METTL14 regulates CD8⁺T-cell activation and immune responses to anti‐PD‐1 therapy in lung cancer

A mouse Lewis lung carcinoma (LLC) cell line and mouse peripheral blood mononuclear cells (PBMCs) were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). LLC cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, USA) at 37 °C under 5% CO₂.

PBMCs were cultured in RPMI-1640 medium (Gibco, USA) supplemented with 10% FBS (Gibco, USA) at 37 °C under 5% CO₂. A MagniSort™ Mouse CD8⁺T-cell Enrichment Kit (Invitrogen, USA) was used for negative selection of mouse CD8⁺ T cells. Then, CD8⁺ T cells were activated with Dynabeads™ Mouse T-Activator CD3/CD28 (Gibco, USA). CD8⁺ T cells and LLC cells were cocultured for 48 h. In brief, activated CD8⁺ T cells were resuspended in serum-free RPMI 1640 medium in the upper chamber of a transwell plate (5 mm), and LLC cells were cultured in the lower chamber of a transwell chamber.

For the knockdown of METTL14, YTHDF1, YTHDF2 and YTHDF3, short hairpin RNAs (shRNAs) against METTL14, YTHDF1, YTHDF2, and YTHDF3 (sh-METTL14, sh-YTHDF1, sh-YTHDF2, and sh-YTHDF3) and a negative control (sh-NC) were purchased from RiboBio (Guangzhou, China). For overexpression of METTL14 and HSD17B6, the cDNA sequences of METTL14 and HSD17B6 were cloned and inserted into the pLV-EF1a-EGFP(2A) Puro vector by RiboBio (Guangzhou, China). LLC cells were seeded at 1 × 10⁵ cells/well in a 24-well plate. When the percentage of LLC cells was approximately 70%, the cells were transfected with 1 μg of shRNA or 50 pmol of plasmid for 24 h or 48 h using Lipofectamine 2000 (Invitrogen, USA) following the manufacturer's instructions.

LLC cells were seeded in 96-well plates at a density of 2 × 10³ cells/well. LLC cells were incubated with 100 μL of CCK8 reagent (Abcam, USA) for 4 h at 24 h, 48 h, 72 h, and 96 h. The optical density was measured at 450 nm using a microplate reader (Bio-Rad, USA).

LLC cells were seeded at 5 × 10³ cells/well in 6-well plates and cultured at 37 °C for 2–3 weeks. The cell culture was terminated when the clone was visible to the naked eye. The supernatant was removed, and the cells were washed twice with PBS. The cells were fixed with 4% paraformaldehyde for 15 min and stained with 0.1% crystal violet staining solution for 10–30 min. Ten random fields of view were observed under a light microscope (Olympus, Japan), and the number of colonies was counted.

For cell migration, LLC cells were cultured in serum-free medium in a transwell upper chamber with an 8 μm pore size, and DMEM medium containing 20% FBS was added to the transwell lower chamber. For cell invasion, LLC cells were cultured in serum-free medium in a transwell upper chamber pretreated with Matrigel (BD Biosciences, USA), and DMEM medium containing 20% FBS was added to the lower chamber. The cells were fixed with 4% paraformaldehyde for 15 min and stained with 0.1% crystal violet for 15 min. The migrated and invasive cells were counted under a microscope (Olympus, Japan) and analyzed statistically.

LLC cells were seeded into 24-well plates at a density of 2 × 10⁵ cells/well and incubated at 37°Cfor 12 h. The wound location was recorded by a sterile tip. After 24 h, the marked wound location was photographed by a microscope (Olympus, Japan) to assess the cell migration ability.

A total of 1 × 10⁵ cells were centrifuged at 1200 rpm for 5 min to remove the supernatant. Then, 50 μL of binding buffer (Beyotime, China) was added to each sample tube, and the sample was resuspended. Then, 5 μL of Annexin V-FITC (Beyotime, China) and 5 μL of PI (Beyotime, China) were added to the sample tube. The solution was gently shaken and incubated for 15 min at room temperature in the dark. Then, 200 μL of 1 × binding buffer (Beyotime, China) was added, and cell apoptosis was detected by a FACSCalibur flow cytometer (BD Biosciences, USA).

The RIP assay was carried out with a Magna RIP Kit (Millipore, USA). LLC cells were lysed with RIP lysate for 5 min on ice, and then the cell lysates were centrifuged at 1500 rpm for 5 min. The supernatant was collected and precleaned with protein A-Sepharose beads. RIP wash buffer was used to wash the magnetic beads. RIP lysates were immunoprecipitated with 5 μg of METTL14, YTHDF2 or IgG antibody (negative control). Then, 900 μL of RIP immunoprecipitation buffer was added to the magnetic beads. Then, 100 μL of the supernatant was added to the bead-to-antibody complex in the previous step for a total volume of 1 mL. The above magnetic bead-antibody complex was resuspended in 150 μL of proteinase K buffer at 55 °C for 30 min. Then, 250 μL of RIP wash buffer was added to the supernatant. Finally, RNA was extracted via the phenol‒chloroform method. Agarose gel electrophoresis was used to detect HSD17B6 expression.

