tRNA m¹A modification regulates cholesterol biosynthesis to promote antitumor immunity of CD8⁺ T cells

CD8⁺ T cells are pivotal in tumor destruction, and the durability, longevity, and function of CD8⁺ T cells determine the efficacy of immunotherapy. As a crucial component of antitumor immunity, CD8⁺ T cells must rapidly and efficiently respond to tumor-associated antigens, necessitating timely metabolic reprogramming upon activation (Liu et al., 2022). The pursuit of innovative approaches to selectively modulate the metabolic pathways in effector CD8⁺ T cells has become a focal point in cancer research. Nevertheless, our understanding of the posttranscriptional functional and metabolic modulation of tumor-infiltrating CD8⁺ T cells remains limited.

The N1-methyladenine (m¹A) modification of RNA, which involves the methylation of adenosine at the nitrogen-1 position, is one of the most prevalent posttranscriptional modifications (Li et al., 2016; Safra et al., 2017). This modification is predominantly found in tRNA and ribosomal RNA (rRNA), with a rare occurrence in messenger RNA (mRNA) (Dominissini et al., 2016; Li et al., 2016, 2017; Peifer et al., 2013; Safra et al., 2017; Zhou et al., 2019). In tRNAs, m¹A modification can occur at multiple sites, typically at position 58 in the tRNA T-loop structure (Li et al., 2017; Anderson and Droogmans, 2005; Dégut et al., 2016; Oerum et al., 2017). Prior in vitro studies have shown that tRNA m¹A can alter RNA secondary structure (Oerum et al., 2017; Barraud et al., 2008, Basavappa and Sigler, 1991) and plays a crucial role in the initiation and elongation of the translation process (Liu et al., 2016; Macari et al., 2016). As a reversible tRNA modification, m¹A is modulated by m¹A “writers,” “erasers,” and “readers” proteins (Li et al., 2016). In mammalian nuclei and cytoplasm, m¹A modifications in tRNA are catalyzed by the writer complex TRMT6/61A, with TRMT61A as the m¹A catalytic core (Wang et al., 2016). The in vivo function of m¹A modification, however, remains largely unexplored.

Recent studies have linked the aberrant expression of m¹A-related enzymes to various malignant tumors. For instance, TRMT6/TRMT61A–mediated tRNA m¹A methylation has been found to accumulate in hepatocellular carcinoma (HCC) patient tumor tissues and is inversely correlated with HCC survival (Wang et al., 2019, 2021). Similar observations have been reported in gastrointestinal, pancreatic, ovarian, breast, and urinary bladder cancers and in lung adenocarcinoma. (Zhao et al., 2019; Li et al., 2019; Yamato et al., 2012; Woo and Chambers, 2019; Tasaki et al., 2011; Shi et al., 2015). Our previous work also observed increased expression of m¹A writers in colorectal cancers (CRCs), which correlates highly with the role of tumor-infiltrating immunosuppressive cells (Chen et al., 2021). We have recently uncovered a novel regulatory mechanism in CD4⁺ T cells, in which tRNA m¹A modification promotes T cell expansion through efficient MYC protein synthesis (Liu et al., 2022). However, the role of tRNA m¹A modification in tumor-associated immunostimulatory cells remains unknown.

In this study, we reveal a novel mechanism by which the antitumor response of mouse CD8⁺ T cells is potentiated by tRNA m¹A methyltransferase TRMT61A through the modulation of cholesterol metabolism. We demonstrate that the expression of the TRMT61A gene in tumor-infiltrating CD8⁺ T cells is positively correlated with T cell–mediated cytotoxicity in CRC patients and the specific deletion of Trmt61a in mouse T cells results in impaired antitumor cytotoxicity of CD8⁺ T cells. We further show that TRMT61A-mediated tRNA m¹A installation is essential for the efficient translation of a key protein, ATP citrate lyase (ACLY), which is essential for cholesterol biosynthesis (Feng et al., 2020; Lim et al., 2022). Collectively, our study elucidates the novel function and mechanism of tRNA m¹A modification as a cholesterol metabolic checkpoint in CD8⁺ T cells, offering insightful strategies for T cell–mediated immunotherapy.

Given the established link between tRNA m¹A modification and CRC progression (Chen et al., 2021), we hypothesized that tumor microenvironment (TME) might modulate m¹A modification to influence tumor-infiltrated CD8⁺ T cell function. We first focused on the tRNA m¹A writer gene Trmt61a and investigated its expression patterns in relation to T cell tumor-killing capacity within the TME. Bioinformatic analysis of public databases revealed that TRMT61A expression was upregulated in tumor tissues compared with adjacent normal tissues across various cancers (Fig. 1 A). Moreover, TRMT61A expression levels correlated inversely with patient survival rates in colon adenocarcinoma (Fig. 1 B), confirming a connection between tRNA m¹A modification and CRC progression.

To directly assess the impact of the TME on Trmt61a expression, we purified CD8⁺ T cells from murine colon tumors. In contrast, we observed a marked reduction in Trmt61a expression in tumor-infiltrated CD8⁺ T cells compared with that in CD8⁺ T cells from spleens and lymph nodes (LNs) of tumor-bearing mice or healthy controls (Fig. 1 C), suggesting a suppressive effect of the TME on m¹A modification in tumor-infiltrating leukocytes (TILs). In vitro, co-culture of CD8⁺ T cells with tumor cells or exposure to tumor cell supernatant further confirmed the suppressive effect of the TME on Trmt61a expression in CD8⁺ T cells (Fig. 1 D). Furthermore, we analyzed public single-cell RNA sequencing (scRNA-seq) data and found that TRMT61A gene expression in TILs was positively correlated with T cell–mediated cytotoxicity in CRC patients (Fig. 1 E), indicating that TME might suppress m¹A modification in TILs and consequently the tumor-killing function of TILs. Consistently, Trmt61a expression was significantly induced upon CD8⁺ T cell activation (Fig. 1 F). Collectively, Trmt61a expression in CD8⁺ T cells is dynamically regulated in response to TME and activation signals.

To elucidate the physiological role of Trmt61a and its mediated m¹A modification in CD8⁺ T cells, we generated Trmt61a^flox/flox mice and crossed them with Cd4^cre mice to achieve T cell–specific deletion of Trmt61a (Trmt61a conditional knockout, Trmt61a-cKO) (Fig. S1 A). Efficient ablation of Trmt61a in CD8⁺ T cells was confirmed at both the mRNA and protein levels, with no effect on the expression of its partner protein Trmt6 (Fig. S1, B and C). Concurrently, global tRNA m¹A modification levels in Trmt61a-KO CD8⁺ T cells were significantly reduced compared with that in wild-type (WT) cells, with no significant change detected in mRNAs that are generally over 200 nt (Fig. 2 A).

