Chemicals and reagents

Rabusertib (S7109), prexasertib (S7178), AZD0156 (S8375), ceftazidime (S3649), and AZD6738 (S7693) were purchased from Selleckchem. MG132 (474,790), cycloheximide (0180), cefepime for in vitro studies (A3737), gemcitabine (G6423), and 4-hydroxytamoxifen (H6278) were purchased from Sigma-Aldrich. Human and mouse recombinant interferon-gamma (IFNγ, 485-MI, 285-IF) and TNF-alpha (TNFα, 410-MT, 210-TA) were from R&D Systems. BML-277 (HY-13946), Pifithrin-α (HY-15484) and nutlin-3a (HY-10029) were purchased from MedChem Express. AZD0156 was dissolved in ethanol, and gemcitabine and ceftazidime were dissolved in water. Penicillin G, meropenem, ceftriaxone, and cefazolin were obtained from Oakdell Pharmacy (7220 Louis Pasteur Dr., San Antonio, TX, 78,229) and dissolved in sterile 0.9% NaCl. Remaining drugs used for in vitro drug studies were dissolved in dimethyl sulfoxide (DMSO). Drugs were aliquoted in individual tubes, stored at -80° C, and thawed as needed to avoid multiple freeze-thaws or made fresh when required. The αPD1 antibody pembrolizumab was purchased from Invivogen, and the αPD1 antibody balstilimab was a kind gift from Agenus. The α-mouse PDL1 antibody (clone 10F.9G2) and IgG isotype control were purchased from BioXCell (BE0101, BE0090). In vivo ceftazidime was from MedChem express.

Cell lines

B16-F10 (herein B16), ID8agg, ES2, and 4T1 we described previously [10, 13]. All of these cell lines were grown in RPMI-1640 containing 5% fetal bovine serum, 1% penicillin/streptomycin, 1% L-glutamate, and 1% HEPES buffer (Gibco, ThermoFisher, “complete medium”). T24 cells were a gift from R. Svatek (UTHSA) and cultured in McCoy’s 5A complete medium (Thermo-Fisher). PDL1− TN61R and PDL1+TN61R Nras mutant mouse melanoma cells (NFH1 and NCH1 respectively) were made from single cell suspensions of tumours from 4-hydroxytamoxifen-induced mice plus UVB (4.5 kJ/m2) and passaged in vitro as polyclonal cell populations. NCH1, NFH1, MDA-MB-231, and Tni cells were grown in DMEM complete medium. All cell lines were periodically tested by Mycoalert PLUS (Lonza Bioscience) and confirmed Mycoplasma-free.

Stable PDL1lo cell lines were generated using shRNAs targeting CD274 or non-targeting controls as we previously described [10]. B16 cells were infected with lentivirus expressing murine control or Raptor-targeting shRNAs from Sigma-Aldrich (TRCN0000077469) and transduced cells were selected in 2 mg/mL neomycin to generate Raptorlo cells. PDL1KO T24 cells were generated by CRISPR/Cas9 as we previously described [25] and PD1lo T24 by using lentiviral particles containing PDCD1 shRNAs (TRCN0000083508) from Sigma-Aldrich. PD1KO and PDL1KO B16 cells were generated using CRISPR/Cas9-GFP expression plasmids (SantaCruz Biotechnology), which included 3 pooled sgRNAs targeting PD1 or PDL1. sgRNA sequences used for CRISPR/Cas9 mediated gene deletion for PD1 were (1) 5'-AGCCCCTCGCCCAAACCAGA-'3; (2) 5'-CGGAGGATCTTATGCTGAAC-3'; 5'-GTGCCTCGGCCATGGGACGT-3'. sgRNA sequences for PDL1 were (1) 5'-GTATGGCAGCAACGTCACGA-3'; (2) 5'-CTGGATCCACGGAAATTCTC-3'; (3) 5'-TCCAAAGGACTTGTACGTGG-3'. Stable PDL1KI B16 was generated as previously described [10]. Plasmids were transfected using Turbofect reagent (Thermo-Fisher) according to the manufacturer’s protocol. Cells were cytometrically sorted by GFP expression, then single cell clones were isolated and analysed for targeted gene disruption using genomic DNA PCR primers spanning exon regions of PD1 or PDL1. Gene targeting was validated by Immunoblotting and/or sequencing.

