Tumor Cell–Intrinsic USP22 Suppresses Antitumor Immunity in Pancreatic Cancer

Immune checkpoint blockade (ICB) has led to improved clinical outcomes in several types of cancer (1, 2). However, the majority of patients treated with ICB fail to respond, leading to overall response rates of 20% to 40% (2). The abundance of tumor-infiltrating T cells is a major factor predicting response to immunotherapy, as T-cell–inflamed tumors are more sensitive to ICB than non-T-cell–inflamed tumors (3). Thus, there is an urgent need to understand the factors regulating intratumoral T-cell infiltration. Studies have shown that the tumor mutational burden does not fully explain the abundance of tumor-infiltrating T cells (4), and there is an increasing agreement that various tumor cell–intrinsic factors, including signaling molecules, secreted factors, and transcriptional regulators, influence the level of T-cell infiltration into tumors and the resulting response to immunotherapy (5).

Pancreatic ductal adenocarcinoma (PDA) is predicted to become the second leading cause of cancer-related death in the United States within the next five years (6). Pancreatic tumors are characterized by T-cell exclusion and an immunosuppressive tumor microenvironment (TME). As a result, PDA is largely resistant to immunotherapy (7). We and others have reported that tumor cell–intrinsic factors–secreted molecules, signaling pathways, and epigenetic status–play a central role in shaping the immune TME in PDA and other types of cancer (5, 8–10). Importantly, heterogeneous mechanisms may contribute to the suppression of antitumor immunity (9). Several studies have demonstrated that a drug combination including gemcitabine, nab-paclitaxel, CD40 agonistic antibody, PD-1 blocking antibody, and CTLA-4 blocking antibody promotes positive outcomes in preclinical mouse models of PDA (9, 11, 12). Importantly, preclinical studies indicate that the abundance of tumor-infiltrating activated CD8⁺ T cells predicts sensitivity to this therapy (9). In this study, we evaluated the behavior of ubiquitin-specific protease 22 (USP22) as a tumor cell—intrinsic factor shaping the immune TME in PDA.

All animal procedures were conducted following NIH guidelines. All mouse protocols were in accordance with, and with the approval of the Institutional Animal Care and Use committee (IACUC) of the University of Pennsylvania (Philadelphia, PA). All wild-type (WT) C57BL/6 mice were purchased from The Jackson Laboratory and/or bred at the University of Pennsylvania (Philadelphia, PA). Kras-LSL-G12D/+;Trp53-LSL-R172H/+, Pdx1-Cre, and Rosa-LSL-YFP (KPCY) mice were bred in-house, backcrossed for over 10 generations with C57BL/6J mice (Jackson Laboratories), and assessed at the DartMouse Speed Congenic Core facility at the Geisel School of Medicine at Dartmouth College.

All mouse pancreatic tumor cell clones were tested by the Research Animal Diagnostic Laboratory (RADIL) at the University of Missouri (Columbia, MO), using the Infectious Microbe PCR Amplification Test (IMPACT) II. Tumor cells were cultured in DMEM (high glucose without sodium pyruvate) with 10% FBS (Gibco) and glutamine (2 mmol/L). The clones were regularly tested using the MycoAlert Mycoplasma Detection Kit (Lonza). Mouse pancreatic tumor cell clones 6419c5, 6694c2, and 6422c1 were generated in our laboratory (9) and were used for less than 18 passages. These tumor cell clones were derived from the KPCY mice (as described above) and, thus, expressed YFP at the time of generation. We used 293T cells (ClonTech, 632180) for lentivirus packaging.

Pancreatic tumor cells (as described above in the “Tumor cells” section) were dissociated into single cells with 0.25% trypsin (Gibco), washed with serum-free DMEM twice, and counted in preparation for subcutaneous or orthotopic implantations. A total of 2.0 × 10⁵ tumor cells were implanted subcutaneously and 5.0 × 10⁴ tumor cells were implanted orthotopically into the pancreas of 6- to 8-week-old female C57BL/6 mice via laparotomy as described previously (9). Tumors were harvested 18–24 days following implantation. Endpoint criteria included tumor volume exceeding 500 mm³, severe cachexia, or weakness and inactivity.

