The Pellino1–PKCθ Signaling Axis Is an Essential Target for Improving Antitumor CD8⁺ T-lymphocyte Function

Tumors are heterogeneous hematologic malignancies, and the complexities of the microenvironment, in tandem with defective signaling milieus, represent a formidable challenge for the development of targeted therapeutics. Therefore, considering the heterogeneous genetic and epigenetic backgrounds inherent in cancer development, identifying a common signaling axis that directly targets both cancer cells and their microenvironment would appear to be a sound antitumor strategy.

The tumor microenvironment (TME) is an interactive supporting network of various components, including blood vessels, cytokines, chemokines, and immune cells, which sustain tumor cells' survival and growth. The TME arises from interactions between tumor cells, immune cells, and cancer-associated stromal cells (1). As a result of these interactions, the TME is associated with immunosuppressive mechanisms, contributing to immune evasion in a number of ways, for example, by secreting cytokines. Prominently among these are TGFβ, forkhead box O3 (FOXO3), COX-2, VEGF, and interleukins (IL6, IL10, and others), and in sum lead to T-cell dysfunction (2).

Therapeutic strategies to counter immunosuppressive mechanisms are already seeing encouraging clinical trial results (3). Among these, immune checkpoint blockade, antibodies targeting CTLA-4 or the programmed cell death protein-1/PD1 ligand 1 (PD-1/PD-L1) axis, which leads to expansion of activated T cells, has had the broadest impact on patients with melanoma, non–small cell lung cancer (NSCLC), urothelial bladder cancer, and renal cell cancer (4–7). This has led to remarkable clinical success, albeit in a small proportion of patients (5, 8). Adoptive T-cell therapy (ACT) with T cells expressing chimeric antigen receptors (CAR), called CAR T cells, have demonstrated remarkable efficacy (9). However, therapy is also more effective if combined with an approach that alters the immune-suppressive TME (3).

CD8⁺ T cells play important roles in the elimination of pathogens, tumors, and transplanted tissues. Prolonged exposure of T cells to their cognate antigen results in T-cell dysfunction, all of which favors tumor progression and invasiveness. This dysfunctional state is related to the coexpression of multiple immune checkpoint inhibitory receptors and loss of effector functions, including the secretion of IL2, TNFα, and IFNγ (10). However, the precise mechanisms underlying T-cell dysfunction have not been fully determined. The antitumor function of CD8⁺ T cells is dependent on two critical factors: (i) differentiation; and (ii) infiltration into the tumor site, which occurs by trafficking and subsequent sequestration of CD8⁺ T cells into the TME (11). Researchers have linked elevated infiltration of cytotoxic CD8⁺ T cells in the TME with positive antitumor effects in many types of cancers (12, 13). Thus, elevations in cytotoxic CD8⁺ T cells in the TME are usually associated with a favorable prognosis in various cancers. Nevertheless, the underlying mechanisms involved in eliciting and maintaining effector response in CD8⁺ T cells remain elusive.

Pellino1 (Peli1, pellino E3 ubiquitin protein ligase 1) is the key regulator of pattern recognition receptor (PPR) signaling and acts as a signal transducer in immune responses (14). Studies have unveiled a critical role for Peli1 in activating toll-like receptors (TLR) and T-cell receptor (TCR) signaling to mediate inflammation and autoimmunity (15–17). Peli1-deficient T cells are hyper-responsive to stimulation via TCR-CD28 and develop spontaneous autoimmunity in mice. Previously, we found that tumorigenesis induced by aberrant Peli1 expression is associated with the promotion of oncogenic signaling through activation of oncogene BCL6 and inactivation of tumor suppressor BubR1 (18, 19). Peli1 expression is also closely associated with poor prognosis in B-cell lymphoma and lung cancer (18, 20). Here, we sought to elucidate the potential significance of Peli1 in regulating the function of effector cells in the TME. Our findings revealed that Peli1 mediated immune-suppressive signaling by counteracting PKCθ-mediated CD8⁺ T-cell function.

All animal experiments were conducted in accordance with a protocol approved by Institutional Animal Care and Use Committee (IACUC 2018–08–29–2, IACUC 2019–10–08–1, and IACUC 2020–11–03–1) of Sungkyunkwan University School of Medicine (SUSM). SUSM is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC International), abiding by the Institute of Laboratory Animal Resources (ILAR) guidelines.

Peli1 knockout (KO) mice were provided from Ho Lee (National Cancer Center, South Korea). With the use of homologous recombination in C57BL/6NTac embryonic stem cells, a cassette, including the coding sequence of LacZ gene followed by a promoter-driven neo marker surrounded by flippase recognition target (FRT) sites, was introduced between exon 1 and exon 2 of Peli1 by the European Mouse Mutagenesis Consortium (EUCOMM) Program. Both neo and exon 2 of Peli1 were flanked by loxP sites. Details of the allele (Peli1 “knock out-first-reporter tagged insertion” allele, referred as Peli1^tm1a) and targeting strategy can be found at EUCOMM. To obtain Peli1-KO mice, mice carrying the Peli1^tm1a allele were crossed with B6.FVB-Tg(Zp3-Cre)3Mrt/J mice to excise the floxed promoter-driven neo selection cassette and obtain Peli1^lacZ/+ mice. After one backcross to eliminate the transgene encoding Cre, the Peli1^lacZ/+ line was intercrossed to generate experimental Peli1^+/+ (referred to as Peli1-WT) and Peli1^lacZ/lacZ (referred to as Peli1-KO) littermates.

pTRE TetO Myc-Peli1 transgenic mice were provided from Ho Lee (National Cancer Center, South Korea). To generate doxycycline-inducible human PELI1-transgenic mice (rtTA-Peli1), we crossed pTRE TetO Myc-Peli1 transgenic mice with R26-M2rtTA mice (B6.Cg-Gt(ROSA)26Sortm1(rtTA*M2)Jae/J; The Jackson Laboratory). These animals were maintained under specific pathogen-free conditions. All experiments were performed on mice homozygous for both the R26-M2rtTA transgene and human PELI1 transgene. To induce expression of the human PELI1 transgene, 4-week-old rtTA-Peli1 mice were provided drinking water containing 2 mg/mL doxycycline (Sigma) and 5% sucrose (Sigma) for the indicated times. Water was protected from light and exchanged every three days.

OT-1 transgenic mice were provided by Sang-Jun Ha (University of Yonsei, South Korea). Peli1-KO mice were crossed with OT-1 mice to obtain Peli1-KO/OT-1 mice. Three to 8 generations were used for experiments. SJL mice were purchased from The Jackson Laboratory, United States. Six-week-old WT C57BL/6 male mice were purchased from ORIENT BIO Inc. CD45.1⁺ recipient mice were purchased from The Jackson Laboratory (United States) and crossed in-house with each other. We used the mice after three generations. All mice were on a C57BL/6J background, and genotyping was performed with Ready-2X-GO polymerase mixture (HelixAmp), using genomic DNA isolated from tail biopsy samples. Eight- to 10-week-old mice were used for experiments, and animals were maintained under specific pathogen-free conditions.

