METTL1 mediates PKM m7G modification to regulate CD155 expression and promote immune evasion in colorectal cancer

Abstract

Background

Colorectal cancer (CRC) is characterized by poor responsiveness to immune evasion and immunotherapy. RNA 7-methylguanine (m7G) modification plays a key role in tumorigenesis. However, the mechanisms by which m7G-modified RNA metabolism affects tumor progression are not fully understood, nor is the contribution of m7G-modified RNA to the CRC immune microenvironment.

Methods

The expression levels of Methyltransferase-like 1 (METTL1) and m7G in human tissues were determined. In this study, the effect of METTL1 on RNA m7G levels was evaluated, the effect of METTL1 on PKM mRNA modification was confirmed, the expression level of the PKM2 protein was detected, and the mechanism involved RT‒qPCR, Western blot, RNA stability analysis and RIP analysis. Lactate and H3K9 lactylation (H3K9la) induced by METTL1/PKM2 were analyzed via the extracellular acidification rate (ECAR) and lactic acid assays. Cut&Run was used to detect METTL1/PKM2-induced CD155 (PVR) transcription. In addition, METTL1 knockout mice were studied in vivo with CD155 blockers.

Results

We demonstrated that m7G RNA METTL1 enhances PKM2 expression by acting on PKM mRNA, leading to tumor progression and increased glycolysis. Specifically, METTL1 mediates m7G methylation of PKM mRNA and enhances the expression of its encoded PKM2, which in turn enhances glycolysis, promotes H3K9la, and activates METTL1 transcription, creating a positive feedback loop. Moreover, increased PKM2 dimer expression and nuclear translocation activated CD155 expression and induced CRC immune evasion.

Conclusions

Our findings reveal a general mechanism by which METTL1/PKM2/H3K9la signaling regulates RNA metabolism and highlight METTL1 targeting as a potential strategy for CRC immunotherapy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-024-05991-1.

Introduction

Colorectal cancer (CRC), a common malignant tumor, has a rising incidence and a high mortality rate annually, some of which are in the middle and late stages when it is first diagnosed [1]. A greater proportion of patients with advanced colorectal cancer experience recurrence after surgery, and the five-year survival rate is only approximately 30% [2]. The inhibitory immune microenvironment is closely related to the therapeutic effect and disease progression of CRC [3, 4]. Poliovirus receptor (PVR, CD155) is an important tumor immune checkpoint molecule, and the expression of CD155 is associated with tumor stage, lymph node metastasis and poor prognosis [5]. Studies have shown that CD155 can influence the development of CRC by regulating the functions of NK cells and T cells [6, 7].

Recent studies have shown that RNA modification, as an important epigenetic modification mode, can regulate RNA metabolism and the tumor immune microenvironment by regulating the interaction between tumor cells and host immune cells. N7-Methylguanosine (m7G) is one of the most common RNA modifications [8, 9]. This modification, which involves the addition of a methyl group to the N7 site of RNA guanine by methyltransferase, is a type of posttranscriptional regulation widely found in eukaryotic cells [10]. METTL1 (methyltransferase-like 1) performs its main catalytic function as a methyltransferase and forms a complex with WDR4 (WD repeat-containing protein 4) to act as a “writer”. The internal m7G modification of mRNAs has been reported [11]. METTL1 can increase the translation of target mRNAs through this modification and promote tumor occurrence and development. In addition, QKI (Quaking), the first recognition protein modified by m7G in mRNA, has been found to dynamically regulate mRNAs modified by m7G in cell stress particles under stress states such as tumor chemotherapy and radiotherapy, thus altering the response of tumor cells to treatment [12]. Nevertheless, the impact of m7G regulatory genes on the biological behavior of CRC has not been clearly elucidated.

Pyruvate kinase (PK) is a key gene involved in anaerobic cell oxidation and an important marker of tumor cell glycolysis; it catalyzes the conversion of phosphoenolpyruvate (PEP) to pyruvate [13]. Pyruvate kinase M2 (PKM2), a protein encoded by PKM, endows cells with a specific glucose metabolic phenotype. PKM2 increases the expression of tumor cells and increases the production of ATP to promote glycolysis, resulting in the accumulation of the intermediate metabolite lactate in tumor cells and metabolic reprogramming, thus providing favorable conditions for the proliferation of tumor cells [14]. In addition, PKM2 can also function as a transcription factor to promote cancer through nonmetabolic pathways. Studies have shown that in tumor cells, due to the increased abundance of protein dimers, PKM2 enhances nuclear translocation, participates in the formation of transcription complexes, regulates the expression of downstream oncogenic genes, and promotes the occurrence and development of tumors [15, 16].

Cancer cells often undergo glycolysis for energy, known as the Warburg effect, and lactate is an important product of this process [17]. Lactate can enter the nucleus and be modified into the lysine lactate (Kla) site on histones, thereby altering chromatin structure and affecting gene function, an epigenetic modification known as histone lactation [18]. This modification, which usually occurs in active gene promoter regions, promotes the binding of transcription factors to DNA and unwinds DNA, thereby increasing gene transcription [19]. Recent studies have shown that histone lactate modification plays a key role in regulating biological processes such as tumorigenesis and tumor cell reprogramming [20].

In our study, we revealed that METTL1 mediates m7G modification of PKM mRNA, enhances PKM2 protein expression and promotes glycolysis and lactate accumulation. Lactate increases the expression of METTL1 in CRC cells through the lactation of histone H3K9, forming a positive feedback loop. METTL1-m7G mediates PKM to activate the expression of the immune checkpoint molecule CD155. The regulatory network involving METTL1/PKM2/H3K9la in RNA metabolism and the tumor immune microenvironment could pave the way for the development of targeted immune therapies for CRC.

Materials and methods

Patient and clinical samples

Paraffin sections of tumor and adjacent normal tissue samples from 52 colorectal cancer patients (surgeries were carried out from 2014–2019) for IHC staining were obtained with informed consent under a protocol approved by the Department of Surgical Oncology, Affiliated Hospital of Jiangnan University (AHJNU). We complied with all relevant ethical regulations. Information about the sex and tumor characteristics of the patients is provided in Table S1.

Cell culture and transfection

Homo sapiens (METTL1 and sh-METTL1) and Mus musculus (Mettl1 and sh-Mettl1) were selected and inserted into vectors to ensure stable overexpression and knockdown, respectively, while an empty plasmid was used as a control. siMETTL1 and sgMettl1 were purchased from GeneChem (Shanghai, China). siRNAs and plasmids were transfected via a Lipofectamine 3000 kit (Invitrogen) according to the manufacturer’s instructions. All sequences are listed in Supplementary Table S2. HCT-8 cells (ATCC, CCL-244), HCT-116 (ATCC, CCL-221), RKO (ATCC, CCL-2577), MC-38 (EK-bioscience, CC-Y2123), 293 T (ATCC, CRL-11268) and CT-26 cells were nurtured in Dulbecco’s modified Eagle’s medium (DMEM), DLD-1 (ATCC, CCL-247) and Caco2 cells were cultured in RPMI 1640 medium. All cells were cultured with 10% fetal bovine serum (FBS; Clark, FB15015) and 100 U/ml penicillin-100 μg/ml streptomycin. The last STR test of the main cell lines involved in this design was commissioned by BIOWING (Shanghai, China) in May 2023.

