Visual Abstract

Protein arginine methyltransferase 5 (PRMT5) participates in the symmetric dimethylation of arginine residues of proteins and contributes to a wide range of biological processes. However, how PRMT5 affects the transcriptional and epigenetic programs involved in the establishment and maintenance of T cell subset differentiation and roles in antitumor immunity is still incompletely understood. In this study, using single-cell RNA and chromatin immunoprecipitation sequencing, we found that mouse T cell–specific deletion of PRMT5 had greater effects on CD8+ than CD4+ T cell development, enforcing CD8+ T cell differentiation into Klrg1+ terminal effector cells. Mechanistically, T cell deficiency of PRMT5 activated Prdm1 by decreasing H4R3me2s and H3R8me2s deposition on its loci, which promoted the differentiation of Klrg1+CD8+ T cells. Furthermore, effector CD8+ T cells that transited to memory precursor cells were decreased in PRMT5-deficient T cells, thus causing dramatic CD8+ T cell death. In addition, in a mouse lung cancer cell line–transplanted tumor mouse model, the percentage of CD8+ T cells from T cell–specific deletion of PRMT5 mice was dramatically lost, but CD8+Foxp3+ and CD8+PDL1+ regulatory T cells were increased compared with the control group, thus accelerating tumor progression. We further verified these results in a mouse colon cancer cell line–transplanted tumor mouse model. Our study validated the importance of targeting PRMT5 in tumor treatment, because PRMT5 deficiency enforced Klrg1+ terminal CD8+ T cell development and eliminated antitumor activity.

Similar to phosphorylation and acetylation protein post-translational modifications, protein arginine methylation plays a critical role in the fields of epigenetics and signal transduction (1, 2). Protein arginine methyltransferase 5 (PRMT5) is the main type II protein arginine methyltransferase, which catalyzes the generation of arginine monomethyl or symmetric dimethyl groups and participates in a variety of important biological functions, including cell proliferation, survival, pre-mRNA splicing, DNA repair, and the regulation of various nuclear transcription factors (35). PRMT5, MEP50, and pICln form a methyl complex that methylates Sm protein, thereby promoting the aggregation of small ribonucleoproteins and regulating RNA splicing (6). PRMT5 regulates various histone methylation modifications and induces gene transcription activation or repression by binding to the cofactor COPR5. It methylates H3R2 to form H3R2me1 or H3R2me2s, recruits the WDR5/MLL methyltransferase complex, and induces H3K4me3 formation, thus mediating the transcriptional activation of genes (7). Furthermore, it participates in the methylation of H4K20me3, H3K9me3, and H3K27me3, which in turn affects the expression of fetal globin genes (8); PRMT5 also symmetrically methylates H4R3me2, thereby mediating the silencing of multiple genes (9).

Naive CD8+ T cells expand and acquire effector functions when activated by infectious and malignant diseases. The pools of effector CD8+ T cells are divided into two main subsets after activation, short-lived effector cells (SLECs) and memory precursor effector cells (MPECs), based on the expression of Klrg1 and CD127 (10). Numerous transcription factors have also been implicated in this cell fate decision: T-bet, Blimp1, Zeb2, and Id2 favor SLECs, whereas Eomes, Bcl-6, STAT3, and Id3 support the formation of MPECs (1114). PRMT5 has been reported to be essential for survival, implicated in stem cell function, and it plays critical roles in hematopoietic and immune cell development (15, 16). Deletion of PRMT5 in all cell types (full-body deletion) is embryonic lethal, and conditional knockout (CKO) of PRMT5 results in a concurrent loss of hematopoietic progenitor cells and severely impaired cytokine signaling (17). To date, a limited number of reports have revealed that T cell–specific knockdown of PRMT5 influences T cell development, but these reports are inconsistent. For example, Inoue et al. reported that T cell–specific deletion of PRMT5 does not influence thymus CD4+CD8+ double-positive (DP) or single-positive (SP) CD4+ or CD8+ T cell development or maturation (18). However, Webb et al. reported that PRMT5 controls thymus CD4+CD8+ DP or SP CD4+ or CD8+ T cell homeostasis and development (19). A selective inhibitor of PRMT5 was found to reduce the clinical score of experimental autoimmune encephalomyelitis (EAE); furthermore, PRMT5 promotes cholesterol biosynthesis and mediates Th17 responses in EAE (19, 20). Inhibitors of PRMT5 promote CD8+ T cell apoptosis by upregulating P53 expression and reducing AKT pathway activity (21). However, how PRMT5 affects the corresponding epigenetic state and is involved in the establishment and maintenance of CD8+ T cell subset differentiation is still incompletely understood.

In recent years, an increasing number of studies have found that PRMT5 is highly expressed in a variety of tumors, both primary tumors and metastases, including pancreatic cancer, glioblastoma, and lung cancer (2225). However, T cell–specific deletion of PRMT5 in antitumor activity remains incompletely explored. In this study, we found that T cell–specific deletion of PRMT5 did not affect thymus CD4+CD8+ DP or SP CD4+ or CD8+ T cell development and maturation. However, in the periphery, T cell–specific deletion of PRMT5 had a much greater impact on CD8+ T cells than on CD4+ T cell development. CD8+ T cells expressed much more PRMT5 than CD4+ T cells in the periphery; in another mechanism, peripheral CD8+ T cells were likely to differentiate into SLECs and become more vulnerable to death. The detailed molecular mechanism showed that T cell–specific deletion of PRMT5 decreased H4R3me2s and H3R8me2s deposition on the Prdm1 loci, thus activating the Prdm1 gene. We further revealed that PRMT5 deficiency in T cells led to a greater reduction in CD8+ T cells and induced CD8+ regulatory T (Treg) cells, thus promoting tumor progression in cancer cell line–transplanted tumor mouse models. We explored the important role of PRMT5-dependent gene silencing in the establishment and maintenance of the peripheral CD8+ T cell lineage and promoted naive CD8+ T cell differentiation into SLEC CD8+ T cells, demonstrating the critical role of PRMT5 in CD8+ T cell development, especially in malignant disease states. Thus, the adverse events that target PRMT5 in tumor treatment should be valued and considered.

Male C57BL/6 mice were purchased from the Shanghai Laboratory Animal Center, Chinese Academy of Sciences (Shanghai, China). CD4-Cre transgenic mice were purchased from The Jackson Laboratory. PRMT5fl/fl mice were obtained from the European Mutant Mouse Archive. PRMT5-floxed mice were crossed with CD4-Cre transgenic mice to generate CD4-Cre+ PRMT5fl/fl mice and CD4-Cre–negative littermate controls. The animals were housed in the animal care facilities of the Shanghai Jiao Tong University School of Medicine, Xin Hua Hospital, under pathogen-free conditions. This study was carried out in accordance with the recommendations in the guidelines of the institutional animal care and use committee of Xin Hua Hospital.

The mouse lung cancer cell line (LLC) and colon cancer cell line (MC38) were purchased from the Cell Resource Center, Shanghai Academy of Biological Sciences, Shanghai, China. All cells were cultured at 37°C in a humidified incubator with 5% CO2 in RPMI 1640 medium (Life Technologies) supplemented with 10% FBS (Life Technologies), 100 U/ml penicillin, and 100 μg/ml streptomycin (Life Technologies).

Single-cell suspensions were harvested from fresh tissues according to standard procedures and surface stained in FACS buffer (PBS with 2% FBS, 1 mM EDTA, and PBS) with mAbs obtained from eBioscience, BD Pharmingen, or BioLegend. For intracellular cytokine staining, cells were stimulated for 5 h with a cell stimulator and inhibitor (eBioscience). After surface staining, the cells were fixed and permeabilized with the Fix/Perm Foxp3 Transcription Factor Staining Buffer Set (eBioscience) and BD Cytofix/Cytoperm buffer (BD Biosciences), respectively, following the manufacturer’s instructions, and then stained with mAbs. Flow cytometric analysis was performed with a FACSCanto II instrument (BD Biosciences) and FlowJo software (BD Biosciences). Related T cell subsets from mouse spleens or thymuses were sorted with a FACSAria II Cell Sorter (BD Biosciences); the postsort purity was routinely >95%. Information for the Abs is as follows: BD Pharmingen: CD274 (PD-L1)-PE (catalog no. 55809), CD3e-allophycocyanin-cyanine 7 (Cy7) (catalog no. 557596), CD4-allophycocyanin-Cy7 (catalog no. 552051), CD44-PE (catalog no. 553134), CD45-PE-Cy7 (catalog no. 552848), CD62L-PE-Cy7 (catalog no. 560516), CD8a-allophycocyanin-Cy7 (catalog no. 557654), CD8a-PE-Cy7 (catalog no. 552877), CD8a-PerCP-Cy5.5 (catalog no. 551162); BioLegend: CD11b-allophycocyanin (catalog no. 101212), CD11c-allophycocyanin-Cy7 (catalog no. 117324), CD132 (common γ-chain)-allophycocyanin (catalog no. 132307), CD19-Pacific Blue (catalog no. 115523), CD1d-FITC (catalog no. 123507), CD25-allophycocyanin (catalog no. 102012), CD279 (PD-1)-allophycocyanin (catalog no. 135210), CD366 (Tim-3)-PerCP-Cy5.5 (catalog no. 119717), CD39-PE-Cy7 (catalog no. 143805), CD3e-PerCP-Cy5.5 (catalog no. 100218), CD3e-Pacific Blue (catalog no. 100214), CD4-PerCP-Cy5.5 (catalog no. 100434), CD4-FITC (catalog no. 100406), CD45-allophycocyanin (catalog no. 103112), CD45R/B220-FITC (catalog no. 103206), CD8a-Brilliant Violet 421 (catalog no. 100737), F4/80-PerCP-Cy5.5 (catalog no. 123128), granzyme B-PE (catalog no. 372208), KLRG1 (MAFA)-FITC (catalog no. 138410), Ly-6C-Brilliant Violet 421 (catalog no. 128031), Ly-6G-PE (catalog no. 127608), NK-1.1-PE (catalog no. 108707); eBioscience: CD127-allophycocyanin (catalog no. 17-1271-82), CD49b (integrin-α2)-PE (catalog no. 12-5971-82), FOXP3-PE (catalog no. 12-5773-82), IFN-γ–allophycocyanin (catalog no. 17-7311-82), TNF-α–eFluor 450 (catalog no. 48-7321-82); MBL: CD1d tetramer (α-GalCer loaded)-PE (catalog no. TS-MCG-1).