EL4 cells were purchased from KCLB (Korea Cell Line Bank), and EL4 and MC38 cells were provided by Seung-Woo Lee (Pohang University of Science and Technology, South Korea) and were maintained in DMEM medium (Welgene) supplemented with 10% FBS (GenDEPOT) and 1% antibiotics [10,000 U/mL penicillin, 10 mg/mL streptomycin, and 25 μg/mL amphotericin B in 0.85% NaCl (Welgene)]. LLC and EG7 cells were provided by Seung-Woo Lee and maintained in RPMI medium (Welgene) supplemented with 10% FBS and 1% antibiotics (as described). Jurkat cell and HeLa cell lines were obtained from the ATCC and maintained in DMEM supplemented with 10% FBS and 1% antibiotics (as described). Sf9 cells were purchased from ATTC and maintained in SF-900-II SFM medium (Gibco) supplemented with 1% antibiotics (as described). Sf9 cells were grown at 27°C in a humidified environment with shaking at 120 rpm and maintained in a 150 mL or 250 mL sterile flask. Titers ranged from 1–8 × 10⁶ cells/mL, and cells with a viability of 96%–100% were used for experiments. All cell lines were expanded and frozen stocks made after 2 to 3 passages. No more than 5 passages for EL4, MC38, LLC and EG7 cells, no more than 8 passages for Jurkat and HeLa cells, and no more than 15 passages for Sf9 cells were used before new stocks were thawed for experimental use. There was no further authentication or Mycoplasma contamination prior to all experiments.

EL4 (5 × 10⁵), EG7 (5 × 10⁵), MC38 (1 × 10⁶), or LLC (3 × 10⁶) cells per 100 μL PBS were injected subcutaneously into the flank of 8-week-old WT, Peli1-KO, Peli1-KO/OT-1, rtTA and rtTA-Peli1 mice. WT mice were used as controls for Peli1-KO and Peli1-KO/OT-1 mice, and rtTA mice were used as controls for rtTA-Peli1 mice. Tumors were monitored until 18–23 days after inoculation; 14–18 days after inoculation, mice were sacrificed, and blood (aorta), spleen, lymph nodes, and tumors were collected. Both female and male mice were employed after randomization to the various experimental groups. Tumor size was determined by caliper measurements 3 times a week, and the tumor volume was calculated according to the formula: volume = 0.5 × length × width² (width is the shorter axis and length is the longer axis).

CD8⁺ T cells, CD4⁺ T cells, and macrophages were depleted during EL4 tumor growth by intraperitoneal injection of 200 μg of anti-IgG2b control (BE0090, Bioxcell), anti-CD8 [(YTS169), BE0117, Bioxcell], 300 μg of anti-CD4 [(GK1.5), BE003–1, Bioxcell], or anti-CSF1R [(AFS98), BE0213, Bioxcell], respectively, every two to three days, starting one day prior to tumor inoculation.

Single-cell suspensions were prepared from tumors, spleens and inguinal lymph nodes. For isolation of TILs, tumors were excised 14–18 days after inoculation. Tumor tissues were minced into small pieces (<1 mm pieces) and digested with 1 mg/mL collagenase Type IV (Sigma), 1 mg/mL dispase (genDEPOT), and 10 μg/mL DNase I (Sigma) in DMEM for 30 minutes at 37°C in a shaker. The dissociated tumor cells were centrifuged at 50 × g for 10 minutes at 4°C, and supernatant was collected. TILs were collected after centrifuging at 1,200 rpm for 4 minutes at 4°C for use in further experiments. Red blood cells were removed using red blood cell lysis buffer (eBioscience). For isolation from spleens and lymph nodes, tissues were passed through 70-mm strainers (Sartorius) with a syringe plunger. Cells were suspended in complete medium [RPMI-1640 (Welgene) supplemented with 10% FBS (GenDEPOT)] and centrifuged for 5 minutes at 300 × g, and red blood cells were removed using red blood cell lysis buffer (eBioscience).

CD8⁺ T cells from spleen and tumors of wild-type (WT) or Peli1-KO mice were sorted using a BD FACS Aria II. EL4 cells were labeled with 1 μmol/L CFSE (eBioscience) and pulsed with 1 μmol/L OVA257–264 (InvivoGen) at 37°C for one hour. 10⁵ CD8⁺ T cells were then cultured with EL4 cells at 1:1 ratio and incubated for 5 hours. Cells were stained with annexin V and analyzed by flow cytometry, as described below (Supplementary Table S1).

EG7 cells are derived from EL4 cells and express chicken ovalbumin (OVA) as a tumor antigen. MC38 is a solid tumor predominantly composed of immunosuppressive cell types, such as monocytic myeloid-derived suppressor cells (M-MDSCs; ref. 21). Congenic CD45.1⁺ SJL (8–10-week-old) recipient mice were injected subcutaneously with 5 × 10⁵ EG7 cells. CD8⁺ T cells were isolated from spleen and lymph nodes (dissociated as described above) of WT OT-1 or Peli1-KO/OT-1 mice using the Magnisort mouse CD8⁺ T cell enrichment kit (Thermo Fisher Scientific). Isolated CD8⁺ T cells were stimulated with plate-coated anti-CD3 (5 μg/mL, eBioscience), soluble anti-CD28 (2 μg/mL, Invitrogen), and OVA257–264 peptide (1 μg/mL, InvivoGen) in RPMI medium supplemented with IL2 (20 ng/mL, R&D Systems) for 48 hours. 5 × 10⁶ WT OT-1 or Peli1-KO OT-1 cells were intravenously transferred into tumor-bearing mice 10 days after tumor cell injection. Tumors were monitored for 23 days postinoculation, and mice were sacrificed on day 23 to collect draining lymph nodes (DLN) and TILs. Tumor size was determined by caliper measurements 3 times a week.

Single-cell suspensions from spleens, lymph nodes, and TILs were prepared in PBS and stained with antibodies listed in Supplementary Table S1. For intracellular staining, cells obtained from tumor tissues or draining lymph nodes (as described above) were cultured for 5 hours with PMA (50 ng/mL, Sigma-Aldrich) plus ionomycin (500 ng/mL, Sigma-Aldrich). 3 μg/mL Brefeldin A (eBioscience) was added during the final 4 hours of incubation. After stimulation, cells were washed and fixed with intracellular fixation buffer (88–8824–00, eBioscience) for 30 minutes in room temperature, followed by cell permeabilization with permeabilization buffer (88–8824–00, eBioscience) for 30 minutes in 4°C. Cells were stained with fluorescently conjugated cytokine antibodies (Supplementary Table S1). Data were obtained using a Canto II flow cytometer (BD Biosciences) and analyzed with FlowJo software (FlowJo, LLC). To purify TILs from EL4 tumors, dissociated cells were sorted for macrophages (F4/80⁺CD11b⁺), CD4⁺ T cells (CD3⁺CD4⁺), CD8⁺ T cells (CD3⁺CD8⁺), and NKT cells (CD3⁺NK1.1⁺) using a BD FACS Aria II (antibodies listed in Supplementary Table S1). All flow cytometry results were obtained by live cell gating and cell doublet exclusion based on FSC/SSC.

WT CD8⁺ T cells were isolated from spleens and lymph nodes using the MagniSort mouse CD8 T-cell enrichment kit (Invitrogen). Naive CD8⁺CD44^loCD62L^hi cells were sorted from WT to >90% purity with a BD FACS Aria II (BD Biosciences). Naive CD8⁺ T cells were resuspended at 1 × 10⁶ cells/mL in complete medium in the presence or absence of anti-CD3 (10 μg/mL, eBioscience) and anti-CD28 (2 μg/mL, BD Biosciences) and incubated on ice for 30 minutes. Rabbit anti-hamster IgG (secondary cross-linking Ab, 1 μg/ml, Sigma-Aldrich) was added, and cells were transferred to 37°C for the indicated times. Stimulations were terminated by washing twice with cold PBS.

