Abstract CD8^+ T play essential roles in antitumor immune responses. However, immunotherapy has limited clinical efficacy in many solid tumors. Here, we performed an epigenetic-wide CRISPR-Cas9 screen in CD8^+ T cells directly under cancer immunotherapy setting and found that Prdm12 is a transcriptional repressor implicated in nociceptive neuron development but uncharacterized within immunological contexts. Prdm12 deletion markedly enhanced in vivo tumor clearance of mouse CD8^+ T cells and promoted activation, effector differentiation marker expression, and cytokine secretion in both murine and human CD8^+ T cells in vitro. Mechanistically, Prdm12 deficiency augmented effector transcriptional programs while inhibiting exhaustion of CGRP-RAMP1 neuroimmune axis facilitation. Additionally, Prdm12 ablation remodeled the chromatin accessibility landscape, with H3K9me3 deposition at loci regulating T cell differentiation (Trib1 and Sgk1) and exhaustion (Rgs1 and Nr4a2). These results together reveal a negative regulatory mechanism for CD8^+ T cells and advance our understanding of cancer immunotherapy by linking neurobiological signaling to immune regulation. __________________________________________________________________ Prdm12 restrains CD8^+ T cell antitumor activity by linking neuroimmune signaling to epigenetic exhaustion programs. INTRODUCTION CD8^+ T cells constitute a cornerstone of cancer immunotherapy ([36]1). Multiple immunotherapeutic modalities, including adoptive cell transfer (ACT), immune checkpoint inhibitors (ICIs), oncolytic viruses, cancer vaccines, and cytokine therapies, have emerged as cornerstone strategies in clinical oncology ([37]2, [38]3). Clinically approved ICIs exert their therapeutic effects by reactivating antitumor CD8^+ T cells through blockade of coinhibitory receptors such as Cytotoxic T-Lymphocyte-Associated Protein 4 (CTLA-4) and Programmed Cell Death Protein 1/Programmed Death-Ligand 1 (PD-1/PD-L1) ([39]4, [40]5). ACT approaches—including tumor-infiltrating lymphocytes (TILs) and genetically engineered T cells expressing transgenic T cell receptors (TCRs) or chimeric antigen receptors (CARs)—have revolutionized treatment paradigms for refractory cancers ([41]6, [42]7). This efficacy stems from the capacity of adoptively transferred T cells to convert immunologically “cold” tumors (exhibiting sparse TIL infiltration) into “hot” tumors (marked by dense TIL accumulation), thereby amplifying ACT therapeutic efficacy ([43]8). However, clinical responses to immune checkpoint blockade (ICB) therapy and CAR T cell modalities remain inconsistent across patient populations. Thus, elucidating cell-intrinsic and tumor microenvironment (TME)–driven mechanisms that constrain T cell functionality is critical to enhancing the efficacy of current immunotherapies. CRISPR-mediated genome editing has further expanded the therapeutic potential of ACT by enabling precision engineering of T cell function ([44]9). Genome-wide CRISPR screening enables systematic identification of genetic modifiers that potentiate or impair cytotoxic T cell antitumor activity ([45]10–[46]12). However, achieving persistent single guide RNA (sgRNA) expression in vivo presents notable technical hurdles. Consequently, in practice, focused sgRNA libraries targeting high-priority candidates offer a strategic alternative for in vivo CRISPR screening, offering opportunities to circumvent tumor-mediated immunosuppression. Understanding the epigenetic regulation of CD8^+ T cells within the TME is critical for advancing cancer immunotherapy ([47]13). Previous studies have demonstrated that epigenetic modifications—including DNA methylation, histone posttranslational modifications, and chromatin remodeling—play essential roles in governing CD8^+ T cell differentiation and functional plasticity ([48]14, [49]15). These processes synchronize effector-associated transcriptional programs with dynamic epigenetic remodeling, marked by shifts in histone modification states that either promote or repress gene transcription ([50]16). Nevertheless, the precise epigenetic determinants orchestrating T cell functional states remain incompletely defined ([51]17). Enhancing T cell function can be achieved by targeting inhibitory transcription factors (TFs) and epigenetic regulators. Ghoneim et al. ([52]18) demonstrated that Dnmt3a-deficient CD8^+ T cells, lacking the gene-specific methylation program characteristic of wild-type T cells, exhibited enhanced effector functions and heightened responsiveness to PD-1 blockade. Similarly, Fraitta et al. ([53]19) reported that TET2-deficient CD19 CAR T cells displayed a remodeled epigenetic landscape, redirecting their differentiation trajectory toward sustained antitumor activity and inducing remission in patients with chronic lymphocytic leukemia. These findings underscore the therapeutic potential of epigenetic modulation to overcome ICB resistance and address limitations of CAR T cell therapies, particularly their durability and functional persistence. Furthermore, epigenetic signatures may serve as predictive biomarkers for clinical outcomes in endogenous or adoptively transferred T cell therapies. Elucidating specific epigenetic programs and their enzymatic regulators enables precision engineering of biomarkers and T cell fate determination, thereby enhancing therapeutic efficacy ([54]20–[55]24). The nervous and immune systems, evolutionarily conserved and likely coevolved, use shared molecular mediators and receptors for bidirectional communication ([56]25). This bidirectional signaling between neurons and immune cells orchestrates neural circuitry, barrier integrity, and antimicrobial defense ([57]26, [58]27). Immunohistochemical studies reveal that immune cells are anatomically colocalized with peripheral nervous system (PNS) nerve fibers and express neuropeptide/neurotransmitter receptors, enabling neural regulation of immune responses in both homeostasis and disease ([59]28, [60]29). Nociceptors, a principal subclass of PNS sensory neurons, are defined by expression of ion channels such as voltage-gated sodium channel Nav1.8 and transient receptor potential vanilloid 1 (TRPV1). These channels mediate nociceptive signaling in response to noxious stimuli, safeguarding tissue integrity. Emerging evidence underscores nociceptors as critical regulators of innate immunity and inflammatory processes ([61]30). Calcitonin Gene-Related Peptide (CGRP), a nociceptor-derived neurotransmitter, serves as a key mediator of neuroimmune cross-talk ([62]31, [63]32). Upon activation, CGRP is released from nociceptor terminals and acts in a paracrine manner on neighboring cells, including malignant cells ([64]33). The CGRP receptor, a G protein–coupled receptor (GPCR) complex comprising RAMP1, requires RAMP1 for ligand binding and signal transduction ([65]34). RAMP1 signaling is implicated in modulating innate and adaptive inflammatory responses ([66]35, [67]36), while CGRP directly regulates immune cell function ([68]37). Notably, RAMP1 signaling substantially impairs CD8^+ T cell tumoricidal activity ([69]38). The transcriptional regulator PRDM12 (PR domain zinc finger protein 12; also termed HSAN8 or PFM9), a member of the PRDI-BF1 protein family, in diverse biological processes including neurogenesis, nociception, oncogenesis, and metabolic regulation, with pathogenic mutations linked to congenital insensitivity to pain in humans, no studies have investigated the impact of Prdm12 deficiency on CD8^+ T cell–mediated antitumor immunity ([70]39). Here, we demonstrate that Prdm12 ablation potently enhances CD8^+ T cell antitumor activity by synergistically mitigating T cell exhaustion through disruption of the CGRP-Ramp1 neuroimmune axis ([71]38, [72]40) and epigenetically reprogramming effector differentiation via H3K9me3-mediated chromatin remodeling. Integrative multiomics analyses revealed that Prdm12 coordinates repression of effector genes (Rgs1, Trib1, Nr4a2, and Sgk1) through promoter-targeted H3K9me3 deposition during exhaustion. Thus, Prdm12 constitutes an epigenetic checkpoint governing CD8^+ T cell functionality through dual neuroimmune and chromatin-regulatory mechanisms. RESULTS In vivo CRISPR knockout screen identifies epigenetic regulators of ACT To systematically interrogate intrinsic epigenetic regulators governing antitumor immunity, we designed a mouse knockout (MKO) library targeting 323 murine epigenetic