified water gel (DietGel).

Murine in vivo tumor growth experiment

0.5×106 B16F10-OVA cells were injected in both flanks of C57BL/6 mice and allowed to establish for 4 days. After 4 days, mice were randomized and administered either PBS or AGN192403 (125 mg/kg in PBS) by daily oral gavage for 6 days. Tumor growth was followed by measuring tumor volume three times weekly by calipers, where tumor volume is calculated with the formula: . The day after the final treatment with AGN192403, sentinel tumors were harvested. The rest of the mice were followed until the total tumor volume exceeded 1,500 mm3. Mice were given ad libitum access to drinking water, chow and a nutritionally fortified water gel (DietGel).

Tumor dissociation and immune cell isolation

Tumors were harvested and cut into ±0.5 mm2 pieces and incubated in 5 mL dissociation medium (RPMI with 10 U/mL DNAse I and 200 U/mL collagenase type IV) at 37°C for 1 hour while shaking. After dissociation, tumor preparations were then mashed subsequently through 100 µm and 70 µm cell strainers (Corning) and washed (addition of buffer, followed by centrifugation at 1,000× g for 5 min) in PBS (Gibco) containing 0.1% BSA (Sigma; isolation buffer) before being resuspended and incubated in red blood cell lysis buffer (155 mM NH4Cl, 10 mM NaHCO3, 0.1 mM EDTA in distilled water; all Sigma) for 5 min. Tumor samples were then washed twice in isolation buffer, before being used in downstream experiments as described above. CD8 cells were isolated from these tumor samples using the Dynabeads FlowComp Mouse CD8 Kit (Thermo Fisher Scientific) following manufacturer’s instructions. RNA sequencing was performed as described above.

Quantification and statistical analysis

The details of the quantifications and statistical analyses performed are described in the respective figure legends. Analyses were performed by Prism (GraphPad Software, V.8.4.3) or R. Unless when otherwise specified, a p value of lower than 0.05 was considered statistically significant. Experiments were repeated at least twice, except for the in vivo ACT experiment, which was performed only once. Raw data on which statistics were performed can be found in online supplemental file 2.