Total RNA from each tissue or from macrophages, CD4⁺ T cells, CD8⁺ T cells, and NKT cells was extracted using a Qiazol or RNeasy Mini Kit (QIAGEN) and was reverse-transcribed with oligo (dT, Abm) and QuantiTect Reverse Transcription kit (QIAGEN) using 2 μg total RNA. qRT-PCR was performed with the Rotor-Gene SYBR Green PCR kit (QIAGEN) using the Rotor-Gene Q 5plex PCR in ABI Prism 7900 detection system. The average gene expression was normalized using Gapdh, and relative expression was calculated using the 2^−ΔΔC_t method. All the samples were run in triplicate. Primer sequences are listed in Supplementary Table S2.

CD8⁺ TILs from EL4 tumor-bearing WT and Peli1-KO mice were sorted into lysis buffer (79306, QIAGEN) using BD FACS Aria cell sorter. Total RNA was isolated using Qiazol reagent (QIAGEN). RNA quality was assessed by Agilent 2100 bioanalyzer using the RNA 6000 Nano Chip (Agilent Technologies), and RNA integrity number (RIN) values were at least 7. RNA quantification was performed using ND-2000 Spectrophotometer (Thermo Inc.). cDNA templates were synthesized using an EasyScript cDNA synthesis kit (Abm), and libraries were constructed using QuantSeq 3′ mRNA-seq library prep kit FWD (Lexogen). High-throughput sequencing (GRCm38/mm10; accession number: GCA_000001635.2) was performed as single-end 75 sequencing using NextSeq 500 (Illumina).

QuantSeq 3′ mRNA-Seq reads were aligned using Bowtie2 (22). Bowtie2 indices were either generated from a genome assembly sequence or the representative transcript sequences for aligning to the genome and transcriptome (GSE42824, GSE41867, GSE9650, GSE15930). The alignment file was used for assembling transcripts, estimating their abundances, and detecting differential expression of genes. Differentially expressed genes (DEG) were determined based on counts from unique and multiple alignments using coverage in Bedtools (23). By default, the threshold comprises an adjusted p-value <0.05 and fold change >1.5. The RT (Read Count) data were processed based on the quantile normalization method using EdgeR within R (R development Core Team) using Bioconductor. Gene classification was based on searches done by DAVID (http://david.abcc.ncifcrf.gov/) and Medline databases (http://www.ncbi.nlm.nih.gov/). Principal component analysis (PCA) was performed on the centered expression matrix of DEGs with a fold change of at least 2 between samples. The functional enrichment of DEGs was analyzed by ranking all genes based on fold-change difference in expression, then applying gene set enrichment analysis (GSEA) (24), to examine enrichment of previously published T-cell gene signatures (25). The FDR-adjusted P values are reported (q-value), with q-values <0.05 considered significant.

The full-length cDNA sequence of the human Peli1 protein was amplified using oligo-dT primers. Peli1 full-length and Peli1 ΔC (1–275) were subcloned into Myc GFP-tagged plasmids, respectively (18). Peli1 ΔC included 280 N-terminal amino acids but lacked the C-terminal RING domain. GFP-tagged Peli1 RING mutants were constructed by QuickChange site-directed mutagenesis kit (Stratagene) (26). Human codon sequences of genes encoding PKCθ and ZAP70 were constructed using molecular cloning by PCR. Full-length cDNA sequences of human PKCθ and ZAP70 were amplified from pDONR233-PKCθ (#23424, Addgene) and pDONR233-ZAP70 (#23887, Addgene). For PKCθ, XhoI restriction site was placed at 5′ and 3′ ends. For ZAP70, EcoRI and SalI restriction sites were placed at 5′ and 3′ ends, respectively. The following primers were used (5′ to 3′): PKCθ_forward (ATTCTCGAGTCGCCATTTCTTCGGATT) and PKCθ_reverse (ATTCTCGAGTCAGGATATCAGCCGCTC) and ZAP70_forward (ATTGAATTCCCAGACCCCGCGGCGCAC) and ZAP70_reverse (ATTGTCGACTCAGGCACAGGCAGCCTC). The two genes were subcloned into each restriction site of the FLAG and GST-tagged expression plasmid, respectively. GST-tagged Peli1 bacmid was constructed using the Baculovirus Expression System with Gateway Technology kit (Invitrogen). The full-length cDNA sequence of human Peli1 was amplified using oligo-dT primers. EcoRI and SalI restriction sites were placed at 5′ and 3′ ends, respectively. The following primers were used (5′ to 3′): Peli1_forward (ATTGAATTCTTTTCTCCTGATCAAGA) and Peli1_reverse (ATTGTCGACTTAGTCTAGAGGTCCTTG). The Peli1 fragment was subcloned into pENTR 1A plasmid (Invitrogen). 100 ng of pENTR 1A-Peli1 plasmid was then mixed with 150 ng of pDEST 10 plasmid (GST-tagged plasmid) and TE buffer to a final volume of 8 μL. 2 μL of LR clonase II (Invitrogen) was added to plasmid mixture and mixed by vortexing. After incubating the reaction at 25°C for 1 hour, 1 μL of the Proteinase K solution was added, and samples were incubated at 37°C for 10 minutes to terminate the reaction. pENTR 1A-Peli1 plasmid was integrated with the pENST 10 plasmid (GST-tagged plasmid) through an LR recombination reaction. 1 μL of LR recombination reaction sample was transformed into DH5α-competent cells using the heat-shock method at 42°C for 30 seconds, followed by shaking at 37°C for 1 hour. After centrifugation at 5,000 rpm for 3 minutes, the transformed cells were plated on LB plates containing 50 μg/mL ampicillin (Sigma), and the plates were inverted and incubated at 37°C overnight. 100 ng of GST-tagged Peli1 plasmids were then transformed into the E.coli strain DH10Bac (Invitrogen). Clones with the integrated expression cassette were selected by blue-white screening on LB plates containing 100 μg/mL X-Gal and 40 μg/mL IPTG in addition to KTG antibiotics (Sigma; 50 μg/mL kanamycin, 10 μg/mL tetracycline, 7μg/mL gentamicin). Bacmid was then isolated from positive clones. HA-tagged Ub (HA-Ub), HA-tagged K48R (HA-K48R), and HA-tagged K63R (HA-K63R) were obtained from Ki-Young Lee (Sungkyunkwan University, South Korea). For transient transfection, cells were electroporated using a microporator (Digital Biotechnology) according to the manufacturer's instructions (https://www.bid-on-equipment.com/laboratory/lab-equipment/363512~digital-bio-mp-100-mini-microporator.htm). HeLa cells (3 × 10⁶ cells/shock) were transfected with plasmids (8 U/shock) and harvested after 48 hours. Proteasome inhibitor, MG132 (25 μmol/L, AG scientific) was added for 5 hours before cell harvest.

HeLa cells were transfected with an expression plasmid encoding GFP, GFP-tagged Peli1 WT, H313A, C336A mutants; Flag-PKCθ or Flag-ZAP70; Myc, Myc-Peli1, or Myc-Peli1 ΔC; HA-Ub, HA-K48R, and HA-K63R in the indicated combinations. At 48 hours posttransfection, cells were collected into two aliquots. One aliquot was used for conventional immunoblotting, as described below. The other aliquot was used for immunoprecipitation with anti-Flag antibodies, as described below. Immunoprecipitates were washed three times with IP buffer (see below), and bound proteins were immunoblotted with the indicated antibodies in Supplementary Table S3.