Plasmids, transfections, and lentivirus transductions

Chek2 expression plasmids or vector only controls for Chk2 re-expression experiments (Chk2RE) were obtained from OriGene (RC201278 and MR208692 for human and mouse, respectively). RCHY1 (PIRH2) siRNA Smartpools or scrambled controls were obtained from Dharmacon (Cat#: L-065323–01-0005). Mouse PDL1 expression plasmid for rescue experiments and further subcloning was from OriGene (Cat#: MG203953). Transient transfections were performed using lipofectamine 3000 reagent (Thermo-Fisher) per the manufacturer’s protocol. Transient induction of differential PDL1 expression into PDL1KO ES2 cells by was by transfecting empty vector (Addgene 72299) or human Flag-PDL1 (Addgene 121446) expression plasmids. After 48-h, cells were collected and lysed for immunoblot analysis.

For PDL1mem expression constructs, a mouse PDL1-tGFP fusion protein construct (OriGene MG203953) was PCR amplified and subcloned into the pDisplay vector (Invitrogen) at the BglII and PstI sites, resulting in N-terminal fusion to the murine Igκ-chain leader sequence and C-terminal fusion to the platelet derived growth factor receptor transmembrane domain, anchoring PDL1 to the extracellular plasma membrane (PDL1mem). To generate intracellular localised PDL1 constructs (PDL1cyto), a myristoylation (myr) tag [52] was incorporated at the N-terminus of PDL1, preventing PDL1 from extracellular surface presentation. To generate the pcDNA-myr-PDL1cyto expression vector, mouse PDL1-GFP CDS was PCR amplified and cloned into a pcDNA6C vector (Invitrogen) using EcoR1-XhoI sites. All constructs were sequence verified.

PDL1 vectors with C-terminal mCherry tag (pLV-mPDL1 aa18-132-mCherry, pLV-mPDL1 aa133-220-mCherry, pLV-mPDL1 aa1-260-mCherry, pLV-mPDL1 aa1-290-mCherry, pLV-mPDL1 aa260-290-mCherry) were designed using pLV-mCherry-vector backbone and were custom synthesized and sequence verified by Vector Builder (Chicago, IL). All PDL1 mutant constructs were CRISPR/Cas9 resistant for stable re-expression in PDL1KO cells.

Lentivirus was produced by transfecting HEK293T cells with targeting plasmid along with packaging vectors pMD-2G, pMDLg/pRRE and pRSV-Rev plasmids (Addgene 12259, 12,253, 12,251) using Turbofect Transfection Reagent (R0531, Thermo Scientific) according to the manufacturer’s transfection protocol. The lentivirus-containing supernatants were collected 24 and 48 h after transfection and concentrated by concentrator (Lenti-X, cat 631,231, Takara). Transduction was performed by adding the lentivirus to PDL1KO B16 cells in the presence of polybrene (10 μg/mL, TR-1003-G, Sigma), followed by selection with puromycin (1 μg/mL) for 14 days. Similarly, PDL1KO B16 cells stably expressing pDisplay-PDL1mem and pcDNA-myr-PDL1int constructs were generated by neomycin selection (2 mg/mL). Pooled clones were selected and expression of recombinant proteins was confirmed by immunoblotting and/or imaging as indicated.

