Pharmacological induction of MHC-I expression in tumor cells revitalizes T cell antitumor immunity

Antigen presentation by major histocompatibility complex class I (MHC-I) is crucial for T cell–mediated killing, and aberrant surface MHC-I expression is tightly associated with immune evasion. To address MHC-I downregulation, we conducted a high-throughput flow cytometry screen, identifying bleomycin (BLM) as a potent inducer of cell surface MHC-I expression. BLM-induced MHC-I augmentation rendered tumor cells more susceptible to T cells in coculture assays and enhanced antitumor responses in an adoptive cellular transfer mouse model. Mechanistically, BLM remodeled the tumor immune microenvironment, inducing MHC-I expression in a manner dependent on ataxia-telangiectasia mutated/ataxia telangiectasia and Rad3-related–NF-κB. Furthermore, BLM improved T cell–dependent immunotherapeutic approaches, including bispecific antibody therapy, immune checkpoint therapy, and autologous tumor-infiltrating lymphocyte therapy. Importantly, low-dose BLM treatment in mouse models amplified the antitumor effect of immunotherapy without detectable pulmonary toxicity. In summary, our findings repurpose BLM as a potential inducer of MHC-I, enhancing its expression to improve the efficacy of T cell–based immunotherapy.


Western blotting
For Western blot analysis, cells were harvested, lysed using radioimmunoprecipitation assay (RIPA) lysis buffer, and boiled at 100°C for 10 min.Cell lysates were separated via SDS-PAGE and subsequently transferred onto nitrocellulose membranes (#66485, Pall) following established protocols.
After blocking with 5% non-fat milk in PBS + 0.1% Tween-20 for 1 h, membranes were incubated overnight with primary antibodies.Protein bands were visualized through Odyssey infrared imaging system.(Fluorescence Chemiluminescence lmaging System, Clinx Science, Shanghai, China) after incubation with horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature.

RNA extraction and quantitative Real Time PCR
Total RNA was extracted using a GeneJET RNA purification KIT (#K0732, ThermoFisher) following the manufacturer's protocol.Subsequently, the extracted RNA (1 μg) was transcribed into cDNA using HiScript II Q RT SuperMix, according to the manufacturer's instructions (Vazyme).Real-time qPCR (RT-qPCR) was performed using AceQ qPCR SYBR Green Master Mix (Vazyme) and gene-specific primers (sequences listed in Supplemental Table 2).Data were normalized to GAPDH and quantified via the 2 -ΔΔCt method.

/ 40
Cell viability and cytotoxicity assay Primary bladder cancer cell lines were trypsinized, counted, and plated in 96-well plates in triplicate.The cells were then incubated overnight.Adherent cells were treated with BLM or the vehicle control (PBS) for three days.Cell viability and cytotoxicity were determined using CellTiter-Glo reagent (Promega) according to the manufacturer's instructions.The absorbance was measured using an EnSpire Multilabel Reader (PerkinElmer).
Cell viability and cytotoxicity assay SK-BR-3 and B16OVA cells underwent treatment with indicated Azacitidine (AZA) or Decitabine (DAC) in complete medium every 24 h for five days.Cells were then trypsinized, counted, and plated in 96-well plates in technical triplicate.They were subsequently incubated overnight.Adherent 3 / 40 cells were subjected to treatment with BLM or a vehicle control (PBS) for three days.Cell viability and cytotoxicity were determined using the CellTiter-Glo reagent (Promega) as per the manufacturer's instructions.Absorbance was measured using the EnSpire Multilabel Reader (PerkinElmer).
RNAi-mediated gene silencing, CRISPR/Cas9-mediated gene knockdown In siRNA knockdown experiments, three on-target siRNAs (GenePharma), one negative control siRNA, and one GAPDH positive control were used.Cells were transfected with siRNA using Lipofectamine RNAiMAX (13778150, Thermo Fisher Scientific).Total protein was extracted 48 h post-transfection for Western blot analysis.
For CRISPR/Cas9 knockdown, 293T cells were co-transfected with PGMD2G, psPAX2, and sgRNA-expressing lentiCRISPR v2 plasmids.The supernatant was collected at 48 and 72 h post-transfection and used to infect cell lines with 1 μg/mL polybrene before puromycin selection.Knockdown efficiency of sgRNAs was determined via western blotting after three days of puromycin selection, and resistant cells were harvested for functional assays.
Sequences of all siRNAs and sgRNAs are in Supplemental Table 3 and 4.