For the Peli1 protein purification using the bacteria expression system, 100 ng of GST-tagged Peli1 plasmids were transformed into BL21 (DE3) competent cells using the heat-shock method at 42°C for 1 minute, followed by shaking at 37°C for 1 hour. After centrifugation at 5,000 rpm for 3 minutes, the transformed cells were plated on LB plates containing 50 μg/mL ampicillin (Sigma), and the plates were inverted and incubated at 37°C overnight. The resistant single cell was picked and amplified in 20 mL LB medium containing 50 μg/mL ampicillin (LA medium) at 37°C overnight. For large scale, the cells in 20 mL liquid culture were inoculated into 500 mL LA medium and incubated at 37°C until OD₆₀₀ reached 0.6. After adding 500 μmol/L stock of IPTG, the cells were incubated at 18°C overnight. The cells were centrifuged at 5,000 rpm for 15 minutes and resuspended in lysis buffer [150 mmol/L NaCl (Sigma), 10 mmol/L Tris-HCl (Sigma), 1 mmol/L EDTA (GenDEPOT), 100 μg/mL lysozyme (Sigma), 1 mmol/L PMSF (Sigma), 1 mmol/L DTT (Sigma) supplemented with a mixture of protease inhibitor cocktail (GenDEPOT; 50 μmol/L PMSF, 10 μmol/L pepstatin A, 20 μmol/L leupeptin, 100 μmol/L benzamidine, and 50 μmol/L bestatin)]. The cells were sonicated on ice for 20 minutes and were centrifuged at 13,000 rpm for 30 minutes. The insoluble fraction was discarded. Lysates were then incubated with Glutathione Sepharose 4B beads (Cytiva) and 1% Trion X-100 (Sigma) at 4°C overnight on a rotator. After centrifugation at 3,000 rpm for 3 minutes, lysates were washed three times in PBS and used for protein interaction experiments.

For the Peli1 protein purification using the baculovirus expression system, 1.5 × 10⁶ Sf9 cells per well in 6-well plates were seeded in 2 mL SF-900-II SFM media containing 10% FBS. 2 μg of GST-tagged Peli1 bacmid was diluted in SF-900-II SFM medium without FBS to obtain a final volume of 100 μL. In a separate tube, 8 μL of Cellfectin II (Invitrogen) was mixed with SF-900-II SFM without FBS to a final volume of 100 μL. The diluted plasmid DNA was transferred to the diluted Cellfectin to form the DNA–Cellfectin complexes. Cellfectin–DNA complex mixture was added to cells. The DNA–Cellfectin mixture were removed from the wells after 5 hours and replaced with 2 mL of prewarmed SF900-II SFM media with 1% antibiotics. The plate was placed into a 27°C incubator for 5 days in a humidified container. After 5 days, the supernatant containing the first virus generation was harvested. The supernatant was sterile-filtered using 0.22 μmol/L filter unit for further virus amplification.

For GST-tagged Peli1 production, 100 mL SF9 cells seeded at 1.5 × 10⁶ cells/mL in a 500 mL sterile flask were infected with 10 mL of GST-tagged Peli1 baculovirus to obtain a 10× final virus dilution. Cells were grown at 27°C for 3 days on a shaker until the cell viability dropped to 70%–80%. Infected cells were harvested by centrifugation at 500 × g for 5 minutes. The infected cell pellets were washed three times with cold PBS and centrifuged at 500 × g for 5 minutes. The infected cells were resuspended in lysis buffer [150 mmol/L NaCl (Sigma), 25 mmol/L Tris-HCl (Sigma), 1% NP40 (Sigma), 1 mmol/L EDTA (GenDEPOT), 1 mmol/L DTT (Sigma) supplemented with a mixture of protease inhibitor cocktail (GenDEPOT; 50 μmol/L PMSF, 10 μmol/L pepstatin A, 20 μmol/L leupeptin, 100 μmol/L benzamidine, and 50 μmol/L bestatin)]. Resuspended cells were incubated on ice for 10 minutes. Lysates were then centrifuged at 15,000 × g for 15 minutes at 4°C, and the insoluble fraction was discarded. The supernatant was incubated with Glutathione Sepharose 4B beads (Cytiva) and 1% Trion X-100 (Sigma) at 4°C overnight on a rotator. After centrifugation at 3,000 rpm for 3 minutes, samples were washed three times in PBS and used for protein interaction experiments.

For the GST pulldown assays, fusion proteins (1 μg) were adsorbed onto glutathione-protein A/G-Sepharose beads (Amersham Biosciences) and incubated with whole-cell extracts (1 mg) from Jurkat cells for 12 hours. Jurkat cells were lysed with immunoprecipitation (IP) [150 mmol/L NaCl (Sigma), 20 mmol/L Tri-HCl (Sigma), 5 mmol/L EDTA (GenDEPOT), 1% Triton X-100 (Sigma), 1 mmol/L phenylmethanesulfonylfluoride (Sigma), 10 mmol/L NaF (Sigma), 1 mmol/L Na₃VO₄ (Sigma), and 1 mmol/L DTT (Sigma) supplemented with a mixture of protease inhibitor cocktail (GenDEPOT; 50 μmol/L PMSF, 10 μmol/L pepstatin A, 20 μmol/L leupeptin, 100 μmol/L benzamidine, and 50 μmol/L bestatin)] and incubated at 4°C for 30 minutes. The cells were then lysed by passing the cell clumps five times through a 27-gauge needle. The lysates were centrifuged at 13,000 rpm for 30 minutes, and the insoluble fraction was discarded. Bound proteins were separated by SDS-PAGE and analyzed by immunoblotting, as described below.

For immunoprecipitation and ubiquitination assays, Jurkat cells and transfected HeLa cells were lysed as described above using IP buffer. Lysates (0.5 mg) were incubated with an anti-Peli1 (1 mg/1 μg) or normal IgG (control, 1 mg/1 μg), anti-Flag (1 mg/1 μg) or normal IgG (control, 1 mg/1 μg) and then incubated with protein A/G agarose beads for 12 hours. Cells were then pelleted, washed three times in immunoprecipitation buffer, and analyzed by immunoblotting, as described below.

Cells and mouse tissues were lysed in IP buffer (as above). Prepared lysates (10–50 μg) were separated on 6%–12% of Mini-Protean TGX gels (Bio-Rad) in 1X Sodium dodecyl sulphate buffer (GenDEPOT) and transferred onto nitrocellulose blotting membranes (ATTA) in transfer buffer (250 mmol/L Tris-HCl, 2 mol/L glycine). Blocking was performed in a buffer of 5% milk (BD) in 1X TBS-T (1X TBS + 0.1% Tween-20) to block nonspecific binding for 1 hour at room temperature. Primary antibodies (Supplementary Table S3) were diluted in blocking buffer (5% milk) and incubated with the membrane overnight at 4°C. Membranes were washed with 1X TBS-T for four times. HRP-conjugated goat anti-mouse IgG (SA001, GenDEPOT) or secondary anti-rabbit IgG (SA002, GenDEPOT) were diluted 1:7,000 in blocking buffer and incubated for 2 hours at room temperature. The target proteins were detected using ECL solution (AB frontier), developed on film (CU-BU new, Agfa) by an automatic X-ray film processor (JP-33, JPI).