For tumor growth kinetics, tumors were measured every 3 days. Tumor length and width were measured with calipers, and tumor volumes were then calculated as length × width²/2. Tumor volumes of 500 mm³ were used as an endpoint for survival analysis. Tumor regressions and were calculated using the initial tumor size at the start of treatment to tumor size 15 days later.

A total of 1.0 × 10⁵ Usp22-WT and Usp22 knockout (Usp22-KO) tumor cells were injected via tail vein into C57BL/6 mice. After 14 days, lungs were harvested, weighed, and subjected to flow cytometry for immune profiling as described below in the flow analysis section. The tail vein injection experiment was performed in one experiment for each using two CRISPR KO clones, with 5 mice per group for analysis.

Gemcitabine (Hospira) and nab-paclitaxel (Abraxane; Celgene) were purchased from the Hospital of the University of Pennsylvania Pharmacy. Gemcitabine (G) was procured as pharmaceutical grade suspension at 38 mg/mL, and further diluted to 12 mg/mL in PBS and administered at 120 mg/kg via intraperitoneal injection. Abraxane (A) was purchased as a pharmaceutical-grade powder resuspended at 12 mg/mL in PBS for intraperitoneal injection at dose of 120 mg/kg. Vehicle control mice received the equivalent to nab-paclitaxel dose of human albumin (huAlb; Sigma). Chemotherapy was injected when tumor was 3 to 5 mm. For anti-CD40 agonist treatment, mice were injected intraperitoneally with 100 μg of either agonistic CD40 rat anti-mouse IgG2a (clone FGK45, endotoxin free) or the isotype control IgG2a (clone 2A3, Bio X Cell) 48 hours after chemotherapy. For checkpoint blockade treatment, mice were injected intraperitoneally with 200 μg of anti–PD-1 (clone RMP 1-14, Bio X Cell) and 200 μg anti–CTLA-4 (clone 9H10, Bio X Cell), starting day 10 (therapy start timepoint), with six and three doses, respectively. Control mice received the isotype control IgG2a (clone 2A3, Bio X Cell) on treatment days. CD4⁺ and CD8⁺ T cells were depleted using intraperitoneal injections of 200 mg anti-CD4 (clone GK1.5, Bio X Cell) and anti-CD8 (clone 2.43, Bio X Cell), three days prior tumor implantation, and three times a week for the duration of the experiment. Control groups received IgG2b isotype control (Bio X Cell). The treatment experiment was performed in one experiment for each using two CRISPR KO clones, with 7 to 8 mice per group for analysis.

RNA sequencing (RNA-seq) data is deposited at GEO (GSE140088). RNA samples were extracted from sorted YFP⁺ tumor cells from subcutaneous tumors 21 days postimplantation using the Qiagen RNeasy Micro Kit following manufacturer's instructions. RNA was sent to Novogene, for library preparation and high-throughput sequencing using Illumina sequencers (Hiseq 2500) to generated paired-end 150 bp data. Fastq files were checked for quality using FastQC (Babraham). Raw counts of gene transcripts were obtained using alignment-independent tool, Salmon (https://combine-lab.github.io/salmon/), using standard settings. The raw count matrix was subsequently imported into R-studio (R version 3.3.3) and used as input for DESeq2 following the vignette of the package for normalization and differential gene expression analysis. Salmon was also used to normalize and quantitate gene expression in transcripts-per-million (tpm) through quasi-alignment. Differentially expressed genes from the DESeq2 analysis were used as input for MSigDB gene set enrichment analysis (GSEA). Differentially expressed genes (P_adj < 0.01 and absolute fold change > 2) were used as input for EnrichR analysis (ENCODE and ChEA Consensus TFs from ChIP-X and ChEA datasets).