Flow Cytometry
For flow cytometry, cells were washed with PBS and resuspended in stain buffer (#554657, BD Pharmingen) along with primary conjugated antibodies at 4°C for 30 min.After washing with Stain buffer, samples were analyzed using flow cytometry (CytoFlex S Flow Cytometer, Beckman, USA) and the data were analyzed using FlowJo™ v10.7 Software (BD Life Sciences, USA).For drug administration, BLM (dissolved in saline, intraperitoneal injection) was administered three times every other day at 3 mg/kg.Fresh melanoma tumor samples were collected from the tumor-bearing mice.Each group contained three mice, and the single cells were mixed and regarded as one sample for further experiments.The scRNA-seq libraries were constructed using the GEXSCOPE® Single-Cell RNA Library Kit (Singleron Biotechnologies) and Singleron Matrix® Automated single-cell processing system (Singleron Biotechnologies), according to the manufacturer's protocol.Libraries were sequenced on an Illumina Novaseq 6000 with 150 bp paired-end reads.
Preprocessing, filtering, and normalization.Subsequent analyses were performed using "Seurat v4" (2), Single-cell gene expression data of all samples were merged, and transcriptomes were filtered for cells with 500-55,000 genes detected, 1000-20,000 UMIs counted, fraction of mitochondrial reads < 30%, and fraction of hemoglobin reads < 5%.After filtering, UMI counts were variance-stabilized using scTransform (3) with 3000 variable features, while regressing out the number of UMIs and the fraction of mitochondrial reads.Functional analysis.InferCNV (v1.16.0) was used for copy number assignment.