Total RNA was isolated from cells or tissues following the standard TRIzol (Takara) protocol. First-strand cDNA was synthesized from total RNA using RT Master Mix (Takara). Transcription levels were detected using SYBR Green–based RT-PCR performed by the ABI StepOne QRT-PCR Detection System (Life Technologies). The mRNA levels were normalized to those of β-actin mRNA. The primer sequences were as follows: Prmt5 forward: 5′-CTGAATTGCGTCCCCGAAATA-3′;reverse: 5′-AGGTTCCTGAATGAACTCCCT-3′; Prdm1 forward: 5′-GCTGCTGGGCTGCCTTTGGA-3′; Prdm1 reverse: 5′-GGAGAGGAGGCCGTTCCCCA-3′; Id2 forward: 5′-ATGAAAGCCTTCAGTCCGGTG-3′; Id2 reverse: 5′-AGCAGACTCATCGGGTCGT-3′; Bhlhe40 forward: 5′-ACGGAGACCTGTCAGGGATG-3′; Bhlhe40 reverse: 5′-GGCAGTTTGTAAGTTTCCTTGC-3′; Tbx21 forward: 5′-AACCGCTTATATGTCCACCCA-3′; Tbx21 reverse: 5′-CTTGTTGTTGGTGAGCTTTAGC-3′; Runx2 forward: 5′-GACTGTGGTTACCGTCATGGC-3′; Runx2 reverse: 5′-ACTTGGTTTTTCATAACAGCGGA-3′; Klrg1 forward: 5′-ACCAACAGCTTGATACAGAGGT-3′; Klrg1 reverse: 5′-ATCCACTGCAAAGCAACTTCA-3′; Actb forward: 5′-TGTCCACCTTCCAGCAGATGT-3′; Actb reverse: 5′-AGCTCAGTAACAGTCCGCCTAG-3′.

Cells were directly lysed and subjected to 10% SDS-PAGE. Immunoblotting was performed by transferring the proteins to nitrocellulose membranes (Schleicher & Schuell Microscience) with a Mini Trans-Blot apparatus (Bio-Rad Laboratories). Membranes were incubated overnight at 4°C with the specific primary Abs PRMT5 (catalog no. 79998S; Cell Signaling Technology), symmetric dimethyl arginine motif (catalog no. 13222S; Cell Signaling Technology), histone H3K4me3 (catalog no. ab8580; Abcam), H4R3me2s (catalog no. 61187; Active Motif), H3R8me2s (catalog no. A-3706; EpiGentek), and β-actin (catalog no. Ac026; ABclonal), all at a dilution of 1:1000. After the membranes were washed, subsequent incubations with appropriate IRDye 800CW– or Alexa Fluor 680–conjugated secondary Abs (1:5000; Abcam) were conducted for 1 h at room temperature, and the signals were visualized with the Odyssey Infrared Imaging System.

CD8+ T cells were harvested and cross-linked with 1% formaldehyde for 10 min at room temperature. Then, glycine was added to the cells and incubated for 5 min at room temperature. Cells were washed with ice-cold PBS and centrifuged. Then, we discarded the supernatant and suspended the cells in lysis buffer supplemented with 5 μl protease inhibitor mixture and 5 μl PMSF. The samples were sonicated, and 50 μl of product was removed to assess the DNA fragment size. The remainder was stored at −80°C. Abs against control IgG, H3R4me2s, and H3R8me2s were used for the ChIP assay. DNA extracted from 10-μl preimmunoprecipitated samples was used as an input control. DNA was amplified by QRT-PCR and normalized to the input. QRT-PCR primers for the Prdm1 peak of H4R3me2s were as follows: forward: 5′-TCTTCCCCACTCCCTACTCA-3′; reverse: 5′-GCTTACAGTTGGTCCCCAAA-3′. QRT-PCR primers for the Prdm1 peak of H3R8me2s were as follows: forward: 5′-ACCTGCAGACTTTCCTGTGG-3′; reverse: 5′-GTGCCCCTGCTAGATGAGAT-3′.

Cells were cross-linked with 1% formaldehyde for 10 min at room temperature and quenched with 125 mM glycine. The fragmented chromatin fragments were precleared and then immunoprecipitated with protein A + G magnetic beads coupled with anti-H4R3me2s and anti-H3R8me2s Abs. After reverse cross-linking, ChIP and input DNA fragments were end repaired and dA tailing using the NEBNext End Repair/dA-Tailing Module (E7442; New England Biolabs) followed by adaptor ligation with the NEBNext Ultra Ligation Module (E7445; New England Biolabs). The DNA libraries were amplified for 15 cycles and sequenced using the Illumina NextSeq 500 system in single-end 1 × 75 sequencing mode.

Raw reads were filtered to obtain high-quality clean reads by removing sequencing adapters, short reads (length <35 bp), and low-quality reads using Cutadapt (version 1.9.1) and Trimmomatic (version 0.35). Then, FastQC was used to ensure a high read quality. The clean reads were mapped to the mouse genome (assembly GRCm38) using Bowtie2 (version 2.2.6) software. Peak detection was performed using the MACS (version 2.1.1) peak finding algorithm with 0.01 set as the p value cutoff. Annotation of peak sites to gene features was performed using the ChIPseeker R package. Genome graphs were generated and viewed using Integrative Genomics Viewer, and we aligned the H4R3me2s, HR8me2s, and input control in CD8+ T cells.

Cells from PRMT5 CKO or control mice were sorted by flow cytometry. Total RNA was extracted from tissue with TRIzol reagent (Invitrogen). The RNA quality was checked using a Bioanalyzer 2200 (Agilent) and kept at −80°C. RNA with RNA integrity number >6.0 was considered suitable for further experiments. cDNA libraries were prepared using the NEBNext Ultra Directional RNA Library Prep Kit, NEBNext Poly(A) mRNA Magnetic Isolation Module, and NEBNext Multiplex Oligos according to the manufacturer’s instructions. The products were purified and enriched by PCR to create the final cDNA libraries, which were then quantified using an Agilent 2200 instrument. The tagged cDNA libraries were pooled in equal ratios and used for 150-bp paired-end sequencing in a single lane of the Illumina HiSeq X Ten. Before read mapping, clean reads were obtained from the raw reads by removing the adaptor sequences and low-quality reads. The clean reads were then aligned to the mouse genome (GRCm38) using HISAT2. High-throughput sequencing was used to obtain gene counts, and the reads per kilobase of transcript per million reads mapped method was used to determine gene expression.

We applied the DESeq2 algorithm to filter the differentially expressed genes, followed by significance analysis with calculation of the p value and false discovery rate (FDR). Genes were defined as differentially expressed according to the following criteria: fold change (FC) >1.0, p < 0.05, FDR < 0.05.

The gene annotation file was retrieved from the Ensembl Genome Browser 90 database (http://www.ensembl.org/index.html). Functional enrichment analyses were performed through Metascape (http://metascape.org/gp/index.html#/main/step1).

The single-cell RNA-sequencing (scRNA-seq) libraries were generated using the 10x Genomics Chromium Controller Instrument and Chromium Single Cell 3′ V3 Reagent Kits (10x Genomics, Pleasanton, CA). Briefly, cells were concentrated to 1000 cells/µl, and ∼8000 cells were loaded into each channel to generate single-cell gel bead-in-emulsions, which resulted in the expected mRNA barcoding of 5000 single cells for each sample. After the RT step, the single-cell gel bead-in-emulsions were disrupted, and the barcoded cDNA was purified and amplified. The amplified barcoded cDNA was fragmented, A-tailed, ligated with adaptors, and index PCR amplified. The final libraries were quantified using the Qubit High Sensitivity DNA assay (Thermo Fisher Scientific), and the size distribution of the libraries was determined using a high-sensitivity DNA chip on a Bioanalyzer 2200 instrument (Agilent). All libraries were sequenced using an Illumina sequencer (Illumina, San Diego, CA) in a 150-bp paired-end run.

scRNA-seq data analysis was performed by NovelBio Bio-Pharm Technology Co. with the NovelBrain Cloud Analysis Platform. We applied fastp with default parameter filtering of the adaptor sequence and removed the low-quality reads to obtain clean data. Then, feature-barcode matrices were obtained by aligning reads to the mouse genome (GRCm38 Ensembl version 92) using CellRanger version 3.1.0. We applied downsampling analysis among samples sequenced according to the mapped barcoded reads per cell of each sample and finally constructed the aggregated matrix. Cells containing over 200 expressed genes and a mitochondrial Unique Molecular Identifier rate <20% passed the cell quality filtering, and mitochondrial genes were removed. The Seurat package (version 2.3.4; https://satijalab.org/seurat/) was used for cell normalization and regression according to the Unique Molecular Identifier counts of each sample and the percentage mitochondria rate to obtain the scaled data. Principal component analysis (PCA) was constructed on the basis of scaled data with the top 2000 most variable genes, and the top 10 PCs were used for t-distributed stochastic neighbor embedding (tSNE) construction and Uniform Manifold Approximation and Projection construction. Using the graph-based cluster method, we acquired the unsupervised cell cluster result based on the top 10 PCA PCs, and we calculated the marker genes using the FindAllMarkers function with the Wilcoxon rank-sum test algorithm under the following criteria: log-normal FC >0.25; p < 0.05; and minimum p > 0.1. To identify the cell type in detail, clusters of the same cell type were selected for further tSNE analysis, graph-based clustering, and marker analysis.

We applied single-cell trajectory analysis using Monocle 2 (http://cole-trapnell-lab.github.io/monocle-release) with DDR-Tree and default parameters. Before Monocle analysis, we selected marker genes from the Seurat clustering result and raw expression counts of the cells that passed the filtering. On the basis of pseudotime analysis, branch expression analysis modeling was applied for branch fate–determined gene analysis.

To assess transcription factor regulation strength, we applied the single-cell regulatory network inference and clustering (pySCENIC version 0.9.5) workflow using the 20-thousand motifs database for RcisTarget and GRNboost.

To characterize the relative activation of a given gene set, we performed QuSAGE (2.16.1) analysis.

LLC cells (150 μl; 1.5 × 106 cells) and MC38 cells (150 μl; 1.5 × 106 cells) were s.c. implanted into the right flanks of PRMT5 CKO or control mice. The tumor length (L) and width (W) were measured daily. Tumor volumes were calculated with the formula V = L× W2/2.

The Student unpaired two-tailed t test was used for comparisons involving two groups. All presented data are the mean ± SEM. Statistical analyses were performed with Prism 7 (GraphPad Software). A p value < 0.05 was considered to indicate a statistically significant difference.

The scRNA-seq, bulk RNA-seq, and ChIP-seq data used in this study are available in the Gene Expression Omnibus database under accession number GSE186863 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE186863).