The short hairpin RNA (shRNA) knockdown vector, pLKO.1.puro, was a gift from Robert Weinberg (Addgene plasmid #8453). The CRISPR vector, lentiCRISPR v2, was a gift from Feng Zhang (Addgene plasmid # 52961). The vector and pVSVg and psPAX2 lentiviral packaging plasmids (Addgene) were cotransfected into 293T cells (ClonTech, 632180) using PEI reagent (Polysciences, 23966-2). Lentiviral particles were collected 48 or 72 hours after transfection and filtered for usage. Tumor cells transduced either with Cas9-Puro (control) or Cas9-guide-Puro (KO; control is from Addgene, #52961, and KOs were cloned following the instruction from Addgene) were selected with 8 μg/mL puromycin (Invitrogen, A1113803). Single-cell clones were picked from bulk KO cell line using single-cell sorting using a BD Jazz FACS sorter into 96-well plates. KO efficiencies were assessed by gene-specific qPCR analysis and immunofluorescence to detect target proteins (staining methods described below). shRNA sequences used were:

• Atxn7l3-sh1-F-

• CCGG-AGGCGAACCGTACGGATTTAT-CTCGAG-ATAAATCCGTACGGTTCGCCT-TTTTTG;

• Atxn7l3-sh1-R-

• AATT-CAAAAA-AGGCGAACCGTACGGATTTAT-CTCGAG-ATAAATCCGTACGGTTCGCCT; Atxn7l3-sh2-F-

• CCGG-TCGAAGATCCAAGTCTCTAAA-CTCGAG-TTTAGAGACTTGGATCTTCGATTTTTG; Atxn7l3-sh2-R-

• AATT-CAAAAA-TCGAAGATCCAAGTCTCTAAA-CTCGAGTTTAGAGACTTGGATCTTCGA.

CRISPR sgRNA sequences used were:

• Usp22-A-TCT-GCG-TGG-ACT-GAT-CAA-CC;

• Usp22-B-AGT-TCC-AGC-TCC-CGT-TTA-GT.

RNA was prepared from cultured tumor cells using RNeasy Mini Kit or RNeasy Micro Kit (Qiagen). cDNA was generated using High-capacity cDNA Reverse Transcription Kit from 1 μg RNA in 20 μL reaction volume and diluted 1:10 for qPCR analysis (Life Technologies). qPCR analysis was performed using 2 μL diluted cDNA with biological (2–3) and technical replicates (2–3) using SsoAdvanced SYBR reagent (Bio-Rad) and Bio-Rad qPCR platform, and results were normalized to the expression of Tbp using the Bio-Rad software. Primer sequences utilized for qPCR were: Usp22-F-CTC-CCC-ACA-CAT-TCC-ATA-CAA-G; Usp22-R-TGG-AGC-CCA-CCC-GTA-AAG-A; Atxn7l3-F-TTG-TCT-GGC-CTG-GAT-AAC-AGC; Atxn7l3-R-CCG-GTG-TAC-TTC-AAA-GCA-GAA-TC; Tbp-F-AGA-ACA-ATC-CAG-ACT-AGC-AGC-A; Tbp-R-GGG-AAC-TTC-ACA-TCA-CAG-CTC.

For the flow cytometric analyses, subcutaneous or orthotopic tumors following 18–24 days of implantation were chopped into small pieces and digested in collagenase (1 mg/mL in DMEM; Sigma-Aldrich) at 37°C for 45 minutes and filtered through a 70-μm cell strainer. Single-cell suspensions were stained with antibodies on ice for 30 minutes and washed twice with PBS with 5% FBS for flow cytometric analysis. No intracellular staining is needed for this analysis. Cells were then analyzed by flow cytometry using BD FACS (BD Biosciences) and FlowJo software (Treestar).