Immunoblots and subcellular fractionations

Immunoblotting was performed as we described [53, 54]. Briefly, cells were lysed using RIPA buffer (20 mmol/L Tris–HCl, pH 8.0, 150 mmol/L NaCl, 1 mmol/L disodium EDTA, 1 mmol/L EGTA, 2.5 mmol/L sodium pyrophosphate, 1 mmol/L β-glycerophosphate, 1% triton-X100 plus Halt protease/phosphatase inhibitor cocktail from Thermo-Fisher (78,442). Plasma membrane/subcellular fractionation kit was obtained from Abcam (ab65400). Protein concentrations were quantified using a Pierce-BCA Kit (23,328). 20–40 µg of lysate plus 4 × loading dye were run on 4% to 15% gradient SDS-PAGE Precast TGX gels (Bio-Rad) and transferred to nitrocellulose membranes (Bio-Rad) using a rapid transblot system (Bio-Rad). Membranes were blocked in TBS (pH 7.4) plus 1% Tween-20 (TBST) and 5% non-fat dry milk, incubated overnight at 4° C with diluted primary antibodies against indicated proteins. Membranes were incubated with either horseradish peroxide–conjugated secondary antibodies (Cell Signaling) for 2 h at room temperature then washed with TBST 3 times for 3 min each followed by incubation with Western Lightening Plus ECL reagent from Perkin Elmer (NEL10400L). ImageJ or ImageLab softwares were used to quantify relative band intensity by normalizing to loading controls. Antibodies against the following targets from Cell Signaling were used: PDL1 (#13,684), ATM (#2873S), Vinculin (#18799S), γH2AXser139 (#9718S), Chk2 (#2662S), GAPDH (#2118S), Ubiquitin (#58395S), Rabbit HRP-conjugated secondary (#7074S), Mouse HRP-conjugated secondary (7076S), p-ATRSer428 (#2853S), p-Chk1ser345 (#2348S), ATR (#13934S), p-Chk2Thr68 (#2197S), p-ATMSer1981 (#5883S), Na+/K+ ATPase (#3010S), PDL1 (#29122S), Chk1 (#2360S), Vimentin (#3932S), Chk2 XP (#6334S), and Chk2 (3440S). Antibodies from Abcam against these targets were: PDL1 (ab213480) and PIRH2 (ab189907). Antibodies from Santa Cruz include those detecting Chk2 (sc-5278) and PIRH2 (sc-374505). The detection antibody against PD1 was from Thermo-Fisher (7A11B1), and RFP detection antibody (dsRED polyclonal) was from Takara (632496). Primary and secondary antibody dilutions were made per the manufacturer’s recommendation. Membrane and cytoplasmic fractionations were performed using the plasma membrane protein extraction kit (Abcam ab65400) following manufacturer’s instructions. Membrane fraction is pure surface plasma membrane. Cytoplasmic and nuclear fractionations were performed using NE-PER nuclear and cytoplasmic extraction reagents (Thermo Scientific 78833). 

Immunoprecipitation

Cells were cultured on 15-cm2 plates, trypsinized, and counted with Vi-Cell XR cell viability analyzer with 10 million cells used per immunoprecipitation sample. After centrifugation at 500 × gravity for 5 min, cell pellets were washed with ice-cold 1 × PBS and lysed in Pierce IP Lysis Buffer (#87,787) supplemented with 1:100 Halt Protease and Phosphatase Inhibitor Cocktail (#78,442) from Thermo-Fisher. Samples were incubated on a rotator for 15 min at 4° C then centrifuged at 13,000 × gravity for 15 min at 4° C. Supernatants were transferred to a new microcentrifuge tube. Input fractions were transferred to a new tube and boiled with equivalent volume of 2 × loading dye for 10 min. Antibodies against Chk2 (sc-5278), PIRH2 (sc-374505), or normal mouse IgG (sc-2025) as indicated for endogenous protein pull-downs were added to the remainder of the supernatants and incubated on a rotator overnight at 4° C. Pierce Protein G Magnetic Beads (Thermo-Fisher) were washed with ice-cold 1 × PBS four times on a magnetic rack, resuspended in Pierce IP Lysis Buffer, and added to immunoprecipitation samples. For co-immunoprecipitation of mCherry or turbo-GFP (tGFP) tagged PDL1 recombinant proteins, we used pre-conjugated RFP or tGFP trap beads (Chromotek). Samples were incubated on a rotator for 2 h at 4° C then washed four times with ice-cold 1 × PBS on a magnetic rack. Beads from each immunoprecipitation sample were resuspended in 1 × loading dye and boiled for 10 min, and supernatants were transferred to a new tube. Samples were subsequently analyzed via immunoblotting as described above.

RT-qPCR

CTRL or PDL1KO cells were harvested with vehicle controls or rabusertib and subjected to total RNA extraction using TRIzol reagent. A Superscript Vilo kit (Invitrogen) was used for reverse transcription of total RNA followed by qPCR of cDNAs using the SYBR green method. The following primers were used for murine Chek2: 5’- CCGAGCTTATTGGGAAAGGC-3’, 5’- AGCCATCTTTACCTCCCCAC -3’ and β-Actin: 5’- CATTGCTGACAGGATGCAGAAGG -3’, 5’- TGCTGGAAGGTGGACAGTGAGG-3’. Data were analyzed by the 2(-ΔΔ C(T)) method.