The codes generated during this study are available from the authors upon request.

To explore the role of PRMT5 in T cells, we first generated PRMT5 CKO mice. PRMT5fl/fl mice were obtained from the European Mutant Mouse Archive; exon 7 of the PRMT5 gene was flanked by two LoxP sites (Supplemental Fig. 1A and 1B); and then Prmt5-floxed mice were crossed with CD4-Cre transgenic mice (The Jackson Laboratory) to generate CD4-Cre+Prmt5fl/fl (PRMT5 CKO) and littermate control mice. T cell PRMT5 mRNA expression in spleens was significantly reduced compared with that in the control group (Supplemental Fig. 1C). We further isolated splenic T cells and stimulated them with anti-CD3 and anti-CD28 Abs in vitro. PRMT5 expression was significantly increased compared with that in unstimulated T cells, but the expression level in PRMT5 CKO mice was significantly reduced compared with that in the control group (Supplemental Fig. 1D); furthermore, the levels of symmetrical protein arginine demethylation and H4R3me2s were greatly decreased in PRMT5 CKO mice (Supplemental Fig. 1D). These data demonstrated that we had successfully constructed PRMT5 CKO mice.

Reports concerning T cell–specific deletion of PRMT5 and effects on thymus T cell development and maturation have been inconsistent (18, 19). We therefore further characterized the thymus T cell subpopulations. First, we measured H3R8me2s, H3R4me2s, and H3K4me3 expression in thymus cells between these two groups. These three histone methylation levels have been reported to be regulated by PRMT5 and thus to activate or repress gene expression (24, 26, 27). H3R8me2s, H3R4me2s, and H3K4me3 expression levels were reduced in PRMT5 CKO mouse thymocytes (Fig. 1A). There were no differences in thymus cell numbers between these two groups (Fig. 2B). In addition, the cell numbers of CD4CD8 T cells (double negative), CD4+CD8+ T cells (DP), CD4+CD8 (CD4+ SP), and CD8+CD8 (CD8+ SP) T cells showed no significant differences between these two groups (Fig. 1C). CD4+ iNKT cells, but not CD8+ iNKT cells, were significantly decreased in PRMT5 CKO mice (Fig. 1D). CD4+CD44+ memory T cells were reduced in PRMT5 CKO mice, but CD8+CD44+ memory T cells did not change (Fig. 1E). Consistent with previous reports (19), the frequency of CD4+Foxp3+ Treg cells was significantly decreased in PRMT5 CKO mice (Fig. 1F). However, there were no significant differences in the proliferation and apoptosis of CD4 and CD8 SP T cells between these two groups (Fig. 1G). These results showed that T cell–specific deletion of PRMT5 did not affect the development or maturation of thymus T cells but was necessary for CD4+ Treg cell and CD4+ iNKT cell development.

FIGURE 1.

PRMT5 deficiency influenced CD4+Foxp3+ Treg and CD4+ iNKT development in the thymus. (A) Western blot analysis of H4R3me2s, H3R8me2s, H3K4me3, and PRMT5 expression in thymocytes from CD4-CrE-PRMT5fl/fl (PRMT5 CKO) and littermate control mice. β-Actin was used as the loading control. One of three similar representative datasets is shown. (B) The absolute number of thymocytes in control (n = 12) and PRMT5 CKO (n = 15) mice. (C) FACS analysis of CD4+CD8+ (DP), CD4+CD8 (CD4+ SP), and CD8+CD4 (CD8+ SP) T cell subsets in control (n = 14) and PRMT5 CKO mice (n = 17). (D) The percentages of CD4+ iNKT and CD8+ iNKT cells in control (n = 5) and PRMT5 CKO (n = 5) mice, respectively. (E) The percentages of CD44+ cells in CD4+ and CD8+ T cells in the thymocytes from control (n = 5) and PRMT5 CKO (n = 5) mice. (F) The percentages of CD4+Foxp3+ Treg cells from control (n = 5) and PRMT5 CKO (n = 8) mice. (G) The percentages of proliferating cells (BrdU+) and apoptotic cells (7-aminoactinomycin D–positive [7-AAD+]) in CD4+ and CD8+ T cells in the thymus from control (n = 5) and PRMT5 CKO (n = 5) mice, respectively. (HJ) FACS was used to isolate DP, CD4+ SP, and CD8+ SP T cell subsets from the thymus of control (n = 3) and PRMT5 CKO mice (n = 3), and bulk RNA-seq was performed. The differentially expressed genes (log2 FC >1.0; p < 0.05) between DP (H), CD4+ SP (I), and CD8+ SP (J) are shown. The bar graphs represent mean ± SEM. Statistical differences were determined by a two-tailed unpaired Student t test. *p < 0.05, **p < 0.01.

FIGURE 1.

PRMT5 deficiency influenced CD4+Foxp3+ Treg and CD4+ iNKT development in the thymus. (A) Western blot analysis of H4R3me2s, H3R8me2s, H3K4me3, and PRMT5 expression in thymocytes from CD4-CrE-PRMT5fl/fl (PRMT5 CKO) and littermate control mice. β-Actin was used as the loading control. One of three similar representative datasets is shown. (B) The absolute number of thymocytes in control (n = 12) and PRMT5 CKO (n = 15) mice. (C) FACS analysis of CD4+CD8+ (DP), CD4+CD8 (CD4+ SP), and CD8+CD4 (CD8+ SP) T cell subsets in control (n = 14) and PRMT5 CKO mice (n = 17). (D) The percentages of CD4+ iNKT and CD8+ iNKT cells in control (n = 5) and PRMT5 CKO (n = 5) mice, respectively. (E) The percentages of CD44+ cells in CD4+ and CD8+ T cells in the thymocytes from control (n = 5) and PRMT5 CKO (n = 5) mice. (F) The percentages of CD4+Foxp3+ Treg cells from control (n = 5) and PRMT5 CKO (n = 8) mice. (G) The percentages of proliferating cells (BrdU+) and apoptotic cells (7-aminoactinomycin D–positive [7-AAD+]) in CD4+ and CD8+ T cells in the thymus from control (n = 5) and PRMT5 CKO (n = 5) mice, respectively. (HJ) FACS was used to isolate DP, CD4+ SP, and CD8+ SP T cell subsets from the thymus of control (n = 3) and PRMT5 CKO mice (n = 3), and bulk RNA-seq was performed. The differentially expressed genes (log2 FC >1.0; p < 0.05) between DP (H), CD4+ SP (I), and CD8+ SP (J) are shown. The bar graphs represent mean ± SEM. Statistical differences were determined by a two-tailed unpaired Student t test. *p < 0.05, **p < 0.01.

Close modal
FIGURE 2.

T cell–specific deletion of PRMT5 had a greater effect on CD8+ than on CD4+ T cells in the periphery and promoted Klrg1+CD127CD8+ T cell differentiation. (A) The percentages of CD11b+, CD3+, NK, and B cells in spleens from control (n = 5) and PRMT5 CKO (n = 5) mice. (B) The frequencies of CD4+ and CD8+ cells in CD3+ T cells from control (n = 8) and PRMT5 CKO (n = 10) mouse spleens. (CH) The percentages of iNKT (C); CD62L+CD44 (Naive), CD62L+CD44+ (central memory), and CD62LCD44+ (effector memory) T cell subsets (D); CD25+Foxp3+ Treg cells (E); Klrg1+ T cells (F); and CD127+Klrg1, CD127+Klrg1+, and CD127Klrg1+ T cells in CD4+ (G) and CD8+ T (H) cells from control (n = 8) and PRMT5 CKO (n = 15) mouse spleen. (I) The ratio of CD127Klrg1+ T cells (SLECs) from CD4+ and CD8+ T cells in PRMT5 CKO to control mice. (J and K) CD8+CD45RB+ T cells isolated by FACS from control (n = 3) and PRMT5 CKO (n = 3) mice and stimulated with anti-CD3 and anti-CD28 Abs for 24 h and 48 h. The percentages of CD127-Klrg1+ T cells (J) and annexin V+ apoptosis cells were measured (K). Data are representative of three independent experiments. The mean ± SEM was graphed. Statistical differences were determined by the two-tailed unpaired Student t test. *p < 0.05, **p < 0.01, ***p < 0.01.

FIGURE 2.

T cell–specific deletion of PRMT5 had a greater effect on CD8+ than on CD4+ T cells in the periphery and promoted Klrg1+CD127CD8+ T cell differentiation. (A) The percentages of CD11b+, CD3+, NK, and B cells in spleens from control (n = 5) and PRMT5 CKO (n = 5) mice. (B) The frequencies of CD4+ and CD8+ cells in CD3+ T cells from control (n = 8) and PRMT5 CKO (n = 10) mouse spleens. (CH) The percentages of iNKT (C); CD62L+CD44 (Naive), CD62L+CD44+ (central memory), and CD62LCD44+ (effector memory) T cell subsets (D); CD25+Foxp3+ Treg cells (E); Klrg1+ T cells (F); and CD127+Klrg1, CD127+Klrg1+, and CD127Klrg1+ T cells in CD4+ (G) and CD8+ T (H) cells from control (n = 8) and PRMT5 CKO (n = 15) mouse spleen. (I) The ratio of CD127Klrg1+ T cells (SLECs) from CD4+ and CD8+ T cells in PRMT5 CKO to control mice. (J and K) CD8+CD45RB+ T cells isolated by FACS from control (n = 3) and PRMT5 CKO (n = 3) mice and stimulated with anti-CD3 and anti-CD28 Abs for 24 h and 48 h. The percentages of CD127-Klrg1+ T cells (J) and annexin V+ apoptosis cells were measured (K). Data are representative of three independent experiments. The mean ± SEM was graphed. Statistical differences were determined by the two-tailed unpaired Student t test. *p < 0.05, **p < 0.01, ***p < 0.01.

Close modal

Next, we isolated thymus DP, CD4+ SP, and CD8+ SP T cells from PRMT5 CKO and control mice and then performed bulk RNA-seq. The data showed that in DP T cells, only 25 genes were downregulated and 29 genes were upregulated in PRMT5 CKO mice (Fig. 1H). In CD4+ SP T cells, among the differentially expressed genes between the two groups, 126 genes were downregulated and 59 genes were upregulated in PRMT5 CKO mice (Fig. 1I). However, the number of differentially expressed genes in CD8+ SP T cells was much lower, with only 20 genes downregulated and 16 genes upregulated in PRMT5 CKO mice (Fig. 1J). RNA-seq data suggested that conditional deletion of PRMT5 in T cells did not greatly affect T cell transcription levels in the thymus.