Immunohistochemistry (IHC) and immunofluorescence (IF)

Paraffin-embedded CRC samples were cut into 0.5-micron-thick slides. To extract the antigen, the sample was soaked in 0.01 M citric acid buffer (pH 6.0) in a pressure cooker for 20 min. For the IF assay, the sample was fixed with 4% paraformaldehyde at room temperature for 20 min and then incubated in an ice bath with 0.05% Triton X-100 in phosphate buffer (PBS) for 5 min. The samples were subsequently incubated with 2% bovine serum albumin (BSA) in PBS for 1 h at room temperature and then with METTL1, PKM2, H3K9la and CD155 antibodies overnight at 4 °C. After that, antibodies with a fluorescent dye or secondary antibodies bound to horseradish peroxidase (HRP) were added to the sample at room temperature for 1 h. Multiple IF stains were performed via multiple fluorescence immunohistochemical stains (abs50013, ABSIN). In brief, the detected cells or tissues were incubated with METTL1, PKM2, H3K9la, and CD155 antibodies and rotated at room temperature for 1 h. The samples were then washed and incubated with the corresponding secondary antibodies one by one. Nuclei were reverted with DAPI, and images were captured via laser scanning confocal microscopy (FV1200, Olympus) or fluorescence microscopy (IX73, Olympus). The fluorescence signal intensity and subcellular distribution were analyzed with ImageJ software (Fiji).

According to previous studies, the IHC scores of METTL1, PKM2, H3K9la and CD155 staining were based on the coverage of tumor areas and the percentage of positive protein staining [21]. Immunoassays were performed with 3,3′-diaminobenzidine tetrachloride hydrochloride (DAB), and hematoxylin was used for reverse nuclear staining. The staining scores of METTL1, PKM2, H3K9la and CD155 were determined according to the intensity and proportion of positive cells in 3 fields randomly selected under the × 40 objective lens. The tumor cell: Positive staining ratios were as follows: 0 (unstained cells), 1 (< 25%), 2 (26–50%), 3 (51–75%), and 4 (> 76%). The dyeing intensity was recorded as 0 (undyed), 1 (light dyed), 2 (medium dyed), or 3 (deep dyed). The formula for calculating the staining index (SI) was as follows: SI = staining intensity × proportion of positive cells. The protein expression signal was recorded with a digital biopsy scanner (OptraScan). The list of antibodies used is shown in Supplementary Table S3.

CCK-8 assays

Hct116 and RKO cells were carefully selected to facilitate the knockdown of METTL1. The cells were carefully placed in 96-well plates and cultivated for 24, 48, 72, or 96 h. At specific intervals, 10 µL of CCK-8 reagent (Vazyme) was carefully added to each well of a 96-well plate, followed by an incubation period of 2 h. The absorbance was measured with a microplate reader (Biotek) at 450 nm.

RNA immunoprecipitation (RIP) and methylated RIP (MeRIP)

Control and METTL1 knockout HCT-116 cells were treated with sgRNA, collected, and washed twice with precooled PBS. The cells were lysed in cold multibody buffer (3 M KCl, 1 M MgCl2, 1 M HEPES–NaOH, 5% NP-40, 0.5 mM dithiothreitol (DTT)). Cleavage was performed with 40 U/μl RNase OUT and 200 units/mL EDTA-free protein inhibitor cocktail for 30 min, the mixture was centrifuged at 4 °C and 13000 rpm, and the supernatant was collected. The supernatant was incubated overnight with 5 μg of anti-METTL1 antibody or an equal amount of IgG and magnetic beads at 4 °C. The RNA‒antibody complex was subsequently resuspended in 150 μl of protease k buffer and washed five times in NT-2. Finally, the mRNA was extracted, and qRT-PCR was performed. m7G-MeRIP was carried out in accordance with a previously published protocol. mRNA was purified from total RNA via the Hieff NGS mRNA separation kit (Yeasen, 12603ES24) [22]. Then, 5 μg of anti-m7G antibody or IgG was mixed with the magnetic beads and incubated at room temperature for 30 min. The purified RNA was mixed with the antibody and magnetic beads and incubated at 4 °C overnight. The complex containing the antibody and mRNA was resuspended in 150 μl of protease k buffer and washed 5 times with NT-2. Finally, RNA was extracted, and qRT‒PCR was performed.

Histone extraction

Approximately 10⁷ cells were collected and resuspended in 200 μl of prelysis buffer and incubated on ice for 30 min. The mixture was subsequently centrifuged at 12,000 rpm at 4 °C for 5 min, after which the supernatant portion (containing acid-soluble protein) was transferred to a new vial. A 0.3 volume of balance-DTT buffer was immediately added to the supernatant. OD readings were used to quantify protein concentrations. BSA can be used as a standard. In this study, the nuclear reference protein used was Histone 3 (ProteeinTech, 68345–1-Ig).

RNA pull-down

PKM RNA was first transcribed in vitro via an SP6/T7 biotin labeling kit (R7061, Beyotime). The amplified RNA was labeled with a Pierce RNA 3′-terminated thiobiotinization kit (Thermo Scientific). The obtained biotin-labeled RNA was subsequently incubated with the cell lysate at 4 °C overnight. Subsequently, 30 µL of streptavidin magnetic beads (Thermo Fisher) was added to the mixture, which was subsequently incubated at room temperature for 2 h. The beads were then washed at room temperature for 2 h with wash buffer with a total volume of 200 μL (D3308, Beyotime) to obtain a protein suspension, which was used for Western blot analysis.

Dot blot

Total RNA was extracted via TRIzol reagent, and mRNA was purified via the Dynabeads mRNA Purification Kit (Invitrogen). The concentration and purity of the mRNA were determined via a NanoPhotometer. The mRNA was subsequently denatured by heating at 95 °C for 5 min, followed by rapid cooling on ice. Next, 200–400 ng of mRNA was added dropwise in an orderly fashion to the Amersham Hybond-N + membrane (RPN303B, Cytiva) and delicately air-dried for 5 min. To establish covalent bonding, the membrane was subjected to a 254 nm UV crosslinking process in an ultraviolet crosslinker (254 nm UV for 5 min); subsequently, it was blocked with 5% skim milk in TBST and incubated overnight at 4 °C with the anti-m7G antibody (Genelily, Gab202202). For further immunodetection, an HRP-conjugated anti-mouse IgG secondary antibody was introduced onto the membrane, which was subsequently incubated at room temperature for 1 h. Accurate signal visualization was achieved via chemiluminescence. Methylene blue (0.02%; Sigma‒Aldrich) staining was performed to ensure that RNA content was evenly distributed throughout the membrane.