In vitro viability assays

Tumour cells were seeded in a 96 well plate. B16 and ID8agg cells were seeded into 96 well plates pre-coated with fibronectin derived from bovine plasma (Sigma Aldrich). All cells were treated with indicated drugs and respective vehicle controls for the indicated time periods. Plates were then incubated with 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyl tetrazolium bromide (MP Bio) MTT reagent (3 mg/ml) for 2 h. Medium was fully decanted and 200 µl of DMSO was added to each well. Plates were read at 490 nm with a BIO-TEK Synergy HT plate reader. Raw data was analyzed using Excel and normalized data plotted using Graph Pad Prism software. Cefepime, ceftazidime, penicillin G, meropenem, ceftriaxone, and cefazolin were used at 80 μM, a concentration safely and effectively achievable in vivo56,57. Pembrolizumab and balstilimab were used at 50 μg/mL. For MTT assays, IFNγ and TNFα were used at 1 ng/mL and 10 ng/mL respectively, added to wells at the same time as rabusertib, and incubated for 96 h.

Endogenous ubiquitination assays

The detection of ubiquitinated protein was conducted as previously described32. Briefly, 6 × 106 cells as indicated were treated with MG132 (2 μM) for 12 h. For T24 experiments, gemcitabine (10 nM) or vehicle (water) was added at the same time as MG132. Cells were lysed in denaturing buffer (50 mM Tris–HCl, 0.5 mM EDTA and 1% SDS) followed by heating at 95° C for 10 min and quenched with nine volumes of quenching buffer (0.5% Triton X-100, 20 mM Tris–HCl (pH 8.0), 137 mM NaCl, 10% glycerol, 2 mM EDTA). Protease inhibitor cocktail was freshly added to all buffers. Cell lysates were incubated on a rotator for 30 min at 4° C then centrifuged at 20,000 × gravity for 15 min at 4° C. Supernatants were subsequently processed using Dynabeads Protein G for Immunoprecipitation (Thermo Fisher) plus anti-Chk2 antibody (Santa Cruz-5278). The final eluate was processed and immunoblot analysis was performed as described above.

Immunofluorescence, confocal microscopy, and image visualization and analysis

For confocal imaging of PDL1mem and PDL1cyto constructs, cells were grown on collagen-treated slides to confluence, fixed with 4% fresh formaldehyde/PBS, permeabilized with ice-cold 100% methanol, and blocked in 5% normal goat serum/buffer for 1 h at RT. Cells were incubated with PDL1 primary antibody at 1:1000 overnight at 4 °C on a rocker. AlexaFluor secondary antibodies were used at 1:500 for 1 h at RT in the dark, followed by DAPI as nuclear counterstain at 0.1 mg/ml for 5 min. Prolong® Anti-Fade mounting agent with #1.5 coverslips was used to seal slides and allowed to dry overnight prior to imaging. Images were taken on Zeiss LSM710 Confocal Microscope using EC Plan-Neofluor 40x/1.30 Oil DIC objective, using same settings between fields of view and cell lines. 16-bit images have dimensions of 1024(x) by 1024(y) and z-dimensions that varied per sample (10–20 z-planes per image).