The following antibodies used were for the analysis: CD279 (PD-1) FITC (29F.1A12; BioLegend 135214), CD335 (NKp46) PE (29A1.4; BioLegend 137604), CD103 PE/Dazzle 594 (2E7; BioLegend 121430), CD3e PE/Cy5 (145-2C11; BioLegend 100310), CD45 AF700 (30-F11; BioLegend 103128), CD8a PE/Cy7 (53-6.7; BioLegend 100722), I-A/I-E (MHCII) PE/Cy7 (M5/114.15.2; BioLegend 107630), Ly-6G V450 (1A8; BD Biosciences 560603), H-2Kb/H-2Db (MHCI) AF647 (28-8-6; BioLegend 114612), F4/80 APC/Cy7 (BM8; BioLegend 123118), CD11b PerCP-Cy5.5 (M1/70; BD Biosciences 550993), CD11c BV605 (N418; BioLegend 117334), Ly-6C BV570 (HK1.4; BioLegend 128030), and CD4 BV650 (RM4-5; BioLegend 100546).

The following gating strategies were used for immune cells: myeloid cells—live CD45⁺CD11b⁺; granulocytic myeloid-derived suppressor cells (gMDSC)/neutrophils—live CD45⁺CD11b⁺Gr1⁺; macrophages—live CD45⁺F4/80⁺CD11b⁺; CD11c⁺ dendritic cells (DC)—live CD45⁺F4/80^– CD11c⁺; CD103⁺ DCs—live CD45⁺CD11b^–F4/80^–CD11c⁺ CD103⁺; T cells—live CD45⁺CD11b^–F4/80^–NKp46^–CD3⁺; CD4⁺ T cells—live CD45⁺CD11b^–F4/80^–NKp46^–CD3⁺CD4⁺; CD8⁺ T cells—live CD45⁺CD11b^–F4/80^–NKp46^–CD3⁺CD8⁺. The flow analysis of immune infiltration of USP22-WT versus KO tumors were performed in two experiments using six CRISPR KO clones, with 4 to 5 mice per tumor cell clone.

Tumor cells were incubated with EdU for 3 hours in DMEM with 10% FBS and Glutamax (Thermo Fisher Scientific 35050061) and then fixed with fixation buffer following the instruction of the reagent (eBioscience 00-5123-43 and 00-5223-56). Samples were further processed for EdU staining as per protocol (Thermo Fisher Scientific, C10086). Samples were stained for EdU and then analyzed by flow cytometry using BD FACS LSR machine (BD Biosciences) and FlowJo software (Treestar).

Usp22-WT and Usp22-KO tumor cells were treated with 100 ng/mL of IFNγ (PeproTech) in DMEM with 10% FBS and Glutamax (Thermo Fisher Scientific 35050061) for 24 hours. Tumor cells were trypsinized from culture plates and resuspended in PBS with 5% FBS for staining of antibodies. Single-cell suspensions were stained with antibodies on ice for 30 minutes and washed twice with PBS with 5% FBS for flow cytometric analysis. Samples were stained for MHCI (BioLegend 114612) and then analyzed by flow cytometry using BD FACS LSR machine (BD Biosciences) and FlowJo software (Treestar).

For CD3, USP22, and Gr1 staining, collected implanted tumor tissues were fixed in Zn-formalin for 24 hours and embedded in paraffin. Sections were deparaffinized, rehydrated, and prepared by antigen retrieval for 6 minutes each, and then blocked in 5% donkey serum (Sigma, D9663) for 1 hour at room temperature, incubated with primary antibodies overnight at 4°C, washed with 0.1% PBST (PBS with Tween-20), incubated with secondary antibodies for 1 hour at room temperature, and then washed and mounted. Slides were visualized and imaged using an Olympus IX71 inverted multicolor fluorescent microscope and a DP71 camera. For CD3 and Gr1 staining quantification, stained cells were counted for CD3⁺ T cells manually in 5–8 fields per sample.