Next, we characterized the cellular composition of the Trmt61a-cKO mice at steady state. We found that the major immune organs, including the thymus, spleen, and inguinal LNs (iLN), displayed similar sizes between WT and Trmt61a-cKO mice. However, the mesenteric LNs (mLN) in Trmt61a-cKO mice were slightly enlarged, suggesting the presence of mild spontaneous colitis (Fig. S1 D). Under steady-state conditions, Trmt61a-cKO mice exhibited similar T cell subset compositions in the thymus (Fig. S1, E–G), indicating that Trmt61a deficiency does not disrupt thymic T cell development. Nevertheless, the numbers of both total CD4⁺ and CD8⁺ T cells, as well as naïve CD4⁺ and CD8⁺ T cells, were increased in the LNs but decreased in the spleen of Trmt61a-cKO mice (Fig. S1, H–L; Fig. S2, A–E; and Fig. 2, B–F), consistent with the observed mild colitis.

Further characterization of CD4⁺ T cell functional populations revealed aberrant T cell activation in the peripheral lymphoid organs of Trmt61a-cKO mice, as evidenced by a significant reduction in the proportions of previously activated T cells in the spleen and LNs (Fig. S2, D and E), along with an increased number of regulatory T (Treg) cells in the mLN (Fig. S2, F–H) compared with WT mice. These alterations indicate a disrupted T cell homeostasis in the peripheral lymphoid organs of cKO mice.

To evaluate the function of TRMT61A in CD8⁺ T cell–mediated antitumor immunity in vivo, we employed the MC38 tumor model. Subcutaneous injection of MC38 adenocarcinoma cells into WT and Trmt61a-cKO mice revealed that tumors in Trmt61a-cKO mice grew faster and achieved a significantly larger size than those in WT mice (Fig. 2, G–I). Furthermore, both the proportion and number of tumor-infiltrating CD8⁺ T cells were reduced in Trmt61a-cKO mice (Fig. 2, J–L). Additionally, the proportion of Ki-67⁺ CD8⁺ T cells and the secretion of granzyme B were significantly diminished in Trmt61a-cKO tumor-infiltrating CD8⁺ T cells (Fig. 2, M–P), although IFN-γ secretion remained unchanged (Fig. S3, A and B).

Given that Trmt61a-cKO mice also exhibit a defective phenotype in CD4⁺ T cells (Liu et al., 2022), we hypothesized that the unaltered IFN-γ production in CD8⁺ T cells could be influenced by the CD4⁺ T cell phenotype in the TME. Despite this, neither the infiltration (Fig. S3, C and D) nor the cytotoxicity (Fig. S3, E–H) of CD4⁺ T cells were affected in Trmt61a-cKO mice. However, the proportion of Treg cells was reduced in Trmt61a-cKO tumor-infiltrating CD4⁺ T cells (Fig. S3, I and J). Given the critical inhibitory role of Treg cells in TILs, this reduction may account for the unchanged IFN-γ production in CD8⁺ T cells.

To exclude the influence of CD4⁺ T cells, we crossed Trmt61a-cKO mice with OT-I transgenic mice, which harbor a transgenic T cell receptor specific for the ovalbumin (OVA) peptide. WT and cKO OT-I CD8⁺ T cells were then transferred into MC38-OVA tumor-bearing Rag1^−/− mice to assess antitumor responses in vivo (Fig. 3 A). Before transfer, there was no significant difference in the secretion of granzyme B and IFN-γ between in vitro–activated live WT and cKO OT-I CD8⁺ T cells (Fig. 3, B–E). However, consistent with the previous tumor model, tumors in Rag1^−/− mice receiving cKO OT-I CD8⁺ T cells grew faster and reached larger sizes than those receiving WT OT-I CD8⁺ T cells (Fig. 3, F–H). Furthermore, the proportion and number of tumor-infiltrating OT-I CD8⁺ T cells were significantly reduced in the Trmt61a-cKO group (Fig. 3, I–K). The secretion of both granzyme B and IFN-γ was also decreased in the tumor-infiltrating OT-I CD8⁺ T cells from the Trmt61a-cKO group (Fig. 3, L–O). These findings underscore the essential role of TRMT61A in maintaining CD8⁺ T cell homeostasis and their antitumor functions.

To further investigate the role of Trmt61a in antigen-specific CD8⁺ T cell responses, we used WT and Trmt61a-cKO OT-I mice to perform the in vitro killing assay. In vitro stimulation of WT-OT-I and Trmt61a-cKO-OT-I naïve CD8⁺ T cells with the OVA peptide followed by co-culture with OVA-expressing MC38 tumor cells revealed that Trmt61a-deficient CD8⁺ T cells exhibited an impaired tumor-killing capability, as evidenced by reduced target cell lysis (Fig. 4, A and B). Consistently, the secretion of granzyme B and IFN-γ by Trmt61a-deficient CD8⁺ T cells was diminished compared with WT cells (Fig. 4, C–F), further substantiating the compromised tumor-killing function of CD8⁺ T cells upon Trmt61a deletion.

To delineate the cellular events contributing to the functional deficits of CD8⁺ T cells, we first assessed the expression of the T cell activation marker CD44 in Trmt61a-deficient CD8⁺ T cells upon in vitro stimulation. We observed a decreased expression of CD44, indicative of activation defects in Trmt61a-deficient CD8⁺ T cells (Fig. 4, G and H). Second, we assessed the proliferation ability of Trmt61a-deficient CD8⁺ T cells by CellTrace labeling and observed impaired proliferation of Trmt61a-deficient CD8⁺ T cells compared with WT controls following in vitro TCR stimulation (Fig. 4, I and J). This in vitro result aligns with the reduced numbers of tumor-infiltrating CD8⁺ T cells observed in vivo in the TME (Fig. 2, G–P). Third, the expression of exhaustion markers TIM-3, PD-1, and LAG-3 did not significantly change in Trmt61a-cKO CD8⁺ T cells (Fig. S4, A–D). Simultaneously, the secretion of effector cytokines granzyme B and IFN-γ was decreased in Trmt61a-deficient CD8⁺ T cells with chronic stimulation (Fig. S4, E–H). Lastly, we evaluated cell apoptosis using annexin V/7-aminoactinomycin D (7-AAD) staining and found no significant differences in early and late apoptosis between WT and Trmt61a-deficient CD8⁺ T cells (Fig. S4, I and J). Altogether, these results demonstrate that Trmt61a deficiency impairs CD8⁺ T cell antitumor function and T cell activation and proliferation but does not affect T cell apoptosis and exhaustion.

To confirm that the diminished tumor-killing capability of CD8⁺ T cells due to TRMT61A deletion was indeed a consequence of reduced tRNA m¹A modification, we reintroduced WT Trmt61a or an m¹A catalytic dead-mutant Trmt61a gene into Trmt61a-cKO CD8⁺ T cells via retroviral transfection. The reduction of RNA m¹A modification in Trmt61a-cKO CD8⁺ T cells was reversed by the transfection of WT Trmt61a but not by the m¹A58 catalytic-dead-mutant Trmt61a (Fig. S4, K and L). Concurrently, the re-expression of both WT Trmt61a and m¹A58 catalytic-dead-mutant Trmt61a restored the protein levels of TRMT61A in Trmt61a-cKO CD8⁺ T cells (Fig. S4 M). Notably, the overexpression of WT Trmt61a, but not the m¹A58 catalytic-dead-mutant Trmt61a, partially rescued the defective tumor-killing function (Fig. 4 K) and the impaired proliferation (Fig. 4 L) of Trmt61a-cKO CD8⁺ T cells in vitro.