We then evaluated the impact of PRMT5 deletion on peripheral immune cell populations. The proportion of CD3+ T cells in the spleen from PRMT5 CKO mice was significantly decreased; however, the proportion of NK cells was significantly increased compared with that in the control group (Fig. 2A). The frequency of CD4+ T cells was decreased, but a robust loss of CD8+ T cells was observed in PRMT5 CKO mice (Fig. 2B). The proportions of iNKT cells in both CD4+ and CD8+ T cells were decreased in PRMT5 CKO mice (Fig. 2C). The proportions of naive T cells (CD62L+CD44) and central memory T cells (CD62L+CD44+) in CD4+ T cells were not significantly changed between these two groups, but the proportion of naive T cells in CD8+ T cells was significantly decreased, and the effector memory CD4+ and CD8+ T cells were significantly increased in PRMT5 CKO mice (Fig. 2D). The proportion of CD4+Foxp3+ Treg cells showed no significant difference between these two groups (Fig. 2E). However, expression of the effector molecule Klrg1 was increased in both CD4+ and CD8+ T cells in PRMT5 CKO mice, especially in CD8+ T cells (Fig. 2F). We then used Klrg1 and CD127 to discriminate SLEC CD8+ T cells (Klrg1+CD127low) and MPEC CD8+ T cells (KlrgCD127+). Interestingly, we found that the proportion of SLEC CD8+ T cells was significantly increased in PRMT5 CKO mice (Fig. 2G and 2H). The increasing ratio of SLEC T cells in PRMT5 CKO mice was much higher in CD8+ T than in CD4+ T cells (Fig. 2I). Moreover, we isolated naive CD8+ T cells from PRMT5 CKO and control mice and stimulated them with anti-CD3 and anti-CD28 Abs in vitro. The percentage of SLEC CD8+ T cells from PRMT5 CKO mice was significantly increased compared with that in the control group (Fig. 2J) after 24 and 48 h of stimulation. Moreover, apoptotic cells were also significantly increased in PRMT5 CKO mice (Fig. 2K). These data suggested that PRMT5 deficiency had much greater effects on CD8+ T than CD4+ T cells and induced SLEC CD8+ T cell development, which might cause CD8+ T cell loss.

To investigate how PRMT5 influenced CD8+ and CD4+ T cell development in the periphery, we analyzed the RNA profiles of purified CD8+ and CD4+ T cells by FACS from PRMT5 CKO and control littermates. PCA showed that CD8+ T cells were more different between the PRMT5 CKO and control groups than CD4+ T cells (Fig. 3A). CD8+ T cells expressed much more PRMT5 than CD4+ T cells in the periphery, and CD8+ T cells exhibited much higher levels of histone methylation than CD4+ T cells, including H3K4me3, H3R8me2s, and H4R3me2s (Fig. 3B and 3C). The number of differentiated genes (upregulation and downregulation) was also increased in CD8+ T cells compared with CD4+ T cells (582 versus 485; log2 FC >1; FDR <0.05) (Fig. 3D and Supplemental Fig. 2A). Notably, the number of upregulated transcripts (496) in CD8+ T cells was higher than the number of downregulated transcripts (86) in PRMT5 CKO mice (Fig. 3D), suggesting the critical role of PRMT5 in gene transcriptional repression. Gene Ontology (GO) analysis of the differential gene expression profiles showed that CD8+ T cells from PRMT5 CKO mice were significantly enriched in cell division, the inflammatory response, and pyroptosis process but were downregulated in MAPK activity and lipid metabolic process (Fig. 3E). GO analysis of the CD4+ T cell differential gene expression profiles showed an enrichment of these cells in the inflammatory response of PRMT5 CKO mice and a downregulation of the positive regulation of IL-10 production (Supplemental Fig. 2B). Moreover, gene set enrichment analysis (GSEA) indicated that CD8+ T cells from PRMT5 CKO mice were enriched in Klrg1hi CD8+ T cell gene and terminal effector gene signatures, whereas the naive gene signature was significantly enriched in control CD8+ T cells (Fig. 3F). A heat map plot showed that naive gene expression levels were lower in PRMT5 CKO mice, whereas terminal differentiation genes were expressed at much higher levels (Fig. 3G). We verified the expression of the Id2, Bhlhe40, Tbx21, Runx2, Klrg1, and Prdm1 genes in CD8+ T cells in vitro by QRT-PCR, and the results were consistent with the RNA-seq data (Fig. 3H). GO tree analysis further suggested that CD8+ T cells from PRMT5 CKO mice were enriched in the cell cycle and apoptosis processes (Supplemental Fig. 2C and 2D). Thus, in comparison with control mice, CD8+ T cells from PRMT5 CKO mice overexpressed a series of genes associated with terminal effector functions that led to cell death.

FIGURE 3.

Peripheral CD8+ T cells from PRMT5 CKO mice express a transcriptional profile enriched for effector genes. (A) PCA of the RNA profiles of purified CD8+ and CD4+ T cells from control (n = 3) and PRMT5 CKO (n = 3) mice. (B) QRT-PCR analysis of PRMT5 expression in CD4+ and CD8+ T cells isolated from control mouse spleens (n = 3). The mean ± SEM was graphed. (C) Western blot analysis of H3R8me2s, H3K4me3, H4R3me2s, and PRMT5 expression in CD4+ and CD8+ T cells isolated from control mouse spleens (n = 3). Actin was used as the loading control. One of the three similar representative datasets is shown. (D) Volcano plot showing the up- and downregulated genes in CD8+ T cells between control and PRMT5 CKO mice. (E) Bar graph of enriched terms of the differentially expressed genes of CD8+ T cells between the control and PRMT5 CKO mice. Blue boxes show the GO terms of interest. (F) GSEA was performed to determine the specifically enriched gene signatures in control and PRMT5 CKO murine CD8+ T cells. NES, normalized enrichment score. (G) Heatmap shows the effector, naive, and memory genes between control and PRMT5 CKO murine CD8+ T cells. (H) QRT-PCR verification of the related genes in CD8+ T cells from the control (n = 3) and PRMT5 CKO mice (n = 3). The mean ± SEM was graphed. Statistical differences were determined by two-tailed unpaired Student t test. *p < 0.05, **p < 0.01. NES, normalized enrichment score.

FIGURE 3.

Peripheral CD8+ T cells from PRMT5 CKO mice express a transcriptional profile enriched for effector genes. (A) PCA of the RNA profiles of purified CD8+ and CD4+ T cells from control (n = 3) and PRMT5 CKO (n = 3) mice. (B) QRT-PCR analysis of PRMT5 expression in CD4+ and CD8+ T cells isolated from control mouse spleens (n = 3). The mean ± SEM was graphed. (C) Western blot analysis of H3R8me2s, H3K4me3, H4R3me2s, and PRMT5 expression in CD4+ and CD8+ T cells isolated from control mouse spleens (n = 3). Actin was used as the loading control. One of the three similar representative datasets is shown. (D) Volcano plot showing the up- and downregulated genes in CD8+ T cells between control and PRMT5 CKO mice. (E) Bar graph of enriched terms of the differentially expressed genes of CD8+ T cells between the control and PRMT5 CKO mice. Blue boxes show the GO terms of interest. (F) GSEA was performed to determine the specifically enriched gene signatures in control and PRMT5 CKO murine CD8+ T cells. NES, normalized enrichment score. (G) Heatmap shows the effector, naive, and memory genes between control and PRMT5 CKO murine CD8+ T cells. (H) QRT-PCR verification of the related genes in CD8+ T cells from the control (n = 3) and PRMT5 CKO mice (n = 3). The mean ± SEM was graphed. Statistical differences were determined by two-tailed unpaired Student t test. *p < 0.05, **p < 0.01. NES, normalized enrichment score.

Close modal

To explore and dissect the heterogeneity of T cells between control and PRMT5 CKO mice in detail, we performed scRNA-seq. We isolated CD3+ T cells from control and PRMT5 CKO mouse spleens and acquired 6392 and 7266 cells, respectively, for sequencing. First, we conducted unsupervised clustering and clustered T cells into 19 clusters, including naïve, central memory, effector memory CD4+ and CD8+ T cells, CD4+ Treg cells, proliferating T cells, CD4+CD8+ DP T cells, and γδ T cells (Supplemental Fig. 3A and 3B). Consistent with the FACS analysis, we found that CD8+ T cells occupied a smaller proportion than CD4+ T cells in PRMT5 CKO mice. However, the proportions of proliferating T cells and γδ T cells were greatly increased in the PRMT5 CKO mice compared with the control mice, and the CD4+ Treg cell frequency showed a slight decrease (Fig. 4A). Interestingly, we found that cluster 7 (C7-CD4) expressed high levels of IFN response-related genes, such as Ifit3, Ifit1, Ifit2, and Irf7, but this cluster was increased in PRMT5 CKO mice; in contrast, cluster 11 (C11-CD4) expressed high levels of effector factors, such as Rorα, Cxcr3, IFN-γ, and IL2, and it was greatly decreased in PRMT5 CKO mice (Supplemental Fig. 3B, Fig. 4B).

FIGURE 4.

scRNA-seq revealed that Klrg1+ terminally differentiated CD8+ T cells were enriched in PRMT5 CKO mice. FACS-sorted CD3+ T cells from control (n = 3) and PRMT5 CKO mouse (n = 3) spleens and scRNA-seq. (A) The percentages (left panel) and ratios (right panel) of immune cell subsets of CD4+ Treg cells, proliferating T cells, γδ+ T cells, CD4+ T cells, CD8+ T cells, and CD4+CD8+ T cells (DP) between control and PRMT5 CKO mouse spleens. (B) The ratios of the related CD4+ T cell clusters between control and PRMT5 CKO mouse spleens. (C) tSNE plot showing the CD8+ T cell reclusters, including naive, central memory (TCM), effector, and proliferative T cells. Each dot corresponds to a single cell colored according to the cell cluster. (D) Expression levels of relative marker genes across CD8+ T cells illustrated as tSNE plots. The expression was measured as the log2 (count + 1). (E) The ratios of T cell subsets between PRMT5 CKO and control mice. (F) The trajectory of CD8+ T cell state transition in two dimensions inferred by Monocle. Each dot corresponds to a single cell colored according to its cluster label. Arrows show the direction of increase of certain cell properties. (G) Heatmap of the t values of area under the curve scores of expression regulation by transcription factors of the indicated clusters, as estimated using SCENIC. (H) Violin plots comparing the indicated gene expression in CD8+ T cell clusters. Expression was measured as the log2 (count + 1). (I) Network analysis of Prdm1 and its target genes. The length of the line represents the coexpression relationship score.