The following primary antibodies: CD3 (Abcam ab5690, 1:100 dilution), USP22 (Abcam ab195289, 1:100 dilution), Gr-1 (eBioscience 14-5931-85, 1:50 dilution), YFP (Abcam, ab6673). Secondary antibodies were purchased from Invitrogen (A-11055, A-21207, A-21209) and were used as 1:250 dilution for all staining.

The DepMap portal (https://depmap.org/portal/) and the CRISPR (Avana) Public 19Q3 dataset were used for this analysis. No samples were excluded from the dataset, and 625 cell lines in total were included in this analysis. Following the DepMap instruction, the dependency scores of genes were downloaded and plotted as dot plots. Pearson correlation was calculated for all the plots using Prism.

Statistical comparisons between two groups were performed using Student unpaired t test. For comparisons between multiple groups, one-way ANOVA with Tukey's HSD post-test was used. For survival comparison between two groups, log-rank P values of Kaplan–Meier curves were determined in GraphPad Prism 8 (GraphPad). On graphs, error bars represent either range or SEM, as indicated in legends. For all figures, P < 0.05 was considered statistically significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

PRISM software and R were used for the data processing, statistical analysis, and result visualization (http://www.graphpad.com). The R language and environment for statistical computing and graphics (https://www.r-project.org) was utilized in this study for the statistical and bioinformatics analysis of RNA-seq data. The R packages used for all analysis described in this manuscript were obtained from the Bioconductor and CRAN.

We previously established an experimental system that recapitulates the heterogeneity of immune cell infiltration in pancreatic cancer (9). This platform consists of a group of primary murine PDA tumors stratified into two main subsets: a T-cell–low group, enriched for myeloid-derived suppressor cells (MDSC), and a T-cell–high group, enriched for T cells and DCs, with reduced MDSCs. We and others have previously demonstrated that several tumor cell–intrinsic factors, including CSF2, CXCL1, CSF3, EPHA2, and PTGS2, can control the abundance of tumor-infiltrating myeloid cells and T cells, which, in turn, affects the sensitivity of PDA tumors to combination immunotherapy (8). In these studies, we found that mechanisms underlying the establishment of the non-T-cell–inflamed TME varied from tumor to tumor. Here, we aimed to identify alternative tumor cell–intrinsic regulators of the immune TME that might lead to novel therapeutic targets for precision immunotherapy.

Ablation of USP22 in liver tumor cells has been shown to increase tumor immunogenicity and promote infiltration of T cells into the resulting liver tumors (13). An examination of our PDA clone library revealed that T-cell–low tumors expressed higher USP22 at the RNA and protein level compared with T-cell–high tumors (Supplementary Fig. S1A and S1B). These findings are consistent with the idea that USP22 promotes the non-T-cell–inflamed TME and suppresses antitumor immunity in PDA. To functionally examine the role of USP22 in regulating the immune TME, we utilized CRISPR/Cas9 strategy to generate six independent Usp22-KO cell clones (Supplementary Fig. S1C) from a well-established T-cell–low PDA tumor cell clone, 6419c5. We then performed flow cytometric analysis to assess the phenotype of immune cells that infiltrated into subcutaneously implanted Usp22-WT and Usp22-KO tumors. USP22 ablation resulted in decrease in total myeloid cells and gMDSCs and an increase in CD3⁺, CD4⁺, and CD8⁺ T cells in the TME (Fig. 1A and B). CD8⁺ T cells in Usp22-KO tumors were more frequently positive for CD44 and PD-1 (Fig. 1C). These changes in myeloid and T-cell populations within the TME were confirmed by immunofluorescence staining of CD3 and Gr1 (Fig. 1D and E). Usp22-KO tumors also exhibited increased natural killer (NK)–cell infiltration but no significant change in the number of macrophages or DCs (Supplementary Fig. S1D).We also observed a minor decrease of tumor growth in vivo and 10% decrease of EdU incorporation in vitro following ablation of USP22 (Supplementary Fig. S1E and S1F), consistent with previous reports suggesting that USP22 regulates cell-cycle progression and tumor growth (14–16). Likewise, T-cell depletion had no effect on the growth of either Usp22-WT or Usp22-KO tumors (Supplementary Fig. S1G and S1H), suggesting that the growth effects of Usp22 ablation are not T-cell–dependent.