regulators, each perturbed by six unique sgRNAs. The library incorporated 50 nontargeting control (NTC) sgRNAs (table S1). The MKO library was cloned into a CRISPR lentiviral vector with an sgRNA expression cassette, and high-titer lentiviral particles were generated for T cell transduction to enable genome-wide screening (fig. S1A and [73]Fig. 1A). To establish a tractable tumor model, we engineered a B16 melanoma cell line stably expressing chicken ovalbumin (OVA) ([74]Fig. 1, A and B). Flow cytometry using an anti–SIINFEKL:H-2kb antibody identified clone 7, exhibiting modest yet stable OVA expression, as the optimal tumor model ([75]Fig. 1B). To assess intrinsic epigenetic regulators in antitumor immunity, we performed in vivo CRISPR knockout (KO) screening coupled with antigen-specific T cell therapy. OT-I TCR transgenic CD8^+ T cells coexpressing Cas9 were adoptively transferred into Rag1^−/− mice engrafted with B16–OVA–green fluorescent protein (GFP) cells expressing cognate antigens ([76]Fig. 1A). Naive CD8^+ T cells isolated from OT-I;Cas9 transgenic mice were ex vivo activated and validated for CD44 and CD69 up-regulation (fig. S1, B and C) before adoptive transfer into Rag1^−/− hosts. Seven days posttransfer, splenic CD8^+ T cells were cocultured with B16 or B16-OVA-GFP cells alongside anti-CD107a antibodies to quantify antigen-specific degranulation. CD107a surface exposure, a well-established marker of cytotoxic granule exocytosis ([77]41), correlates with antigen-driven T cell activation. The results demonstrated that OT-I;Cas9 CD8^+ T cells exhibited marked degranulation in the presence of the OVA antigen, confirming their antigen specificity (fig. S1D). To confirm antigen-specific antitumor activity, Rag1^−/− mice bearing subcutaneous B16 (control) or B16-OVA-GFP tumors received phosphate-buffered saline (PBS) or OT-I;Cas9 CD8^+ T cells 10 days postinoculation. OT-I;Cas9 CD8^+ T cell treatment notably suppressed tumor growth in B16-OVA-GFP–bearing mice relative to PBS-treated controls (fig. S1E). These results validate the antigen-driven specificity of our model and establish a platform for investigating T cell immunotherapy mechanisms. Fig. 1. In vivo CD8^+ T cell KO screen identified genes regulating tumor infiltration. [78]Fig. 1. [79]Open in a new tab (A) An overview of the in vivo CRISPR KO screening workflow designed to identify epigenetic regulators of antigen-coupled T cell (Ag-CT) responses. (B) Evaluation of antigen presentation using a SIINFEKL:H-2kb antibody in B16–ovalbumin (OVA) clonal cell lines. (C) Representation of tumor growth kinetics over time in Rag1^−/− mice implanted with B16-OVA cells, following distinct therapeutic interventions. Treatments included PBS (n = 4), adoptive transfer of OT-I;Cas9 CD8^+ T cells infected with vector (n = 5) or MKO (n = 5). The red arrow denotes the time point of adoptive transfer. Data are presented as means ± SEM. Significance levels are indicated as follows: **P < 0.01, ***P < 0.001, and ****P < 0.0001, determined by two-way analysis of variance (ANOVA). (D) Results from the MAGeCK analysis of the in vivo tumor infiltration screening data, with the top-ranked genes listed according to their robust rank aggregation (RRA) scores. (E) Venn diagram of the three enrichment criteria used to identify the key genes that regulate tumor infiltration [P < 0.05 (n = 18), ≥2 independently enriched sgRNAs (n = 249), and top rank ≥ 20 (ranked by RRA score; n = 20)]. (F) Distribution of all sgRNAs in the epigenetic KO library (top). Distribution of sgRNA targeting. FC, fold change; HT, high-throughput sequencing; DPI, days postinjection. We subsequently conducted an in vivo CRISPR screen to assess T cell tumor infiltration dynamics. OT-I;Cas9 CD8^+ T cells were transduced with the MKO sgRNA library and adoptively transferred into Rag1^−/− mice engrafted with B16-OVA-GFP tumors ([80]Fig. 1A). Ten days posttransfer, tumor growth was markedly reduced in mice receiving MKO-transduced T cells compared to that in vector controls, confirming enhanced therapeutic efficacy (fig. S1E). Flow cytometry of splenic, lymph node, and tumor-derived single-cell suspensions confirmed successful engraftment of adoptively transferred CD8^+ T cells (fig. S2A). High-throughput sequencing of the sgRNA library in TILs quantified the relative abundance of sgRNA-transduced OT-I;Cas9 CD8^+ T cells. Data were analyzed using MAGeCK, a model-based algorithm for genome-scale CRISPR screen analysis ([81]42). Applying stringent criteria (enrichment fold change of individual sgRNAs ≥ 2, adjusted P < 0.05), we prioritized the top 20 enriched genes per tumor, retaining sgRNAs meeting these thresholds. This integrated approach identified Prdm12, Mettl13, Thumpd2, Arid1a, Bmi1, and Cbx3 as top-enriched candidates ([82]Fig. 1, D to F; and fig. S2, B and C). Prior work demonstrated that Arid1a deletion sustains effector programs in tumor-infiltrating T cells while suppressing exhaustion-associated genes, augmenting CD8^+ T cell antitumor activity ([83]43). Similarly, Bmi1 inhibition drives CD8^+ T cell effector differentiation while impairing memory formation ([84]44). Notably, Cbx3 ablation in CD8^+ T cells enhances transcription initiation and chromatin remodeling, up-regulating Prf1, Gzmb, and Ifng and potently suppressing tumor growth ([85]45, [86]46). However, Prdm12, a top-enriched candidate, remains understudied in CD8^+ T cell biology. These findings implicate Prdm12 as a putative negative regulator of antitumor immunity. Prdm12 KO mouse CD8^+ T cells have notable immunotherapeutic potential To investigate the role of Prdm12 in immunotherapy, sgRNAs targeting Prdm12 were synthesized and electroporated into OT-I;Cas9 T cells using CRISPR-Cas9–mediated gene editing. These Prdm12 KO OT-I;Cas9 T cells were adoptively transferred into mice bearing B16-OVA-GFP tumors ([87]Fig. 2A). Prdm12 mRNA expression levels were quantified using quantitative polymerase chain reaction (qPCR; [88]Fig. 2B), and protein expression was assessed by Western blotting ([89]Fig. 2C), confirming successful Prdm12 down-regulation. Notably, mice treated with Prdm12 KO OT-I;Cas9 CD8^+ T cells exhibited markedly inhibited tumor growth compared to those receiving control (NTC) T cells ([90]Fig. 2D). Furthermore, endpoint tumor weights were markedly reduced in the Prdm12 KO group ([91]Fig. 2E). In parallel, a second sgRNA targeting Prdm12 yielded consistent results (fig. S3, A to C). To further characterize the antitumor response, flow cytometry revealed a marked increase in TILs frequency within tumors and splenic CD8^+ T cell abundance in the Prdm12 KO group ([92]Fig. 2, F and G; and fig. S3, D and E). Collectively, these findings demonstrate that Prdm12 KO enhances the antitumor efficacy and tumor-infiltrating capacity of CD8^+ T cells in vivo. Fig. 2. Prdm12 KO improves ACT efficacy. [93]Fig. 2. [94]Open in a new tab (A) A schematic representation outlining the experimental design. (B) Quantitative polymerase chain reaction (qPCR) analysis quantifying Prdm12 mRNA levels in OT I CD8^+ T cells that were transduced with either Prdm12-targeting sgRNAs or an NTC sgRNA. (C) Western blot analysis evaluating Prdm12 protein levels in OT-I CD8^+ T cells after transduction with sgRNA targeting Prdm12 or an NTC sgRNA. (D) Tumor growth curves of B16-OVA tumors in Rag1^−/− mice following different treatments: PBS (control), adoptive transfer of OT I;Cas9 CD8^+ T cells electroporated with sgRNA targeting Prdm12 (sgPrdm12), or NTC sgRNA (sgNTC). The black arrow indicates the time of adoptive transfer of the sgPrdm12- or sgNTC-transduced OT I;Cas9 CD8^+ T cells. (E) Tumor size of Rag1^−/− mice bearing B16-OVA tumors was measured at designated time points after T cell transplantation. (F) Flow cytometry plots illustrating the gating strategy for CD8^+ T cells in tumors and spleens in mice in mice treated with sgNTC or sgPrdm12-targeted OT I;Cas9 CD8^+ T cells. FITC, fluorescein isothiocyanate. (G) Frequencies of CD8^+ T cell in tumors and spleens as determined by the flow cytometry analysis shown in (F). Data are representative of two independent experiments (means ± SEM) with at least five mice per group. **P < 0.001, ***P < 0.001, and ****P < 0.001 versus control (two-tailed Student’s t test and two-way ANOVA analysis). SSC-A, side scatter-area. Immunological characteristics of Prdm12 KO mouse CD8^+ T cells Although no prior studies have explored