To further clarify the physiological role of m¹A in CD8⁺ T cells, we also constructed the Trmt6^flox/flox mice (Fig. S4 N), which was the partner of Trmt61a in the same m¹A writer complex, then crossed with Cd4^cre mice to specifically delete Trmt6 gene in T cells. The knockout efficiency of Trmt6 in CD8⁺ T cells of Trmt6^flox/floxCd4^cre (Trmt6-cKO) mice was verified at both mRNA and protein levels (Fig. S4, O and P). Indeed, Trmt6-deficient CD8⁺ T cells phenocopy the defective proliferation of Trmt61a-deficient CD8⁺ T cells, which further supported the importance of m¹A modification in CD8⁺ T cells (Fig. S4, Q and R).

Collectively, these results reveal that tRNA m¹A is indispensable for CD8⁺ T cell–mediated tumor immunity.

As demonstrated by our m¹A MS analysis, tRNA m¹A levels were significantly reduced in cKO CD8⁺ T cells, while mRNA m¹A levels remained largely unchanged (Fig. 2 A). Consequently, we focused on tRNA and excluded the possibility that mRNA serves as a direct target of TRMT61A-mediated m¹A modification in CD8⁺ T cells. Previous in vitro studies have established that tRNA plays a pivotal role in both the initiation and elongation phases of protein translation. Therefore, the direct downstream targets of TRMT61A are likely the genes that exhibit reduced protein levels without corresponding changes in mRNA levels.

Therefore, to comprehensively explore the molecular mechanisms underlying the impaired tumor-killing function of CD8⁺ T cells, we first performed RNA-seq analysis on naïve and activated CD8⁺ T cells from Trmt61a-cKO mice and their WT littermate counterparts. The sequencing results indicated that, compared with WT naïve CD8⁺ T cells, only a limited number of genes were significantly dysregulated in cKO naïve CD8⁺ T cells (Fig. 5 A), suggesting a minimal impact of TRMT61A on the transcriptional profile of naïve quiescent CD8⁺ T cells. However, in line with the observed phenotypic alterations, TCR-activated Trmt61a-deficient CD8⁺ T cells exhibited a substantial number of significantly downregulated and upregulated genes (Fig. 5, B and C). Intriguingly, the downregulated genes in activated Trmt61a-deficient CD8⁺ T cells were predominantly associated with the steroid/cholesterol biosynthesis pathway (Fig. 5 D).

Cholesterol, an integral component of cell membranes, is synthesized through an energetically demanding process involving over 20 enzymes, with 3-hydroxy-3-methylglutaryl coenzyme A reductase and squalene epoxidase (SQLE) as the rate-limiting steps (Luo et al., 2020). Our analysis revealed that several enzymes within the steroid/cholesterol biosynthesis pathway were downregulated at the transcriptional level in Trmt61a-deficient CD8⁺ T cells (Fig. 5, E and F), implying a potential role for TRMT61A in promoting cholesterol production within CD8⁺ T cells.

Additionally, we noticed an increase in the expression of the sterol O-acyltransferase 2 (Soat2) gene in Trmt61a-deficient CD8⁺ T cells, suggesting a potential involvement of cholesterol esterification in m¹A-regulated mechanism. To address this point, we first performed BODIPY staining to assess the lipid droplet content. However, no significant differences in lipid droplet levels were observed between WT and Trmt61a-deficient CD8⁺ T cells, regardless of whether they were in a naïve or activated state (Fig. S5, A and B). Next, to assess changes in the absolute content of cholesterol esters in CD8⁺ T cells following Trmt61a knockout, we conducted MS lipidomics analysis. Notably, the absolute cholesterol ester content did not significantly change after Trmt61a knockout in CD8⁺ T cells (Fig. S5, C and D). Furthermore, our RNA-seq data revealed that the mRNA level of Soat1 (also known as Acat1) remained unchanged (data not shown). The unaltered cholesterol esters and Acat1 mRNA levels were consistent with previous findings that ACAT1 was the primary enzyme responsible for cholesterol esterification in CD8⁺ T cells, while ACAT2 played a nearly dispensable role (Yang et al., 2016). Collectively, these results did not support the involvement of cholesterol esterification in our mechanism.

To further narrow down the direct target genes of TRMT61A, we conducted MS–based proteomics analysis on in vitro–activated CD8⁺ T cells from Trmt61a-cKO mice and their WT littermates, aligned with RNA-seq data collected at the same activation time point. This analysis revealed 55 upregulated and 130 downregulated proteins in Trmt61a-deficient CD8⁺ T cells. Notably, there was a significant downregulation of proteins most associated with the terpenoid backbone biosynthesis pathway, which is upstream of the steroid/cholesterol biosynthesis pathway (Fig. 5, G and H). Additionally, bioinformatic analysis of scRNA-seq data from TILs further supported these findings, demonstrating a positive correlation between TRMT61A expression and cholesterol metabolism in CRC patients (Fig. 5 I). Collectively, these results suggest that m¹A modification may function as a key regulator of cholesterol metabolism in activated CD8⁺ T cells.

Previous studies have shown that cholesterol is essential for clonal expansion (Kidani et al., 2013; Bensinger et al., 2008) and antitumor response CD8⁺ T cells (Yang et al., 2016; Yan et al., 2023). In line with our RNA-sequencing and proteomics data, we found no significant differences in cholesterol levels between quiescent naïve WT and Trmt61a-deficient CD8⁺ T cells (Fig. 6, A–C). However, upon TCR activation, a significant reduction in cholesterol content was observed in Trmt61a-deficient CD8⁺ T cells compared with WT cells (Fig. 6, A–C). Interestingly, supplementation with cholesterol was found to restore the compromised tumor-killing function (Fig. 6 D) and granzyme B secretion (Fig. 6, E and F) as well as the impaired proliferation (Fig. 6 G) of Trmt61a-cKO CD8⁺ T cells in vitro. Moreover, it was the overexpression of WT Trmt61a, but not m¹A58 catalytic-dead Trmt61a, that largely reversed the cholesterol reduction of Trmt61a-deficient T cells in vitro, as evidenced by filipin III fluorescence staining (Fig. 6, H and I). Taken together, our data suggest that tRNA m¹A modification regulates cholesterol biosynthesis to promote antitumor immunity of CD8⁺ T cells.

In a final effort to pinpoint the direct regulatory targets of tRNA m¹A modification in CD8⁺ T cells, we examined the intersection of genes that were unchanged at the transcriptional level and decreased at the protein level in Trmt61a-deficient CD8⁺ T cells. According to our RNA-seq and proteomics data (Fig. 5), this analysis yielded 76 genes hindered in the process of protein synthesis, many of which were significantly associated with metabolic pathways (Fig. 7, A and B). Among these, ACLY emerged as a key enzyme reported to generate acetyl-CoA from citrate, thereby linking carbohydrate metabolism with cholesterol biosynthesis (Feng et al., 2020; Lim et al., 2022) (Fig. 7 C). Notably, we validated by quantitative reverse transcription PCR (RT-qPCR) and western blot that while the mRNA levels of ACLY in Trmt61a-deficient CD8⁺ T cells remained unaltered compared with that in WT cells, the protein levels were diminished (Fig. 7, D and E), suggesting that Acly might be translationally regulated by m¹A modification.