Glycolysis extracellular acidification assay

The effects of METTL1 on glycolysis were assessed with a fluorescent glycolysis assay (ab197244, Abcam). HCT-116 (2 × 10⁴ cells) (with or without knockdown) and RKO (2 × 10⁴ cells) (with or without overexpression) cells were seeded in black-walled black bottom 96-well plates. Adherent cells were subjected to overnight CO2 purge, and some wells were treated with 2DG (5 mM) for 48 h. The cells were incubated in glycolysis assay buffer at 37 °C, and fluorescence at Ex/Em = 380/615 nm was recorded at 30 s intervals for 1 h via a microplate reader (Fluoroskan FL, Thermo). The glycolytic activity was normalized to the number of cells.

Western blotting

Protein extracts were obtained from the cells used in this study via an NP40 lysis kit (P0013F, Beyotime). Quantification was performed via a BCA protein assay kit (P0012, Beyotime), followed by electrophoresis on an SDS‒PAGE gel. Protein samples were then transferred to nitrocellulose filter membranes (Cytiva) in equal proportions, blocked with 5% skim milk (Sangon) in TBST, and incubated with primary antibodies overnight at 4 °C. The membranes were then incubated with HRP-conjugated secondary antibodies, developed with chemiluminescent reagents, and visualized via a Molecular Imaging System (Tanon).

Single-cell RNA (scRNA) sequencing and spatial transcriptome (ST) analysis

The datasets for spatial scRNA sequencing of CRC and paired normal tissues were obtained from BioSample (https://www.ncbi.nlm.nih.gov/biosample/) (SAMN34566856 and SAMN34566857). To capture gene expression information on ST slides, the Visium Spatial platform using 10 × Genomics was used (Homo sapiens: GRch38) [23]. The datasets for ST of CRC samples were obtained from the BioSample (https://www.ncbi.nlm.nih.gov/biosample/) SAMN33698432. To capture gene expression information on ST slides, the Visium Spatial platform using 10 × Genomics was used [24]. This involved the use of spatially barcoded RNA oligonucleotides according to the default protocol. The quality of the raw sequencing reads from the STs was assessed, and the reads were aligned via Space Ranger (version 2.0.0). The resulting gene‒spot matrices obtained from processing the ST-Visium and scRNA samples were analyzed via the Seurat package (version 4.3.0) in R. Loci with a minimum threshold of 200 genes were selectively selected, whereas genes with fewer than eight read counts or fewer than four loci were ignored. Point alignment was achieved by implementing normalized LogVMR functions. For dimensionality reduction and clustering, principal component analysis (PCA) was employed with a resolution of 2.0, resulting in the identification of distinct clusters, which was performed on the basis of the scRNA-seq or ST signature via the default parameters in Seurat. The SpatialFeaturePlot function in Seurat was used to depict spatial feature expression.

RNA stability assay

To assess RNA stability, control and sgMETTL1 cells were treated with actinomycin D (2 μg/mL, RASP-101, Selleck) for 6 h. As described previously, total RNA was extracted via an RNA extraction kit and analyzed via PCR-gel electrophoresis. The results were normalized to the measurements obtained at 0 h.

Tumor xenograft

Six-week-old male BALB/c nude mice and C57BL/6 mice were purchased from SLAC Laboratory Animal Center Shanghai (Shanghai, China). METTL1 wild-type (WT) and METTL1 knockout (KO, heterozygous ±) C57BL/6 mice were purchased from Caygen Biology (S-KO-03182), and genes were identified via PCR via mouse tail DNA before experimental inclusion. All the above experimental animals were randomly divided into two groups with 5 animals in each group. HCT-116 (stable shNC cells or stable shMETTL1 cells), RKO (stable NC cells or stable METTL1 cells), and MC-38 (stable NC cells or stable METTL1 cells) cells were used in this study. A total of 1 × 10⁶ cells diluted in 100 µL of phosphate-buffered saline were injected into each mouse. For the METTL1 inhibitor in vivo combination treatment experiment, tumor-bearing mice were inoculated with tumors for one week, METTL1-WDR4-IN-1 (HY-162080, MedChemExpress) 2 mg/kg and a PKM2 inhibitor (Shikonin, HY-N0822, MedchemExpress) were intraperitoneally injected into the mice (10 mg/kg), or a DMSO control (60313ES60, Yeasen) was used. The mice (5 per group) were maintained until death, or the tumor reached a maximum subcutaneous area of 2000 mm³. The tumor volume was assessed weekly over a span of 4 weeks, after which the mice were euthanized, and the tumor weights were measured. The investigators were not blinded to the group allocation. The mouse experiments were conducted in accordance with the guidelines and regulations of the Jiangnan University Animal Ethics Committee (ethical approval was provided by JN. No. 20231030m0960420).

Flow cytometry

To prepare a single-cell suspension, the collected tumors were digested with collagenase IV at a concentration of 1 mg/ml, hyaluronidase at a concentration of 0.2 mg/mL, and DNase I at a concentration of 0.15 mg/ml. Cell digestion, isolation and selection were performed according to previous methods [22]. In brief, the mixture was filtered through a cell strainer with a pore size of 70 mm. The cells were then washed with cold PBS. The red blood cells were subsequently lysed via the MultiScience product, whereas the dead cells were labeled via the Invitrogen Fixable Viability Dye Kit. The cells were washed again with a staining buffer composed of PBS supplemented with 2% FBS and 1 mM EDTA. The cell suspension was then resuspended in the same staining buffer supplemented with a CD16/CD32 blocking antibody from MultiScience and incubated for 20 min. Subsequently, specific antibodies targeting membrane molecules at appropriate dilutions were added to the suspension and incubated for 30 min on ice in the absence of light. Finally, intracellular staining was conducted via a FoxP3/Transcription Factor Staining Buffer Kit (IC001, MultiScience).

Distinct immune cell populations were identified and gated in this study: EpCAM + cell populations (Invitrogen), CD45 + cell populations, CD4 + T cells (CD3 + CD4 +), CD8 + T cells (CD3 + CD8 +), NK (natural killer) cells (CD3- NK1.1 +) and activated NK cells (NK1.1 + CD107a +). The addition of antibodies was then performed, followed by an overnight incubation at 4 °C. To ensure precise compensation adjustments, we used either single-stained samples or UltraComp eBeads Compensation beads from Invitrogen. Flow cytometry data were acquired via a Beckman Coulter DxFlex or multicolor BD Fortessa flow cytometer, and subsequent analysis was performed via FlowJo software (V10).

Chromatin cleavage under targets and release via nuclease (Cut&Run)-qPCR

Chromatin coprecipitation analysis was conducted using an anti-PKM2 antibody (15822–1-AP, ProteinTech), following the guidelines stipulated in the Cut&Run Assay Kit (HD101, Vazyme), as provided by the manufacturer. Fold enrichment was determined via qRT‒PCR and expressed as a proportion of the input chromatin (percentage of input). Detailed information regarding the primers used is provided in Table S3.