FIGURE 4.

scRNA-seq revealed that Klrg1+ terminally differentiated CD8+ T cells were enriched in PRMT5 CKO mice. FACS-sorted CD3+ T cells from control (n = 3) and PRMT5 CKO mouse (n = 3) spleens and scRNA-seq. (A) The percentages (left panel) and ratios (right panel) of immune cell subsets of CD4+ Treg cells, proliferating T cells, γδ+ T cells, CD4+ T cells, CD8+ T cells, and CD4+CD8+ T cells (DP) between control and PRMT5 CKO mouse spleens. (B) The ratios of the related CD4+ T cell clusters between control and PRMT5 CKO mouse spleens. (C) tSNE plot showing the CD8+ T cell reclusters, including naive, central memory (TCM), effector, and proliferative T cells. Each dot corresponds to a single cell colored according to the cell cluster. (D) Expression levels of relative marker genes across CD8+ T cells illustrated as tSNE plots. The expression was measured as the log2 (count + 1). (E) The ratios of T cell subsets between PRMT5 CKO and control mice. (F) The trajectory of CD8+ T cell state transition in two dimensions inferred by Monocle. Each dot corresponds to a single cell colored according to its cluster label. Arrows show the direction of increase of certain cell properties. (G) Heatmap of the t values of area under the curve scores of expression regulation by transcription factors of the indicated clusters, as estimated using SCENIC. (H) Violin plots comparing the indicated gene expression in CD8+ T cell clusters. Expression was measured as the log2 (count + 1). (I) Network analysis of Prdm1 and its target genes. The length of the line represents the coexpression relationship score.

Close modal

Next, we reclustered CD8+ T cells on the basis of distinct transcription profiles, grouping the unsupervised clusters into four major subset categories: naive, central memory, effector, and cycling cells (Fig. 4C, Supplemental Fig. 3C). As expected, the naive cells grouped in a homogeneous category characterized by the highest expression of Sell (CD62L); the central memory cells were enriched in Il7r, Cxcr3, and Ly6c genes (cluster 5 and cluster 11); the effectors were characterized by Zeb2, Klrg1, and Gzmb (cluster 9); and the cycling subset was defined by cell cycle genes, including Birc5 and Stmn1 (cluster 12) (Fig. 4D, Supplemental Fig. 3C). The proportions of effector and proliferative T cells were greatly increased in PRMT5 CKO mice (Fig. 4E).

We then analyzed CD8+ T cell development trajectories and found that most cells from each cluster aggregated on the basis of expression similarities and presented a branched structure, with cluster 9 effector T cells positioned at the opposite end of both naive clusters and the proliferating cluster (Fig. 4F). Cluster 9 aggregated much more in PRMT5 CKO mice than in control mice (Fig. 4F). We further applied the SCENIC method to explore the transcription factors that might regulate the development of different CD8+ T cell clusters. Consistent with the phenotype, cluster 9 expressed much lower levels of memory-related genes, such as Bach2, Bcl2, and Sox4, but high levels of the terminal differentiation-related genes Tbx21 and Prdm1 (Fig. 4G and 4H). We conducted network analyses to identify Prdm1 downstream genes and revealed Prdm1-related terminal effector program genes, including Zeb2, Ccl3, Gzma, Bhlhe40, Cxcr3, Il18rap, and Klrg1 (Fig. 4I). GO analysis showed that Prdm1 downstream genes were associated with T cell activation, differentiation, and cell death (Supplemental Fig. 3D).

The effector CD8+ T cell pool consists of two subsets: The majority are terminal effectors that are destined to die, whereas the minority subset of MPECs survive to give rise to the pool of long-lived memory T cells (28). Recent work has demonstrated that Klrg1+ effector CD8+ T cells lose Klrg1, differentiate into memory T cell lineages, and convey enhanced protective immunity (29). Thus, we reclustered the effector CD8+ T cells and further divided them into three clusters: Cluster 0 expressed high levels of the Klrg1, S1pr5, Zeb2, and Cx3cr1 genes and represented the terminal effector cluster; cluster 1 expressed high levels of IL7r, Ly6e, and IL2rb but low levels of Klrg1 and represented long-lived memory T cells; and cluster 2 showed intermediate expression levels of these genes between cluster 0 and cluster 1 (Supplemental Fig. 3E and 3F). The percentage of cluster 0 was much higher in PRMT5 CKO mice than in control mice (Supplemental Fig. 3G). Thus, these results provided unequivocal evidence that PRMT5-deficient CD8+ T cells tended to undergo terminal effector differentiation and cause CD8+ T cell number loss in PRMT5 CKO mice.

Epigenetic programming is critical for the differentiation of T cells. To elucidate the epigenetic states associated with the terminal effector gene loci and to investigate the contribution of PRMT5-dependent chromatin changes, we performed ChIP-seq for the repressive markers H4R3me2s and H3R8me2s in CD8+ T cells isolated from control and PRMT5 CKO mice. The total number of H4R3me2s islands was not significantly changed between these two groups, but the number of H3R8me2s target islands increased in control mice compared with PRMT5 CKO mice (Fig. 5A). Similar to the common peaks, variable peaks caused by H4R3me2s and H3R8me2s were enriched in distal intergenic regions or introns and were less likely to be found in promoters (Fig. 5B). These findings suggested that PRMT5 might modify the deposition of different histones and regulate the expression of different genes, thus affecting the differentiation of CD8+ T cell subsets.

FIGURE 5.

H4R3me2s and H3R8me2s epigenetic modifications induced Blimp1 expression in PRMT5 CKO mice. CD8+ T cells were isolated from control (n = 5) and PRMT5 CKO (n = 5) mice and were analyzed by ChIP-seq using Abs against H4R3me2s and H3R8me2s. (A) Venn diagrams summarizing the H4R3me2s and H3R8me2s Ab binding peaks and the overlapping peaks between control and PRMT5 CKO mice. (B) Distribution of H4R3me2s and H3R8me2s ChIP-seq peaks in CD8+ T cells. (C) Venn diagram of PRMT5 repressed, H4R3me2s, and H3R8me2s target genes, which indicated nine overlapping genes. (D) Representative Prdm1 genomic region deposition of H3R8me2s in control and PRMT5 CKO mouse CD8+ T cells. Normalized input ChIP-seq reads. (E) Representative Prdm1 genomic region deposition of H4R3me2s in control and PRMT5 CKO mouse CD8+ T cells. Normalized input ChIP-seq reads. (F) CD8+ T cells isolated from control and PRMT5 CKO mice were analyzed by ChIP-qPCR using Abs against H3R8me2s and H4R3me2s, respectively, Prdm1 loci was measured. Data are the mean ± SEM of three independent experiments. *p < 0.05.

FIGURE 5.

H4R3me2s and H3R8me2s epigenetic modifications induced Blimp1 expression in PRMT5 CKO mice. CD8+ T cells were isolated from control (n = 5) and PRMT5 CKO (n = 5) mice and were analyzed by ChIP-seq using Abs against H4R3me2s and H3R8me2s. (A) Venn diagrams summarizing the H4R3me2s and H3R8me2s Ab binding peaks and the overlapping peaks between control and PRMT5 CKO mice. (B) Distribution of H4R3me2s and H3R8me2s ChIP-seq peaks in CD8+ T cells. (C) Venn diagram of PRMT5 repressed, H4R3me2s, and H3R8me2s target genes, which indicated nine overlapping genes. (D) Representative Prdm1 genomic region deposition of H3R8me2s in control and PRMT5 CKO mouse CD8+ T cells. Normalized input ChIP-seq reads. (E) Representative Prdm1 genomic region deposition of H4R3me2s in control and PRMT5 CKO mouse CD8+ T cells. Normalized input ChIP-seq reads. (F) CD8+ T cells isolated from control and PRMT5 CKO mice were analyzed by ChIP-qPCR using Abs against H3R8me2s and H4R3me2s, respectively, Prdm1 loci was measured. Data are the mean ± SEM of three independent experiments. *p < 0.05.

Close modal

Among the 454 genes that were upregulated in vivo in PRMT5 CKO cells compared with control CD8+ T cells (Fig. 3D), 37 and 42 genes were less decorated by H4R3me2s and H3R8me2s, respectively (Fig. 5C). Nine genes were shared between these three groups (Msc, Lrrn3, Map9, Prdm1, Runx2, Lrig3, Arhgap20, Dync2h1, and Diaph3), including the terminal effector CD8+ T cell fate decision–related gene Prdm1 (Blimp1). In PRMT5 CKO mice, compared with the controls, H4R3me2s and H3R8me2s deposition were both decreased on the Prdm1 loci, mainly in introns and distal intergenic regions in CD8+ T cells (Fig. 5D and 5E). The transcriptional repressor Blimp1 has been reported to be required for terminal CD8+ T cell differentiation and represses the acquisition of central memory T cell properties (30). Consistent with these results, ChIP-PCR verified that the levels of the alternative repressive markers H4R3me2s and H3R8me2s were decreased on Prdm1 loci in PRMT5 CKO CD8+ T cells, which again supported Prdm1 expression (Fig. 5F). These results indicated that in PRMT5-deficient CD8+ T cells, Blimp1 failed to acquire the repressive mark of H4R3me2s and H3R8me2s, resulting in defective silencing and the promotion of terminal effector CD8+ T cell differentiation.

To further determine the effect of T cell depletion of PRMT5 on tumor immunity, we first implanted syngeneic LLC lung cancer cells s.c. into PRMT5 CKO mice and littermate controls and then measured tumor size at a series of time points after implantation. As expected, PRMT5 CKO mice developed larger tumor volumes and weights than the control animals (Fig. 6A and 6B and Supplemental Fig. 4A). In addition, the absolute numbers of infiltrating CD45+ immune cells and CD4+ and CD8+ T cells were reduced in PRMT5 CKO mice (Fig. 6C and D). The percentages of CD8+ and NK cells in CD45+ cells were decreased, but the percentages of CD11b+ cells, neutrophils, and monocytes were increased, and the percentages of B cells, dendritic cells, macrophages, and CD4+ T cells were not significantly different between these two groups (Fig. 6E). Moreover, we noted a markedly reduced ratio of CD8+ T cells among the tumor-infiltrating immune cells in PRMT5 CKO mice compared with naive mice (Fig. 6F), suggesting that the tumor inflammatory environment might cause much more CD8+ T cell loss in PRMT5 CKO mice. Strikingly, the percentage of CD4+Foxp3+ Treg cells was decreased in PRMT5 CKO mice compared with that in control mice, but the percentage of Foxp3+CD8+ T cells was significantly increased (Fig. 6G).