Cryopreserved ER and PR human ovarian section slides were thawed for 30 min at RT, fixed for 20 min in 4% paraformaldehyde, and washed three times in 1 × PBS. Samples were permeabilized with 0.1% Triton X-100 for 5 min and blocked in Background Sniper (Biocare Medical, cat #BS966) for 15 min at RT. Antibodies against PDL1 (Abcam #ab205921 or R&D #MAB1561), γH2AX (Millipore #05–636), and Chk2 (Abcam #ab109413) were diluted with Da Vinci Green (Biocare Medical #PD900L) antibody diluent and incubated overnight at 40 C. After three washes, appropriate fluorophore-conjugated secondary antibodies (AlexaFluor488 or AlexaFluor568, Invitrogen) diluted in Fluorescence Antibody Diluent (Biocare Medical #FAD901L) were incubated for one hour at room temperature. Samples were washed and stained with DAPI (Invitrogen #D1306) and HCS CellMask Deep Red stain (Invitrogen #H32721) for 10 min then mounted in Prolong Gold Anti-fade Mountant (Invitrogen, cat #9071). Stained tissue was imaged by laser scanning confocal microscopy using a Zeiss LSM880 and 20X (0.8 N.A.) Plan/Apo objective with a pinhole aperture of 1–1.5 AU. Images were captured and processed using identical microscope settings (e.g., laser transmission and gain). Quantifications of marker expression from representative 2X zoom images (minimum n = 3 patients) were performed using Imaris (Bitplane, v9.0) image analysis software Cells module. The segmentation algorithm detected nuclei via the DAPI channel and the cell body via the far-red channel for CellMask Deep Red, with cell splitting based on one nucleus per cell and cell body signal intensity thresholding using Imaris Image Visualization and Analysis Software 9.9. Statistical analysis was performed by a paired, two-tailed student’s t-test with unequal variance. Indirect immunofluorescence of nuclei for Chk2/PDL1 colocalisation was performed as described [57]. Chk2 foci imaging was obtained from pre-extracted nuclei and represent the chromatin-bound fraction. Cells were seeded and grown on glass coverslips for 24 h and if specified, treated with either Chk1i (10 nM) for 12 h or MG132 (5 μM) for 2 h. Cells were left intact or cell nuclei were pre-extracted as specified for each experiment with nuclear extraction buffer (NEB; 10 mM PIPES (pH 6.8), 100 mM NaCl, 300 mM sucrose, 3 mM MgCl2, 1 mM EGTA (pH 8.0), 0.5% Triton X-100) for 2 min at room temperature then fixed with 4% paraformaldehyde for 10 min at 4° C. Nuclei were blocked in 5% BSA and 0.3% Triton X-100 in PBS, immunoblotted with a primary antibody (1:500 in dilution buffer; 1% BSA and 0.3% Triton X-100 in PBS), followed by secondary antibody (2 μg/mL in dilution buffer). DNA was counterstained with DAPI. Slides were viewed on an Olympus FV3000 confocal microscope. Primary antibodies include rabbit αChk2 (Abcam), mouse αPDL1 or mouse αGFP (Sigma). Secondary antibodies include α-mouse (Abcam ab150103, Alexa Fluor 647), α-rabbit (Abcam ab150081, Alexa Fluor 488), α-mouse (Abcam ab150117, Alexa Fluor 488), and α-rabbit (Santa Cruz sc362292, CFL-647). Nuclear foci quantification was performed using CellProfiler.

Flow cytometry

Flow cytometry to detect expression of surface proteins in cultured cells was performed as we previously described [10, 11]. For cell death analysis with Annexin V, cells were treated with 1 μM Chk1 inhibitor (rabusertib) for 48 h. Cells were stained using the Dead Cell Apoptosis Kit with Annexin V for flow cytometry (Thermo-Fisher V13242) according to the manufacturer protocol. For sorting on PDL1 expression, CTRL T24 cells were stained with αPDL1 antibody (BioLegend 374,508) and sorted on the BD FACS Aria II. For immune analyses, mice were sacrificed using cervical dislocation under isofluorane anesthesia. tumours were collected and strained through 40 micron cell strainers into RPMI-1640 medium. 5 × 106 cells were collected from single-cell suspensions (counted with Vi-Cell-XR cell counter, Beckman Coulter), resuspended in 200 μL of 1 × PBS, and transferred to U-bottom 96 well plates. Cells were stained with Live/Dead Cell Stain for UV excitation (Thermo-Fisher Scientific) for 30 min 4 °C protected from light, followed by incubation with anti-CD16/32 (Biolegend) for 30 min at 4 °C protected from light. Surface antibodies were used at 1:100 and include αCD45 (58–0451-82) from Invitrogen, CD3 (TONBO 80–0032-U100), αCD4 (BD Biosciences 612,952), αCD8 (BD Biosciences 563,786), PD1 (BD Biosciences 566,515), αCD11b (BioLegend 101,226), αCD11c (TONBO 60–0114-U100), αF4/80 (Biolegend 123,135), αGr1 (BioLegend 108,440), NK1.1 (TONBO 20–5941-U100), and αB220 (BD Biosciences 612,972). After surface staining, cells were washed with FACS buffer (1 × PBS with 2% FBS), then activated with the Cell Activator Cocktail (Biolegend in CR10 medium (RPMI-1640 with 10%FBS, L-glutamine, sodium pyruvate, non-essential amino acids, penicillin/streptomycin, HEPES). After 6 h at 37° C, cells were and permeabilized using the Foxp3/Transcription Factor Staining Buffer Kit (TONBO Biosciences). Following permeabilization, cells were incubated with intracellular antibodies diluted to 1:100 for 30 min at 4° C. Intracellular antibodies include αFoxP3 (ThermoFisher 15–5773-82), αIFNγ (BD Biosciences 612,769), and αGranzyme B (BioLegend 515,408). After intracellular antibody incubation, cells were washed and fixed. Data were acquired using a Cytek Aurora flow cytometer and analyzed with FloJo software V.10.7.1. For sorting of cultured cells, ES2 cells were stained with Ghost UV 450 (Tonbo 13–0868-T100) Live/Dead stain for 30 min at 4 °C, washed with FACS buffer, then stained for 30 min at 4 °C with isotype control or αPDL1 (Fisher 50–112-8926) then were sorted into PDL1lo or PDL1hi populations using a BD FACSAria Fusion cell sorter.