Chromatin immunoprecipitation (ChIP) sequencing analysis

We downloaded fastq files specifically chosen for PKM2 Cut&tag (PRJNA1000492) from SRA (https://www.ncbi.nlm.nih.gov/sra/) [25]. Bowtie2 was used to align the reads to the reference genome. The aligned reads from both the IP and input libraries were utilized for MACS2 peak calling, an approach that identifies PKM2 peaks on chromatin in either bed or bigwig format, allowing smooth visualization via IGV software. To discover both de novo and known motifs, MEME was applied, followed by localization of the identified motif in relation to the summit of each peak. The called peaks were annotated by intersecting them with the gene architecture via the R package ChIPseeker.

Histone Kla appraisal and testing

The isolated and collected histone proteins or peptides were sent to PTM Bio (Hangzhou, China) for HPLC‒MS detection of Kla modifications. The lactation of lysine, the oxidation of methionine and the acetylation of the N-terminus of the peptide were fixed as variable modifications, and the mass tolerance of MS was set to ± 0.02 Da. After the protein expression levels of the samples were quantified, we normalized the ratio of all the quantitative Kla peptides to the corresponding protein levels. The pan-Kla antibody (PTM-1401RM) and the identified H3K9la-specific antibody (PTM-1419RM) were used for the ChIP and WB assays.

Polysome profiling

HCT-116 shMETTL1 and stable NC cells were cultured in a 10 cm dish and subsequently subjected to polysome profiling. Before the experiment began, CHX (C4859, Sigma‒Aldrich) was added to the medium at a concentration of 100 mg/mL, and the cells were incubated with CHX for 7 min to block active mRNA translation. The medium was removed, and the samples were rinsed with precooled 4 °C phosphate buffer (PBS), which also contained 100 mg/mL CHX. The cells were subsequently centrifuged at 500 × g for 5 min, suspended in 500 mL of cleavage buffer (containing Triton X-100, 1% protease and 40 U/mL RNase) and placed on ice for 30 min. The lysate was then centrifuged at 15000 × g at 4 °C for 15 min. After that, 400 mL of the supernatant was taken and distributed into a sucrose density gradient ranging from 10 to 50% (A502792, Sanger Bio). A Beckman Coulter L-100XP ultracentrifuge was used, and the mixture was centrifuged at 45,000 rpm at 4 °C for 2 h. Finally, the centrifuged sample was divided into 20 components via a fraction collector (LCMS-2050, Shimadzu). The absorption value of each component at the A260 nm wavelength was recorded.

Bioinformatics analysis

The RNA-seq transcriptome data and corresponding clinical information of the cancers were downloaded and analyzed from the TCGA cohort via the UCSC Xena database (https://xenabrowser.net/datapages/). The identification of infiltrating immune cells in the TCGA cohort of CRC patients via CIBERSORT software was performed locally via R software 4.0.1. The RNA-seq data (RPKM format) of the TCGA-COAD and TCGA-READ cohorts were analyzed to obtain the abundance ratio matrix of CIBERSORT 22 immune cells. Correlation analysis was subsequently conducted on the contents of the NK cell type score in the CRC samples. Next, Kaplan–Meier analysis for overall survival was performed on the basis of the cutoff value of the NK cell score. We also used the Tumor Immune Estimation Resource (TIMER2.0) database (http://timer.cistrome.org/) to analyze the correlation of ovarian cancer-infiltrating NK cells with gene expression. The sequencing data generated in this study have been deposited in the SRA repository under accession codes SUB14773935 (for MeRIP-seq), PRJNA902758 (for QKI RIP-seq) [12], and SUB14773857 (for RNA-seq). All study data are included in the article and/or supporting information.

Statistical analysis

Pearson correlation analysis was conducted to evaluate the correlation between two variables, while Student’s t test was performed to compare two variable groups. The data obtained from the Western blot, IF and molecular expression analyses and other quantitative values were compared via ordinary one-way analysis of variance (ANOVA), with multiple comparisons included when necessary. The protein expression results among the different groups are presented in quantification bar graphs, detailing the means accompanied by their respective standard deviations. The hazard ratio (HR) was determined via the Cox proportional hazards model, with a reported confidence interval of 95%. Additionally, Kaplan‒Meier survival curves were generated. The associations between METTL1, PKM2, and H3K9la and clinicopathological features were examined via Pearson’s χ2 test or Fisher’s exact test. Comparisons of mouse models among different groups were performed via t tests and one-way ANOVA, with multiple comparisons included when necessary. The results are expressed as the means ± SDs, and statistical significance was set at p < 0.05. Statistical analysis was performed via SPSS (v22) and the R package. Significance was determined at a two-sided p value < 0.05.

Results

High METTL1 expression is associated with tumor progression in patients with CRC

Tumorigenesis and progression are closely related to abnormal RNA modification levels in tumor cells. Therefore, we examined the expression levels of the main RNA modification markers (m1A, m5C, m6A and m7G) in colorectal cancer (CRC) tissues, with a special focus on the expression level of m7G in CRC (Fig. 1A). m7G levels are regulated mainly by the RNA modification writer METTL1. Accordingly, we explored the expression level of METTL1 in CRC tissues and found that METTL1 expression was significantly higher in CRC tissues than in normal tissues adjacent to cancer (Fig. 1B and Fig. S1A). Studies performed on cohorts of CRC patients at AHJNU and TCGA showed that METTL1 expression was associated with poor patient prognosis in both single and multifactorial analyses (Fig. 1C, D and Fig. S1B-C). Further survival analyses revealed that individuals with high levels of METTL1 expression presented worse survival outcomes than patients with lower METTL1 expression levels did (Fig. 1E, Fig. S1D and Fig. S1E).

Fig. 1.

METTL1 expression affects CRC progression. A The expression levels of major RNA methylation markers (m1A, m5C, m6A, and m7G) in CRC and normal tissues were analyzed via IHC; scale bar, 60 μm. B The expression level of METTL1 in 52 pairs of CRC and paired normal tissues was analyzed via IHC, and the difference in METTL1 protein expression between CRC and normal tissues was analyzed (P = 0.0002). Scale bar, 60 μm. C and D Univariate and multivariate Cox regression analyses were performed to analyze the relationship between METTL1 expression and TNM stage in the tumor tissue of the AHJNU CRC patient cohort. E Kaplan‒Meier survival analysis of CRC patients in the high- and low-METTL1 expression groups, P = 0.0102. F The baseline level of METTL1 expression in different types of CRC cells was detected by Western blotting. G and H qPCR and Western blotting were used to detect METTL1 knockdown in HCT-116 cells (P < 0.001, P < 0.001, P = 0.038) and METTL1 overexpression in RKO cells (P = 0.0091). The results are presented as the means ± SDs. *P < 0.05; **P < 0.01; ***P < 0.001 (Student’s t test; log-rank test). CRC, colorectal cancer; AHJNU, Affiliated Hospital of Jiangnan University

In the human CRC cell line HCT-116, we observed a high level of METTL1 expression, whereas low METTL1 expression was detected in the RKO cell line (Fig. 1F). Therefore, in this study, we constructed a stably transfected HCT-116 cell line for METTL1 knockdown experiments and overexpressed METTL1 in the RKO cell line for subsequent functional studies (Fig. 1G and Fig. 1H). In addition, we constructed a mouse CRC cell line, MC-38, for stable transfection and knockdown (Fig. S1F).