FIGURE 6.

T cell–specific deletion of PRMT5 promoted tumor progression. LLC lung cancer cells were s.c. implanted into PRMT5 CKO (n = 7) mice, along with littermate controls (n = 5), and (A) tumor size was measured at a series of time points after implantation. (B) Tumor weights in the two groups. (C) The absolute number of CD45+ immune cells in the two groups. (D) The absolute numbers of CD4+ and CD8+ T cells in the two groups. (E) The percentages of different infiltrating immune cell subsets in the two groups. (F) Reduced ratios of PRMT5 CKO to the control in naive and tumor states. (G) Percentages of Foxp3+CD25+ Treg cells in CD4+ and CD8+ T cells between the two groups. (H) Percentages of PD-L1+ cells in CD4+ and CD8+ T cells between the two groups. (I) Percentages of TNFα+ and IFN-γ+ cells in CD4+ and CD8+ T cells between the two groups. (J, K) Percentages of Klrg1+CD127, Klrg1+CD127+, and Klrg1CD127+ T cells in CD4+ and CD8+ T cells from spleens (J) and tumor tissues (K). (LN) MC38 colon cancer cells were s.c. implanted into PRMT5 CKO and littermate control mice (n = 4 each group). (L) Tumor weights. (M) Percentages of Klrg1+CD127, Klrg1+CD127+, and Klrg1CD127+ T cells in CD8+ T cells from tumor tissues. (N) The percentages of PD-L1+, CD39+, and TIM3+ cells in CD8+ T cells between these two groups. Data are representative of two independent experiments. Bar graphs show the level of mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; two-tailed t test.

FIGURE 6.

T cell–specific deletion of PRMT5 promoted tumor progression. LLC lung cancer cells were s.c. implanted into PRMT5 CKO (n = 7) mice, along with littermate controls (n = 5), and (A) tumor size was measured at a series of time points after implantation. (B) Tumor weights in the two groups. (C) The absolute number of CD45+ immune cells in the two groups. (D) The absolute numbers of CD4+ and CD8+ T cells in the two groups. (E) The percentages of different infiltrating immune cell subsets in the two groups. (F) Reduced ratios of PRMT5 CKO to the control in naive and tumor states. (G) Percentages of Foxp3+CD25+ Treg cells in CD4+ and CD8+ T cells between the two groups. (H) Percentages of PD-L1+ cells in CD4+ and CD8+ T cells between the two groups. (I) Percentages of TNFα+ and IFN-γ+ cells in CD4+ and CD8+ T cells between the two groups. (J, K) Percentages of Klrg1+CD127, Klrg1+CD127+, and Klrg1CD127+ T cells in CD4+ and CD8+ T cells from spleens (J) and tumor tissues (K). (LN) MC38 colon cancer cells were s.c. implanted into PRMT5 CKO and littermate control mice (n = 4 each group). (L) Tumor weights. (M) Percentages of Klrg1+CD127, Klrg1+CD127+, and Klrg1CD127+ T cells in CD8+ T cells from tumor tissues. (N) The percentages of PD-L1+, CD39+, and TIM3+ cells in CD8+ T cells between these two groups. Data are representative of two independent experiments. Bar graphs show the level of mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; two-tailed t test.

Close modal

Diskin et al. reported the prevalence of PD-L1 expression on T cells in a preinvasive autochthonous model of pancreatic ductal adenocarcinoma and murine melanoma (31). Furthermore, PD-L1+ T cells suppress neighboring T cells and induce an alternative macrophage program via PD-1 ligation (31). In addition, PD-L1 expressed on CD8+ T cells enhances the suppression of the CD8+ T cell–mediated immune response (32). PRMT5 knockdown melanoma cells exhibit higher expression of PD-L1 (33). These reports suggest that PRMT5 may regulate PD-L1 expression and that PD-L1–expressing CD8+ T cells have immunoregulatory activity, so we aimed to further detect PD-L1 expression in PRMT5-deficient T cells. Interestingly, PD-L1 expression was greatly increased in PRMT5 CKO mice (Fig. 6H), suggesting that PD-L1+CD8+ T cells might contribute to the progression of tumors in PRMT5 CKO mice. In addition, the inflammatory cytokines TNF-α and IFN-γ were greatly decreased in CD4+ T and CD8+ T cells from PTMT5 CKO mice (Fig. 6I and Supplemental Fig. 4B). The expression levels of the exhaustion-related molecules CD39, PD1, and TIM3 were not significantly different (Supplemental Fig. 4C). We further measured SLEC T cells in the tumor model. Similar to the case in naive mice, the percentages of SLEC T cells in spleens were greatly increased in PRMT5 CKO mice (Fig. 6J). Notably, in tumor tissues, SLECs in CD4+ T cells were not increased in PRMT5 CKO mice but were greatly increased in CD8+ T cells (Fig. 6K). However, CD4+ and CD8+ T cell proliferation showed a downward trend in PRMT5 CKO mice (Supplemental Fig. 4D). Moreover, we implanted syngeneic MC38 colon cancer cells s.c. into PRMT5 CKO mice and littermate controls, and the data were consistent with the LLC mouse tumor model, including the tumor weights, percentages of infiltrating CD8+ T cells, SLEC CD8+ T cells, and CD8+PD-L1+ T cells (Fig. 6L6N and Supplemental Fig. 4E–4G). Interestingly, the exhaustion-related molecules TIM3 and CD39 were increased in CD8+ T cells from PRMT5 CKO mice (Fig. 6N), in contrast to observations in the LLC mouse tumor model, suggesting that different types of tumor microenvironments (TMEs) might influence CD8+ T cell phenotypes. These data suggested that T cell–specific deletion of PRMT5 induced many more Klrg1+CD8+ effector T cells and CD8+ Treg cells, thus causing substantial loss and impaired function of CD8+ T cells, which might be responsible for tumor progression in the TME.

Several mouse models have been established to investigate the importance of PRMT5 in hematopoiesis, including bone marrow hematopoietic stem and progenitor cell, T cell, B cell, and erythroid cell (15, 17, 34). However, how PRMT5 affects the differentiation of CD8+ T cell subsets in the periphery, as well as the role in antitumor activity and related molecular mechanisms, has not been explained in detail. Our study uncovered a mechanism by which PRMT5 affected H4R3me2s and H3R8me2s deposition on the Prdm1 loci and enforced terminal short-life effector CD8+ T cell differentiation, causing substantial CD8+ T cell loss. Our study further illustrated the role of T cells carrying a specific deletion of PRMT5 in tumor progression and revealed that T cell–specific loss of PRMT5 had a greater impact on CD8+ than CD4+ T cell differentiation, inducing more Klrg1+CD8+ T cells and Foxp3+CD8+ and PDL1+CD8+ Treg cells in cancer line–transplanted mouse models, thus eliminating antitumor activity.

Persistent expression of Klrg1, which is enforced by Blimp1, has been associated with terminal differentiation. Recently, Garg et al. reported that Klrg1+ Treg cells express high amounts of Blimp1, whereas Klrg1 Treg cells essentially lack Blimp1 (35). Our work showed the high frequency of Foxp3+CD8+ T cells in PRMT5 CKO mice in a cancer cell line–transplanted tumor model, indicating the critical role of Blimp1 in the regulation of Klrg1 and Foxp3 expression in PRMT5 CKO CD8+ T cells. Moreover, SCENIC analysis revealed that Prdm1 was exclusively expressed in Klrg1+ T cells, supporting the important roles of Blimp1 in regulating CD8+ T cell subset differentiation in PRMT5-deficient T cells. Blimp1 is regulated by inflammatory cytokines through STAT signaling (36, 37). PRMT5 has been reported to control H4R3me2s in mouse embryonic fibroblasts (38) and to increase H3R8me2s modification, thus epigenetically decreasing IL-12 expression in the innate response (39). We provide novel mechanisms by which CD8+ T cell depletion of PRMT5 reduces H4R3 and H3R8 methylation, thus decreasing deposition on the Blimp1 loci and thereby activating Blimp1 expression, which might induce SLEC CD8+ and Foxp3+CD8+ T cell development.

Tanaka et al. (40) reported that PRMT5 is required for T cell survival and proliferation by maintaining cytokine signaling. Similarly, Snyder et al. (41) demonstrated that PRMT5 regulates the T cell–mediated IFN response. Specific loss of PRMT5 in the CD4+ Th cell compartment suppresses Th17 differentiation and protects mice from developing EAE (19). Webb et al. pointed out that the NF-κB–mammalian target of rapamycin–MYC axis drives PRMT5 protein induction after T cell activation (42). Our results are consistent with these previous studies because PRMT5 expression was elevated after T cell activation; T cell–specific deletion of PRMT5 led to peripheral T cell lymphopenia in mice; and PRMT5 was required for the survival and proliferation of T cells. Moreover, our research using scRNA-seq, RNA-seq, and epigenomic profiling further explored the effect of PRMT5 on T cell subset differentiation and the detailed underlying regulatory mechanisms. First, we revealed that PRMT5 had greater effects on CD8+ than CD4+ T cell development due to high PRMT5 expression on CD8+ T cells. Second, PRMT5 depletion induced naive CD8+ T cells to differentiate into SLEC CD8+ T cells directly and caused cell death. Third, T cell deficiency of PRMT5 activated Prdm1 by decreasing H4R3me2s and H3R8me2s deposition on its loci, which promoted the differentiation of SLEC CD8+ T cells. Finally, our work revealed the tumor-promoting ability of T cells carrying a specific deletion of PRMT5 in two cancer cell line transplanted tumor models. Overall, our study provides an in-depth genomic analysis of PRMT5-deficient T cells and defines the central roles of PRMT5 in T cell physiology and homeostasis.