To validate the direct regulation of the m¹A modification on ACLY, we conducted a codon-switch experiment. By altering the codons within the Acly cDNA that are decoded by tRNAs with high m¹A modification to their synonymous counterparts decoded by tRNAs with low m¹A modification, we created a mutant Acly (ACLY-Mut) plasmid in which the original authentic CAG (leucine), CAC (valine), and GCC (glycine) codons were replaced by cognate synonymous AAG (leucine), AAC (valine), and ACC (glycine) codons (Fig. 7 F). Strikingly, overexpression of this ACLY-Mut plasmid, but not ACLY-WT plasmid, was sufficient to restore the impaired expression of ACLY in Trmt61a-deficient CD8⁺ T cells in vitro (Fig. 7 G), confirming that the translation of Acly mRNA is directly influenced by tRNA m¹A modification through codon decoding. Consistent with this, overexpression of the ACLY-Mut plasmid, but not the ACLY-WT plasmid, replenished the defective cellular cholesterol content in Trmt61a-deficient CD8⁺ T cells (Fig. 7, H and I). Moreover, ACLY-Mut constructs could partially rescue the proliferation (Fig. S5, E and F) and cytotoxicity (Fig. S5, G and H) of Trmt61a-deficient CD8⁺ T cells to a greater extent than ACLY-WT constructs. These results suggest that TRMT61A-mediated tRNA m¹A modification is crucial for the efficient translation of ACLY, thereby ensuring the proper progression of cholesterol biosynthesis and the promotion of antitumor immunity in CD8⁺ T cells.

Our study sheds light on the pivotal role of tRNA m¹A modification as a metabolic regulator that enhances the tumor-killing function of CD8⁺ T cells through the modulation of cholesterol biosynthesis. The discovery that Trmt61a, the m¹A writer gene, is declined in tumor-infiltrating CD8⁺ T cells, upregulated upon TCR activation, and correlates with cytotoxicity in CRC patients points to a potential role in the immune response to cancer. The deletion of Trmt61a in CD8⁺ T cells resulted in a diminished capacity to reject tumors, both in vivo and in vitro, highlighting the importance of m¹A modification in the antitumor immune response.

As the central subject of our study, Trmt61a^fl/flCd4^Cre mice showed baseline differences in T cell counts in peripheral immune organs and CD8⁺ T cell effector functions after in vitro TCR activation. These basal phenotypes could impact the in vivo antitumor response of CD8⁺ T cells. To eliminate the impact of these basal phenotypes, we transferred an equal number of live WT and cKO OT-I CD8⁺ T cells into MC38-OVA tumor-bearing Rag1^−/− mice, thereby eliminating the baseline differences in T cell counts. Additionally, the effector cytokine secretion of in vitro–activated live WT and cKO OT-I CD8⁺ T cells showed no differences before the transfer (Fig. 3, B–E). However, after transferring into MC38-OVA tumor-bearing Rag1^−/− mice, the secretion of effector cytokines in tumor-infiltrating cKO OT-I CD8⁺ T cells was decreased compared with that in WT OT-I CD8⁺ T cells. This finding confirmed that the impaired in vivo antitumor response of cKO CD8⁺ T cells was triggered by stimulation from the TME, rather than by the basal phenotype in the Trmt61a^fl/flCd4^Cre mice.

The mechanistic insight that tRNA m¹A promotes cholesterol biosynthesis by enhancing the translation of ACLY is a novel finding with significant implications. Cholesterol is a critical component of cellular membranes and is essential for the clonal expansion and cytotoxic functions of CD8⁺ T cells (Kidani et al., 2013; Bensinger et al., 2008; Yang et al., 2016; Yan et al., 2023). However, there remain some controversial issues that cholesterol induces CD8⁺ T cell exhaustion (Ma et al., 2019; Hu et al., 2024). Our findings that Trmt61a-deficient CD8⁺ T cells exhibit reduced cholesterol levels and impaired proliferation and cytotoxicity are consistent with the concept that cholesterol metabolism is a key determinant of T cell function in the TME. Concurrently, the unaltered tumor-infiltration and function of Trmt61a-deficient CD4⁺ T cells correspond to a previous report that CD4⁺ T cells are more tolerant to cholesterol deficiency than CD8⁺ T cells, which have a higher demand for rapid clonal expansion (Yan et al., 2023).

The observation that cholesterol supplementation can rescue the defective functions of Trmt61a-deficient CD8⁺ T cells underscores the potential therapeutic applicability of modulating cholesterol metabolism in cancer immunotherapy. This approach could provide a novel strategy to enhance the efficacy of T cell–based immunotherapies, which have shown limited success in solid tumors due to the complex immunosuppressive TME (Maalej et al., 2023).

In the context of the last decade’s advances in T cell–based immunotherapies, such as CAR-T cell therapy, our findings offer a fresh perspective on the role of RNA epigenetic regulation in modulating T cell metabolism and function. The reversibility of m¹A modification and its dynamic regulation by writers, erasers, and readers proteins present an attractive target for therapeutic intervention. By targeting the m¹A machinery, we may be able to modulate T cell metabolism and enhance their antitumor activity, presenting a promising avenue for the treatment of both cancer and autoimmune diseases.

The interplay between RNA epigenetic regulation and cholesterol metabolism in T cells is a burgeoning field with much to explore. Interestingly, the m¹A writer gene TRMT61A is hampered by potential signals from TME and tumor culture medium, likely to be specific metabolites, cytokine, or tumor DNA from tumor cells. This observation highlights a potential interaction between tumor metabolism and RNA epigenetic modifications in T cells. Understanding the crosstalk mechanisms between these processes may reveal innovative metabolic intervention strategies for tumor immunotherapy. Furthermore, elucidating how tRNA m¹A modification cooperates with other RNA modifications to regulate T cell function will be an exciting area of future research.

In conclusion, our study highlights the importance of tRNA m¹A modification as a crucial translational checkpoint in CD8⁺ T cells, with significant implications for cancer immunotherapy. By uncovering the role of TRMT61A in regulating cholesterol metabolism and antitumor immunity, we pave the way for novel therapeutic strategies that could potentially transform the landscape of T cell–based cancer treatments.