the role of Prdm12 loss in CD8^+ T cell–mediated antitumor immunity, we characterized the immunological features of CD8^+ T cells after Prdm12 ablation. Prdm12-ablated CD8^+ T cells exhibited elevated expression of the activation marker CD69 ([95]Fig. 3A). When cocultured with B16-OVA-GFP cells, sgPrdm12-treated OT-I;Cas9 CD8^+ T cells displayed a marked increase in degranulation, as indicated by elevated CD107a surface expression, compared to sgNTC group cells ([96]Fig. 3B). CX3CR1 expression, a chemokine receptor linked to T cell cytotoxicity and migration ([97]47, [98]48), was notably up-regulated in sgPrdm12-treated cells compared to that in the sgNTC group ([99]Fig. 3C). To further investigate the differentiation state of sgPrdm12 and sgNTC-transduced CD8^+ T cells, we analyzed their phenotypes. sgPrdm12-transduced CD8^+ T cells displayed an increased proportion of naive T cells (TN; CD62L^+CD44^−) and reduced central memory T cells (TCM; CD62L^+CD44^+), while no notable differences were observed in effector memory (TEM; CD62L^−CD44^+) T cell frequency (fig. S4, A and B). Furthermore, Prdm12 KO notably increased the proportion of CD127^hi memory precursor (TMP) cells while decreasing KLRG1^hi short-term effector (TEFF) cells ([100]Fig. 3, D and E). Additionally, terminal TEFF-like populations (TEFF-like; CD39^+Ly108^−) were markedly increased, and exhausted precursor populations (TEX-Pre; CD39^−Ly108^+) were notably reduced in sgPrdm12-treated samples compared to the sgNTC group ([101]Fig. 3, F and G). The cytokine production capacity of Prdm12-ablated CD8^+ T cells was subsequently assessed. Following stimulation with anti-CD3/28 (1 μg/ml), Prdm12-ablated CD8^+ T cells exhibited elevated production of effector cytokines interleukin-2 (IL-2), interferon-γ (IFN-γ), and tumor necrosis factor–α (TNF-α; [102]Fig. 3, H and I). These data confirm that Prdm12 modulates T cell activation and negatively regulates CD8^+ T cell effector functions, thus establishing its role as a regulator of primary CD8^+ T cell activity. Fig. 3. Immunological characterization of Prdm12 KO mouse CD8^+ T cells. [103]Fig. 3. [104]Open in a new tab (A) Flow cytometry analysis of CD69 expression, a key immune marker indicative of T cell function. (B) Representative flow cytometry histograms showing the CD107a degranulation assay for Prdm12 KO CD8 T cells. (C) Flow cytometry plots (left) and statistical analysis (right) demonstrating the CX3CR1^+ T cell population of sgNTC and sgPrdm12 CD8^+ T cells. (D and E) Flow cytometry plots (D) and statistical analysis (E) depicting the KLRG1^+CD127^− terminal effector (TE) and KLRG1^−CD127^+ memory precursor (MP) populations among sgNTC and sgPrdm12 groups, gated on CD8^+ T cells. (F and G) Flow cytometry plots (F) and corresponding statistical analysis (G) showing the distribution of Ly108^+CD39^− TEX precursor cells and Ly108^−CD39^+ CD8^+ TEFF-like cells. (H and I) Flow cytometry analysis plots (H) showing IL-2 (top), IFN-γ (middle), and TNF-α (bottom) production by sgPrdm12 and sgNTC CD8^+ T cells, with and without anti-CD3/28 stimulation. Quantification bar plot (I) summarizing the data from (H). Data are representative of three independent experiments (means ± SEM). **P < 0.01, ***P < 0.001, and ****P < 0.0001 by unpaired two-sided t test. APC, allophycocyanin. Prdm12 KO globally alters gene expression in mouse CD8^+ T cells To investigate the underlying mechanisms, bulk RNA sequencing (RNA-seq) was performed comparing Prdm12-deficient and control CD8^+ T cells. Principal components analysis demonstrated distinct clustering patterns between groups (fig. S4C). Pearson correlation analysis confirmed high reproducibility of normalized RNA-seq data (fig. S4D). A total of 135 genes were markedly up-regulated (sgPrdm12-high), and 1245 genes were down-regulated (sgPrdm12-low) in Prdm12 KO CD8^+ T cells ([105]Fig. 4A and table S2). Up-regulated genes included effector cytokines (e.g., Gzmb and Gzmd) and immune-related TFs such as Eomes ([106]Fig. 4, A and B). Down-regulated genes encompassed markers of T cell exhaustion (e.g., Nr4a1, Klf2, and Tcf7) and genes interacting with neurotransmitters, such as Ramp1 ([107]Fig. 4B). Five genes (e.g., Gzmb, Nr4a1, Klf2, Tcf7, and Ramp1) were validated via qPCR, confirming differential expression in Prdm12 KO cells ([108]Fig. 4C). Gene Ontology (GO) analysis of up-regulated genes identified enrichment in pathways such as antiviral defense, cytokine response, chemokine/type II IFN production, T cell cytotoxicity, and T cell activation regulation ([109]Fig. 4D). Down-regulated pathways were linked to cell differentiation, RNA polymerase II transcription, adhesion/migration, inflammatory response, guanosine triphosphatase activity modulation, and IL-2 suppression ([110]Fig. 4E). These analyses support the association of Prdm12 with T cell activation and its regulatory role in CD8^+ T cell effector functions. These results indicate that Prdm12 critically regulates immune responses and CD8^+ T cell–mediated antitumor efficacy. Collectively, these findings position Prdm12 as a key modulator of CD8^+ T cell antitumor activity and a promising therapeutic target to improve immunotherapy outcomes. Fig. 4. Transcriptome analysis of Prdm12 KO mouse CD8^+ T cells by mRNA-seq. [111]Fig. 4. [112]Open in a new tab (A). A volcano plot illustrating the differential expression of genes between Prdm12 KO and NTC CD8 T cells, quantified through bulk mRNA sequencing (three biological replicates per condition). (B) A heatmap depicting the differential gene expression profiles of Prdm12 KO versus NTC CD8^+ T cells, as determined by RNA sequencing. (C) qPCR validation of genes significantly up-regulated or down-regulated, including Ramp1, Nr4a1, Klf2, Tcf7, and Gzmb. Data are representative of three independent experiments (means ± SEM). **P < 0.01 and ****P < 0.0001 by an unpaired two-sided t test. (D and E) DAVID Gene Ontology (GO) analysis of the up-regulated (D) and down-regulated (E) gene sets in Prdm12 KO CD8^+ T cells. Prdm12 KO prevents CGRP/Ramp1-driven mouse CD8^+ T cell exhaustion and enhances cytotoxicity Cytotoxic CD8^+ T cells in humans and mice express at least 10 neuropeptide receptors, including RAMP1, a component of the CGRP receptor. Nociceptor-derived neuropeptides impair immunity against bacteria ([113]49) and fungi ([114]50) while promoting exhaustion in cytotoxic CD8^+ T cells. Given that nociceptors directly interact with CD8^+ T cells in vitro and their neuropeptides suppress antibacterial immunity ([115]51, [116]52). Furthermore, it has been demonstrated that the CGRP-RAMP1 signaling axis promotes CD8^+ T cell exhaustion ([117]38). Interestingly, our transcriptomic data suggest that Prdm12 KO decreases Ramp1 expression, which indicates a potential link between neural signaling and immune regulation. To confirm this speculation, we incubated in vitro–activated CD8^+ T cells with CGRP (100 nM) for 4 days. We confirmed that CGRP stimulation up-regulated Prdm12 and Ramp1 expression ([118]Fig. 5A). Furthermore, CGRP-treated wild-type CD8^+ T cells showed elevated PD-1^+TIM-3^+ CD8^+ T cell populations but reduced CD107a^+ expression ([119]Fig. 5, B and D). These effects were absent in Prdm12 KO CD8^+ T cells ([120]Fig. 5, C and E). To further investigate CGRP’s impact on CD8^+ T cell activation states, we performed differentiation experiments (fig. S5). Wild-type CD8^+ T cells exposed to CGRP exhibited a reduced KLRG1^−CD127^+ memory cell population and an expanded KLRG1^+CD127^− terminal effector population (fig. S5, A and B). Additionally, Ly108^+CD39^− exhausted precursor cells increased, whereas Ly108^−CD39^+ effector-like cells decreased (fig. S5, E and F). Conversely, Prdm12 KO cells reversed these trends, showing reduced exhaustion and enhanced cytotoxicity (fig. S5, C, D, G, and H). Furthermore, rescuing Ramp1 expression in Prdm12 KO cells suppressed the cytotoxicity of mouse CD8^+ T cells toward B16-OVA tumor cells in a 24-hour coculture assay, thereby weakening tumor cell killing (fig. S5, I to L). Collectively, our findings demonstrate that the CGRP-Prdm12/Ramp1 axis drives CD8^+ T cell exhaustion, characterized by reduced cytolytic molecule expression and elevated coexpression of exhaustion markers, consistent with prior evidence that CGRP restricts CD8^+ T cell activity ([121]40). Consequently, Prdm12 KO CD8^+ T cells resist CGRP-induced exhaustion, thus