Human resected tissue specimens of 70 diffuse large B-cell lymphoma (DLBCL) cases, and five spleens, and five lymph nodes as a control were obtained from the Asan Medical Center (Seoul, South Korea). Cases in which the fixation was well-maintained by formalin, were large enough to perform immunostaining multiple times, and had no necrosis (in which nonspecific antigen expression could be minimized) were included in this study, and cases that did not satisfy all inclusion criteria were excluded. Finally, 63 DLBCL cases, five spleens, and five lymph nodes were included in this study. The quality of immunostaining was confirmed by staining tonsil tissue. This study was approved by the Asan Medical Center Institutional Review Board (no. 2019–0701) and conducted in accordance with the Declaration of Helsinki. Consent was waived by decision of the Institutional Review Board. IHC and Opal multiplex IF staining were performed using tissue microarrays (TMA) obtained from formalin-fixed, paraffin-embedded tissue blocks. The IHC staining for Peli1 and CD8 was performed using a Benchmark automatic immunostaining device (Roche Tissue Diagnostics) and an UltraViewTM Universal DAB Detection Kit (Ventana Medical Systems), according to the manufacturer's instructions. Four-micron–thick sections were immunostained with primary antibodies against Peil1 (1:1000, LS-B8877; LS Bio) and CD8 (1:100, 108M-96; Cell Marque). All stained slides were observed using BX53 upright microscope (Olympus).

The tyramide signal amplification (TSA)-based Opal method was used for immunofluorescence (IF) staining. Sectioned slides were deparaffinized in xylene, rehydrated in ethanol, washed in distilled water, and fixed in 10% neutral buffered formalin (NBF) for 10 minutes. All multiplexed staining was performed with the Opal 7 Immunology Discovery Kit (OP7DS2001KT, Akoya Biosciences). Antigen retrieval was performed in AR buffer using microwave treatment (MWT). After MWT, the slides were incubated for 15 minutes and covered with blocking buffer (ARD1001EA, Akoya Biosciences) for 10 minutes. Primary antibodies against Peli1 (1:1,000, LS-B8877, LS Bio) and CD8 (1:100, 108M-96, Cell Marque), and secondary antibodies (ARH1001EA, Akoya Biosciences) were incubated for 1 hour. TSA visualization was performed using Opal fluorophores (Opal 620 and Opal 540; 1:100). MWT was further performed to remove the anti-Opal fluorophore complex with AR buffer. All multiplexed staining was performed by repeating MWT through TSA visualization. Hematoxylin and eosin (H&E) staining and three IHC staining slides were scanned using a PANNORAMIC 250 Flash III scanner (3DHistech Ltd.) at 40× magnification and were then independently reviewed for quality and consistency by a trained pathologist (SJS) before being considered for digital image analysis. Opal multiplex IF images of stained slides were acquired with the Vectra Polaris (Akoya Biosciences) whole-slide scanner using the manufacturer's protocol (https://www.akoyabio.com/phenoptics/mantra-vectra-instruments/vectra-polaris/). Manual annotation was carried out for tumor regions of interest (ROI) of all staining slides based on H&E staining. The edges of the tissue were excluded from annotation because they were likely affected by staining artifacts. The CaseViewer v.2.4, DensitoQuant application (3DHistech Ltd.) was used for digital image analysis. DensitoQuant identifies the positive stain based on an automatic color separation method through which individual positive pixels are counted and classified as negative (< one-quarter of maximal intensity), weak- (one-quarter ∼ one-half of maximal intensity), moderate- (one-half ∼ three-quarters of maximal intensity), and strong (> three-quarters of maximal intensity) based on the intensity and threshold ranges. For each marker, the percentage of positive density was calculated as a ratio of the total annotated area. Statistical analyses were performed using R3.0.2. Differences in positive density between DLBCL, spleen, and lymph nodes were evaluated using Wilcoxon rank sum test. Statistical significance was defined as P < 0.05.

The single-cell transcriptome of CD8⁺ T cells of tumor tissues and peripheral blood from patients with hepatocellular carcinoma (HCC) was obtained from the Gene Expression Omnibus (GEO) of the US National Center for Biotechnology Information (NCBI; accession number: GSE98638, https://www.ncbi.nlm.nih.gov/geo/), which was deposited by Zhang and colleagues (27). The obtained normalized gene expression values were converted to transcripts per million (TPM) and log-transformed (log₂(TPM+1)). The abundances of PELI1 and genes encoding well-known makers of either naïve (CCR7, TCF7, LEF1 and SELL) or exhausted T cells (PDCD1, ENTPD1, CXCL13, CTLA4, HAVCR2, CCL3, LAG3, and SIRPG) were visualized on Uniform Manifold Approximation and Projection (UMAP) plots. Spearman correlation (ρ and P value) of PELI1 abundance with selected genes was computed and visualized in scatter plots. The analysis and visualization of single-cell transcriptomes were conducted with RStudio (https://www.rstudio.com/) with R version 4.1.0 (https://www.r-project.org), and installed R packages “dplyr”, “naniar”, “stringr”, “ggpubr”, “ggplot2”, “egg”, “RColorBrewer”, “viridis”, “M3C”, “umap” were applied for the visualization of UMAPs and correlation analysis (28, 29).

Data are presented as means ± SEM. All data were analyzed using GraphPad Prism 4.5 software (GraphPad Software). A P < 0.05 was considered statistically significant. Each experiment was carried out three or more times, and reproducible results were obtained. Therefore, representative data are shown in figures.

Peli1 is a receptor signal-responsive E3 ubiquitin ligase that can be activated by various receptor-mediated signals, particularly in B cells, macrophages, and T cells. Thus, aberrant regulation of these receptor-mediated signaling pathways can cause unbalanced activation or expression of Peli1, which in turn triggers forced signal cascades. Indeed, Peli1 is upregulated in patients with B-cell lymphoma (18) and lung cancer (20), and is associated with poor outcomes in those patients (18). Accordingly, we hypothesized that a mouse model, in which Peli1 is overexpressed or deleted, would be appropriate for the identification of Peli1's role in tumor immunity. To examine the expression of Peli1 in tumor-infiltrating leukocytes (TIL), EL4 cells (murine T cell leukemia/lymphoma cells) were injected subcutaneously into C57BL6 mice, and subsequently, spleen and tumors were isolated from tumor-bearing mice (Fig. 1A; Supplementary Fig. S1). Inhibitory receptor PD-1 was expressed by a higher percentage of CD8⁺ TILs (Fig. 1B). Peli1 was significantly upregulated in CD8⁺ TILs, but not CD4⁺ and NKT TILs compared with spleen (Fig. 1C). Peli1 mRNA expression was also increased in CD8⁺ TILs from LLC (murine lung cancer cells) and MC38 (murine colorectal cancer cells) subcutaneous tumors (Fig. 1C; Supplementary Table S2). However, for macrophage TILs, Peli1 expression increased in EL4 tumors, but there was no change in LLC and MC38 tumors. In NKT TILs, Peli1 was increased in EL4 and LLC tumors (Fig. 1C). These results indicate that Peli1 is increased in the TME, particularly in CD8⁺ TILs.

We next compared the expression of Peli1 and CD8 proteins in diffuse large B-cell lymphoma (DLBCL) tumor tissues from 63 patients and 10 normal spleen and lymph node samples (Fig. 1D). IHC analyses showed that Peli1 was expressed on human DLBCL CD8⁺ TILs, as well as tumor cells. The expression of Peli1 and the infiltration of CD8⁺ TILs were heterogeneous in DLBCL tissues. Overall, the density of Peli1 expression was upregulated in DLBCL tissues compared with both normal spleen and lymph node tissues (Fig. 1E). Although we did not assess other immune cell subsets, our evidence suggests that Peli1 is upregulated in CD8⁺ TILs.