Trmt61a-floxed mice were generated at the Shanghai Biomodel Organism Science & Technology Development Co., Ltd. using the CRISPR-Cas9–based genome-editing system. The left LoxP site is located 1,640 base pairs (bp) upstream of Trmt61a (ENSMUSG00000060950) exon 1 (positions 111,642,865–111,642,899 on the positive strand of chromosome 12), while the right LoxP site is located 237 bp downstream of exon 1 (positions 111,645,633–111,645,667 on the positive strand of chromosome 12). The guide RNA (gRNA) and donor oligonucleotides used are listed as follows. For Trmt61a left-side loxP: 5′-TGC​ATG​AAC​CAT​GTT​GTC​GG-3′ and 5′-AGC​ACA​CCT​TTA​AGC​ACA​GTA​TTT​GTT​AGG​CAG​AGC​AGG​CAG​ATC​TCT​TGG​GTC​TGA​GAC​CAG​CCT​GAT​CTA​CAT​AGT​GAG​TTC​CAG​GCA​GCC​AAG​GCT​ATA​TAG​GGA​GGC​TGT​CTG​AAA​GAC​AAA​ATA​TAG​CCC​TGC​ATG​AAC​CAT​GTT​GTC​GGT​ATG​ATG​TTA​CAC​AAA​TTA​GCT​TTC​TGT​TCC​CCA​TCT​GTA​AAG​TGG​TCA​TAA​AAT​TGT​AAG​GAA​GAG​GAT​TCA​ATG​ACA​TGA​GAG-3′; for Trmt61a right side loxP: 5′-GTG​CCC​TAT​GAG​GTC​GGA​GC-3′ and 5′-TGG​GAG​CTA​ACC​TGG​GTT​GGA​GAA​GAA​AGA​GGA​GGA​GGT​GCC​CTA​TGA​GGT​CGG​AGC​TGG​GAT​TGC​TTG​ATG​CTG​GCC​AAG​GTG​CTG​ACT​GCT​TTT​TTT​CTA​GGC​CAG​CTG​CCC​CTT​CCT​GGA​GGG​AGG​CTG​GAT​ACT​CCA​GAG​CTC​ATT​GCA​GAG​AGA​AGT​CTG​TCC​CAC​ACT​TAC​AGC​CAG​GCC​CTT​CCC​TAG​TTC​CAT​CGA​AGA​TTA​GCA​CTG​TCA​CTG​ATA​CGA​GGC​CAG​GTG​GTA-3′.

Trmt6-floxed mice were generated at Cyagen Biosciences (Suzhou) Inc. using the CRISPR-Cas9–based genome-editing system. The left LoxP site is situated 598 bp upstream of Trmt6 (ENSMUSG00000037376) exon 2 (positions 132,653,463–132,653,497 on the reverse strand of chromosome 2) and the right LoxP site is located 401 bp downstream of exon 2 (positions 132,654,369–132,654,403 on the reverse strand of chromosome 2). The gRNA and donor oligonucleotides used are listed as follows. For Trmt6 left-side loxP: 5′-AAG​AGA​CTG​AGA​TCT​CCG​ATA​GG-3′ and 5′-GCT​TGT​CTT​TGA​AGT​TGC​TCT​AAG​AGA​CTG​AGA​TCT​CCG​ATA​GGA​AGG​CTA​ATG​CCT​GAC​CCT​TGG​CAG​TAC​TTC​ATT​AGT​TCT​ACA​TCC​ATT​TCC​AAT​GTG​TAG​ATT​GCC​AAT​AAT​GTT​TAT​TCT​GAC​ACA​GGC​TTT​TTG​GAA​TTT​TGC​TTT​TCT​AAT​AGA​GTA​GCC​AAT​TAG​ACA​GA-3′; for Trmt6 right side loxP: 5′-ATG​CTA​GAG​AAT​TAG​CCC​AAC​GG-3′ and 5′-CTT​ATC​TCA​AAA​AGG​ATC​TTT​CAG​GAT​GCT​AGA​GAA​TTA​GCC​CAA​CGG​TAA​AAG​TAC​TTG​TTG​TCC​TCT​GAG​AAG​ATC​CAG​GTT​CAG​TTC​TGA​ACA​CCC​ACA​TGG​TGG​CTC​ACA​ACC​ATC​TGT​TAC​TTC​AGT​TCC​AGG​GAT​CTG​ATG​CCC​TTT​CTG​AGC​TCA​GGC​ACC​AGG​CAT​GCA​TCT​GGT​ACT​CAT​ACA​TGC​ACA​CAG​GCA​AAA​CAC​TCA​A-3′.

Trmt6/61a^flox/flox mice were crossed with Cd4^Cre mice to obtain conditional knockout mice. Cd4^Cre mice were purchased from the Jackson Laboratory and were fully backcrossed to C57BL/6 mice (over >10 generations). Trmt6/61a^flox/flox mice without Cd4^Cre were used as WT littermate controls for Trmt6/61a^flox/floxCd4^Cre mice.

OT-I TCR-transgenic mice were kindly provided by Prof. Qiang Zou (Shanghai Jiao Tong University School of Medicine, Shanghai, China). The OT-I mice were crossed with Trmt61a^flox/floxCd4^Cre mice to produce Trmt61a^flox/floxCd4^Cre OT-I mice.

For all experiments, 6–8-wk-old sex-matched mice were used, unless otherwise indicated. All the mice were bred and maintained under specific pathogen–free conditions at the animal facility of Shanghai Biomodel Organism Science & Technology Development Co., Ltd. Animal procedures were approved by the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University School of Medicine.

MC38 and MC38-OVA (MC38-expressing OVA) murine colon cancer cells (gender: female) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) medium supplemented with 10% fetal bovine serum (FBS). All cells were grown at 37°C with 5% CO₂. 8-wk-old mice were injected subcutaneously in the abdomen with 5 × 10⁵ MC38 tumor cells. The challenged mice were monitored for tumor size expressed as tumor area. To minimize individual variations, six age- and sex-matched mice in each group were used. Analysis of tumor size was done in a blinded fashion and lethality was defined as tumor size reaching 225 mm².

For the adoptive transfer in MC38-OVA tumor-bearing Rag1^−/− mice, MC38-OVA tumor cells (5 × 10⁵) were injected subcutaneously into Rag1^−/− mice. The splenocytes isolated from Trmt61a^fl/fl or Trmt61a^fl/flCd4^Cre OT-I mice were pre-activated with 1 μg/ml of SIINFEKL peptide and 10 ng/ml of IL-2 for 3 days, and then with only 10 ng/ml of IL-2 for another 1 day. Then, 3 days after tumor transplant, an equal number of activated live OT-I CD8⁺ T cells (1 × 10⁶) were i.v. transferred into MC38-OVA tumor-bearing Rag1^−/− mice. The tumor size was measured every 3 days and analyzed at day 18.

Thymus, spleen, peripheral, and mLN were collected and pressed through a 200-gauge mesh. Spleen cells were prepared by lysing the erythrocytes with red blood cell lysis buffer (Thermo Fisher Scientific). Monoclonal antibodies (mAbs) against CD3 (clone 145-2C11), CD4 (clone RM4-5), CD8 (clone 53-6.7), CD44 (clone IM7), CD62L (clone MEL-14), and granzyme B (clone QA16A02) were purchased from BioLegend. The cells were incubated with anti-CD16/CD32 (clone 93; BioLegend) to block the Fc receptors first and then stained with other antibodies at 4°C for 30 min. For intracellular cytokine staining, the cells were stimulated with phorbol 12-myristate 13-acetate (50 ng/ml; Sigma-Aldrich), ionomycin (1 μg/ml; Sigma-Aldrich), and GolgiPlug (1 μl/ml; BD Biosciences) for 4–6 h. After surface antibody staining, the cells were fixed and permeabilized using the Fixation/Permeabilization Solution Kit 5 (BD Biosciences), followed by staining with antibodies against the intracellular molecules. For intracellular transcription factor staining, the cells were fixed/permeabilized using the FoxP3/Transcription Factor Buffer Set (BioLegend). Dead cells were excluded using the Zombie NIR Fixable Viability Kit or Zombie Aqua Fixable Viability Kit (Biolegend). Data were collected using a BD Fortessa X20 flow cytometer (BD Biosciences). The data were analyzed using FlowJo software (Tree Star).