preserving antitumor efficacy. Fig. 5. CGRP modulates the activation of CD8^+ T cells. [122]Fig. 5. [123]Open in a new tab (A) qPCR validation of Prdm12 and Ramp1 expression in wild type CD8^+ T cells exposed to CGRP. (B and C) Cells were stimulated with CGRP (100 nM) for 96 hours. Wild-type CD8^+ T cells exhibited an elevated frequency of PD-1^+TIM-3^+ cells, but Prdm12 KO reversed this result. (D and E) Upon 96-hour stimulation with CGRP (100 nM), wild-type CD8^+ T cells exhibited a reduced frequency of CD107a^+ cells, but Prdm12 KO reversed this result. Data are representative of three independent experiments (means ± SEM). Not significant (n.s.), P > 0.05; **P < 0.01, ***P < 0.001, and ****P < 0.0001 by an unpaired two-sided t test. Prdm12 reshapes the epigenetic profile of mouse CD8^+ T cells In studies of neural crest development in the Xenopus laevis, Rienzo et al. ([124]39) demonstrated that PRDM12 suppresses neural crest development by modulating histone H3K9 trimethylation (H3K9me3) deposition at promoter regions of neural crest markers, including Foxd3, Slug, and Sox9. To investigate H3K9me3 deposition and chromatin accessibility in Prdm12-mediated antitumor mechanisms, genome-wide Cleavage Under Targets and Tagmentation (CUT&Tag) sequencing for H3K9me3 and ATAC-seq (assay for transposase-accessible chromatin) were performed in Prdm12 KO and control CD8^+ T cells. OT-I CD8^+ T cells isolated from OT-I mice were activated with anti-CD3/CD28 antibodies and electroporated 72 hours later before being subjected to CUT&Tag. Pearson correlation analysis indicated high reproducibility of the CUT&Tag data after normalization (fig. S6A). We observed elevated H3K9me3 signals in Prdm12 KO CD8^+ T cells ([125]Fig. 6A and fig. S6B). The increased chromatin accessibility is associated with enhanced transcriptional activity ([126]Fig. 6B), which was also associated with reduced H3K9me3 marks ([127]Fig. 6C). An integrated analysis of down-regulated genes in the Prdm12 KO transcriptome and loci with increased H3K9me3 peaks was conducted ([128]Fig. 6D). H3K9me3 peaks were enriched at loci of key genes, including Rgs1, Trib1, Nr4a2, and Sgk1. Notably, previous reports have shown that RGS1 is up-regulated via type II IFN–signal transducer and activator of transcription 1 signaling. This up-regulation impedes T cell trafficking to tumors by inhibiting calcium influx and extracellular signal–regulated kinase/Akt kinase activation. RGS1 KO in adoptively transferred cytotoxic T lymphocytes (CTLs) enhances their infiltration, survival, and antitumor efficacy in breast and lung tumor models ([129]53). Nuclear Receptor Subfamily 4 Group A (NR4A) TFs (e.g., NR4A2) impair CAR T cell function in solid tumors ([130]54) and drive T cell exhaustion ([131]55). TRIB1 suppresses effector T cell responses in chronic disease contexts. Sgk1-deficient CD8^+ T cells result in reduced Foxo1 phosphorylation and increased Foxo1 nuclear translocation, which promotes the formation of early memories, resulting in longer-lived memory T cells. These cells demonstrate enhanced recall responses upon reinfection and show improved tumor rejection capabilities ([132]56). We found that increased chromatin accessibility correlated with increased transcription; however, a subset of regions showed reduced accessibility despite elevated transcription ([133]Fig. 6B). Loss of Prdm12 resulted in an enrichment of H3K9me3 peaks in the promoter regions of Rgs1, Trib1, Nr4a2, and Sgk1 in T cells ([134]Fig. 6E). This H3K9me3 increase coincided with reduced chromatin accessibility at these loci in Prdm12 KO cells ([135]Fig. 6E), identifying Rgs1, Trib1, Nr4a2, and Sgk1 as potential targets of Prdm12-mediated H3K9me3. CUT&Tag-qPCR validated increased H3K9me3 enrichment at promoters/exons of Rgs1, Trib1, Nr4a2, and Sgk1, while reverse transcription–qPCR confirmed their reduced expression in Prdm12 KO cells ([136]Fig. 6, F and G). H3K9me3 marks in T cells localized predominantly to promoters and exons, with secondary deposition in introns and intergenic regions. Substantial H3K9me3 peak alterations occurred in promoter/exon regions and intronic/intergenic regions (fig. S6C). TF motif analysis of H3K9me3 CUT&Tag data revealed enriched binding sites for JunB, Atf3, and Batf in Prdm12 KO CD8^+ T cells versus controls ([137]Fig. 6H). Activator Protein-1 (AP-1) TFs regulate T cell development, function, and exhaustion, including in CAR T cells. AP-1 family members like JunB are critical for cytotoxic CD8^+ T cell responses, enabling malignant cell elimination ([138]57). Atf3 facilitates CD8^+ T cell infiltration into colorectal tumors, amplifying antitumor immunity ([139]58). Batf drives effector-like differentiation in human CAR T cells, enhancing tumor rejection ([140]59). The data indicate that Prdm12 coordinates with TFs JunB, Atf3, and Batf to control TEFF differentiation. Collectively, Prdm12 serves as an epigenetic brake on T cell differentiation. Its ablation sustains a TEFF-favorable chromatin state, enhancing CD8^+ T cell effector function via H3K9me3-mediated repression of Rgs1, Trib1, Nr4a2, and Sgk1. Fig. 6. Prdm12 reshapes the epigenetic profile of CD8^+ T cells. [141]Fig. 6. [142]Open in a new tab (A) Heatmaps of H3K9me3 and opening chromatin at Prdm12-enriched regions in wild-type (WT) and Prdm12 KO CD8^+ T cells (peak center, ±5 kb for H3K9me3 and ATAC-seq). TSS, transcription start site. (B) Scatterplot exhibiting Pearson correlation analysis of peak accessibility of the nearest genes versus their differential expression levels. (C) Scatterplot showing the inverse correlation between H3K9me3 levels and gene expression levels at promoter/gene regions in Prdm12 KO CD8^+ T cells. R^2, coefficient of determination. (D) Venn diagram displaying the overlap between gene sets with increased H3K9me3 levels and gene sets with decreased RNA-seq levels in Prdm12 KO CD8^+ T cells. (E) H3K9me3 binding signals from CUT&TAG assays, along with open chromatin region signals detected by ATAC-seq for the WT and Prdm12 KO CD8^+ T cells at the Rgs1, Trib1, Nr4a2, or Ramp1 loci. (F) Chromatin immunoprecipitation–qPCR analysis of H3K9me3 levels in the promoter region of Rgs1, Trib1, Nr4a2, and Sgk1 in WT and Prdm12 KO CD8^+ T cells. IgG, immunoglobulin G. (G) qPCR analysis of Ramp1, Trib1, Nr4a2, and Sgk1 expression in WT and Prdm12 KO CD8^+ T cells. (H) Top three motifs in the H3K9me3 CUT&TAG peaks of T cells upon loss of Prdm12. Data are representative of three independent experiments (means ± SEM). *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by an unpaired two-sided t test. Characterization of PRDM12 in human T cells Meta-analyses further reveal PRDM12 up-regulation across multiple solid tumors, including colorectal, breast, renal, and lung cancers, underscoring its potential as a prognostic biomarker ([143]60). However, the role of PRDM12 in human CD8^+ T cell–mediated ACT remains unexplored. To evaluate the clinical relevance and translational potential of our findings, the immune profiles of human CD8^+ T cells were assessed. To investigate antitumor cytotoxic activity in PRDM12-deficient human CD8^+ T cells, in vitro coculture experiments were performed. In vitro–activated CD8^+ T cells were transduced with lentiviruses encoding Prostate-Specific Membrane Antigen (PSMA) -specific CAR and engineered into PSMA–CAR T cells using ribonucleoprotein (RNP) electroporation. These PSMA–CAR T cells were cocultured with PSMA^+ PC3 cells for 6 to 8 hours, and cytotoxicity was quantified via flow cytometry. PRDM12 KO CD8^+ T cells exhibited notably elevated CD107a and GZMB expression, indicating enhanced cytotoxic capacity ([144]Fig. 7, A and B). To further assess PRDM12’s clinical relevance, a detailed analysis of PRDM12 expression, CTL infiltration, and patient prognosis was performed using the TIMER 2.0 database. In patients with melanoma, PRDM12 expression exhibited a notable negative correlation with CD8^+ T cell infiltration [correlation coefficient (r) = −0.224, P = 1.6 × 10^−6] and metastatic cases (r = −0.158, P = 3.74 × 10^−3) ([145]Fig. 7, C and D, left). High cytotoxic CD8^+ T cell infiltration conferred an overall survival benefit only in patients with low PRDM12 expression. Conversely, high PRDM12 expression abrogated or reversed the survival advantage associated with CD8^+ T cell infiltration ([146]Fig. 7, C and D, right). Collectively, these findings demonstrate that PRDM12 