To determine whether depletion of Peli1 in CD8⁺ T cells enhances antitumor responses, we generated Peli1-KO mice (Supplementary Fig. S2). We next used a specific antibody to deplete CD8⁺ and CD4⁺ T cells in EL4 tumor–bearing mice and confirmed depletion in blood, lymph nodes, and spleens (Fig. 2A). The CD8⁺ T cell-depleted group showed similar tumor growth as the IgG-treated control group in WT mice (Fig. 2B). However, tumor growth in Peli1-KO mice was significantly attenuated in IgG-treated but not in the CD8⁺ T cell–depleted group, suggesting a potential inhibitory role of Peli1 on antitumor CD8⁺ T-cell responses (Fig. 2B). Depletion of CD4⁺ T cells has been shown to reduce tumor growth in WT mice, but there was little difference in Peli1-KO mice (Fig. 2C). Next, a CSF1R antibody [to deplete tumor-associated macrophages (TAM); Fig. 2D] attenuated tumor growth in both WT and Peli1-KO mice (Fig. 2E). However, there was no significant difference between WT and Peli1-KO mice with TAM depletion. Depletion of CD8⁺ T cells led to significant reduction of IFNγ and CXCL9, key cytokines for the maintenance of CTLs, in EL4 tumor–bearing Peli1-KO mice (Fig. 2F). Together, these results suggest that Peli1 counteracts antitumor CD8⁺ T-cell function.

We first checked that Peli1-KO mice had no changes in immune cell populations, including CD4⁺ T cells, CD8⁺ T cells, regulatory T cells (Treg), dendritic cells (DC), and macrophages (Supplementary Fig. S3A and S3B). For memory-phenotype T cells (CD44⁺CD122⁺), effector (CD62L^loCD44^hi) and memory (CD62L^hiCD44^hi) T-cell populations were similar between WT and Peli1-KO mice (Supplementary Fig. S3C and S3D). Next, we additionally injected MC38 and LLC cells into WT and Peli1-KO to examine the role of Peli1 in regulating tumor immunity more broadly. Peli1-KO mice significantly inhibited the growth of tumors (Supplementary Fig. S4A and S4B), indicating that Peli1 deficiency delayed tumor development. We next analyzed accumulation and activation of CD8⁺ T cells in spleen and TILs from WT and Peli1-KO mice. The percentages and absolute cell numbers of CD8⁺ TILs in Peli1-KO mice were increased compared with WT mice (Fig. 3A; Supplementary Fig. S4C). Peli1-deficient CD8⁺ TILs consistently had increased expression of Ki67, a marker of cell proliferation, and expression of antiapoptotic genes, Bcl2 and Bcl2l1 (Fig. 3B and C). Peli1 deficiency induced T-cell activation of genes encoding effector cytokines and IFNγ, as well as cytotoxic molecules, granzyme B (GzmB) and inducible nitric oxide synthase (iNOS; Supplementary Fig. S4D). Peli1 deficiency also increased CXCL9 and CCL20 transcripts, which are potent chemoattractants for CD8⁺ T cells, in the EL4 tumors (Supplementary Fig. S4E). However, Peli1-KO mice did not have altered immune suppression because the percentage of Foxp3⁺ Tregs remained unchanged between EL4 tumor–bearing WT and Peli1-KO mice (Supplementary Fig. S4F and S4G). Although the population of TAMs was similar between WT and Peli1-KO mice, Peli1 KO mice showed a reduced population of CD206^hiMHC^lo M2-like macrophages (Supplementary Fig. S4H).

To explore further whether Peli1 modifies the effector function of CD8⁺ TILs, we compared the functionality of these cells in WT and Peli1-KO mice after PMA and ionomycin ex vivo stimulation. Effector function was significantly improved in Peli1-KO mice, as indicated by higher production of IFNγ and GzmB (Fig. 3D and E). The mRNA expression of GzmB and perforin were also increased in Peli1-deficient CD8⁺ TILs (Fig. 3F). We also performed ex vivo cytolytic assays using WT and Peli1-KO CD8⁺ TILs. Peli1-KO CD8⁺ TILs displayed enhanced apoptosis in tumors compared with WT CD8⁺ TILs; this was not observed in cells from the spleen (Fig. 3G). These data indicate that Peli1 deficiency increases the antitumor activity of CD8⁺ TILs by conferring higher tumor-specific cytolytic activity.

We next compared the relative number of IFNγ⁺CD8⁺ T cells in draining lymph nodes (DLN) from EL4 tumor–bearing WT and Peli1-KO mice (Supplementary Fig. S5A). Peli1-KO CD8⁺ T cells had significantly improved effector function, with higher production of IFNγ, IL2, and TNFα, and a higher percentage of polyfunctional CD8⁺ T cells (Supplementary Fig. S5A and S5B). Long-term persistence of tumor antigens and/or the suppressive TME drive the progression of antitumor effector CD8⁺ T cells into a dysfunctional, “exhausted” state (30). Therefore, we questioned whether depletion of Peli1 alters the ratio between effector and exhausted T-cell populations. We found that the population of effector CD8⁺ T cells was significantly increased in the DLNs and TILs of Peli1-KO mice (Fig. 4A; Supplementary Fig. S5C), whereas exhausted CD8⁺ T cells were diminished in EL4 tumor–bearing Peli1-KO TILs (Fig. 4B). The ratio of effector/exhausted CD8⁺ TILs from EL4 tumor–bearing Peli1-KO mice were higher than EL4 tumor–bearing WT mice (Fig 4C). We also analyzed the expression of immune checkpoint inhibitory receptors (Fig. 4D; Supplementary Fig. S5D). The expression of PD-1, TIM-3, CTLA-4, LAG3, and KLRG1 in CD8⁺ TILs was significantly downregulated in EL4 tumor–bearing Peli1-KO mice compared with the WT mice. Peli1 expression was also enriched in exhausted CD8⁺ T cells (Supplementary Fig. S5E). Collectively, these data support the finding that Peli1 deficiency enhances effector functions of CD8⁺ TILs and reduces exhausted CD8⁺ TILs.

To further expand our understanding of Peli1 in exhausted CD8⁺ TILs compared with naïve CD8⁺ T cells, we obtained and analyzed single-cell transcriptome data of human CD8⁺ T cells from patients with hepatocellular carcinoma (NCBI GEO: GSE98638; Supplementary Fig S6A; ref. 27). Expression of PELI1 and genes encoding well-known markers of either naïve (CCR7, TCF7, LEF1, and SELL) or exhausted T cells (PDCD1, ENTPD1, CXCL13, CTLA4, HAVCR2, CCL3, LAG3, and SIRPG) indicated that PELI1 expression was elevated in exhausted compared with naïve CD8⁺ T cells (Supplementary Fig S6B). To further estimate the association between PELI1 and exhausted CD8⁺ T cells, we computed Spearman correlation (ρ and P values). PELI1 expression was consistently, positively, and significantly correlated with T-cell exhaustion markers. In contrast, its expression did not significantly correlate with naïve CD8⁺ T-cell markers, except for TCF7, which had a significant negative correlation (Supplementary Fig S6C). Taken together, this single-cell transcriptomic analysis of human CD8⁺ T cells in liver cancer further suggest that Peli1 is enriched in exhausted CD8⁺ TILs.

To examine whether transfer of Peli1-deficient CD8⁺ T cells could provide a therapeutic effect in murine transplanted tumor models, EG7 cells stably expressing OVA were subcutaneously injected into recipient, congenic CD45.1⁺ SJL mice. WT or Peli1 KO-OT-1 cells (CD45.2⁺) were intravenously transferred into EG7 tumor–bearing mice after tumor injection (Fig. 5A). Whereas WT OT-1 cell transfer led to similar tumor growth compared with the nontransferred group, Peli1-KO OT-1 cell transfer delayed tumor growth (Fig. 5B). Clearly, the transfer of Peli1-deficient CD8⁺ T cells was effective in suppressing tumor growth.