Supplemental Tables
Supplemental Table 1.Results of high throughput screening of the FDA-approved drugs.MFI was normalized to IFN-γ-DMSO treated group.Supplemental Table 2 Melanoma and colon cancer mouse models B16F10 cells (2 × 10 5 /mouse) or MC38 cells (5 × 10 5 /mouse) were subcutaneously inoculated into C57BL/6 mice (purchased from Shanghai JieSiJie Laboratory Animal Company Limited, 6-8 weeks old).For drug administration, BLM (dissolved in saline, intraperitoneal injection) was administered every alternate day at 3 mg/kg.Anti-PD-L1 monoclonal antibody ((#BE0101, clone 10F.9G2, Bio X Cell; dissolved in PBS, intraperitoneal injection) was administered at 200 μg on days 7 and 12. Mouse weight and tumor volume were measured every two days.Cell lines SU-DHL-4, T-47D, BT549, and MB49 cell lines were obtained from the Cell Bank of the Chinese Academy of Sciences.The cell lines were maintained in RPMI 1640 medium (#C11875500BT, Gibco, Thermo Fisher).B16F10 and B16OVA cell lines were acquired from the Cell Bank at the Chinese Academy of Sciences and cultured in Dulbecco's Modified Eagle Medium (#10566016, Gibco, Thermo Fisher).The HEK293T cell line was kindly provided by Dengke K. M's Lab at California University, San Francisco, and was cultured in Dulbecco's Modified Eagle Medium (#C1199500BT, Gibco, Thermo Fisher).The MC38 cell line, generously shared by Jun O. Liu's Lab at Johns Hopkins University, Baltimore, and was maintained in RPMI 1640 medium (ATCC modification) (#A1049101, Gibco, Thermo Fisher).The SK-BR-3 cell line was procured from ATCC and cultured in McCoy's 5A medium (ATCC modification) (#16600082, Gibco, Thermo Fisher).The MDA-MB-231 cell line was obtained from the Cell Bank of the Chinese Academy of Sciences and cultured in Leibovitz's L-15 (#11415064, Gibco, Thermo Fisher).All culture media were supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin.The cells were maintained in 95% humidified air with 5% CO2 at 37°C.Tumor-infiltrating cell analysis At the conclusion of the mouse experiment, tumor samples were collected and digested with collagenase I (C0130, Sigma).The digested tumor cells were then filtered through a 70-μm cell strainer to obtain single-cell suspensions.These cells were subsequently lysed with Red Blood Cell Lysis Buffer (#B541001, Sangon Biotech, Shanghai) to remove erythrocytes.Tumor cells were washed with staining buffer and subjected to flow cytometric staining.Reagents and antibodiesThe chemicals used in this study were bleomycin sulfate (#T6116, Topscience), BAY11-7082 (#HY-13453, MedChemExpress), KU60019 (#HY-12061, MedChemExpress), AZD6738 (#HY-19323, MedChemExpress), and NU7441 (#S2638, Selleckchem).For Mouse T cell activation, OVA Peptide (257-264) TFA were procured from MedChemExpress, and Mouse Recombinant IL-2 was purchased from STEMCELL Technologies.For human T cell activation, anti-human CD3ε (clone UCHT-1) and anti-human CD28 (clone CD28.2) antibodies were obtained from BioLegend.Human IL-2 Recombinant Protein was acquired from Thermo Fisher Scientific (200-02-1MG, PeproTech®).The antibodies used in this study were listed in Supplemental Table5 and 6.CCLE AnalysisWe retrieved RNA-Seq data from the DepMap Portal, which contains information on 1,406 cancer cell lines.This dataset provides processed gene expression data quantified in accordance with the Genotype-Tissue Expression (GTEx) pipelines.TCGA Data AnalysisTranscriptomic profiles of tumor data across all TCGA datasets (33 cancer types) were obtained using UCSCXenaTools.Gene expression levels were quantified as logarithmic Transcripts Per Kilobase Million (TPM).The IOBR package facilitated the selection of specific pathways, with correlations between the BLM signature score and the indicated pathway scores assessed using Spearman correlations(1).Bulk RNA-seq AnalysisFor comparative analysis of transcription profiles between control and BLM-treated SU-DHL-4 cells, SU-DHL-4 cells were exposed to either vehicle control or 10 μM BLM for 24 h.Total RNA was isolated according to the manufacturer's protocol, and samples from each experiment were submitted to Novogene, Inc. for sequencing.Illumina HiSeq 2000 was employed for paired-end 150 bp sequencing, with subsequent mapping of sequencing reads 8 / 40 to the hg38 genome using hisat2.Differentially expressed genes were statistically analyzed using DESeq2.Single-cell RNA-sequencing Single-cell RNA sample collection and sequencing B16F10 cells (2 × 10 5 /mouse) were subcutaneously inoculated into C57BL/6 mice (purchased from Shanghai JieSiJie Laboratory Animal Company Limited, 6-8 weeks old).