MeRIP was measured by the MeRIP™ m6A Transcriptome Profiling Kit (RiboBio, China) according to the manufacturer’s instructions. Briefly, 18 μg of total RNA was added to 2 μL of 10 × RNA fragmentation buffer, and the reaction was performed at 70 °C for 7 min. The segmented RNA was added to 2 μL of 0.5 M EDTA. The reaction products were added to 18 μL of 3 M sodium acetate (pH 5.2),1 μL of glycogen (20 mg/mL) and 600 μL of absolute ethanol and precipitated overnight at -20 °C. The precipitated product was centrifuged at 12,000 × g at 4 °C for 30 min and washed twice with 75% ethanol. RNA was dissolved in 50 μL of nuclease-free water. Then, 500 μL of the MeRIP reaction solution was added to the prepared anti-m6A magnetic beads and incubated for 2 h. After the IP, 100 μL of the eluent buffer was added to the sample, which was gently blown and mixed to fully resuspend it. The eluted RNA was recovered, and the IP group RNA was captured by a m6A antibody. After immunoprecipitation, a portion of the RNA was analyzed via qRT‒PCR.

To detect RNA stability in cells, LLC cells were transfected with METTL14 overexpression lentivirus and sh-YTHDF2 for 24 h, and then 5 mg/mL actinomycin D (Sigma, MA, USA) was added to the cells. RNA was isolated using TRIzol reagent (Invitrogen, NY, USA) at 0, 2, 4, and 6 h, analyzed by real-time PCR, and normalized to GAPDH. The t1/2 of HSD17B6 was calculated.

The levels of IFN-γ and TNF-α were measured using an IFN-γ ELISA Kit (Beyotime, China) and a TNF-α ELISA Kit (Beyotime, China), respectively,according to the manufacturer’s instructions.

The animal experiments were approved by the Animal Care and Use Committee ofNanjing Medical University (IACUC-1706007). Eighty male BALB/c mice (SPF grade 18–22 g) were purchased from Beijing Weitong Lihua Laboratory Animal Technology (Beijing, China). Male BALB/c mice were housed with fed ordinary diet and sufficient drinking water at 18–22 °C and a relative humidity of 40%-70%.

After one week of acclimation, the subcutaneous tumor model was constructed, and a total of 30 male BALB/c mice were divided into 6 groups: 1). sh-Ctrl, sh-METTL14; 2) sh-Ctrl, sh-Ctrl + anti-PD-1, sh-METTL14, sh-METTL14 + anti-PD-1. The other 20 mice were divided into sh-Ctrl and sh-METTL14 groups, and the survival conditions within 40 days of modeling were recorded.

LLC cells were transfected and prepared into 1 × 10⁶ cells/ml single-cell suspensions in PBS after the corresponding transfection treatments. LLC cells were inoculated subcutaneously into BALB/c mice in a sterile environment, and the LLC lung cancer mouse model was successfully established after nodules with a diameter greater than 5 mm³ appeared subcutaneously in the vaccinated mice. For mice in the PD-1 antibody treatment group, 10 mg/kg anti-PD-1 (Abcam, USA) was injected intraperitoneally on days 11, 14, 17, 20 and 23. An equal amount of IgG was injected as a control.

The removed tumor tissues were fixed in 4% paraformaldehyde for 24 h. Then, the tissues were embedded in paraffin and sectioned. The paraffin-embedded tumor sections were dewaxed and treated with 3% H₂O₂ to deactivate endogenous peroxidase. After blocking with goat serum for 1 h, the sections were incubated with an anti-CD8antibody (1:500, Abcam, USA) at 4 °C overnight. After the sections were incubated with goat anti-mouse IgG H&L (HRP) at 37 °C for 30 min, they were stained with DAB and counterstained with hematoxylin. Images of stained sections were observed under a light microscope (Nikon, Japan).

After the mice were sacrificed, the tumor tissues were collected,minced into small pieces of 1 mm³ and placed in Petri dishes. Collagenase (Sigma, USA) was added to the tissue, which was subsequently digested at 37 °C for 20 min to isolate the cells. The isolated cells were incubated with Alexa Fluor® 647 Fluorescent Anti-CD8 (1:1000, Abcam, USA), Alexa Fluor®488 Fluorescent Anti-CD4 (1:1000, Abcam, USA), and FITC Fluorescent Anti-CD45 (1:1000, Abcam, USA) at4°C in the dark for 30 min. The cell precipitate was resuspended in 400 μL of PBS and detected by a FACSCalibur Flow Cytometer (BD Biosciences, USA).