critically regulates CD8^+ T cell survival, differentiation, and tumor-killing capacity. Fig. 7. Characterization of PRDM12 in human T cells. [147]Fig. 7. [148]Open in a new tab (A). Representative flow cytometry plots (left) and statistical analysis (right) showing the CD107a degranulation assay for WT and PRDM12 KO human CD8^+ T cells. (B) Flow cytometry plots (left) and corresponding statistical analysis (right) showing the GZMB production of WT and PRDM12 KO human CD8^+ T cells. (C and D) The expression level of PRDM12 is inversely correlated with the infiltration levels of CD8^+ T cells in patients with skin cutaneous melanoma (SKCM) or SKCM metastasis (left). TIDE analyses linking PRDM12 expression to CD8^+ T cell dysfunction signatures revealed potential survival benefits in patients with SKCM or SKCM metastasis (right). Data are representative of three independent experiments (means ± SEM). ****P < 0.0001 by an unpaired two-sided t test. TPM, transcripts per million; HR, hazard ratio. DISCUSSION The therapeutic potential of T cells largely stems from their rapid proliferation, effector cytokine secretion, and potent cytolytic activity. However, the efficacy of ACT for solid tumors is constrained by barriers that impede lymphocyte infiltration into the TME. These barriers, such as T cell exhaustion and limited tumor-specific T cell frequency in patient repertoires, compromise antitumor T cell infiltration, local expansion, and persistence ([149]61, [150]62). Previous CRISPR screens have identified negative regulators of CD8^+ T cell expansion, including PTPN2 ([151]63), Regnase-1 ([152]11), SOCS1 ([153]64), and Roquin ([154]65). Conversely, overexpression of BATF, c-JUN, or IRF4 enhances CD8^+ T cell proliferation ([155]11, [156]65, [157]66). Using an in vivo CRISPR screen with a MKO library targeting epigenetic regulators, we identified Prdm12 as a key epigenetic modulator of CD8^+ T cell antitumor immunity in melanoma. Mechanistically, (i) the neurotransmitter CGRP promotes the expression of Prdm12 and Ramp1 in CD8^+ T cells, thereby driving CD8^+ T cell exhaustion through the CGRP-Ramp1 axis. In contrast, Prdm12 KO rescued CGRP-induced T cell exhaustion and restored CD8^+ T cell cytotoxicity; (ii) Prdm12 KO modulated chromatin accessibility at Rgs1, Trib1, Nr4a2, and Sgk1 loci via H3K9me3 while suppressing their expression, thereby inhibiting exhaustion transition and enhancing CD8^+ T cell effector function ([158]Fig. 8). This study reveals the central regulatory role of Prdm12 in T cell effector differentiation and offers critical evidence for advancing immunotherapy or neuroimmunotherapy strategies through Prdm12 targeting. Fig. 8. Illustration summary of Prdm12 regulation of CD8^+ T cell antitumor responses. [159]Fig. 8. [160]Open in a new tab During cancer immunotherapy, in Prdm12-deficient CD8^+ T cells, Prdm12 KO leads to a marked reduction in Ramp1 expression, which blocks neural signal transmission via the CGRP-RAMP1 pathway, preventing CD8^+ T cells from entering a dysfunctional state and thereby enhancing their cytotoxicity against tumor cells. In the parallel, Prdm12 KO remodeled epigenetic accessibility and increased H3K9me3 labeling of effector-related genes (Rgs1, Trib1, Nr4a2, and Sgk1) in T cells, thereby enhancing T cell effector function and reducing T cell exhaustion. Motif analysis revealed enriched TF binding sites for JunB, Atf3, and Batf in Prdm12 KO CD8^+ T cells. To investigate the antitumor capabilities of CD8^+ T cells, an in vivo tumor infiltration CRISPR screen was used in a melanoma model to identify genes modulating T cell effector function. These findings demonstrate that targeted genetic disruption of epigenetic regulators enhances T cell function, potentially improving ACT therapeutic efficacy. In the B16-OVA tumor model, Prdm12 KO markedly enhanced the antitumor efficacy of OT-I CD8^+ T cells in vivo. Prdm12 KO CD8^+ T cells also exhibited up-regulated IFN-γ and TNF-α production in vitro. These results suggest that Prdm12 KO CD8^+ T cells may confer enhanced persistence and efficacy in clinical applications. Genetic or pharmacological targeting of Prdm12 represents a promising strategy to enhance T cell–based immunotherapies for melanoma and other solid tumors. Collectively, these findings establish a mechanistic framework and actionable targets to optimize ACT-based therapies, advancing their clinical potential for cancer treatment. CD8^+ T cell differentiation critically regulates the elimination of intracellular pathogens and malignant cells in cancer ([161]67). However, the mechanisms defining the differentiation, breakpoints, and cell-intrinsic and cell-extrinsic regulators that promote optimal T cell differentiation into a specific lineage while inhibiting other lineages have not been fully understood. Upon antigen stimulation, naive CD8^+ T cells proliferate and differentiate into various memory and effector cell subsets. These cells exhibit long-term persistence and mediate cytotoxicity against tumor- and virus-infected cells ([162]16, [163]68). Key differentiation subsets include TEFF, TEX, TMP, TCM, TEM, memory (TMEM), and tissue-resident memory (TRM) CD8^+ T cells ([164]69, [165]70). T cell exhaustion is marked by reduced effector function, impaired proliferation, and up-regulated inhibitory checkpoint receptors ([166]20). Chronic tumor antigen exposure drives exhaustion of CD8^+ TILs, compromising their antitumor efficacy ([167]71). TCM exhibits sustained persistence in vivo ([168]72) and superior efficacy in adoptive immunotherapy compared to effector memory subsets ([169]16, [170]20, [171]73). Some studies have elucidated transcriptional networks governing fate decisions among TEFF, TMEM, and TEX CD8^+ T cell subsets. These TFs often promote one lineage while actively suppressing alternative lineages. For instance, Tox drives TEX differentiation while inhibiting TEFF commitment ([172]74), TCF-1 enhances TMEM or TEX formation while suppressing TEFF ([173]75), and Blimp-1, T-bet, and Id2 promote TEFF at the expense of TMEM ([174]76). Our in vitro findings demonstrate that Prdm12 regulates T cell activation, differentiation dynamics, migratory potential, and cytotoxic function. During early differentiation, Prdm12 KO cells displayed a pronounced expansion of CD62L^+CD44^− TN population and CD62L^+CD44^+ TCM population, implicating Prdm12 in sustaining T cell stemness. Prdm12 deficiency reduced the KLRG1^+CD127^− short-TEFF population while expanding the KLRG1^−CD127^+ TMP population. A marked increase in TEFF-like CD39^+Ly108^− cells and a decrease in TEX-Pre CD39^−Ly108^+ cells were observed. In addition, CGRP released by nociceptive neurons serves as a critical immunomodulator, binds Ramp1 receptors on CD8^+ T cells to up-regulate immune checkpoint molecules, including PD-1, Lymphocyte-Activation Gene 3 (LAG-3), and T-cell Immunoglobulin and Mucin-domain containing-3 (TIM-3), ultimately driving T cell exhaustion ([175]38). Moreover, CGRP is also involved in T cell fate determination, and CGRP induces an effective antiviral T helper 1 cellular response when acting on Ramp3-expressing T cells ([176]77). Our RNA-seq analysis revealed significant down-regulation of Ramp1 expression in Prdm12 KO CD8^+ T cells. In contrast, CGRP-treated CD8^+ T cells exhibited elevated expression of Prdm12 and Ramp1, promoting cell differentiation skewing toward terminal effector and exhausted precursor subsets. However, Prdm12 KO markedly attenuated CGRP-induced exhaustion, restored cytotoxic activity, and shifted differentiation toward TEFF-like CD39^+Ly108^− population and KLRG1^−CD127^+ TMP population. These data underscore Prdm12’s critical role in CD8^+ T cell differentiation, cytotoxicity, and neuroimmune cross-talk. Collectively, these findings provide a mechanistic basis for leveraging neuroimmune-regulated T cells to enhance antitumor immunity. Mechanistically, Prdm12 KO elevates H3K9me3 deposition at promoters of Rgs1, Trib1, and Nr4a2, silencing these genes and consequently enhancing CD8^+ T cell effector functions. Elevated RGS1 expression in CTLs correlates with diminished tumor infiltration and inferior clinical outcomes ([177]53). TEX CD8^+ T cells exhibit impaired capacity to resolve chronic infections or malignancies. Reprogramming TEX cells into an effector-like state represents a promising therapeutic strategy. Trib1 KO induces transdifferentiation of TEX cells into