METTL1 expression levels regulate CRC progression and immune evasion

To investigate the role of METTL1 in CRC development and the immune microenvironment, we performed CCK8 and ECAR experiments. METTL1 promoted the proliferation of CRC cells as well as the rate of extracellular acidification, and these effects were reversed when METTL1 expression was inhibited (Fig. 2A and B). Notably, in the TCGA and AHJNU cohort studies, METTL1 and hypoxia markers were coexpressed in the tumor tissues of CRC patients (Fig. S1E and Fig. S2A). At the in vivo level, METTL1 overexpression promoted tumor growth, while tumor size and volume were suppressed when METTL1 expression was knocked down (Fig. 2C and D). Similar results were observed in the same transplantation model, where upregulation of Mettl1 increased the tumor-forming capacity of MC-38 tumor cells, whereas downregulation of Mettl1 reduced the tumor-forming capacity of MC-38 tumor cells in C57BL/6 mice (Fig. 2E and F). To investigate whether METTL1 interferes with the immune response in CRC, we performed flow cytometry analysis of tumor tissues. The results showed that METTL1 reduced the degree of CD8 + T cell infiltration in tumor tissues (Fig. S2B). Moreover, METTL1 effectively inhibited the activation of NK cells (Fig. 2G). IHC analysis of tissue sections revealed that METTL1 overexpression reduced the level of CD16 + cell infiltration (Fig. 2H). On the basis of these findings, we hypothesized that METTL1 not only plays a role in tumor growth but also exacerbates CRC progression by promoting immune evasion. Single-cell sequencing analysis was performed on patient CRC samples. We observed that METTL1 was predominantly expressed in tumor cells (epithelial cells), whereas its expression in CD8 + and NK cells was relatively low (Fig. 2Iand J). Interestingly, we observed increased glycolytic-related activity in tumor tissue (Fig. S2C-S2E). Indeed, upregulated expression of glycolytic signature molecules (LDHA, HK2, SLC2A1, and HIF1A) was associated with decreased expression of NK and T-cell effectors (NCAM1, FCGR3A, PRF1, GZMA, and GZMB) (Fig. S2F). On the basis of these findings, we further deduced that the role of METTL1 in CRC cells may be related to immunosuppression rather than METTL1 in NK or CD8 + T cells.

Fig. 2.

METTL1 regulates CRC progression and immune escape in vitro and in vivo. A The effect of METTL1 on the proliferation ability of CRC cells was detected via a CCK8 assay (P < 0.001, P = 0.029). B An ECAR assay was used to detect the effect of METTL1 on the extracellular acidification capacity of CRC cells (P < 0.001, P < 0.001). C and D Effects of stable knockdown of METTL1 on the tumorigenesis of HCT116 cells (P = 0.0072, P = 0.0063) and overexpression of METTL1 on the tumorigenesis of RKO cells (P = 0.025, P = 0.031). E Stable knockdown and upregulation of Mettl1 in MC-38 cells were detected by qPCR and Western blotting (P < 0.001, P < 0.001). F The effect of Mettl1 overexpression on the tumorigenicity of MC-38 mouse CRC cells was detected in vivo (P = 0.026). G Flow cytometry was used to detect the effect of Mettl1 on the proportion of NK1.1 + CD107a + NK cells among lymphocytes infiltrating tumor tissue (P = 0.0057). H IHC was used to determine the ratio of Mettl1 expression to activated NK cell marker (CD16) expression in infiltrating tumor tissue (P = 0.041). Scale bar, 60 μm. I TSNE plot showing the clusters of scRNA-seq data in CRC and normal samples. Each dot is a cell colored according to its analyzed cell type. J Violin plot showing the expression levels of METTL1 in various cell types in CRC tissue (P = 0.033). The results are presented as the means ± SDs. *P < 0.05; **P < 0.01; ***P < 0.001. ECAR: Extracellular acidification rate; TSNE: t-distributed stochastic neighbor embedding

METTL1 was shown to modify PKM via m7G and upregulate PKM2 expression

To elucidate the process by which METTL1 promotes glycolysis and its effect on tumor immune evasion, we constructed the HCT-116 sgMETTL1 cell line and performed MeRIP sequencing experiments with m7G, which revealed that at the global RNA level, the peaks associated with m7G modification were abundantly clustered in the internal sequences of RNAs (Fig. 3A and Fig. S3A-S3B). Given that METTL1 promotes glycolysis in CRC, we further screened and investigated this mechanism and revealed that METTL1 is involved in m7G modifications inside PKM, the RNA molecule of PKM2, and that the peaks of these modifications are consistent with the m7G modification element “GANGAN (N = A/U/G)” sequence feature (Fig. 3B). Moreover, a RIP‒qPCR assay confirmed that the METTL1 protein could bind to the RNA sequence of PKM (Fig. 3C).

Fig. 3.

METTL1 enhances PKM2 expression via m7G modification. A HCT-116 cells with METTL1 knockout were constructed, m7G MeRIP sequencing was performed, and METTL1-m7G-modified RNA was analyzed on the basis of the metagene distribution profile. B Integrative genomics viewer (IGV) analysis revealed that the m7G peaks of the RIP-seq data were distributed in the PKM2 transcripts. C m7G-modified molecular formula and qPCR were used to analyze the associations of PKM2 mRNA levels with m7G (P = 0.013) and METTL1 (P = 0.002) binding. D Changes in PKM2 expression in HCT-116 cells with stable METTL1 knockdown and in RKO cells with METTL1 overexpression. E PKM2 protein levels in colorectal cancer cells treated with 25 mM MG132 were detected by Western blotting. F The level of PKM2 mRNA bound to the METTL1 protein was detected by RNA pull-down. G IP detection revealed that METTL1/PKM2 affects the binding of QKI and G3BP1. H Polysome profiling was used to measure the level of mRNA binding to polyribosomes in HCT116 cells after METTL1 knockdown. I The effects of METTL1 on the expression levels of m7G and PKM2 were detected by Western blotting and dot blotting. *P < 0.05; **P < 0.01

To investigate whether METTL1-m7G has a role in PKM2 protein expression, we first analyzed the binding of RNA by QKI, a known reading protein of m7G that is able to participate in the methylation process of PKM (Fig. S3C). By examining the effects of METTL1-m7G and QKI recognition on PKM2 protein expression, we found a significant increase in PKM2 expression in PKM-encoded proteins (Fig. 3D). By treating cells with the proteasome inhibitor MG132, we further confirmed that METTL1-m7G was able to upregulate PKM2 expression in cells (Fig. 3E). To determine whether QKI, the m7G Reader of RNA, affects the translation of PKM2, we used an RNA pull-down assay (Fig. 3F) and found that QKI could bind to PKM. To explore whether MeTTL1-M7g acts on other glycolysis-related molecules to increase PKM2 expression, we conducted RNA pull-down detection and found that PKM more significantly binds the METTL1 and QKI proteins. These findings suggest that the regulation of PKM by METTL1-m7G increases the expression of PKM2 (Fig S2G). Stress granule-associated G3BP1 did not bind to QKI in HCT-116 cells, with no significant difference in binding level (Fig. 3G). Polysome profiling further verified that there was no increase in translation in the METTL1-knockdown group (Fig. 3H). In the absence of a substantial change in translation, METTL1-m7G stabilized PKM RNA and interfered with its degradation (Fig S3D). In turn, METTL1 was detected at the protein level to increase the cellular m7G level and promote the upregulation of the PKM2 protein (Fig. 3I).