To expand these observations to other tumors, we generated Usp22 KOs from two other T-cell–low tumor cell clones, 6422c1 and 6694c2. Flow analysis of the implanted tumors showed that loss of tumor cell–intrinsic USP22 decreased the abundance of gMDSCs and increased the number of CD3⁺ and CD4⁺ T cells in the TME of both 6422c1 and 6694c2 tumors (Supplementary Fig. S1I–S1L). A nonsignificant trend toward increased CD8⁺ T cells was seen in Usp22-KO 6422c1 tumors (Supplementary Fig. S1H, right) and no difference seen in Usp22-KO 6694c2 tumors (Supplementary Fig. 1J, right). These results suggested that different T-cell–low tumors depended on USP22 to regulate CD8⁺ T-cell infiltration to different degrees, further highlighting the importance of identifying patient-specific regulators of the immune TME.

Next, we assessed the infiltration of immune cells in tumors arising from 6419c5 Usp22-WT and Usp22-KO tumors that were orthotopically injected into the pancreas. Flow cytometry and immunofluorescence staining revealed that Usp22 deletion resulted in a significant decrease of total myeloid cells and gMDSCs and an increase of CD3⁺, CD4⁺, and CD8⁺ T cells in the TME, reproducing the findings from subcutaneous tumors (Fig. 2A–D; Supplementary Fig. S2A). Collectively, these results indicated that tumor cell–intrinsic USP22 establishes a non-T-cell–inflamed immune TME dominated by myeloid cells and a paucity of tumor-infiltrating T cells.

Studies have demonstrated the significant role of both adaptive and innate immune cells in regulating the metastasis of tumor cells, including pancreatic tumor cells (9, 17). Using the orthotopic model, we observed that mice injected with 6419c5 Usp22-KO tumors harbored fewer lung metastases compared with mice injected with 6419c5 Usp22-WT tumors (Fig. 2E). To functionally examine the effect of tumor cell–intrinsic USP22 on lung metastatic colonization, we performed tail vein injection of Usp22-WT and Usp22-KO tumor cells into immunocompetent mice and observed that Usp22-KO tumor cells resulted in decreased metastatic burden compared with the Usp22-WT tumor cells (Fig. 2F). T-cell depletion had no effect on metastatic burden in mice bearing Usp22-WT tumors but led to an increased metastatic burden in mice bearing Usp22-KO tumor cells (Fig. 2F; Supplementary Fig. S2B). Metastatic lung lesions from Usp22-KO tumors exhibited fewer infiltrating myeloid cells and more T cells (Fig. 2G; Supplementary Fig. S2C). Similar results on metastatic burden and immune infiltration were seen when another T-cell–low tumor cell clone, 6694c2, was introduced by tail vein injection (Supplementary Fig. S2D and S2E). These results suggest that Usp22-KO tumor cells have a reduced ability to form metastases secondary to T-cell surveillance.

We previously reported that relative to T-cell–high tumors, T-cell–low tumors have decreased sensitivity to combination immunotherapy with GAFCP (gemcitabine, G; nab-paclitaxel, A; anti-CD40 agonist, F; anti–CTLA-4, C; and anti–PD-1, P). A similar combination therapy, including chemotherapy, checkpoint blockade, and CD40 agonist is currently tested in two clinical trials for patients with PDA (NCT02588443, NCT03214250). Given that loss of tumor cell–intrinsic USP22 resulted in increased T-cell infiltration, we assessed the sensitivity of 6419c5 Usp22-WT and Usp22-KO tumors to this combination. USP22-deficient PDA tumor cells exhibited a significantly better response to GAFCP therapy, resulting in tumor regressions and prolonged survival (Fig. 3A–C). Collectively, these experiments suggest that USP22 expression in PDA tumors cells suppresses antitumor immunity and confers resistance to immunotherapy.