KLR^+ TEFF subsets. Trib1 KO synergizes with anti–PD-L1 blockade to enhance viral clearance ([178]78). These findings demonstrate that Trib1 serves as a critical regulator of CD8^+ T cell exhaustion and that targeting Trib1 represents a therapeutic strategy to augment KLRG1^+ effector differentiation and enhance the efficacy of checkpoint blockade therapies. In addition, T cell exhaustion is further driven by Nuclear Factor of Activated T cells (NFAT)–mediated induction of NR4A TFs (Nr4a1/Nur77, Nr4a2/Nurr1, and Nr4a3/Nor1) ([179]79). CAR T cells deficient in all three NR4A factors (Nr4a-TKO) drive tumor regression and extend survival in murine models. Nr4a-TKO CAR T cells exhibit effector-like phenotypes and up-regulated stemness/memory-associated genes in Nr4a1/2-deficient CD8^+ T cells. These findings underscore NR4A inhibition as a promising immunotherapeutic strategy ([180]54). In summary, Prdm12 epigenetically represses Rgs1, Trib1, and Nr4a2 via deposition of H3K9me3, reprogramming CD8^+ T cell differentiation trajectories and functional states and potently suppressing antitumor immunity. These findings highlight Prdm12 as a promising therapeutic target for restoring CD8^+ T cell function and overcoming immune evasion in cancer. Although this study analyzed the relationship between PRDM12 expression and patient survival using public databases and demonstrated its cytotoxic role in human CD8^+ T cells in vitro ([181]Fig. 7), it has limitations. In vitro experiments cannot fully replicate the complex in vivo microenvironment, and the lack of patient-derived tumor organoids for human-relevant target validation is a constraint. Furthermore, the absence of functional analyses using clinical samples hinders clinical translation. Nevertheless, the core mechanisms validated in vivo and in vitro identify Prdm12 as a previously unidentified potential therapeutic target for T cell–based immunotherapy. Future research should elucidate Prdm12’s role in distinct CD8^+ T cell subsets (naive, effector, and exhausted); use CAR T cell technology to assess its contribution to antitumor immunity in solid tumor models; systematically evaluate therapeutic potential using CRISPR-mediated Prdm12 KO, RNP delivery, or combination with ICIs; and leverage humanized patient-derived xenograft models for functional validation and clinical translation. Modulating Prdm12 expression in T cells via epigenetic approaches may also enhance CAR T cell efficacy, particularly for refractory malignancies. MATERIALS AND METHODS Mice Six- to 8-week-old Rag1^−/− mice were purchased from GemPharmatech Co. Ltd. Rosa26-Cas9 mice and OT-I TCR transgenic mice (OT-I mice) were provided by D. Li’s and B. Du’s laboratories at East China Normal University (ECNU), respectively. OT-I;Cas9 mice were generated by breeding OT-I mice with Rosa26-Cas9 mice. The animals were housed under specific pathogen–free conditions at the Animal Center of ECNU. Experimental protocols were approved by the ECNU Animal Care and Use Committee (ID: m20210238) and conducted in compliance with the animal care guidelines set forth by the Ministry of Science and Technology of the People’s Republic of China. Animals were randomly assigned to experimental groups. T cell CRISPR library cloning We designed and synthesized sgRNA libraries targeting 323 epigenetically regulated genes with six unique sgRNAs per gene and included 50 NTCs. All sgRNA sequences are detailed in table S1. sgRNAs were subcloned into the lentiCRISPR v2 vector (Addgene, no. 52963) using the Gibson Assembly Master Mix [New England Biolabs (NEB)] to generate MKO libraries that were amplified in Escherichia coli DH5α (Sangon Biotech) The MKO library was amplified in E. coli DH5α (Sangon Biotech) with >100-fold coverage, and the sequence was verified by Illumina sequencing. Viral library production MKO library lentiviruses were generated by cotransfecting human embryonic kidney (HEK) 293T cells, which were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S), with MKO plasmid, psPAX2 (Addgene, no. 12260), and pMD2.G (Addgene, no. 12259). At 48 hours posttransfection, supernatant was collected, filtered, and concentrated by ultracentrifugation at 20,000 rpm for 2.5 hours at 4°C, then resuspended in DMEM, and stored at −80°C. Isolation and culture of CD8^+ T cells Spleens from OT-I/Cas9 mice were isolated and mashed through a 100-μm filter and washed again with PBS (2% FBS). Erythrocytes were lysed with Ammonium-Chloride-Kohler (ACK) buffer for 5 min at room temperature, then washed, and filtered through a 40-μm filter. Naive CD8^+ T cells were isolated using a Naive Mouse CD8^+ T Cell Isolation Kit (catalog no. 130-096-543, Miltenyi) and resuspended in mouse CD8^+ T cell medium [cRPMI; RPMI 1640, 10% FBS, 2 mM l-glutamine, P/S (100 U/ml; Thermo Fisher Scientific), and 49 nM β-mercaptoethanol (catalog no. M6250-10ML, Sigma-Aldrich)] at 1 × 10^6 cells/ml. For in vivo, the medium was supplemented with IL-2 (2 ng/ml), IL-7 (2.5 ng/ml), IL-15 (50 ng/ml), and anti-CD28 (1 μg/ml). For in vitro, IL-2 (2 ng/ml), IL-12p70 (2 ng/ml), and anti-CD28 (1 μg/ml) were added. Cells were cultured on plates precoated with anti-CD3 (5 μg/ml). Human CD8^+ T cells were isolated from human peripheral blood mononuclear cells (PBMCs), which were obtained from peripheral blood donated by a healthy adult at Shanghai Zhaxin Integrated Traditional Chinese and Western Medicine Hospital. PMBCs were purchased from Saily Bio (Shanghai, China, catalog no. 200425). CD8^+ T cell isolation was performed using human CD8 MicroBeads (Miltenyi Biotec, no. 130-045-201). Cells were activated with TransAct (Miltenyi Biotec, no. 130-111-160) and cultured at 1 × 10^6 cells/ml in X-Vivo medium (Lonza) supplemented with 10% FBS (Gibco), 1% P/S (Gibco), and IL-2 (100 U/ml) (R&D Systems) in a humidified environment at 37°C with 5% CO[2]. All experimental protocols involving human samples were approved by the Ethics Committee of Shanghai Zhaxin Integrated Traditional Chinese and Western Medicine Hospital (approval no. 202004) and the University Committee on Human Research Protection of East China Normal University (approval no. HR 204-2020). Informed consent was obtained from the donor before sample collection. T cell transduction and virus titration T cells were activated with CD3/28 for 3 days and then transferred to fresh medium with concentrated virus. Viral titer was determined by puromycin selection 3 days posttransduction. Library viral transduction of T cells OT-I;Cas9 CD8^+ T cells were transduced with the concentrated MKO lentiviral library at a low MOI to ensure single viral particle infection per cell. Each infection sample contained ~5 × 10^6 cells, and a minimum of three replicates were used per experiment to generate the T cell library. Gene editing in mouse OT-I;Cas9 CD8^+ T cells and human CD8^+ T cells Mouse CD8^+ T cells were electroporated using P3 Primary Cell Nucleofection Solution (catalog no. V4XP-3032, Lonza) according to the manufacturer’s instructions. RNP complexes were generated by mixing 300 pmol of sgRNA (sgNTC/sgPrdm12) synthesized by Synthego with 300 pmol of Cas9 protein and incubated for 10 min at room temperature before electroporation. Cells were resuspended in P3 Primary Nuclei Infection Solution at a ratio of 1 × 10^6 per 20 μl and then transferred to cuvettes for electroporation using the EH-115 program. After electroporation, cells were resuspended in IL-2 (2 ng/ml) and IL-12p70 (2 ng/ml) medium and incubated for 48 hours for in vitro experiments. The electrotransformation protocol for human CD8^+ T cells was as described for murine cells, with the difference that the electrotransformation procedure was EO-100 program, with a 100-μl system and sgPRDM12 sequence: ggtgatcgccccggagcacg. In vitro stimulation of cytotoxic CD8^+ T cells with CGRP Mouse CD8^+ T cells were isolated, stimulated for 48 hours, and subsequently treated with CGRP (0.1 μM; HY-P0203, MedChemExpress, MCE) for an additional 96 hours. Surface expression of PD-1, TIM-3, and CD107a was assessed by flow cytometry. T cell differentiation status was evaluated on the basis of marker expression following in vitro stimulation. Flow cytometry and analysis Surface protein staining Prdm12 KO CD8^+ T cells, splenocytes, or tumor cell suspensions were stained with surface marker antibodies in 0.1% bovine serum albumin/PBS on ice for 30 min, then washed, and resuspended in fluorescence-activated