Primary T cells were isolated from the spleen and LNs of age-matched WT and conditional knockout mice (6–8 wk old). Naïve CD8⁺ or total CD8⁺ T cells were purified using the MagniSort Mouse Naïve CD8 T or CD8 T cell Enrichment Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions, respectively. The purified T cells were stimulated with plate-bound anti-CD3 (5 μg/ml) and anti-CD28 (2 μg/ml) antibodies in replicate wells of 96-well plates. The proliferation of the CD8⁺ T cells was determined and analyzed by CellTrace dilution during the stimulation to naïve CD8⁺ T cells with antibodies against CD3 and CD28 for 72 h. The activation of the cells was determined by the expression of CD44 using flow cytometry when naïve CD8⁺ T cells were stimulated with antibodies against CD3 and CD28 for 24 h. APC Annexin V Apoptosis Detection Kit with 7-AAD (BioLegend) was used to analyze the apoptosis of the CD8⁺ T cells when CD8⁺ T cells were stimulated with antibodies against CD3 and CD28 for 24 h.

Purified naïve CD8⁺ T cells were stimulated with 1 μg/ml of SIINFEKL peptide and 10 ng/ml of IL-2 for 3 days in replicate wells of 96-well plates for 3 days and then cultured without OVA peptide for an additional 2 days. On the day before the co-culture of CD8⁺ T and tumor cells, MC38-OVA tumor cells were planted in replicate wells of 24-well plates (1 × 10⁵ cells per well) and adherently cultured overnight. Stimulated CD8⁺ T cells were then co-cultured with the MC38-OVA cells in different proportions for 16 h. The tumor-killing ability of CD8⁺ T cells was indicated by the proportion of apoptotic MC38-OVA tumor cells using the APC Annexin V Apoptosis Detection Kit with 7-AAD (BioLegend).

The purified naïve CD8⁺ T cells were stimulated with plate-bound anti-CD3 (5 μg/ml) and anti-CD28 (2 μg/ml) antibodies in replicate wells of a 96-well plate. A total amount of 3 μg RNA per sample was used as input material for the RNA sample preparations. Sequencing libraries were generated using the NEBNextUltra RNA Library Prep Kit for Illumina (NEB) following the manufacturer’s recommendations, and index codes were added to attribute sequences for each sample. Clustering of the index-coded samples was performed on a cBot Cluster Generation System using the TruSeq PE Cluster Kit v3-cBot-HS (Illumina) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina HiSeq platform and 125- to 150-bp paired-end reads were generated.

The libraries were sequenced on an Ilumina Novaseq 6000 platform and 150-bp paired-end reads were generated. The clean reads were mapped to the reference genome using HISAT2. Fragments per kilobase of exon model per million reads mapped of each gene were calculated and the read counts of each gene were obtained by HTSeq-count. Principal component analysis (PCA) was performed using R (v 3.2.0) to evaluate the biological duplication of samples. Differential expression analysis was performed using the DESeq2. Q value <0.05 and fold change >1.5 or fold change <0.67 were set as the threshold for significant differentially expressed genes (DEGs). Hierarchical cluster analysis of DEGs was performed using R (v 3.2.0) to demonstrate the expression pattern of genes in different groups and samples. Based on the hypergeometric distribution, KEGG pathway enrichment analysis of DEGs was performed to screen the significant enriched term using R (v 3.2.0). Gene set enrichment analysis was performed using GSEA software.

100 ng of small RNA (<200 nt) or large RNA (>200 nt) was hydrolyzed with 0.2 μl benzonase (Thermo Fisher Scientific), 0.25 μl phosphodiesterase I (Thermo Fisher Scientific), and 0.25 μl bacterial alkaline phosphatase (Thermo Fisher Scientific) in a 20 μl solution containing 4 mM NH₄OAc at 37°C overnight. After complete hydrolysis, the products were dissolved in acetonitrile and then applied to ultra-performance liquid chromatography–MS/MS (UPLC-MS/MS). The nucleosides were separated on a 3-μm HILIC column (Atlantis HILIC Silica column; 2.1 × 150 mm) and then detected using a triple quadrupole mass spectrometer (AB SCIEX QTRAP 6500⁺) in the positive ion multiple reaction monitoring mode. The mass transitions of m/z 282.0 to 150.1 (m¹A), m/z 282.0 to 150.1 (m⁶A), and m/z 268.0 to 136.0 (A) were monitored and recorded. Concentrations of nucleosides in RNA samples were deduced by fitting the signal intensities into the standard curves.

Retroviral transduction was performed by transfecting HEK293T cells with an mPGK-retrovirus (with GFP) plasmid encoding WT or D181A-mutant TRMT61A along with packaging plasmids (the sequence was analyzed using SnapGene v.4.1.6). Purified Trmt61a-deficient naïve CD8⁺ T cells were activated in a 96-well plate for 24 h and then infected with the packaged retrovirus in the presence of 8 mg/ml of polybrene by spinning at 800 g for 180 min. 72 h after retrovirus infection at 37°C, the proliferation of the cells was determined and analyzed by transfected GFP⁺ cell numbers, counted by flow cytometry using Precision Count Beads (cat. no. 424902; BioLegend).

Cellular cholesterol content was measured using the Cholesterol Cell-Based Detection Assay Kit (Cayman) and the Amplex Red Cholesterol Assay Kit (Invitrogen). For the detection assay, cells were stained with filipin III and then analyzed by flow cytometry. For cholesterol quantification, sterols were extracted with a sterol extraction kit (Sigma-Aldrich) and then analyzed using the Amplex Red assay.

For polyclonal CD8⁺ T cells, purified naïve CD8⁺ T cells were initially stimulated with 2.5 μg/ml anti-CD3/CD28 antibodies and 10 ng/ml of IL-2 for 2 days. Following stimulation, the T cells were cultured at a density of 1 × 10⁶ cells per ml in RPMI-1640 medium supplemented with 10% FBS, 2 mM L-glutamine, 5 μM β-mercaptoethanol, and 10 ng/ml IL-2. For chronic stimulation conditions, the cells were transferred to fresh plates with plate-bound anti-CD3 antibody (2.5 μg/ml) every 2 days. For acute conditions, anti-CD3 was omitted. The cells were used for examination until day 8.

The coding DNA sequences of WT-Acly and mut-Acly were constructed into an mPGK-retrovirus (with GFP) plasmid. Retrovirus was produced in 293T cells using the standard protocols. Transfection was performed using X-tremeGENE HP DNA Transfection Reagent (Roche). Purified WT and Trmt61a-deficient naïve CD8⁺ T cells were activated in a 48-well plate for 24 h and then infected with the packaged retrovirus in the presence of 8 mg/ml polybrene by spinning at 800 g for 180 min. GFP⁺ cells were sorted for immunoblot analysis 72 h after infection at 37°C.