Supplemental Figure 1 .
HTFCS system was used to identify the candidate drugs for promoting MHC-I expression.SU-DHL-4 were plated in round-bottom 96-well plates.FDA-approved drugs, 500 U/ml IFN-γ, and DMSO were added to the cell plates, respectively.After 48 h, cells were stained with anti-HLA-A/B/C antibody W6/32-APC and analyzed by IntelliCyt iQue Screener PLUS.Dot plot showing the results of HTFCS using an FDA-approved drugs library.Red dots indicate cabazitaxel, etoposide, entinostat, CI994, and Bleomycin sulfate.Supplemental Figure 2. BLM promoted MHC-I expression in different types of tumor cells.(A) Western blot analysis of HLA-A expression in SK-BR-3 cells after the indicated BLM concentrations or BLM treatment times.qRT-PCR analysis of the antigen presentation gene expression after BLM treatment for 48 h.(B-D) Western blot analysis of HLA-A expression in T-47D (B), MDA-MB-231 (C) and BT549 (D) cells after the indicated BLM treatment concentrations.(E) Western Blot analysis of the HLA-A, HLA-B, HLA-C expression in SK-BR-3 cells after the indicated BLM concentrations for 48 h.(F) SK-BR-3 cells pretreated with BLM (10 μM) for 48 h had their media removed.Cells were then washed with PBS to remove remaining drugs.Following 24 h or 48 h, protein samples were collected.(G) TCGA RNA-seq data analysis 12 / 40 reveals the correlation of BLM-treated signature with MHC Class I pathways.Plots display Spearman correlation and estimated statistical significance for the indicated pathway among different human cancer types.Each dot represents a human cancer type in TCGA.Data are shown as mean ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001 compared to the vehicle group by unpaired t test.13 / 40 Supplemental Figure 3. Antigen presentation pathway was activated after BLM treatment.(A-C) Cell surface H-2K b in B16F10 (A), MC38 (B), and MB49 (C) cell lines incubated with the indicated concentrations of BLM for 24 h (up panel).Quantification of mean fluorescence intensities (MFI) of H-2K b from (A-C), n=3 per group (down panel).(D) Western blot analysis of B2M expression in B16F10 cells after the indicated BLM concentrations or BLM treatment times.(E) qRT-PCR analysis of the antigen presentation gene expression in B16F10 cells after BLM treatment for 48 h.Data are shown as mean ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001 compared to the vehicle group by one-way ANOVA (A-C) and unpaired t test (E).and D) Heatmaps of the interaction strength (C) and the interaction numbers (D) between control and BLM tumor samples.Blue color indicated an increase in the displayed communication in BLM compared to control, whereas the red color indicated a decrease.(E) The CytoTRACE scores were compared among melanoma subclusters (Mel0-Mel5).(F) The differentiation state of melanoma cells by CytoTRACE analysis.The color showed the state of differentiation; from less differentiated (red) to more differentiated (blue) states.(G) Pseudotime trajectory analysis among melanoma subclusters (Mel0-Mel5) was computed by monocle3 package.(H) The CytoTRACE scores were compared between control and BLM group.Data are shown as mean ± SD. ***P < 0.001 compared to the vehicle group by unpaired t test.analysis of levels of γH2AX S139, pATM S1981, ATM, pChk2 T68, p53, and p53 S15 in SK-BR-3 cells.SK-BR-3 cells were harvested after treatment with BLM for indicated times.(E) Western blot analysis of levels of γH2AX S139, pATM S1981, ATM, pChk2 T68, p53, and p53 S15 in SK-BR-3 cells.SK-BR-3 cells were harvested after treatment with BLM for indicated concentrations.(F) Western blot analysis of the HLA-A expression in SK-BR-3 cells.Cells were pre-treated with 10 μM NU7441 for 6 h, then followed with 10 μM BLM for 12 h.(G) Western blot analysis of the expression of HLA-A and STING proteins in STING depleted SK-BR-3 cells after incubation with BLM for 48 h.(H) Western blot analysis of the expression of HLA-A and TRAF6 proteins in TRAF6 depleted SK-BR-3 cells after BLM treatment for 48 h.(I) Western blot analysis of the expression of HLA-A and p53 proteins in TP53 knockdown SK-BR-3 cells after incubation with BLM for 48 h.The knockdown efficiency is shown in the left panel.(J and K) Western blot and qRT-PCR analysis of MHC-I expression levels in, SK-BR-3 (J), and T47D (K) cells after BLM treatment in the presence of 2 or 5 ng/ml IFN-γ.Data are shown as mean ± SD. ***P < 0.001 compared to the vehicle group by one-way ANOVA.TILs at a 1:5 ratio.(E) BCC101 or BCC102 tumor cells were pre-treated with indicated concentrations of BLM for 24 h.Cell viability was measured by the CellTiter-Glo reagent.Data are shown as mean ± SD. 32 / 40 . Primers used in the qRT-

Table 6 .
Antibodies used for flow cytometric analysis.

Table 7 .
Clinical information of patients.