METTL1 regulates H3K9la through PKM2 to form a positive feedback loop

As demonstrated in our study, METTL1 was able to increase the expression level of PKM2, which promotes cellular glycolysis, and in combination with our previous findings, we found that METTL1 could regulate lactate accumulation in RKO cells via PKM2 by detecting the lactate content in culture supernatants (Fig. 4A). Further protein assays revealed that METTL1 could regulate the expression of the hypoxia marker molecules GLUT1, HK2, HIF-1α and LDHA in CRC cells via PKM2, which implied that METTL1/PKM2 could promote glycolysis (Fig. 4B).

Fig. 4.

Interaction of the METTL1/PKM2/H3K9la positive feedback loop. A METTL1 regulates PKM and promotes lactic acid accumulation in CRC cells. B METTL1 regulates PKM and affects the expression of glycolysis marker proteins in tumor cells. C Western blot analysis of protein lactate expression in cancer cells treated with 25 mM L-lactic acid. D Histone extraction and lactation modification abundance and site detection. E Western blot analysis of the effects of METTL1 on the regulation of total lactated histone and H3K9-lactated expression of PKM. F Western blot analysis was performed to detect the expression of METTL1 and H3K9la after treatment with 500 nm rotenone or 25 mm L-lactic acid for 24 h. G A Cut&Tag sequencing dataset (PRJNA885248) was used to analyze the peak of H3K9la immunoprecipitated from METTL1 chromatin at lactate. H Chromatin coimmunoprecipitation PCR and qPCR were used to analyze the promoter region enriched in METTL1 by H3K9la. Lactation of H3K9 significantly upregulated the expression of METTL1. I Schematic diagram of the METTL1/PKM/H3K9la positive cycle regulatory mechanism. * P < 0.05; ** P < 0.01; *** P < 0.001

Indeed, according to existing studies, cellular lactate accumulation can modify histones by lactonization, which in turn can produce biological effects. On the basis of these findings, we found that the level of panlysine lactate (Pan-Kla) was elevated in a time-dependent manner by the addition of lactate, whereas the inhibition of METTL1 decreased the expression level of Pan-Kla (Fig. 4C). Considering the potential biological effects of histone lactylation, we extracted histones from HCT-116 cells and found that the level of lactylated lysine at position 9 of histone H3 (H3K9la) was significantly elevated (Fig. 4D and Fig. S2H). Further protein assays revealed that METTL1 regulation of H3K9la occurred through PKM2 (Fig. 4E). In addition, we found that the expression of PKM2, Pan-Kla and H3K9la could be promoted by lactate and the hypoxia inducer rotenone, and interestingly, METTL1 expression also increased with lactate and hypoxia (Fig. 4F). Considering that H3K9la may promote gene transcription by facilitating chromatin dehelicalization, we analyzed the publicly available Cut&Tag dataset of H3K9la, and the results revealed that H3K9la could bind to the chromatin region where METTL1 is located in response to the promotion of lactic acid; thus, we hypothesized that H3K9la activates METTL1 transcription (Fig. 4G). We purified the H3K9la protein, conducted a ChIP‒qPCR experiment with H3K9la, and found that H3K9la was significantly bound to the promoter region of METTL1 (Fig. 4H). Taken together, these experimental results suggest that METTL1-m7G regulates PKM2 expression and promotes histone lactation and H3K9la through glycolysis and lactate accumulation and that H3K9la subsequently activates METTL1 transcription, forming a positive feedback loop of METTL1/PKM2/H3K9la, which ultimately promotes the growth of CRC (Fig. 4I).

Targeting METTL1 and PKM2 effectively inhibits CRC growth

To investigate the possible benefits of METTL1/PKM2 in the treatment of CRC via the modulation of CD155, we first constructed a mouse model of Mettl1 gene knockout and performed genetic testing (Fig. 5A). Since both Mettl1 and Lyz-Cre are located on chromosome 10, the genotype of the knockout mouse is heterozygous ( ±). We conducted in vivo therapeutic experiments on Mettl1-KO and Mettl1-WT mice intrabitoneally injected with Shikonin (Fig. 5B).

Fig. 5.

Targeting METTL1 and PKM2 inhibits CRC growth. A C57BL/6 knockout mice (±) with Mettl1 knockout were constructed, and wild-type C57BL/6 knockout mice were used as controls. The genotypes of the mice included in the study were identified via PCR‒PAGE. B Mettl1 combined with Shikonin (a PKM2 inhibitor) in the treatment of a CRC model. Tumor growth in C57BL/6 mice with subcutaneous Mettl1-KO or Mettl1-WT tumors; n = 5 per group. Shikonin or DMSO was intraperitoneally injected as a control. C Tumor tissue of C57BL/6 mice (with or without Mettl1 KO) treated with shikonin; n = 5 per group. D and E Weight and volume analysis of the tumor tissue in each group. F Tumor tissue from C57BL/6 mice treated with METTL1-WDR4-IN (2 mg/kg) or/or shikonin (10 mg/kg), saline or DMSO was used as a control; n = 5 per group. D and E Weight and volume analysis of the tumor tissue in each group. The results are expressed as the mean ± SD. N.S., p≥ 0.05; *P < 0.05; **P < 0.01. (Student’s t test; the survival rate was relatively high. ANOVA or Fisher precision test). KO: knockout, WT: wild type

The experimental results revealed that the tumor mass and volume in the Mettl1-KO mice were significantly lower than those in the Mettl1-WT mice (Fig. 5C). In Mettl1 WT mice, there was no statistically significant difference in the inhibitory effect of PKM2 inhibitors on tumor growth. However, when Mettl1 was knocked out, Shikonin effectively inhibited tumor mass and volume (Fig. 5D and E). We explored the efficacy of METTL1 and PKM2 as potential therapeutic targets for CRC. We used METTL1-WDR4-IN-1 combined with Shikonin for in vivo treatment of CRC. Both METTL1-WDR4-IN-1 and Shikonin inhibited tumor formation in C57BL/6 mice (Fig. 5F). However, the combination of METTL1-WDR4-IN-1 and Shikonin had a more significant anticancer effect, and the combined effect of METTL1-WDR4-IN-1 and Shikonin significantly inhibited the weight and volume of the tumors (Fig. 5G and H). METTL1 was found to promote CRC development in a PKM2-dependent manner.