Total RNA was isolated from CD8⁺ T cells with TRIzol reagent (Invitrogen) according to the manufacturer’s instructions and then reverse-transcribed with the TransScript All-in-One First-Strand cDNA Synthesis SuperMix (no. AT341-02; TransGen). iTaq Universal SYBR Green Supermix (no. 1725124; Bio-Rad) was used for RT-qPCR. Actin was set as an internal control to calculate mRNA relative amounts. The primer sequences used for RT-qPCR are as follows:

• Actin (forward, 5′-GGC​TGT​ATT​CCC​CTC​CAT​CG-3′; reverse, 5′-CCA​GTT​GGT​AAC​AAT​GCC​ATG​T-3′),

• Trmt61a (forward, 5′-CGC​ACG​CAG​ATC​CTC​TAC​TC-3′; reverse, 5′-GGA​ACT​CTA​CTG​TGT​GTA​GGT​GG-3′),

• Trmt6 (forward, 5′-CAT​CGG​CCA​TAG​TTA​CGG​CTC-3′; reverse, 5′-TCT​GTC​TTG​TCA​CGG​AAT​GTT​G-3′),

• Cdk2 (forward, 5′-GCG​ACC​TCC​TCC​CAA​TAT​CG-3′; reverse, 5′-GTC​TGA​TCT​CTT​TCC​CCA​ACT​CT-3′),

• Cdkn1b (forward, 5′-TCA​AAC​GTG​AGA​GTG​TCT​AAC​G-3′; reverse, 5′-CCG​GGC​CGA​AGA​GAT​TTC​TG-3′),

• Myc (forward, 5′-ATG​CCC​CTC​AAC​GTG​AAC​TTC-3′; reverse, 5′-CGC​AAC​ATA​GGA​TGG​AGA​GCA-3′).

Total protein from CD8⁺ T cells was extracted in radioimmunoprecipitation assay buffer (P0013E; Beyotime) supplemented with a protease and phosphatase inhibitor cocktail (no. 78443; Thermo Fisher Scientific). Antibodies against TRMT61A (cat. no. PA5-76553; Thermo Fisher Scientific), TRMT6 (cat. no. 16727-1-AP; Proteintech), and ACLY (cat. no. A3719; ABclonal) were diluted in 5% nonfat milk buffer with a concentration of 1:1,000 and incubated at 4°C overnight. After washing with 0.1% phosphate-buffered saline (PBS) with Tween-20 buffer three times, horseradish peroxidase–conjugated secondary antibodies (no. WBKLS0500; Merck Millipore) were added to the membranes and incubated at room temperature for 1 h. The signal was detected by enhanced chemiluminescence with pico-enhanced chemiluminescence using ChemiDoc MP (Bio-Rad). β-Actin (no. 3700S; Cell Signaling Technology) and GAPDH (no. 5174S; Cell Signaling Technology) were used as internal control.

The concentration of purified RNA samples was determined with NanoDrop and the RNA was serially diluted to 50 and 10 ng/μl using RNase-free water. The RNA samples were then denatured at 95°C for 3 min, followed by immediate cooling on ice to prevent the re-formation of secondary structures of mRNA. 2 μl of mRNA was directly dropped onto the Hybond-N⁺ membrane (Amersham). After drying it at room temperature, the spotted RNA was cross-linked to the membrane in a UVP Crosslinker CL-1000 (Analytik Jena) twice using the Autocrosslink mode (1,200 μJ [×100]; 25–50 s). The membrane was washed with TBST (0.1% Tween-20 in 1× TBS) and then incubated in a blocking buffer (0.5% nonfat milk in TBST) for 1 h at room temperature. The membrane was kept overnight with anti-m¹A antibody (cat. no. D345-3; MBL), diluted at 1:1,000 in the blocking buffer. Excess antibody was removed through washing with TBST. This was followed by incubation with HRP-conjugated AffiniPure IgG antibody (Goat anti-rabbit) for 1 h at room temperature. The m¹A dot blot was visualized using a chemiluminescence reagent. The entire process was conducted in an RNase-free environment.

Proteins were precipitated with cold acetone. The pellet was subsequently dissolved in 8 M urea and 100 mM Tris-HCl, pH 8.5. TCEP (final concentration is 5 mM) (Thermo Fisher Scientific) and iodoacetamide (final concentration is 10 mM) (Sigma-Aldrich) were added to the solution and incubated at room temperature for 30 and 20 min for reduction and alkylation, respectively. The protein mixture was diluted four times and digested overnight with trypsin at 1:50 (wt/wt), the reaction was stopped using formic acid (FA), and the peptide mixture was desalted using MonoSpin C18 column (Shimadzu-GL).

The peptide mixture was analyzed by a homemade 30-cm-long pulled-tip analytical column (75 μm ID packed with ReproSil-Pur C18-AQ, 1.9 μm resin; Dr. Maisch GmbH) and the column was then placed in-line with an Easy-nLC 1200 nano HPLC system (Thermo Fisher Scientific) for MS analysis. The analytical column temperature was set at 55°C during the experiments. The mobile phase and elution gradient used for peptide separation were as follows: 0.1% FA in water as buffer A and 0.1% FA in 80% acetonitrile as buffer B, 0–1 min, 2–10% B; 1–81 min, 10–35% B; 81–96 min, 35–60% B; 96–111 min, 60–100% B; 111–120 min, 100% B. The flow rate was set at 300 nl/min.

Data-dependent tandem MS/MS analysis was performed using a Q Exactive Orbitrap mass spectrometer (Thermo Fisher Scientific). Peptides eluted from the LC column were directly electrosprayed into the mass spectrometer with the application of a distal 2.5-kV spray voltage. A cycle of one full-scan MS spectrum (m/z 300–1,800) was acquired followed by top 20 MS/MS events, sequentially generated on the first to the 20th most intense ions selected from the full MS spectrum at a 28% normalized collision energy. Full scan resolution was set to 70,000 with automated gain control target of 3e6. MS/MS scan resolution was set to 17,500 with an isolation window of 1.8 m/z and automated gain control target of 5e5. The number of microscans was one for both MS and MS/MS scans and the maximum ion injection time was 50 and 100 ms, respectively. The dynamic exclusion settings used were as follows: charge exclusion, 1 and >8; exclude isotopes, on; and exclusion duration, 15 s. MS scan functions and LC solvent gradients were controlled by the Xcalibur data system (Thermo Fisher Scientific).

The acquired MS/MS data were analyzed against a UniProtKB Mouse using MaxQuant software. Trypsin was defined as the cleavage enzyme, oxidations (M) and acetylation (protein N term) were specified as variable modifications, and carbamidomethyl (C) was specified as the fixed modification. A database search was performed with a mass tolerance of 20 ppm for precursor ions for mass calibration. Six amino acids were required as the minimum peptide length and label free quantification (LFQ) intensity was used as relative quantification of protein. Q value <0.05 and fold change >1.5 or fold change <0.67 were set as the threshold for significantly differentially expressed proteins. Hierarchical cluster analysis was performed using R (v 3.2.0) to demonstrate the expression pattern of proteins in different groups and samples. Based on the hypergeometric distribution, KEGG pathway enrichment analysis of DEGs was performed to screen the significant enriched term using R (v 3.2.0).