Small cell lung cancer data sets

44 human SCLC cell lines of the ASCL1 subtype (SCLC-A) [28] were classified based on their PDL1 expression into high and low high groups by a median-centered analysis. IC50 values of DDR inhibitors in the SCLC-A cell lines were determined using a 96-h cell viability assay and calculated as previously described [58] and compared between the PDL1 high and low groups by t-test. Grubbs test was used to identify potential outliers.

Protein co-purifications

MBP-PDL1-FLAG (hereby PDL1) and MBP-Chk2-Strep (hereby Chk2) were synthesized and cloned into Series-438A vectors for expression in insect cells [59]. These vectors were transformed into the E. coli strain DH10Bac for bacmid generation. Bacmids were PCR verified and transfected into Sf9 (Gibco) insect cells using Cellfectin II Reagent (Gibco) following the manufacturer’s instructions for production and amplification of recombinant baculovirus. For co-expression of PDL1 and Chk2, 1 mL of each virus stock was used to infect 50 mL of Tni (Expression Systems) insect cells for 48 h at 27˚ C with 150 rpm shaking. Cells were harvested by centrifugation at 1000 × gravity for 10 min and stored at -80˚ C.

We modified a method to test protein complex formation previously described [59]. Briefly, cell pellets were lysed in 20 mL buffer A (25 mM Tris–HCl pH 7.5, 10% glycerol, 0.5 mM EDTA, 200 mM KCl, 1 mM DTT, and 0.01% Igepal CA-630) supplemented with a 4 × protease inhibitor mix (aprotinin, chymostatin, leupeptin, and pepstatin A at 3 μg/mL each), 2 µl benzonase, and 1 mM MgCl2. Lysates were further processed with 10 strokes in a Dounce homogenizer and sonication at 30% amplitude using a pulse (10 s on, 30 s off). Clarification of the lysates was accomplished by centrifugation at 45,000 rpm for 45 min in a Type 70 Ti fixed-angle rotor (Beckman Coulter) at 4˚ C. The supernatant was incubated with 1 mL Anti-FLAG M2 affinity gel (Sigma) for 1 h at 4˚ C with gentle agitation. The immobilized complex was washed with 50 mL of buffer A before elution in buffer A supplemented with 0.4 µg/mL 1 × FLAG peptide across 4 fractions of 1 column volume each. The presence of PDL1 and Chk2 was verified by SDS-PAGE and Coomassie staining, and the elution fractions were pooled before incubation with 500 µL Strep-Tactin Sepharose (IBA) for 1 h at 4˚ C with gentle agitation. The beads were washed with 50 mL buffer A and eluted in buffer A supplemented with 2.5 mM desthiobiotin in 4 fractions (1 complete volume each). Lysate, flow-through, input, and elutions were run on SDS-PAGE gels, stained with Coomassie blue, and imaged on a ChemiDoc MP Imaging System (Bio-Rad).