We next compared the transferred CD8⁺ T cells in DLNs and TILs between WT and Peli1-KO OT-1 recipient groups. The percentage of CD8⁺ T cells in DLNs and TILs was increased in Peli1-KO OT-1–transferred mice (Fig. 5C and D). Peli1-KO OT-1 transferred mice exhibited significantly increased IFNγ⁺GzmB⁺CD8⁺ T cells in DLNs and TILs (Fig. 5E and F). The number of inhibitory receptor PD-1^hi cells of CD8⁺ T cells in Peli1-KO OT-1 recipients was lower than in the WT mice (Fig. 5G), confirming that Peli1 affected the antitumor responses of CD8⁺ TILs. Together, these data demonstrate that Peli1 deficiency leads to enhanced cytolytic activity and infiltration of CD8⁺ T cells in the TME.

To further delineate the molecular nature of Peli1-deficient CD8⁺ T cells in TME, we compared the transcriptional profiles of CD8⁺ TILs between EL4 tumor–bearing WT and Peli1-KO mice. We identified 262 transcripts that were differentially expressed between WT and Peli1-deficient CD8⁺ T cells (Fig. 6A). 152 and 110 genes were up- and downregulated, respectively, in the CD8⁺ TILs of Peli1-KO mice compared with WT mice (P < 0.05, fold-change ≥ 2; Fig. 6B). Genes involved in the positive regulation of cytokine production and immune activation were upregulated in the Peli1-deficient CD8⁺ TILs (Fig. 6B and C), and Peli1 deficiency consistently showed enhanced gene set signatures of effector and cytotoxic functions of CD8⁺ TILs. Peli1 deficiency also led to upregulation of transcription factors such as IRF7, CEBPB, STAT2, and Jun and downregulation of inhibitory receptor LAG3 (Fig. 6B). Gene-set enrichment analysis (GSEA) also revealed that Peli1-deficient CD8⁺ TILs were enriched in gene set signatures for cytokine and inflammatory response signaling (Fig. 6C). The gene sets, defined as upregulated in exhausted T cells in melanoma relative to activated CD8⁺ T cells in acute infection (31, 32), were enriched in WT CD8⁺ TILs, whereas Peli1-deficient CD8⁺ TILs displayed phenotypes of activated and effector T-cell phenotypes (Fig. 6D and E). The gene sets involved in CD8⁺ T-cell exhaustion of chronic lymphocytic choriomeningitis virus infection were inversely correlated with Peli1-KO CD8⁺ T cells (Fig. 6F). Together, these results suggest that Peli1-mediated tumor development is closely linked to T-cell exhaustion through attenuation of effector CD8⁺ TILs with a hyperactivated phenotype.

PKCθ is a central kinase in TCRT cell receptor downstream signal transduction (33, 34) and has been reported to inhibit the suppressive function of Tregs. Thus, PKCθ-deficient mice were are resistant to experimental autoimmune encephalomyelitis (EAE) (35). These mice presented with less T-cell infiltration in the spinal cords and diminished production of proinflammatory cytokines IFNγ and TNF following EAE induction. A similar role is observed in other autoimmune syndromes, such as collagen-induced arthritis (36) and colitis (37), implicating PKCθ in autoimmune diseases. However, whether that is the role of PKCθ pathway in antitumoral T-cell responses remains unclear.

Chronic antigen stimulation, such as that induced by tumors and viral infection, associates with exhausted T cells with inhibited TCR signaling (38, 39). Therefore, to examine how Peli1 impacted TCR signaling, we compared the expression of proteins involved in the activation of TCR signaling. Specifically, we examined CD8⁺ T cells from DLNs and TILs in EL4 tumor-bearing WT and Peli1-KO mice. DLNs are the primary sites for tumor antigen presentation to T cells, and they determine the effect of antitumor immune responses (40). First, we confirmed that Peli1-KO mice had consistent suppression growth of established EL4 tumors (Supplementary Fig. S7A). Although Peli1 had no effect on proximal TCR signaling, its deficiency upregulated PKCθ, PLCγ, c-Rel, TRAF3, phospho-AKT, phospho-JNK, and phospho-ERK1/2 (Supplementary Fig. S7B). We also assessed EL4 tumor growth in doxycycline-inducible Peli1 transgenic mice (rtTA-Peli1, Supplementary Fig. S7C; ref. 41). RtTA-Peli1 mice treated with doxycycline formed larger tumors than the control mice. In contrast, rtTA, rtTA-Peli1 mice without doxycycline treatment, and rtTA mice treated with doxycycline showed similar patterns of tumor growth (Supplementary Fig. S7D). Peli1 regulated downstream of TCR signaling, as its overexpression significantly downregulated PKCθ, PLCγ, phospho-ERK1/2, and c-Rel. (Supplementary Fig. S7E). We also isolated CD8⁺ TILs from EL4 tumor–bearing WT and Peli1-KO mice and stimulated them ex vivo with anti-CD3. Peli1-KO CD8⁺ TILs showed activated TCR signaling compared to the WT CD8⁺ TILs, especially for expression of PKCθ (Fig 7A). PKCθ was most prominently increased in CD8⁺ T cells isolated from Peli1-KO mice but exhibited significant downregulation in CD8⁺ T cells isolated from Peli1 transgenic mice. Together, these data imply that Peli1 regulates downstream of TCR signaling in CD8⁺ TILs through a PKCθ-mediated pathway.

We hypothesized that Peli1 might directly interact with downstream TCR signaling molecules, including PKCθ. We demonstrated that bacteria-expressed Peli1 interacts with PKCθ and Zap70; however, we were unable to detect interactions with PLCγ, ITCH, or Cbl-b (Supplementary Fig. S8A). Baculovirus-expressed Peli1 interacted with both PKCθ and Zap70 (Supplementary Fig. S8B). Immunoprecipitation demonstrated that Peli1 formed a complex with PKCθ, but not Zap70, in stimulated Jurkat T cells (Fig. 7B). Peli1 was found to also bind PKCθ in CD8⁺ T cells and showed increased binding after stimulation (Fig. 7C).

Peli1 contains a forkhead-associated (FHA) domain that helps in binding to its protein substrates and to a RING-like domain for its E3 ubiquitin ligase activity (17). To reveal whether the Peli1 RING-like domain mediates the ubiquitination of PKCθ, we employed GFP-fused-Peli1 WT and two RING mutants [H313A (HA) and C336A (CA; ref. 26)] with HA-Ub. Ubiquitination of PKCθ was induced only by Peli1 WT expression, but not by Peli1 RING mutants (Fig. 7D; Supplementary Fig. S8D). We confirmed that overexpression of Peli1 WT, but not Peli1 mutant (ΔC, deletion form of RING domain), clearly led to the ubiquitination of PKCθ (Fig. 7E), demonstrating that ubiquitination of PKCθ depended on Peli1 E3 ligase activity. In contrast, the ubiquitination of Zap70 was not affected by Peli1 expression (Supplementary Fig. S8C). We also found that Peli1 promoted K48-linked ubiquitination of PKCθ (Fig. 7F and G). These results clearly demonstrated that Peli1 associated with PKCθ and that its RING domain mediated K48-linked ubiquitination of PKCθ.