The driving role of PRMT5 in many types of solid and hematologic cancers has led to efforts to develop PRMT5-selective inhibitors, some of which have now been tested in clinical trials for safety and efficacy (43); however, the effects of PRMT5 on the immune system, especially on CD8+ T cells, cannot be ignored. Recent studies have reported the suppression of CD8+ T cells after treatment with PRMT5 inhibitors (EPZ015666 and DST-437) (21, 44), and our research provides more information relevant to the application of PRMT5 inhibitors because PRMT5 deletion not only significantly inhibits the survival of CD8+ T cells but also induces CD8+ Treg cells, which cause obvious adverse effects. However, the limitation of our study is that it is still unclear whether migration or stimulators in the TME may cause increased infiltration of SLEC CD8+ T cells and CD8+ Treg cells in PRMT5 CKO mice. An inducible system of PRMT5 in the TME may offer more insight in future studies. The present and other reports (17, 40) showed the presence of lymphopenia in PRMT5-specific deletion mice, the induction of T cell proliferation by lymphopenia, and the transient acquisition by naive T cells of a memory-like phenotype (45, 46). Recently, Inoue et al. used a bone marrow reconstruction model in which irradiated Rag1−/− mice were reconstituted with a 1:1 mixture of bone marrow cells from PRMT5 CKO (CD45.2+) and wild-type (CD45.1+) mice (18). PRMT5 CKO bone marrow cells generated fewer CD4+ T cells and CD8+ T cells than wild-type bone marrow cells, which indicated that PRMT5 regulated the number of peripheral CD4+ T cells and CD8+ T cells in a cell-intrinsic manner. In addition, we found that naive CD8+ T cells from the PRMT5 CKO group were more easily activated, more easily differentiated into SLECs, and more able to cause cell apoptosis than those from the control group when stimulated in vitro (Fig. 2I and 2J). These data indicated that the high level of SLEC CD8+ T cells in the PRMT5 CKO mice might not be caused primarily by lymphopenia, but that a T cell deficiency of PRMT5 contributed to the SLEC program initially, leading to increased cell loss. Overall, our findings provide a basis for how PRMT5 controls the differentiation and function of CD8+ T cell subsets, and targeting PRMT5 in tumor treatment should receive considerable attention, because PRMT5 deficiency enforces Klrg1+ terminal CD8+ T cell development and eliminates antitumor activity.

We thank NovelBio Biotechnology of Shanghai for performing the scRNA-seq and data analyses.

This work was supported by grants from the National Natural Science Foundation of China (81571525, 81873863, and 82071753 to Y.Z.; 81672363 to L.S.), Shanghai Municipal Education Commission Gaofeng Clinical Medicine Grant Support (20161315 to Y.Z.), the Medicine and Engineering Cross Research Foundation of Shanghai Jiao Tong University (Project YG2017ZD02 to L.S.), and the Shanghai Science and Technology Committee (20ZR1448900 to H.N.).

L.S., H.N., and Y.Z. designed the experiments. Y.Z., Z.C., S.C., B.Z., L.H., and N.C. performed the experiments. Y.Z., Z.C., S.C., B.Z., Y.M., G.X., and J.Y. analyzed the sequencing data. L.S., H.N., and Y.Z. wrote the manuscript, and all authors contributed to writing and providing feedback on the manuscript.

The single-cell RNA-sequencing, bulk RNA-sequencing, and chromatin immunoprecipitation sequencing data presented in this article have been submitted to the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE186863) under accession number GSE186863.

The online version of this article contains supplemental material.