USP22 has been reported to regulate the immune TME by controlling the protein stability of PD-L1 in liver tumor cells (13). Immunofluorescent staining revealed that USP22 protein is principally located in the nucleus of PDA tumor cells (Fig. 1E; Supplementary Fig. S1B; Supplementary Fig. S3A), consistent with data from the human protein atlas (18, 19). We, therefore, hypothesized that USP22 might influence tumor immunity through nuclear functions independent of its effects on PD-L1 protein stability.

USP22 is a major component of the SAGA/STAGA complex, a multiprotein complex involved in transcriptional regulation (14). As the major deubiquitylase in the deubiquitylation module of this complex, USP22 is thought to regulate transcription by controlling ubiquitin levels of two histones—H2A and H2B—as well as other transcription factors including c-MYC, thereby regulating their stabilization and/or activity (14, 20–25). Analyzing the data from the Cancer Dependency Map portal (DepMap), we found USP22 shared dependencies with other factors in the deubiquitylation module of the SAGA/STAGA complex, including ATXN7, ATXN7L3, KAT2A, TADA1, TAF5L, and TAF6L (Fig. 4A). To further explore whether these other components of the SAGA/STAGA complex regulated the immune TME, we used shRNA to generate 6419c5 Atxn7l3-knockdown cell lines and performed flow analysis on the resulting tumors. Knockdown of Atxn7l3 resulted in a decrease in total myeloid cells and gMDSCs as well as an increase in CD3⁺, CD4⁺, CD8⁺ T cells, NK cells, and DCs (Fig. 4B; Supplementary Fig. S3B and S3C). Likewise, Atxn7l3-knockdown tumors had a higher frequency of PD-1⁺ CD8⁺ T cells (Supplementary Fig. S3C). These results were confirmed by knocking down Atxn7l3 in a second T-cell–low clone, 6694c2 (Fig. 4C; Supplementary Fig. S3D and S3E). These data suggested that at least one other component of the deubiquitylation module of the nuclear SAGA/STAGA complex, ATXN7L3, may regulate the immune TME of PDA tumors.

On the basis of these findings, we hypothesized that ablation of tumor cell–intrinsic USP22 would result in significant transcriptional changes in tumor cells, thereby contributing to USP22's regulation of the immune TME. We used a YFP-lineage marker to sort the cancer cells from implanted tumors and then performed RNA-seq. Usp22-WT and Usp22-KO tumor cells exhibited global differences in transcription (Supplementary Fig. S4A). GSEA of genes whose expression changed with Usp22 loss revealed an enrichment of hallmark gene sets associated with naturally arising T-cell–low and T-cell–high tumor cells (9). Specifically, Usp22-KO tumor cells were enriched for IFN response signatures, whereas Usp22-WT tumor cells were enriched for MYC, TGFβ, and cell-cycle signatures (Fig. 4D and E; Supplementary Fig. S4B–S4D). We checked the expression of a group of factors regulating function and trafficking of myeloid cells and regulatory T cells and found that most of them were not significantly differentially expressed between Usp22-WT and Usp22-KO tumor cells. We found Csf3 was significantly decreased in Usp22-KO tumor cells. In our previous publication (9), we identified CSF3 as one tumor cell–intrinsic regulator promoting the establishment of a non-T-cell–inflamed TME. One article (13) has demonstrated that USP22 can regulate PD-L1 at the posttranslational level to control tumor immunity in liver cancer. Here, we assessed the expression of both PD-L1 and MHCI on tumor cell surface with or without treatment of IFNγ and found that PD-L1 was not differentially expressed between Usp22-WT and Usp22-KO tumor cells (Supplementary Fig. S4E). However, consistent with our RNA-seq result, Usp22-KO tumor cells had increased expression of MHCI on cell surface compared with Usp22-WT tumor cells (Supplementary Fig. S4E). We then used the list of genes that were differentially expressed between Usp22-WT and Usp22-KO tumor cells as input for EnrichR (26, 27). EnrichR is an analytic tool that uses information from the Encyclopedia of DNA Elements (ENCODE) and ChIP-X Enrichment Analysis (ChEA) datasets to identify potential transcriptional regulators. This analysis revealed an enrichment of an IRF8 signature in Usp22-KO tumor cells (Fig. 4F; Supplementary Fig. S4F). This analysis nominated SUZ12 and EZH2, two key components of the polycomb repressive complex 2 (PRC2 complex), as transcriptional regulators active in Usp22-WT tumor cells whose activity is lost following Usp22 deletion (Supplementary Fig. S4F). Together, these evidences suggest potential mechanisms underlying USP22-mediated regulation of the immune TME of pancreatic cancer. Collectively, these data demonstrated that loss of tumor cell–intrinsic USP22 reprogramed the tumor cell transcriptome, resulting in a T-cell–high phenotype at molecular and cellular levels.