cell sorting (FACS) buffer for FACS analysis. Intracellular cytokine staining Prdm12 KO OT-I-CD8^+ T cells or control OT-I-CD8^+ T cells were initially stained with surface markers for 30 min, washed, and then fixed and permeabilized using the Cytofix/Cytoperm Plus Fixation/Permeabilization kit (BD Biosciences, catalog no. 554714) according to the manufacturer’s instructions. Intracellular cytokines (IFN-γ, TNF-α, or IL-2) were stained for 30 min, followed by two washes with Cytoperm/Wash Buffer. The cells were lastly resuspended in magnetic-activated cell sorting (MACS) buffer for FACS analysis. Antibodies from BioLegend were used. Anti-mouse antibodies used Anti-CD8 (catalog no. 100706), anti-CD107a (catalog no. 121611), anti-Cx3CR1 (catalog no. 149047), anti-GzmB (catalog no. 396413), anti–TNF-α (catalog no. 506303), anti–IFN-γ (catalog no. 505829), anti–PD-1 (catalog no. 135207), anti–LAG-3 (catalog no. 125209), anti-CD44 (catalog no. 103008), anti-CD62L (catalog no. 161211), anti-CD39 (catalog no. 143805), anti-CD108 (catalog no. 134609), anti-mouse/human KLRG1 (catalog no. 138429), anti-CD127 (catalog no. 135013), and anti–SIINFEKL-H-2kb (catalog no. 141605). Anti-human antibodies used Anti-CD3 (catalog no. 300329), anti-CD8a (catalog no. 300906), anti-CD107a (catalog no. 328615), and anti-GzmB (catalog no. 372207). All antibodies were sourced from BioLegend. Data were analyzed on a Fortessa LSR II (Becton Dickinson) and processed with FlowJo X 9.9.6 Tree Star, Ashland, OR. Statistical analysis used unpaired two-sided t tests, with significance *P < 0.05, **P < 0.01, and ***P < 0.001. Data are shown as means ± SEM. In vitro degranulation assay Antigen-specific validation assay OT-I;Cas9 CD8 T cells are activated with anti-CD3(5 μg/ml)/CD28(1 μg/ml) for 3 days. B16 or B16-OVA cells are incubated with SIINFEKL (1 ng/ml) for 4 hours in RPMI 1640 medium [RPMI 1640, 10% FBS, P/S (100 U/ml; Thermo Fisher Scientific), 2 mM l-glutamine, and 49 nM β-mercaptoethanol] supplemented with IL-2 (2 ng/ml) and IL-12p70 (2 ng/ml). OT-I;Cas9 CD8 T cells (5 × 10^5) are added to SIINFEKL-pulsed tumor cells. During the coculture, IL-2, IL-12, anti-CD107a phycoerythrin (PE) (1:400 dilution), and monensin (BioLegend, catalog no. 420701) are added. After 2-hour, cells are harvested, stained, and analyzed by flow cytometry. Degranulation assay On day 0, OT-I CD8 T cells are stimulated with anti-CD3/CD28. On day 3, cells are electroporated with sgPrdm12 or sgNTC. T cells were cultured overnight, while B16-OVA cells (5 × 10⁵/ml) were seeded in 24-well plates and pulsed with 1 ng/ml SIINFEKL peptide in RPMI-1640 medium for 4 hours. Subsequently, 5 × 10⁵ T cells were cocultured with the SIINFEKL-pulsed B16-OVA cells for 2 hours at 37°C in medium supplemented with 2 nM monensin, anti-CD107a antibody (1:400), IL-2, and IL-12. Cells were then harvested and analyzed by flow cytometry to detect CD107a^+ CD8^+ T cells. Adoptive transfer of Prdm12 KO OT-I-CD8^+ T cells Naive OT-I CD8 T cells were stimulated with anti-CD3/CD28 for 3 days, electroporated with sgPrdm12 or sgNTC, and incubated for another 3 days. On day 6, 5 × 10^6 cells were injected into Rag1^−/− mice bearing B16-OVA tumors. Tumor sizes were measured every 2 to 3 days, and volume was calculated as length × width^2 × 0.5. At experiment end, mice were euthanized, and tumors and spleens were collected; tumor weight was recorded. Lentivirus expressing OVA-GFP production Lentiviruses expressing GFP-OVA and Ramp1 were produced by cotransfecting HEK293T cells with lenti-OVA-GFP/lenti-Ramp1 plasmid (from B. Du’s lab, at ECNU), psPAX2 (Addgene, no. 12260), and pMD2.G (Addgene, no. 12259). Viral supernatants were collected as previously described. Generation of the B16-OVA-GFP monoclonal cell line B16 melanoma cells were cultured in RPMI 1640 with 10% FBS (Atlas Biologicals, no. FS-0500-AD) and P/S (100 U/ml; Gibco, no. 15070063) and then transduced with OVA-GFP lentivirus. Three days posttransduction, cells were diluted to 10 cells/ml and seeded into 96-well plates. After 2 weeks, clonal GFP-positive B16-OVA cells were identified by fluorescence microscopy, and OVA expression was confirmed using an anti–SIINFEKL:H-2kb antibody. Clone 7, with low and uniform OVA expression, was selected for further experiments. Antigen specificity testing for OT-I T cells Naive CD8^+ T cells from OT-I;Cas9 mice were cultured with anti-CD3/CD28, IL-2 (BioLegend, catalog no. 575404), IL-7 (5 ng/ml) (BioLegend, catalog no. 577802), and IL-15(50 ng/ml) (BioLegend, catalog no. 566302) for 3 days, and, then, 5 × 10^6 cells were injected into Rag1^−/− mice harboring B16-OVA tumors. Seven days postinjection, spleens were harvested, and CD8^+ T cells were purified and cocultured with B16 or B16-F10-OVA cells for 2 hours in cRPMI with IL-2, monensin, and anti-CD107a. Cells were stained and analyzed by flow cytometry. B16 or B16-OVA-GFP cells were inoculated into Rag1^−/− mice, followed by PBS or OT-I;Cas9 CD8^+ T cell injection. Tumor volumes were measured at designated times. Tumor transplantation and tissue processing B16-OVA-GFP cells (5 × 10^6) were injected subcutaneously into Rag1^−/− mice. Naive OT-I;Cas9 CD8^+ T cells were cultured with anti-CD3/CD28, IL-2, IL-7, and IL-15 for 3 days and electroporated with sgRNAs, and 5 × 10^6 cells were injected into tumor-bearing Rag1^−/− mice [E:T (effector to target ratio) = 1:1]. Tumor volumes and weights were measured, and survival was analyzed. Spleens and tumors were isolated at the experiment’s end. Spleen processing Spleens were mashed in 2% FBS/PBS and filtered through a 70-μm strainer, and red blood cells were lysed. Lymphocytes were collected by washing and centrifuging. Tumor processing Tumors were collected, fragmented, and digested with 1 mg/ml collagenase IV (Yeasen) for 60 min at 37°C, filtered, washed, and resuspended in RPMI 1640 medium. For FACS, cells were stained with anti-CD3 and anti-CD8 antibodies for 15 min on ice and analyzed using BD flow cytometer. Statistical analysis For statistical comparison of tumor growth curves, multiple t tests were performed [Benjamini, Krieger, and Yekutieli false discovery rate (FDR) method] on each time point. gDNA extraction Genomic DNA (gDNA) from viral library-infected, sgNTC-edited, or sgPrdm12-edited CD8^+ T cells or tumor cells with TILs was isolated using a TIANamp Genomic DNA Kit. Cells were lysed by resuspending in buffer GA with proteinase K and buffer GB at 70°C for 10 min. After adding ethanol, the mixture was loaded onto an adsorption column and centrifuged at 12,000 rpm. Buffer GD removed proteins, followed by washing with buffer PW. Purified gDNA was eluted with ddH2O. sgRNA readout and high-throughput sequencing For the readout of the sgRNA library, two rounds of PCR reactions were used. First PCR (PCR1): A region encompassing the sgRNA cassette was amplified using primers specific to the T cell CRISPR vector: forward: 5′-AATGGACTATCATATGCTTACCGTAACTTGAAAGTATTTCG-3′; and reverse: 5′-CTTTAGTTTGTATGTCTGTTGCTATTATGTCTACTATTCTTTCCC-3′. Each gDNA sample was divided into several 50-μl reactions, with 2 μg of gDNA included in each reaction. The reaction mix for each contained 5 μl of 10× buffer for KOD-Plus-Neo (TOYOBO), 200 ng of gDNA, 2.5 μl of each PCR1 primer (10 pmol/μl), 1 μl of KOD-Plus-Neo, 5 μl of 2 mM deoxyribonucleotide triphosphates (dNTPs), 3 μl of 25 mM MgSO[4], and water to bring the total volume to 50 μl. The PCR conditions were set as follows: initial denaturation at 94°C for 2 min; 15 cycles of 98°C for 10 s, 60°C for 20 s, and 68°C for 20 s; and final extension at 72°C for 2 min. Second PCR (PCR2): After PCR1, all reactions for each sample were pooled and used for amplification with barcoded PCR2 primers: forward: 5′-GGAGTGAGTACGGTGTGCAATGGACTATCATATGCTTA-3′; and reverse: 5′-GAGTTGGATGCTGGATGGCTTTAGTTTGTATGTCTGTT-3′. One microliter of each PCR1 product was used for the second PCR for indexing. Each reaction mixture contained: 1 μl of PCR1 product, 5 μl of 10× buffer for KOD-Plus-Neo, 2.5 μl of each PCR2 primer (10 pmol/μl), 1 μl of KOD-Plus-Neo, 5 μl of 2 mM deoxyribonucleotide triphosphates (dNTPs), 3 μl of 25 mM MgSO[4], and 50 μl of water. The PCR cycling conditions for the indexing PCR were as follows: initial denaturation at 94°C for 2 min; 15 cycles of 98°C for 10 s, 60°C for 20 s, and 68°C for 20 s; and final extension at 72°C for 2 min. Sample purification and sequencing After PCR2, the