Mice, in vivo tumour challenges, and in vivo treatments

Wild type (WT) C57BL/6 J (Bl6), Rag2KO on Bl6 background, and NOD.CgPrkdcscidIl2rgtm1Wj1/SzJ [non-obese diabetic/severe combined immunodeficiency (NOD/SCID)/interleukin (IL)-2Rγ KO, NSG] mice were purchased from Jackson Laboratory or bred in our in-house facility and maintained under specific pathogen free conditions. Mice were given food and water ad libitum. 8 to 10-week old age- and sex matched mice were used. All animals were approved by the UT Health San Antonio Institutional Animal Care and Use Committee or the Dartmouth College Institutional Animal Care and Use Committee. For in vivo 4T1 experiments in NSG mice, CTRL or PDL1lo 4T1 cells (0.5 × 106 cells) were injected into mammary fat pads of BALB-C or NSG mice with Matrigel Membrane Matrix (1:1, Corning). Mice were randomized, and rabusertib in 5% DMSO was administered intraperitoneally 3 times total over the course of the experiment, on day 3 (10 mg/kg) 10 and 17 (1 mg/kg). For in vivo B16 experiments, CTRL or PDL1KO B16 cells (0.5 × 106) were injected into WT mice subcutaneously. 0.2 × 106 CTRL or PDL1KO B16 cells were injected into NSG mice, and 0.4 × 106 PDL1KO B16 cells per flank were injected subcutaneously into Rag2KO mice. Mice were treated with 25 mg/kg rabusertib or vehicle intraperitoneally on day 3, followed by daily injections of 5 mg/kg rabusertib or vehicle. Tumour volumes were measured using Vernier calipers calculated as (length x width2)/2. For in vivo Nras mutant melanoma experiments, CD274fl/fl TN61R and PDL1 replete CD274+/+TN61R (PDL1− NFH1 and PDL1+ NCH1, respectively) melanoma cells were challenged into WT Bl6 mice (0.5 × 106, subcutaneously) and treated with 25 mg/kg rabusertib intraperitoneally starting day 3 and every four days following, with 5 mg/kg rabusertib every other day. NSG mice received 0.2 × 106 NCH1 or NFH1 cells per flank subcutaneously. Ceftazidime was given 80 mg/kg intraperitoneally daily.

Melanocyte-specific PDL1 knockout genetically engineered mouse melanoma model

In our murine autochthonous melanoma model, mice lack PDL1 only in melanocytes, the melanoma cell-of-origin, and can develop melanomas de novo when induced. To generate these mice, a transgenic mouse with CD274 flanked by LoxP sites (CD274LoxP) on the 129 background from Taconic was crossed onto a BL6 background and then subsequently crossed for 7 generations with a well-established Tyrosinase:CreERT2 inducible NrasQ61R (TN61R) mouse melanoma model29. These melanocyte PDL1-deficient mice (CD274fl/fl TN61R) or PDL1 replete littermate controls (CD274+/+TN61R) were induced by painting 4-hydroxytamoxifen (25 mg/mL in DMSO) on their skin on postnatal day 3 and 4. Mice were induced with 4-hydroxytamoxifen ± ultraviolet B (UVB) radiation on postnatal day 5 using a 312nm 2X8 Watt UVB lamp (Spectroline). We obtained tumour cells from autochthonous melanomas after 4-hydroxytamoxifen induction plus 4.5 kJ/m2 UVB radiation and cultured them to create stable cell lines.

RNA-seq and bioinformatics analysis

Total RNA from CTRL, PDL1KO, PDL1mem, and PDL1cyto B16 cells were purified using an RNeasy kit (Qiagen), and RNA quality was ensured on an Agilent Bioanalyzer. 50-base pair single read sequencing was conducted using Illumina HiSeq 2000 at the UT Health San Antonio Genome Sequencing Core Facility. The entire differential expression (DE) analysis and gene set enrichment analysis (GSEA) were performed on R Studio. R package DESeq2 [60] was used to compare differential gene expression analysis. Converting gene IDs to Ensembl IDs was performed with AnnotationDbi. MSigDB was used to obtain the Mus musculus gene set. Significant genes were selected by removing those that have a log2FC value of -2 < x < 2 post-DE analysis. The fgsea R package was used for enrichment, and ggplot2 was used to visualize GSEA data. Leading edge genes from each pathway (most significant genes driving pathway enrichment) were converted from Ensembl ID to gene symbols using the following website: https://www.syngoportal.org/convert. Selected genes differentially expressed in PDL1KO/CTRL cells were plotted using R package EnhancedVolcano.

Statistics

Statistical analyses were done using GraphPad Prism 9 software. Data were shown as either mean ± standard deviation or ± standard error of the mean where indicated. Unpaired t-test was used to determine significance when comparing measurement data from two individual groups. Analysis of variance was used to determine significance when comparing continuous outcomes across multiple experimental groups. Outliers in data sets were identified by Grubb’s test and removed from analysis. P < 0.05 adjusted for multiple comparisons was considered statistically significant.