Understanding the determinants of T-cell infiltration into tumors has important therapeutic implications. We and others have shown that regulatory factors operating in tumor cells play a central role in determining the makeup of their surrounding microenvironments (5, 8, 28). Here, we reported that USP22, a highly conserved component of the nuclear multiprotein SAGA/STAGA complex, is one such factor. These findings expand our understanding of tumor-intrinsic epigenetic factors that shape the TME (29, 30) and are in line with a report describing a role for USP22 in regulating the immune microenvironment in liver cancer (13).

The SAGA/STAGA complex promotes transcription, an activity mediated in part through a deubiquitylation module that contains USP22 and its binding partners ATXN7L3, ATXN7, and ENY2 (24, 31–34). We found that Atxn7l3 knockdown phenocopied Usp22 loss, supporting the notion that USP22-dependent effects on the immune microenvironment are mediated at least, in part, by the SAGA/STAGA complex. Deletion of Usp22 associated with the decreased expression of genes enriched for consensus target gene signatures of SUZ12 and EZH2, two components of the repressive PRC2 complex. These findings raise the possibility that the SAGA/STAGA and PRC2 complexes act coordinately to regulate the activation and repression of genes that control the immune makeup of PDA tumors.

Usp22 is overexpressed in multiple tumor types, has been shown to regulate cell-cycle activity in association with cancer progression (14, 15, 35–38), and has been reported to play a role in antitumor immunity through the stabilization of PD-L1 in tumor cells (13). This study suggests that transcriptional regulation may be another mechanism by which Usp22 controls the immune TME, thereby converting cells that are fully resistant to immunotherapy to a sensitive state. To date, no small-molecule inhibitors of USP22 have been reported, but given mounting evidence in multiple contexts that USP22 promotes tumor progression, the development of such inhibitors would be desirable. Such compounds may be useful in the context of novel immunotherapy combinations, such as those described in this study and currently under clinical investigation for PDA (NCT03214250).

No potential conflicts of interest were disclosed.

Conception and design: J. Li, R.J. Norgard, B.Z. Stanger

Development of methodology: J. Li, R.J. Norgard, B.Z. Stanger

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Li, S. Yuan, F. Yan, T. Yamazoe, B.Z. Stanger

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Li, S. Yuan, R.J. Norgard, F. Yan, A. Blanco, B.Z. Stanger

Writing, review, and/or revision of the manuscript: J. Li, F. Yan, B.Z. Stanger

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Li, R.J. Norgard, F. Yan

Study supervision: B.Z. Stanger

This work was supported by grants from the NIH (R01-CA229803, to B.Z. Stanger) and the Blavatnik Family Foundation Fellowship (to J. Li). We thank members of the Stanger laboratories for technical help and scientific discussions.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.