Total RNA was extracted from cells and tissues using TRIzol reagent (TIANGEN, China). Then, the RNA was reverse-transcribed to cDNA using a PrimeScript™ RT kit (Takara, Japan). cDNA (50 pg) was amplified by an ABI Real-Time PCR System (Applied Biosystems, USA) using SYBR Premix Ex TaqII (Takara, USA). The results were calculated using the 2^−ΔΔCt method, and GAPDH was used as a control. The sequences of the primers used are shown in Table 1.

Total protein was extracted from cells and tissues using RIPA buffer (Beyotime, China). Then, the protein concentration was detected by a BCA assay kit (Santa Cruz, USA). Protein samples (40 μg) were separated by SDS‒PAGE and then transferred to PVDF membranes. After blocking, the membranes were incubated with primary antibodies overnight at 4 °C and with a goat anti-mouse IgG H&L (HRP)-preadsorbed secondary antibody for 1 h at room temperature. Finally, enhanced chemiluminescence (ECL, Thermo Fisher, MA, USA) was used to visualize the membrane. Protein band analysis was conducted with ImageJ software. The following antibodies were used: anti-Granzyme B (1:1000, ab283315, Abcam, USA); anti-METTL14 (1:1000, ab220030, Abcam, USA); anti-HSD17B6 (1:1000, orb539880, Biorbyt, USA); anti-Perforin (1:1000, ab47225, Abcam, USA); anti-GAPDH (1:2000, ab8245, Abcam, USA); and goat anti-mouse IgG H&L (HRP)-preadsorbed secondary antibody (1:5000, ab47827, Abcam, USA).

The online database GEPIA (http://​gepia.​cancer-pku.​cn/​index.​html) showed that lung cancer patients with high METTL14 expression had shorter overall survival (OS) than patients with low METTL14 expression (Fig. 1A). In addition, the disease-free survival (DFS) of lung cancer patients with high METTL14 expression was lower than that of patients with low METTL14 expression (Fig. 1B). These online prediction results indicated that the level of METTL14 was closely related to the survival of lung cancer patients.

To explore the role of METTL14 in lung cancer progression, LLC cells were transfected with sh-METTL14 for 24 h, and changes in LLC cell function were studied. CCK-8 assays showed that METTL14 knockdown significantly decreased LLC cell viability (P < 0.001, Fig. 2A). In addition, clone formation assay suggested that METTL14 knockdown reduced the proliferation of LLC cells (P < 0.01, Fig. 2A-B). Transwell assays verified the decreased invasion and migration of LLC cells treated with sh-METTL14 (P < 0.01, Fig. 2D-E). Moreover, a wound healing assay verified that LLC cell migration decreased after METTL14 knockdown (P < 0.05, Fig. 2F-G). The LLC cell apoptosis rate also significantly increased after METTL14 knockdown (P < 0.01, Fig. 2H-I). These results consistently indicated that METTL14 suppression significantly inhibited lung cancer progression in vitro.

The effect of METTL14 knockdown was further verified in a syngeneic mouse model, and its effect on PD-1 treatment was studied. Western blot analysis revealed that knockdown of METTL14 decreased METTL14 protein levels in tumor tissues (Fig. 3A). The tumor volume was recorded every 5 days during modeling. After 25 days of modeling, the tumor volume in the sh-METTL14 group was significantly smaller than that in the sh-Ctrl group (P < 0.01, Fig. 3B-C). In addition, the 40-day survival rate of the mice treated with sh-METTL14 was significantly greater than that of the mice in the sh-Ctrl group (Fig. 3D). Subsequently, a syngeneic mouse model was treated with sh-METTL14 and an anti-PD-1 antibody to observe the effect of METTL14 on immunotherapy efficacy. At 25 days after surgery, the tumor volume in the anti-PD-1 antibody treatment group was significantly reduced compared with that in the sh-Ctrl group, and the effect in the sh-METTL14 treatment group was similar to that in the anti-PD-1 group (P < 0.05, P < 0.01; Fig. 3E-F). The tumor volume was significantly lower in the sh-METTL14 plus anti-PD-1 group than in the anti-PD-1 alone group (P < 0.001, Fig. 3E-F). These results demonstrated that METTL14 knockdown effectively enhanced the inhibitory effect of PD-1 treatment on transplanted tumors.