Abbreviations used in this article

ChIP

chromatin immunoprecipitation

ChIP-seq

chromatin immunoprecipitation sequencing

CKO

conditional knockout

Cy7

cyanine 7

DP

double positive

EAE

experimental autoimmune encephalomyelitis

FDR

false discovery rate

GO

Gene Ontology

GSEA

gene set enrichment analysis

iNKT

invariant NKT

MPEC

memory precursor effector cell

PCA

principal component analysis

PRMT5

protein arginine methyltransferase 5

QRT-PCR

quantitative real-time PCR

RNA-seq

RNA sequencing

SCENIC

single-cell regulatory network inference and clustering

scRNA-seq

single-cell RNA sequencing

SLEC

short-lived effector cell

SP

single positive

TME

tumor microenvironment

Treg

regulatory T

tSNE

t-distributed stochastic neighbor embedding

1.
Blanc
R. S.
,
S.
Richard
.
2017
.
Arginine methylation: the coming of age.
Mol. Cell
65
:
8
24
.
2.
Di Lorenzo
A.
,
M. T.
Bedford
.
2011
.
Histone arginine methylation.
FEBS Lett.
585
:
2024
2031
.
3.
Hamard
P. J.
,
G. E.
Santiago
,
F.
Liu
,
D. L.
Karl
,
C.
Martinez
,
N.
Man
,
A. K.
Mookhtiar
,
S.
Duffort
,
S.
Greenblatt
,
R. E.
Verdun
,
S. D.
Nimer
.
2018
.
PRMT5 regulates DNA repair by controlling the alternative splicing of histone-modifying enzymes.
Cell Rep.
24
:
2643
2657
.
4.
Radzisheuskaya
A.
,
P. V.
Shliaha
,
V.
Grinev
,
E.
Lorenzini
,
S.
Kovalchuk
,
D.
Shlyueva
,
V.
Gorshkov
,
R. C.
Hendrickson
,
O. N.
Jensen
,
K.
Helin
.
2019
.
PRMT5 methylome profiling uncovers a direct link to splicing regulation in acute myeloid leukemia.
Nat. Struct. Mol. Biol.
26
:
999
1012
.
5.
Stopa
N.
,
J. E.
Krebs
,
D.
Shechter
.
2015
.
The PRMT5 arginine methyltransferase: many roles in development, cancer and beyond.
Cell. Mol. Life Sci.
72
:
2041
2059
.
6.
Antonysamy
S.
2017
.
The structure and function of the PRMT5:MEP50 complex.
Subcell. Biochem.
83
:
185
194
.
7.
Cao
L.
,
G.
Wu
,
J.
Zhu
,
Z.
Tan
,
D.
Shi
,
X.
Wu
,
M.
Tang
,
Z.
Li
,
Y.
Hu
,
S.
Zhang
, et al
2019
.
Genotoxic stress-triggered β-catenin/JDP2/PRMT5 complex facilitates reestablishing glutathione homeostasis.
Nat. Commun.
10
:
3761
.
8.
Rank
G.
,
L.
Cerruti
,
R. J.
Simpson
,
R. L.
Moritz
,
S. M.
Jane
,
Q.
Zhao
.
2010
.
Identification of a PRMT5-dependent repressor complex linked to silencing of human fetal globin gene expression.
Blood
116
:
1585
1592
.
9.
Majumder
S.
,
L.
Alinari
,
S.
Roy
,
T.
Miller
,
J.
Datta
,
S.
Sif
,
R.
Baiocchi
,
S. T.
Jacob
.
2010
.
Methylation of histone H3 and H4 by PRMT5 regulates ribosomal RNA gene transcription.
J. Cell. Biochem.
109
:
553
563
.
10.
Youngblood
B.
,
J. S.
Hale
,
H. T.
Kissick
,
E.
Ahn
,
X.
Xu
,
A.
Wieland
,
K.
Araki
,
E. E.
West
,
H. E.
Ghoneim
,
Y.
Fan
, et al
2017
.
Effector CD8 T cells dedifferentiate into long-lived memory cells.
Nature
552
:
404
409
.
11.
Yu
B.
,
K.
Zhang
,
J. J.
Milner
,
C.
Toma
,
R.
Chen
,
J. P.
Scott-Browne
,
R. M.
Pereira
,
S.
Crotty
,
J. T.
Chang
,
M. E.
Pipkin
, et al
2017
.
Epigenetic landscapes reveal transcription factors that regulate CD8+ T cell differentiation. [Published erratum appears in 2017 Nat. Immunol. 18: 705.]
Nat. Immunol.
18
:
573
582
.
12.
Chen
Z.
,
Z.
Ji
,
S. F.
Ngiow
,
S.
Manne
,
Z.
Cai
,
A. C.
Huang
,
J.
Johnson
,
R. P.
Staupe
,
B.
Bengsch
,
C.
Xu
, et al
2019
.
TCF-1-centered transcriptional network drives an effector versus exhausted CD8 T cell-fate decision.
Immunity
51
:
840
855.e5
.
13.
Milner
J. J.
,
C.
Toma
,
B.
Yu
,
K.
Zhang
,
K.
Omilusik
,
A. T.
Phan
,
D.
Wang
,
A. J.
Getzler
,
T.
Nguyen
,
S.
Crotty
, et al
2017
.
Runx3 programs CD8+ T cell residency in non-lymphoid tissues and tumours. [Published erratum appears in 2018 Nature 554: 392]
Nature
552
:
253
257
.
14.
Xin
A.
,
F.
Masson
,
Y.
Liao
,
S.
Preston
,
T.
Guan
,
R.
Gloury
,
M.
Olshansky
,
J. X.
Lin
,
P.
Li
,
T. P.
Speed
, et al
2016
.
A molecular threshold for effector CD8+ T cell differentiation controlled by transcription factors Blimp-1 and T-bet.
Nat. Immunol.
17
:
422
432
.
15.
Tee
W. W.
,
M.
Pardo
,
T. W.
Theunissen
,
L.
Yu
,
J. S.
Choudhary
,
P.
Hajkova
,
M. A.
Surani
.
2010
.
Prmt5 is essential for early mouse development and acts in the cytoplasm to maintain ES cell pluripotency.
Genes Dev.
24
:
2772
2777
.
16.
Zhang
T.
,
S.
Günther
,
M.
Looso
,
C.
Künne
,
M.
Krüger
,
J.
Kim
,
Y.
Zhou
,
T.
Braun
.
2015
.
Prmt5 is a regulator of muscle stem cell expansion in adult mice.
Nat. Commun.
6
:
7140
.
17.
Liu
F.
,
G.
Cheng
,
P. J.
Hamard
,
S.
Greenblatt
,
L.
Wang
,
N.
Man
,
F.
Perna
,
H.
Xu
,
M.
Tadi
,
L.
Luciani
,
S. D.
Nimer
.
2015
.
Arginine methyltransferase PRMT5 is essential for sustaining normal adult hematopoiesis.
J. Clin. Invest.
125
:
3532
3544
.
18.
Inoue
M.
,
K.
Okamoto
,
A.
Terashima
,
T.
Nitta
,
R.
Muro
,
T.
Negishi-Koga
,
T.
Kitamura
,
T.
Nakashima
,
H.
Takayanagi
.
2018
.
Arginine methylation controls the strength of γc-family cytokine signaling in T cell maintenance.
Nat. Immunol.
19
:
1265
1276
.
19.
Webb
L. M.
,
S.
Sengupta
,
C.
Edell
,
Z. L.
Piedra-Quintero
,
S. A.
Amici
,
J. N.
Miranda
,
M.
Bevins
,
A.
Kennemer
,
G.
Laliotis
,
P. N.
Tsichlis
,
M.
Guerau-de-Arellano
.
2020
.
Protein arginine methyltransferase 5 promotes cholesterol biosynthesis-mediated Th17 responses and autoimmunity.
J. Clin. Invest.
130
:
1683
1698
.
20.
Webb
L. M.
,
S. A.
Amici
,
K. A.
Jablonski
,
H.
Savardekar
,
A. R.
Panfil
,
L.
Li
,
W.
Zhou
,
K.
Peine
,
V.
Karkhanis
,
E. M.
Bachelder
, et al
2017
.
PRMT5-selective inhibitors suppress inflammatory T cell responses and experimental autoimmune encephalomyelitis.
J. Immunol.
198
:
1439
1451
.
21.
Strobl
C. D.
,
S.
Schaffer
,
T.
Haug
,
S.
Völkl
,
K.
Peter
,
K.
Singer
,
M.
Böttcher
,
D.
Mougiakakos
,
A.
Mackensen
,
M.
Aigner
.
2020
.
Selective PRMT5 inhibitors suppress human CD8+ T cells by upregulation of p53 and impairment of the AKT pathway similar to the tumor metabolite MTA.
Mol. Cancer Ther.
19
:
409
419
.
22.
Qin
Y.
,
Q.
Hu
,
J.
Xu
,
S.
Ji
,
W.
Dai
,
W.
Liu
,
W.
Xu
,
Q.
Sun
,
Z.
Zhang
,
Q.
Ni
, et al
2019
.
PRMT5 enhances tumorigenicity and glycolysis in pancreatic cancer via the FBW7/cMyc axis.
Cell Commun. Signal.
17
:
30
.
23.
Yan
F.
,
L.
Alinari
,
M. E.
Lustberg
,
L. K.
Martin
,
H. M.
Cordero-Nieves
,
Y.
Banasavadi-Siddegowda
,
S.
Virk
,
J.
Barnholtz-Sloan
,
E. H.
Bell
,
J.
Wojton
, et al
2014
.
Genetic validation of the protein arginine methyltransferase PRMT5 as a candidate therapeutic target in glioblastoma.
Cancer Res.
74
:
1752
1765
.
24.
Chen
H.
,
B.
Lorton
,
V.
Gupta
,
D.
Shechter
.
2017
.
A TGFβ-PRMT5-MEP50 axis regulates cancer cell invasion through histone H3 and H4 arginine methylation coupled transcriptional activation and repression.
Oncogene
36
:
373
386
.
25.
Xiao
W.
,
X.
Chen
,
L.
Liu
,
Y.
Shu
,
M.
Zhang
,
Y.
Zhong
.
2019
.
Role of protein arginine methyltransferase 5 in human cancers.
Biomed. Pharmacother.
114
:
108790
.
26.
Pal
S.
,
R. A.
Baiocchi
,
J. C.
Byrd
,
M. R.
Grever
,
S. T.
Jacob
,
S.
Sif
.
2007
.
Low levels of miR-92b/96 induce PRMT5 translation and H3R8/H4R3 methylation in mantle cell lymphoma.
EMBO J.
26
:
3558
3569
.
27.
Chiang
K.
,
A. E.
Zielinska
,
A. M.
Shaaban
,
M. P.
Sanchez-Bailon
,
J.
Jarrold
,
T. L.
Clarke
,
J.
Zhang
,
A.
Francis
,
L. J.
Jones
,
S.
Smith
, et al
2017
.
PRMT5 is a critical regulator of breast cancer stem cell function via histone methylation and FOXP1 expression.
Cell Rep.
21
:
3498
3513
.
28.
Kaech
S. M.
,
S.
Hemby
,
E.
Kersh
,
R.
Ahmed
.
2002
.
Molecular and functional profiling of memory CD8 T cell differentiation.
Cell
111
:
837
851
.
29.
Herndler-Brandstetter
D.
,
H.
Ishigame
,
R.
Shinnakasu
,
V.
Plajer
,
C.
Stecher
,
J.
Zhao
,
M.
Lietzenmayer
,
L.
Kroehling
,
A.
Takumi
,
K.
Kometani
, et al
2018
.
KLRG1+ effector CD8+ T cells lose KLRG1, differentiate into all memory T cell lineages, and convey enhanced protective immunity.
Immunity
48
:
716
729.e8
.
30.
Rutishauser
R. L.
,
G. A.
Martins
,
S.
Kalachikov
,
A.
Chandele
,
I. A.
Parish
,
E.
Meffre
,
J.
Jacob
,
K.
Calame
,
S. M.
Kaech
.
2009
.
Transcriptional repressor Blimp-1 promotes CD8+ T cell terminal differentiation and represses the acquisition of central memory T cell properties.
Immunity
31
:
296
308
.
31.
Diskin
B.
,
S.
Adam
,
M. F.
Cassini
,
G.
Sanchez
,
M.
Liria
,
B.
Aykut
,
C.
Buttar
,
E.
Li
,
B.
Sundberg
,
R. D.
Salas
, et al
2020
.
PD-L1 engagement on T cells promotes self-tolerance and suppression of neighboring macrophages and effector T cells in cancer.
Nat. Immunol.
21
:
442
454
.
32.
Park
H. J.
,
J. S.
Park
,
Y. H.
Jeong
,
J.
Son
,
Y. H.
Ban
,
B. H.
Lee
,
L.
Chen
,
J.
Chang
,
D. H.
Chung
,
I.
Choi
,
S. J.
Ha
.
2015
.
PD-1 upregulated on regulatory T cells during chronic virus infection enhances the suppression of CD8+ T cell immune response via the interaction with PD-L1 expressed on CD8+ T cells.
J. Immunol.
194
:
5801
5811
.
33.
Kim
H.
,
H.
Kim
,
Y.
Feng
,
Y.
Li
,
H.
Tamiya
,
S.
Tocci
,
Z. A.
Ronai
.
2020
.
PRMT5 control of cGAS/STING and NLRC5 pathways defines melanoma response to antitumor immunity.
Sci. Transl. Med.
12
:
eaaz5683
.
34.
Tan
D. Q.
,
Y.
Li
,
C.
Yang
,
J.
Li
,
S. H.
Tan
,
D. W. L.
Chin
,
A.
Nakamura-Ishizu
,
H.
Yang
,
T.
Suda
.
2019
.
PRMT5 modulates splicing for genome integrity and preserves proteostasis of hematopoietic stem cells.
Cell Rep.
26
:
2316
2328.e6
.
35.
Garg
G.
,
A.
Muschaweckh
,
H.
Moreno
,
A.
Vasanthakumar
,
S.
Floess
,
G.
Lepennetier
,
R.
Oellinger
,
Y.
Zhan
,
T.
Regen
,
M.
Hiltensperger
, et al
2019
.
Blimp1 prevents methylation of Foxp3 and loss of regulatory T cell identity at sites of inflammation.
Cell Rep.
26
:
1854
1868.e5
.
36.
Sancisi
V.
,
G.
Gandolfi
,
D. C.
Ambrosetti
,
A.
Ciarrocchi
.
2015
.
Histone deacetylase inhibitors repress tumoral expression of the proinvasive factor RUNX2.
Cancer Res.
75
:
1868
1882
.
37.
Manzotti
G.
,
F.
Torricelli
,
B.
Donati
,
V.
Sancisi
,
M.
Gugnoni
,
A.
Ciarrocchi
.
2019
.
HDACs control RUNX2 expression in cancer cells through redundant and cell context-dependent mechanisms.
J. Exp. Clin. Cancer Res.
38
:
346
.
38.
Girardot
M.
,
R.
Hirasawa
,
S.
Kacem
,
L.
Fritsch
,
J.
Pontis
,
S. K.
Kota
,
D.
Filipponi
,
E.
Fabbrizio
,
C.
Sardet
,
F.
Lohmann
, et al
2014
.
PRMT5-mediated histone H4 arginine-3 symmetrical dimethylation marks chromatin at G + C-rich regions of the mouse genome.
Nucleic Acids Res.
42
:
235
248
.
39.
Xu
J.
,
X.
Xu
,
B.
Wang
,
Y.
Ma
,
L.
Zhang
,
H.
Xu
,
Y.
Hu
,
J.
Wu
,
X.
Cao
.
2017
.
Nuclear carbonic anhydrase 6B associates with PRMT5 to epigenetically promote IL-12 expression in innate response.
Proc. Natl. Acad. Sci. USA
114
:
8620
8625
.
40.
Tanaka
Y.
,
Y.
Nagai
,
M.
Okumura
,
M. I.
Greene
,
T.
Kambayashi
.
2020
.
PRMT5 is required for T cell survival and proliferation by maintaining cytokine signaling.
Front. Immunol.
11
:
621
.
41.
Snyder
K. J.
,
N. C.
Zitzer
,
Y.
Gao
,
H. K.
Choe
,
N. E.
Sell
,
L.
Neidemire-Colley
,
A.
Ignaci
,
C.
Kale
,
R. D.
Devine
,
M. G.
Abad
, et al
2020
.
PRMT5 regulates T cell interferon response and is a target for acute graft-versus-host disease.
JCI Insight
5
:
e131099
.
42.
Webb
L. M.
,
J.
Narvaez Miranda
,
S. A.
Amici
,
S.
Sengupta
,
G.
Nagy
,
M.
Guerau-de-Arellano
.
2019
.
NF-κB/mTOR/MYC axis drives PRMT5 protein induction after T cell activation via transcriptional and non-transcriptional mechanisms.
Front. Immunol.
10
:
524
.
43.
Yuan
Y.
,
H.
Nie
.
2021
.
Protein arginine methyltransferase 5: a potential cancer therapeutic target.
Cell Oncol (Dordr).
44
:
33
44
.
44.
Fong
J. Y.
,
L.
Pignata
,
P. A.
Goy
,
K. C.
Kawabata
,
S. C.
Lee
,
C. M.
Koh
,
D.
Musiani
,
E.
Massignani
,
A. G.
Kotini
,
A.
Penson
, et al
2019
.
Therapeutic targeting of RNA splicing catalysis through inhibition of protein arginine methylation.
Cancer Cell
36
:
194
209.e9
.
45.
Sauce
D.
,
M.
Larsen
,
S.
Fastenackels
,
A.
Roux
,
G.
Gorochov
,
C.
Katlama
,
D.
Sidi
,
J.
Sibony-Prat
,
V.
Appay
.
2012
.
Lymphopenia-driven homeostatic regulation of naive T cells in elderly and thymectomized young adults.
J. Immunol.
189
:
5541
5548
.
46.
Goldrath
A. W.
,
L. Y.
Bogatzki
,
M. J.
Bevan
.
2000
.
Naive T cells transiently acquire a memory-like phenotype during homeostasis-driven proliferation.
J. Exp. Med.
192
:
557
564
.

The authors have no financial conflicts of interest.

This article is distributed under The American Association of Immunologists, Inc., Reuse Terms and Conditions for Author Choice articles.

Supplementary data