Finally, to determine whether Peli1 contributed to ubiquitination-mediated PKCθ degradation in T cells, we used rtTA-Peli1 transgenic mice (41). We stimulated CD3⁺ T cells isolated from WT and rtTA-Peli1 mice and demonstrated a reduction of PKCθ in Peli1-overexpressing T cells compared with WT (Fig. 7H). Taken together, these results demonstrated that Peli1 could directly bind to PKCθ and induced its degradation by K48-linked ubiquitination in T cells.

The data suggest that Peli1 in the TME supports immune suppression and tumor growth, and thus, its inhibition reinvigorated antitumor immunity. Peli1 inhibited proximal TCR signal transduction via activation of PKCθ and suppresses survival, proliferation, and effector functions of CD8⁺ T cells, leading to reductions in GzmB and IFNγ (Fig. 7I). Tumor growth and CD8⁺ T-cell inhibition depended on Peli1 expression and manifested without any extrinsic treatment. These results were confirmed in a transplanted tumor mouse model, with development of cancer dependent on the expression of Peli1 protein. In sum, these results indicate that Peli1 signaling is a key mediator in controlling antitumor CD8⁺ T-lymphocyte function.

Cancer cells escape from T cell-mediated cytotoxicity by exploiting immunosuppressive mechanisms, including the inhibitory immune checkpoint molecules (12, 13). Here, we revealed a novel immunosuppressive signaling pathway by which Peli1 deregulated the proliferation, prosurvival, and cytotoxic functions of CD8⁺ T lymphocytes. TLRs are ubiquitously expressed in immune cells and also play an important role in the TME (42). TLRs induce beneficial antitumor effects by eliciting inflammatory cytokine expression and cytotoxic T-lymphocyte responses (43, 44). Tumor-associated antigens act as pathogen-associated molecular patterns (PAMP) and can activate TLRs to induce tumor gene–related programmed cell death, including apoptosis, autophagy, and necroptosis (44). There is also evidence that the aberrant regulation of TLR activation promotes tumor development, as TLR-induced inflammation accelerates cancer cell growth in the TME (44). Studies have unveiled a critical role for Peli1 in activating TLR signaling to mediate inflammatory signaling (15, 17). Therefore, it is very possible that the TLR-Peli1 signaling axis is a substantial mechanism in the TME.

Constitutive expression of PD-1 by tumor-specific T cells leads to impaired T-cell function and tumor escape upon binding to its ligand PD-L1, which is expressed by tumor cells and/or immune infiltrating cells within TME (3, 45). PD-1 expression constitutes a form of immune adaptation to chronic stimulation, leading to a physiologic limitation of immune responses to favor peripheral tolerance, and in this context, PD-1 can be considered a marker of dysfunctional T lymphocytes (46). We found that PD-1 expression was increased in CD8⁺ T cells of tumor-bearing Peli1-overexpressing mice, whereas it was significantly diminished in tumor-bearing Peli1-deficient mice compared with controls. Our cytokine profiling and RNA-sequencing analyses revealed enhanced effector T-cell responses in Peli1-deficient mice. Peli1 deficiency also led to the increased expression of a series of T-cell effector cytokines, including IL2, IFNγ, and TNFα. We also found that gene sets associated with increased exhausted T cells, relative to activated and effector CD8⁺ T cells, were enriched in WT CD8⁺ TILs, whereas Peli1-KO CD8⁺ TILs displayed gene sets similar to activated and effector T cells. This finding indicates that Peli1 deficiency protects CD8⁺ TILs from exhaustion and permits the retention of their activated phenotype. We also observed increased expression of prosurvival and antiapoptotic proteins, Bcl2 and Bcl2l1, and increased cytotoxic T lymphocytes in Peli1-deficient mice. These mice also showed elevated GzmB and perforin secretion, suggesting that Peli1 is an immune-suppressive regulator by inhibiting the proliferation and cytotoxic effector function of CD8⁺ T cells in the TME.

Strategies for enhancing the function of effector T cells by targeting immunoregulatory membrane receptors, CTLA-4 and PD-1, have been successful in subsets of patients with melanoma, NSCLC, urothelial bladder cancer, and renal cell cancer (3, 4, 6, 7). Nevertheless, it has become apparent that if antagonists to these T-cell checkpoint inhibitory molecules overcome some of the immune-suppressive effects of the TME, there may be other, more fundamental inhibitory reactions occurring to explain why most patients, especially those with microsatellite stable colorectal cancer, ovarian cancer, prostate cancer, and pancreatic ductal adenocarcinoma, rarely exhibit objective responses to these therapies (3, 5). These exceptions indicate that it will require the development of more effective anticancer immunotherapies to overcome immune-suppressive TME. Peli1 is upregulated in a variety of human cancers and in immune-suppressive microenvironments, and thus, seems to regulate a fundamental oncogenic signaling axis by directly targeting both cancer cells and elements of the TME. Considering our previous studies (18–20), Peli1 could be a key therapeutic target for each component of the TME that plays a role in cancer initiation, progression, and invasion. Although we have not investigated whether Peli1 in tumors affects other immune cells, Peli1 deficiency provides an antitumor microenvironment by triggering the apoptotic cell death of tumor cells and promoting CD8⁺ T cell–mediated killing.

Antigenic stimulation of T cells triggers PKCθ phosphorylation and activates effector molecules that transduce signals into distinct subcellular compartments and activate transcription factors, such as NFκB, NFAT, and AP-1, which are essential for the induction of T cell–mediated responses (34). Thus, it is not surprising that PKCθ plays a critical role in the regulation of T-cell tumor immunity. Several studies have described the role of PKCθ in cancers, for example, PKCθ is aberrantly expressed in gastrointestinal stromal tumors, breast cancer, and Ewing sarcoma (47–49). Genomic data from open database analytics has revealed a variety of genetic alterations for PKCθ, including mutation, amplification, and deletions, in patients with cancer (50). Peli1 is responsive to various T-cell receptor signals, and thus, Peli1 functions as a critical negative regulator for TCR-proximal signaling for PKCθ recruitment. Thus, regulation of Peli1's intracellular pool is crucial for the amplification of TCR signaling upon TCR engagement in the TME. It will be interesting to determine whether targeting Peli1 may be an effective strategy to provide antitumor immune responses by triggering the apoptotic cell death of tumor cells and stimulating CD8⁺ T cell–mediated killing.

No disclosures were reported.

J. Park: Conceptualization, data curation, formal analysis, validation, investigation, methodology. S.-Y. Lee: Data curation, formal analysis, validation, investigation, methodology. Y. Jeon: Resources, formal analysis, investigation. K.-M. Kim: Resources, formal analysis, investigation. J.-K. Lee: Conceptualization, resources, formal analysis. J. Ko: Resources, formal analysis, investigation. E.-J. Park: Resources, investigation, methodology. J.-S. Yoon: Investigation, methodology. B.E. Kang: Formal analysis, methodology. D. Ryu: Formal analysis, methodology. H. Lee: Resources, data curation, validation, investigation. S.-J. Shin: Resources, data curation, validation, investigation. H. Go: Conceptualization, resources, data curation, formal analysis, validation, investigation. C.-W. Lee: Conceptualization, data curation, formal analysis, supervision, funding acquisition, validation, writing–original draft, project administration, writing–review and editing.

We thank the members of our laboratory for their technical support. S.-Y. Lee was supported by a Global PhD Fellowship from National Research Foundation of Korea (2017H1A2A1046184). This work was supported by grants (2017R1A2B3006776 and 2019M3A9H1103755) from the National Research Foundation (NSF) funded by the Ministry of Education, Science, and Technology (MEST), Republic of Korea.

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.