METTL1 promotes checkpoint CD155 expression via PKM2 nuclear translocation in CRC cells

Recent studies have shown that PKM2, as a cotranscription factor, can promote the development of cancer through nonmetabolic pathways [26]. In cancer cells, the content of the PKM2 protein dimer increases, which enhances its nuclear localization ability, thus participating in the formation of transcription complexes, regulating the expression of tumor-promoting genes, and thus promoting tumor formation and growth [16]. We further confirmed that METTL1 has a regulatory effect on PKM2 protein expression and nuclear translocation through karyoplasmic separation experiments; that is, PKM2 expression and nuclear transfer can be significantly reduced by knocking out METTL1 and PKM2 (Fig. 6A). The results of the PKM2 cross-linking experiment revealed that in colon cancer cells with METTL1 knockout, the expression of the PKM2 protein tetramer was decreased, and the expression of the dimer was significantly reduced compared with that in the control group. After upregulation of METTL1, the protein expression of the PKM2 dimer was significantly increased (Fig. 6B). In addition, an immunofluorescence assay revealed that METTL1 promoted the expression of PKM2 in the nucleus (Fig. 6C).

Fig. 6.

METTL1 promotes CD155 expression through PKM2 nuclear translocation. A Subcellular localization of PKM2 in METTL1-knockdown CRC cells and Western blot analysis of whole-cell lysates (WCLs), cytoplasmic extracts and nuclei. B METTL1-knockdown and METTL1-overexpressing HCT116/RKO cells were first crosslinked with glutaraldehyde, and then Western blotting was performed to analyze the dimeric and tetrameric expression levels of PKM2. C The effects of METTL1 on PKM2 protein signaling and nuclear localization in RKO cells were detected by IF. D Transcriptome and qRT‒PCR analyses of the effects of METTL1 upregulation on immune checkpoint molecule and PVR (CD155) expression levels. E The peak value of chromatin binding in PVR was analyzed via ChIP‒qPCR sequencing of PKM2 in the database PRJNA800819, and the promoter region of PKM2-enriched PVR was analyzed via ChIP‒qPCR. F Analysis of the distribution of cell types and PVR expression in CRC tissues in the spatial transcriptomic sequencing dataset (SAMN33698432). G IHC was used to detect the expression and localization of METTL1, CD16 and CD155 in the tumorigenic tissues of mice with high or low METTL1 expression. H The effects of METTL1 on the protein expression of H3K9la, PKM2 and CD155 in CRC (Epithelial) cells were detected via flow cytometry of EpCAM + cells and Western blotting. *P < 0.05; **P < 0.01. IF: immunofluorescence

To explore downstream molecules that METTL1 may affect by regulating PKM2, we performed transcriptome sequencing analysis using stably transfected cell lines with METTL1, and the results revealed that METTL1 significantly increased the transcription level of the immune checkpoint molecule CD155, which was also verified by qPCR (Fig. 6D, Fig. S2H and Fig. S3F). On this basis, we used the chromatin immunoprecipitation method and ChIP‒qPCR to determine that the PKM2 signal was enriched in the promoter region of CD155 (Fig. 6E). The gel analysis of m7G-binding mRNAs revealed that the level of m7G in PVR (CD155) mRNA was low; thus, METTL1-m7G may promote PVR expression through PKM2 rather than directly acting on PVR mRNA (Fig. S3G).

To further investigate the effect of CD155 on tumor immune escape in CRC cells, we first analyzed the public ST dataset and found that CD155 was expressed mainly in tumor (Epithelial) cells (Fig. 6F). In an in vivo model of MC-38 overexpression, we found that Mettl1 promoted CD155 expression in tumor tissues and inhibited the proportion of CD16 + NK cells (Fig. 6G). For further verification, we analyzed the pathway by which METTL1/PKM2 regulate CD155 expression in CRC cells and detected the protein in epithelial cells (EpCAM) in tumor tissues. The results showed that overexpression of METTL1 promoted the expression of PKM2, CD155 and H3K9la (Fig. 6H).

METTL1/PKM2/H3K9la axis-mediated regulation of CD155 expression is elevated in CRC and associated with poor prognosis

Consistent with these findings, we confirmed that METTL1 is associated with poor prognosis in CRC patients and can significantly increase the expression level of CD155 (Fig. S3H and Fig. 6D). The expression levels of METTL1, PKM2, H3K9la, and CD155 were detected via IHC. PKM2, H3K9la, and CD155 were also significantly increased in CRC patients with high METTL1 expression (Fig. 7A and B). Correlation analysis of these molecules in CRC tissues revealed that METTL1 was associated with PKM2 (R² = 0.481, P value = 0.0003), H3K9la (R² = 0.293, P value = 0.0035) and CD155 (R² = 0.6053, P value < 0.0001) (Fig. 7C–E). In the TCGA-COAD dataset, we also observed a similar phenomenon, where CD155 expression was significantly correlated with METTL1 expression (R² = 0.210, P value = 5.5e − 05; Fig. S4A), indicating that METTL1-m7G modification may promote CD155 upregulation. In addition, the correlation coefficient between CD155 and QKI was R² = 0.083 (P = 0.083, Fig. S4B), and the correlation coefficient between CD155 and G3BP1 was R² = 0.410 (P = 4.4e-16, Fig. S4C).

Fig. 7.

METTL1/PKM2/H3K9la axis regulation of CD155 is associated with poor prognosis in CRC patients. A and B The protein expression levels of PKM2, H3K9la and CD155 in CRC tissues from the high METTL1 expression group (n = 26) and the low METTL1 expression group (n = 26) were detected via IHC. Scale line, 60 μm. The group with high METTL1 expression was compared with the group with low METTL1 expression. The intensities of PKM2, H3K9la and CD155 in cancer cells were analyzed via a two-tailed paired t test. C, D and E Correlation analysis of the IHC scores between the expression of METTL1 and that of PKM2, H3K9la and CD155. F In the proposed model, METTL1 mediates PKM m7G modification, regulates CD155 expression, and promotes immune evasion in colorectal cancer. *P < 0.05; **P < 0.01; ***P < 0.001

In the TCGA database, we observed that CD155 expression levels in CRC (COAD and READ) were significantly higher in tumor tissue than in normal tissue (Fig. S4D). In addition, in the AHJNU study cohort, increased PKM2 (log-rank P value = 0.039, Fig. S4E) and CD155 (log-rank P value = 0.0027, Fig. S4F) protein levels were associated with poor prognosis according to survival analysis.