Next, the cellular mechanism by which METTL14 knockdown enhances the therapeutic effect of PD-1 was investigated. ELISA revealed that METTL14 inhibition significantly increased the intratumor IFN-γ concentration(P < 0.01, Fig. 4A). However, the serum IFN-γ level did not change between the sh-Ctrl group and the sh-METTL14 group (P > 0.05, Fig. 4B). Similarly, the intratumor TNF-α level was increased by METTL14 inhibition (P < 0.01, Fig. 4C). However, the serum TNF-α level did not change between the sh-Ctrl group and the sh-METTL14 group (P > 0.05, Fig. 4D). These changes in intracellular cytokines may be due to changes in the infiltrating components of immune cells. Therefore, the infiltrating fraction of immune cells in the transplanted tumors was subsequently analyzed by flow cytometry. In sh-METTL14-treated mice, the level of intratumoral infiltration of CD8⁺ T cells was significantly increased (P < 0.001, Fig. 4E), but the levels of CD4⁺ T cells and CD45⁺ cells were not significantly changed (P > 0.05, Fig. 4E). IHC staining revealed that the number of CD8 protein-positive cells was increased in the sh-METTL14-treated group (Fig. 4F). These results suggested that reducing METTL14 levels increased CD8⁺ T-cell infiltration.

After syngeneic mice were subjected to sh-METTL14 treatment, the transplanted tumors were sequenced to analyze the DEGs. The expression of differentially expressed genes was plotted as heatmaps (Fig. 5A) and volcano plots(Fig. 5B). The five genes with the most differential expression between the sh-METTL14 group and the sh-Ctrl group were selected for RT‒qPCR. Compared to those in the sh-Ctrl group, the levels of MYH11, HSD17B6, and A2M were significantly upregulated in the sh-METTL14 group, while the levels of LRRK2 and Cxcl15 were downregulated in the sh-METTL14 group (P < 0.05, P < 0.01, P < 0.001, Fig. 5C). To further explore the effect of METTL14 on DEGs, the expression levels of the above five DEGs were detected after METTL14 knockdown and overexpression in cells; HSD17B6 was significantly upregulated after METTL14 knockdown (P < 0.001, Fig. 5D), and METTL14 overexpression suppressed HSD17B6 expression (P < 0.001, Fig. 5D). The changes in LRRK2 and Cxcl15 were opposite to those in HSD17B6, but MYH11 and A2M did not change significantly after METTL14 overexpression (Fig. 5D). After that, three genes (HSD17B6, LRRK2 and Cxcl15) whose expression significantly changed were selected for detection of m6A methylation levels via the MeRIP method, and HSD17B6 was found to be the most sensitive to changes in METTL14 levels(P < 0.001, Fig. 5E). RIP assays further revealed the binding relationship between METTL14 and HSD17B6 mRNA (Fig. 5F). Furthermore, the knockdown of YTHDF2 significantly altered HSD17B6 mRNA levels (Fig. 5G). Next, the binding relationship between YTHDF2 and HSD17B6 mRNA was also verified by the RIP method (Fig. 5H). The mRNA half-life of HSD17B6 was determined after actinomycin D treatment of LLC cells. METTL14 overexpression significantly accelerated the degradation of HSD17B6(P < 0.001, Fig. 5I), while the knockdown of YTHDF2 inhibited the degradation of METTL14 (P < 0.05, Fig. 5I).

The online database TIMER (http://​timer.​cistrome.​org/​) was used to predict the purity of HSD17B6 in tumors and the correlation of HSD17B6 with CD8⁺T-cell infiltration. HSD17B6 was negatively correlated with lung cancer tumor purity and positively correlated with CD8⁺ T cells (Fig. 6A). Next, LLC cells and activated mouse PBMCs were cocultured in vitro to mimic immune cell infiltration in the tumor microenvironment. RT‒qPCR revealed that the HSD17B6 level in LLC cells was significantly increased after coculture(P < 0.01, Fig. 6B). In the coculture system, HSD17B6 overexpression in LLC cells significantly increased the HSD17B6 mRNA level (P < 0.01, Fig. 6B), while METTL14 overexpression significantly inhibited the HSD17B6 overexpression-induced increase in the HSD17B6 mRNA level (P < 0.001, Fig. 6B). The knockdown of YTHDF2 significantly increased the level of HSD17B6 (P < 0.01, Fig. 6B). The HSD17B6 protein and mRNA levels showed similar trends(Fig. 6C). The levels of the CD8⁺ T-cell effector Granzyme B and perforin-activated PBMCs in the coculture system were detected by western blotting. The changes in these two effectors were also similar to the changes in HSD17B6 expression (Fig. 6D). Finally, the levels of IFN-γ and TNF-α in the PBMC supernatant were detected by ELISA. Consistently, the trends of IFN-γ and TNF-α were the same as those of HSD17B6 (Fig. 6E-F).