BODIPY493/503 dye (cat. no. D2191; Thermo Fisher Scientific) was added to RPMI-1640 medium containing 5% FBS to a final concentration of 0.5 μg/ml for the detection of intracellular lipid droplets. Equal numbers of cells were resuspended and incubated at 37°C for 30 min. Following incubation, the cells were washed twice to remove excess dye. Standard flow cytometry surface staining procedures were then performed.

The samples were thawed on ice. 1 ml of the extraction solvent (MTBE: MeOH = 3:1, V/V) containing internal standard mixture was added. After whirling the mixture for 15 min, 200 μl of ultrapure water was added, vortexed for 1 min, and centrifuged at 12,000 rpm for 10 min. 500 μl of the upper organic layer was collected and evaporated using a vacuum concentrator. The dry extract was dissolved in 200 μl reconstituted solution (ACN: IPA = 1:1, V/V).

The sample extracts were analyzed using an LC-electrospray ionization-MS/MS system (UPLC, ExionLC AD; MS, QTRAP 6500⁺ System). The chromatographic columns from Thermo Accucore C30 (2.6 μM, 2.1 mm × 100 mm i.d.) were used. The solvent system was as follows: A, acetonitrile/water (60/40, V/V, 0.1% FA, and 10 mmol/liter ammonium formate); B, acetonitrile/isopropanol (10/90, V/V, 0.1% FA, and 10 mmol/liter ammonium formate). The gradient program was t = 0 min: A/B (80:20, V/V); t = 2 min: A/B (70:30, V/V); t = 4 min: A/B (40:60, V/V); t = 9 min: A/B (15:85, V/V); t = 14 min: A/B (10:90, V/V); t = 15.5 min: A/B (5:95, V/V); t = 17.3 min: A/B (5:95, V/V); t = 17.5 min: A/B (80:20, V/V); t = 20 min: A/B (80:20, V/V); and the flow rate was 0.35 ml/min, with a temperature of 45°C. Subsequently, the effluent was alternatively connected to an ESI-triple quadrupole-linear ion trap (QTRAP)-MS. Linear ion trap and triple quadrupole (QQQ) scans were acquired on a QTRAP-MS, QTRAP 6500⁺LC‒MS/MS System, equipped with an ESI Turbo Ion-Spray interface, operating in positive and negative ion modes, and controlled by Analyst 1.6.3 software (Sciex). Lipid contents were detected by LC‒MS/MS (AB Sciex QTRAP 6500), the service was provided by Metware, Co., Ltd. (http://www.metware.cn/), applying the lipidomics platform based on an in-house database (MWLDB, v 3.0).

No statistical methods were used to predetermine the sample sizes, but our sample sizes were similar to those reported in previous publications. Samples or mice were grouped according to the treatments or genotypes and thus were not randomized. Data collection and analysis were not performed blind to the conditions of the experiments. No data were excluded from the analyses. No tests for normality were performed; instead, the data distribution was assumed to be normal. Unless indicated otherwise, unpaired Student’s t test and one-way analysis of variance (ANOVA) were used to compare pairs of groups using GraphPad Prism 6. Data are expressed as the mean ± SEM. P values < 0.05 were considered to indicate statistical significance: ****P < 0.0001, ***P < 0.001, **P < 0.01, and *P < 0.05. All in vivo experiments were conducted with at least two independent cohorts.

Gene expression data were normalized from FPKM of tumor samples and adjacent normal samples obtained from The Cancer Genome Atlas (Weinstein et al., 2013). Normalized gene expression values were then log₂-transformed. Wilcoxon signed-rank test assessed the different expressions between tumor tissues and adjacent normal tissues. Kaplan–Meier survival analysis based on the R package survival was performed for patients with high and low expression according to the optimal cutoff point of the expression level by maxstat package. The processed scRNA-seq data based on Smart-seq for human colorectal tumor-infiltrating lymphocytes were obtained from a previous study (GSE146771) (Zhang et al., 2020). T cells were divided into two groups according to the expression and silencing of TRMT61A expression, and the Wilcoxon signed-rank test was used to evaluate the differences in T cell–mediated cytotoxicity and cholesterol metabolism between the two groups.

Fig. S1 shows the generation and characterization of Trmt61a-cKO mice. Fig. S2 shows the CD4⁺ T cell functional populations of Trmt61a-cKO mice under steady state. Fig. S3 shows the analysis of CD4⁺ T cell infiltration or function from WT and Trmt61a-cKO mice in the tumor model. Fig. S4 shows the analysis of apoptosis or exhaustion of CD8⁺ T cells from WT and Trmt61a-cKO mice. Fig. S5 shows the detection of lipid droplets and cholesterol esters of CD8⁺ T cells from WT and Trmt61a-cKO mice and the analysis of proliferation and cytokine secretion of CD8⁺ T cells after overexpression of m¹A codon-switched ACLY. Table S1 shows the MS proteomics data of activated WT and Trmt61a-cKO CD8⁺ T cells. Changed proteins are listed separately in Table S1, with those associated with the cholesterol biosynthesis pathway highlighted in red.

Data related to this study are available within the main text, main figures, and the supplementary materials. The bulk RNA-seq datasets are available through the GEO database at accession code GSE275602. The MS proteomics data are presented in Table S1. The public scRNA-seq data underlying Fig. 1 E and Fig. 5 I were derived from the GEO database at accession code GSE146771.

We thank M. Yang, K. Mao, Y. Zhou, and X. Cai for technical and administrative assistance; the Qiang Zou laboratory for donating the OT-I TCR-transgenic mice; all members of the H.-B. Li laboratory for discussions and suggestions; and Core Facility of Basic Medical Sciences, Shanghai Jiao Tong University School of Medicine for their support.

This work was supported by the Ministry of Science and Technology of China (2021YFA1100800 to H.-B. Li; 2020YFA0803400 to R.-J. Liu), the National Natural Science Foundation of China (82325024/82341017/82350112/82030042/32070917 to H.-B. Li), the Chongqing International Institute for Immunology (2023YJC01 to H.-B. Li), Shanghai Municipal Health Commission (2022JC001/2022XD047 to H.-B. Li), and the Innovative Research Team of High-Level Local Universities in Shanghai (SHSMU-ZDCX20212501 to H.-B. Li).

Author contributions: S. Miao: conceptualization, data curation, formal analysis, investigation, methodology, project administration, validation, visualization, and writing—original draft, review, and editing; H. Li: methodology; X. Song: data curation, formal analysis, visualization, and writing—review and editing; Y. Liu: conceptualization, investigation, and validation; G. Wang: data curation, formal analysis, software, and visualization; C. Kan: data curation; Y. Ye: formal analysis, methodology, software, visualization, and writing—original draft; R.-J. Liu: methodology; H.-B. Li: conceptualization, funding acquisition, project administration, resources, supervision, and writing–original draft, review, and editing.