Discussion

CRC cells exhibit significant immune escape properties. Therefore, our study revealed that METTL1 plays a key role in modulating the tumor immune microenvironment [27]. METTL1 promotes m7G modification within PKM mRNA sequentially and increases the stability of PKM mRNA and the expression of the PKM2 protein, which further leads to the accumulation of lactate and an increase in H3K9la levels. In addition, METTL1/PKM2 triggered an abnormal increase in CD155 in tumor cells, which inhibited the activation of NK and T cells in the microenvironment. These findings suggest that high expression of CD155 mediated by METTL1 may accelerate the development of CRC by promoting immune evasion.

At present, many types of epigenetic modifications, such as N1-adenosine (m1A), 5-cytosine (m5C), N6-adenosine (m6A) and m7G, have been identified on RNA molecules. These RNA epigenetic modifications are reversible chemical modifications catalyzed by specific enzymes that affect the structure, stability, and biological function of RNA, thereby controlling gene expression and cell function. Studies have shown that in mammalian RNA molecules, m7G modification is promoted primarily by the METTL1 protein [28]. Recent studies have revealed that METTL1 is abnormally expressed in many cancer types and is involved in the regulation of tumor formation and development through m7G-dependent mechanisms [29]. To explore whether METTL1 promotes the progression of colorectal cancer (CRC) through m7G modification of key mRNAs, we used RIP-seq analysis combined with RIP-qPCR detection and found that there was significant m7G signal enrichment within the PKM mRNA. Further studies revealed that METTL1 mediated the m7G modification of PKM mRNA and promoted the expression of the PKM2 protein.

Pyruvate kinase is an important gene involved in anaerobic cell oxidation and a key marker of tumor cell glycolysis. PKM2 endows cells with a specific glucose metabolic phenotype, replacing other normal acid kinase phenotypes (PKM1, PKL, and PKR). This study revealed that PKM2 expression is increased in tumor cells and that increased ATP production promotes the process of glycolysis, resulting in the accumulation of the intermediate metabolite lactate in tumor cells and metabolic reprogramming, thus providing favorable conditions for tumor cell proliferation [30]. These results suggest that METTL1 increases PKM2 expression through m7G modification, which leads to significant lactate accumulation in CRC cells.

Lactate is a key product of glycolytic metabolism and plays a key regulatory role in tumor proliferation, metastasis and the immune microenvironment. Lactate plays an important role in epigenetic regulation by entering the nucleus and being modified to a lysine site on histones, thereby altering chromatin structure and affecting gene function via histone lactylation [31]. This modification usually occurs in the active gene promoter region and can promote the binding of transcription factors to DNA, unwinding DNA, and enhancing gene transcription. Recent studies have shown that histone lactylation plays a key role in regulating biological processes such as tumorigenesis and tumor cell reprogramming [32]. The results showed that METTL1-m7G enhanced the ability of PKM2 to promote glycolysis, resulting in lactate accumulation, and further increased H3K9 levels, resulting in H3K9la promoting METTL1 transcription and expression. These findings suggest that a METTL1/PKM2/H3K9la positive feedback regulatory loop exists.

PKM2 not only promotes tumor growth by enhancing aerobic glycolysis in tumor cells but also acts as a transcription factor to promote cancer through nonmetabolic pathways. PKM2 changes from a tetramer to a dimer under the induction of several factors and then enters the nucleus to act on chromatin and regulate related biological processes [14, 15]. In this study, METTL1 was shown to bind the RNA of PKM and catalyze its m7G methylation. For the relevant reader, we detected only the binding PKM RNA of QKI effectively. The upregulated expression of the PKM2 protein at the protein level showed that QKI did not reduce the translation efficiency of PKM in the absence of stress stimulation. PKM2 enhances nuclear heterogeneity, participates in the formation of a transcription complex, regulates the expression of downstream oncogenes, and promotes the development of tumors. After screening the downstream immune checkpoint effector clusters of METTL1/PKM2, CD155 was found to have the most significant response. CD155 is a transmembrane adhesion protein in the immunoglobulin superfamily that can act as a ligand to interact with activator/suppressor receptors on immune cells and affect the immune response. TIGIT has a high affinity for CD155 and is an inhibitory receptor shared by NK cells and T cells. The combination of CD155 and TIGIT can directly trigger immunosuppressor signals, promote IL-10 production in NK cells, reduce IL-12 secretion, and then inhibit T-cell activation. In addition, the expression of CD155 in tumor cells can also induce the internalization and degradation of CD226 on the surface of NK cells, leading to immune escape and promoting tumor progression.

Conclusion

In summary, this study demonstrated that METTL1-mediated m7G modification of PKM mRNA enhances PKM2 protein expression and promotes glycolysis and lactate accumulation. Lactate increases the expression of METTL1 through H3K9la, forming a positive feedback regulatory axis of H3K9la, METTL1 and PKM2. In addition, H3K9la promotes METTL1-m7G-mediated PKM2 through the regulation of CD155 transcription and the activation of the METTL1/PKM2/CD155 molecular axis, thereby driving immune evasion in CRC.

Abbreviations

CRC

Colorectal cancer

METTL1

Methyltransferase like 1

m7G

N7-methylguanosine

PKM2

Pyruvate kinase M2 type

H3K9la

H3K9 lactylation

TSNE

T-distributed stochastic neighbor embedding

KO

Knockout

WT

Wild type

ECAR

Extracellular acidification rate

DAPI

4′,6-diamidino-2-phenylindole

PVR

Poliovirus receptor

IHC

Immunohistochemistry

Author contributions

Fang Wang, Hairong Sun and Xiaowei Qi designed the study and obtained the funding. Yong Mao acted as a guarantor for the overall content. Fang Wang and Fang Zheng together with Yanyan Feng performed in vitro assays. Chen Yang, Hejia Xu and Yang Yan contributed to the conduct and group allocation of animal experiments and flow cytometry. Guifang Li and Chen Yang performed in vivo treatment and data analysis. Fang Wang and Fang Zheng performed scRNA, ST-seq and RNA sequencing analyses. Xiaowei Qi and Zilong He performed the diagnosis and analysis of pathology. Fang Wang, Fang Zheng, and Dongyan Cai developed the protocol and coordinated tissue collection. Chen Yang and Hejia Xu wrote the draft, and all authors revised and approved the final manuscript.

Funding

This study was supported by the Wu Jieping Medical Foundation (320.6750.2023-05-57), the National Natural Science Foundation of China (81502042, 32370584), the Medical Research Project Plan of Research Hospital Affiliated Hospital of Jiangnan University (YJZ202302), the Wuxi City Social Development Science and Technology Demonstration Project (N20201005), Wuxi Science and Technology Innovation and Entrepreneurship Project (No.2024343) and Wuxi Science and Technology Development Healthcare Guiding Program Project (YF2020001).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The application for collected human colorectal cancer tissues followed the Declaration of Helsinki and were approved by the Ethics Committee of Jiangnan University Affiliated Hospital (LS2020058, September 2020). The application for animal experiments follows the Basel Declaration and was approved by the Animal Experiment Ethics Committee of Jiangnan University (JN.No20231030m0960420, February 2023).

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no conflict of interest.