samples were purified using a Tiangel Purification Kit (TIANGEN) and quantified with a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific). The purified PCR products, together with the MKO plasmid samples, were then subjected to high-throughput sequencing on a HiSeq 4000 instrument (Illumina) via the Hi-TOM platform at Xi’an CyanSnow Gene. CRISPR screen data processing Raw data were preprocessed using MAGeCK’s “count” function to obtain sgRNA read counts. Reads were trimmed with Illumina HiSeq Analysis software and aligned to the MKO library, with counts normalized against total reads across all samples. Log[2] fold changes between biological replicates and control plasmids were calculated for each sgRNA. A Bowtie index of the sgRNA library was generated using Bowtie 1.1.2. MAGeCK’s alpha-robust rank aggregation (RRA) scores identified top genes based on the average enrichment of their six gene-specific sgRNAs in TILs versus plasmids [log[2] ratio (TILs/plasmid)], with P < 0.05 significance. P values were derived from z-scores using the normal distribution, and FDRs were adjusted for multiple hypothesis testing using the Benjamini-Hochberg method in R. Top genes met three criteria: 1) The presence of at least two significant sgRNAs in each sample. 2) Significant sgRNA enrichment, defined as P < 0.05. 3) A rank of 50 or higher (ranked by RRA score). Western blot OT-I;Cas9 CD8^+ T cells (5 × 10^6) were electroporated with NTC or sgPrdm12. Protein samples were prepared by adding loading buffer to cell pellets, incubating on ice for 20 min, and heating at 95°C for 5 min. Protein concentration was determined using the Bicinchoninic Acid (BCA) assay. Samples (20 μg) were separated by SDS–polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane using a semidry transfer apparatus (Bio-Rad). The membrane was blocked with 5% nonfat milk in Phosphate-Buffered Saline with Tween-20 (PBS-T) for 1 hour at room temperature and then incubated with primary antibody overnight at 4°C. After PBS-T washes, the membrane was incubated with a horseradish peroxidase–conjugated secondary antibody for 1 hour at room temperature, washed again, and visualized. PRDM12 expression was detected using an Abmart antibody (catalog no. MA9505), and glyceraldehyde-3-phosphate dehydrogenase (Santa Cruz Biotechnology, catalog no. sc-32233) served as the loading control. Quantitative polymerase chain reaction Total RNA was extracted from Prdm12 KO, PD-1 KO, control (CTRL), and OT-I-CD8^+ T cells, as well as Prdm12 KO and control TILs, using TRIzol (Invitrogen, catalog no. 15596026). RNA concentration and purity were measured by NanoDrop (Thermo Fisher Scientific, USA). One microgram of RNA was reverse transcribed to cDNA using PrimeScript RT Master Mix (TAKARA). qPCR was performed using Hieff qPCR SYBR Green Master Mix (Low Rox Plus) (Yeasen) on a real-time PCR system (Applied Biosystems), and relative expression was calculated using the 2^−ΔΔCt method (Livak method), normalized to β-actin. All samples were analyzed in technical triplicates. Data were analyzed using GraphPad Prism, with significance determined by unpaired two-sided t tests (*P < 0.05, **P < 0.01, and ***P < 0.001). Results are shown as means ± SEM. Primers are listed in table S3. RNA-seq and data processing Total RNA from Prdm12 KO and CTRL OT-I-CD8^+ T cells was extracted using TRIzol. mRNA was isolated with Dynabeads Oligo(dT) (Thermo Fisher Scientific), fragmented using the Magnesium RNA Fragmentation Module (NEB). First-strand cDNA synthesis was performed using SuperScript II Reverse Transcriptase (Invitrogen). Second-strand cDNA synthesis was carried out using E. coli DNA polymerase I (NEB), ribonuclease H (NEB), and deoxyuridine triphosphate (dUTP) solution (Thermo Fisher Scientific). The double-stranded cDNA was end repaired, A tailed, and ligated with Illumina-compatible adapters using the NEBNext Ultra II Directional RNA Library Prep Kit (NEB, USA). Adapter-ligated cDNA fragments were size-selected [300 to 500 base pairs (bp)] using AMPure XP beads (Beckman Coulter, USA) and subsequently amplified by PCR. The average insert size for the final cDNA libraries was 300 ± 50 bp. Libraries were quantified using a Qubit Fluorometer (Thermo Fisher Scientific, USA) and validated. Last, 2 × 150-bp paired-end sequencing (PE150) was performed on an Illumina NovaSeq 6000 sequencer (LC-Bio Technology Co. Ltd., Hangzhou, China). Raw mRNA-seq data were filtered using Cutadapt to remove adapters, polyA/G sequences, >5% unknown bases (N), and low-quality reads (>20% bases with Q ≤ 20). Sequence quality was checked with FastQC. This yielded 1.5 Gbp of cleaned paired-end reads, which were aligned to the mouse genome using HISAT2. DESeq2 identified differentially expressed genes (FDR < 0.05, fold change > 2). These genes underwent GO and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis using clusterProfiler in R. ATAC-seq and data processing Prdm12 KO and CTRL OT-I-CD8^+ T cells were prepared for ATAC-seq using the TruePrep DNA Library Prep Kit V2 for Illumina (no. TD501, Vazyme). Cells were lysed, and nuclei were incubated with transposition mix for 30 min at 37°C. Transposed DNA was purified with VAHTS DNA Clean Beads and library-prepped using the TruePrep DNA Library Prep Kit V2 for Illumina (TD502, Vazyme), followed by 12-cycle PCR amplification. Post-PCR products were purified again. Libraries were sequenced as 150-bp paired-end reads on an Illumina NovaSeq 6000 to a depth of 200 million reads per sample by Novogene. ATAC-seq FASTQ files were cleaned using Trimmomatic and aligned to the mouse genome (mm10) with BWA-MEM. Samtools filtered out unmapped, unpaired, and mitochondrial reads. Peak calling was conducted using MACS2 with an FDR q value of <0.01, merged using BedTools, and read counts were quantified by BedTools coverage. Differentially accessible regions were determined via DESeq2 normalization (FDR < 0.05). Motif enrichment analysis was carried out using the MEME suite ([182]http://meme-suite.org/meme/) on peaks that exhibited differential accessibility between the sgPrdm12 and the sgNTC group. TF binding site predictions were made using a known motif discovery approach. Normalized BedGraph files were generated using scaling factors, which were subsequently converted into bigWig files for signal visualization in the Integrated Genomics Viewer (IGV v2.3.61). CUT&TAG sequencing and data processing Prdm12 KO and CTRL OT-I-CD8^+ T cells were processed using NovoNGS CUT&Tag 3.0 Kit (Novoprotein). Briefly, 1 × 10^5 cells were incubated with ConA beads for 10 min at room temperature, followed by overnight incubation with anti-H3K9me3 (no. AB8898, Abcam) antibody at 4°C and a 1-hour incubation with secondary antibody. Cells were then treated with ChiTag transposome for 1 hour at room temperature and tagmentation buffer for 1 hour at 37°C. The reaction was stopped by adding stop buffer and incubating at 50°C for 10 min. DNA was extracted using extraction beads, PCR amplified to generate libraries, and purified before sequencing on Nova-PE150. Raw data were filtered using Trimmomatic and aligned to the mouse genome (mm10) with BWA-MEM. MACS2 identified peaks. A hypergeometric test evaluated overlaps between H3K9me3-bound genes and DEGs. Scaling factors normalized BedGraph files, which were converted to bigwig format for downstream analysis and visualization in IGV (version 2.3.61). Immune infiltration analysis The TISIDB database ([183]http://cis.hku.hk/TISIDB/) was used to evaluate the correlations between Prdm12 expression and TILs in melanoma. Analysis of CTL function in patient survival The clinical outcomes associated with PRDM12 expression in patients with melanoma were assessed using the Kaplan-Meier Plotter, in conjunction with the TIDE tool ([184]80) ([185]http://tide.dfci.harvard.edu/query/). Melanoma datasets were analyzed for evidence of T cell dysfunction, stratified according to the bulk tumor expression levels of PRDM12. Statistics and reproducibility Unless otherwise stated, data are presented as means ± SD from biological replicates (independent animals or independent experiments), with the number of samples (n) indicated in the figures or specified in the figure legends. For the statistical comparison between two groups, an unpaired two-sided t test was used. Significance levels are denoted as follows: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Acknowledgments