Tissue samples

A total of 41 paired fresh lung adenocarcinoma tissues and matched adjacent nontumor samples were collected from the First Affiliated Hospital of Soochow University between 2019 and 2022. The sample inclusion criteria were: (1) The samples were diagnosed LUAD based on their pathological and histological feature according to the Revised International System for Staging Lung Cancer. (2) The patient did not receive chemotherapy, radiotherapy, immunotherapy, and target therapy before surgery. (3) All the patients have signed informed consent, and relevant clinicopathological data are collected with the patient's consent. The sample exclusion criteria were: (1) Pathological examination did not confirm the diagnosis of LUAD. (2) The patient received preoperative radiotherapy and chemotherapy, immunotherapy, targeted therapy or other adjuvant therapy. (3) The patient did not consent to the collection of samples and clinicopathological data. Upon excision, the tissue samples were immediately frozen, preserved in an RNA stabilization solution, and stored at −80 ℃. The study was authorized by the Ethics Committee of the First Affiliated Hospital of Soochow University. Detailed information on the participants is presented in Table S1.

Cell lines and cell culture

Human lung cancer cell types H1299 (CL-0165), H1650 (CL-0166), PC-9 (CL-0668), H1975 (CL-0298), HCC827 (CL-0094), A549 (CL-0016) (lung adenocarcinoma cell lines), H226 (CL-0396) (lung squamous carcinoma cell line), HEK293T (human embryo kidney cell), and BEAS-2B (human bronchial epithelial cell line) were acquired from the Procell Life Science and Technology Co. Ltd. (Wuhan, China). Short tandem repeat (STR) profiling was performed to authenticate all cell lines. Cells were cultivated in RPMI-1640 or DMEM media (Procell, Wuhan, China) augmented with 10% FBS (Gibco, CA, USA) and penicillin/streptomycin (100 mg/ml, Beyotime Biotechnology, Shanghai, China) at 37 ℃ in a 5% CO2 humid incubator.

Transient transfection

Pre-designed sequences of short interfering RNA (siRNA) targeting various coding regions of SRPK1, METTL3, IGF2BP1, IGF2BP2, IGF2BP3, YTHDF1, YTHDF2, and hnRNPA1 were prepared by GenePharma (Suzhou, China). Table S2 enlists the target sequences. Scrambled siRNA was utilized as a negative control. The cells were then transiently transfected with siRNA by Lipofectamine 2000 (Invitrogen, CA, USA). We harvested the cell for subsequent experiments 72 h after.

Establishment of stable cell lines

METTL3 and SRPK1 overexpression lentiviruses and control lentiviruses were acquired from GeneChem Corporation (Shanghai, China). The lentiviruses were infected in the cells per the manufacturer’s guide. Then to establish a stable cell line for the subsequent analyses, cell selection was performed with 1 µg/ml puromycin (Sigma‒Aldrich, St Louis, MO, USA).

Total RNA extraction and real-time PCR

RNA was isolated, cDNA was prepared and real-time quantitative reverse transcription polymerase chain reaction (qRT–PCR) was carried out as described in the literature [31]. Table S3 enlists the primers utilized. The relative expression levels were quantified via the ΔΔCt method. β-Actin was employed as an endogenous control.

Western blotting

This assay was performed per previous studies [31]. The utilized antibodies included anti-METTL3 (ab195352, Abcam, London, UK), anti-SRPK1 (sc-100443, Santa Cruz, CA, USA), anti-IGF2BP2 (11601-1-AP), anti-PKM1 (19987-1-AP), anti-PKM2 (15822-1-AP), anti-PCNA (60097-1-Ig), DYKDDDDK tag (66008-4-Ig), anti-PKM (25659-1-AP), anti-HK1 (19662-1-AP), anti-HK2 (22029-1-AP), anti-LDHA (19987-1-AP), anti-PFKFB3 (13763-1-AP), anti-hnRNPA1 (15821-1-AP, Proteintech, IL, USA), and anti-β-Actin (CW0096M, Cowin Bio, Jiangsu, China) antibodies. Furthermore, goat anti-mouse IgG, HRP-conjugated (CW0102) and anti-rabbit IgG, HRP-conjugated (CW0103, Cowin, Jiangsu, China) antibodies were utilized as secondary antibodies. The electrochemiluminescence reagent (Thermo Fisher Scientific, MA, USA) was used for bands development, imaged using a ChemiDoc XRS + (Bio-Rad, CA, USA), and lastly quantified with ImageJ software (National Institutes of Health, MD, USA).

Cell counting kit-8 (CCK-8) assay

CCK-8 assay was carried out as previously described [31]. For assessing, cell viability, CCK-8 (Boster, Wuhan, China) was utilized per the kit’s instructions.

5-Ethynyl-2′-deoxyuridine (EdU) assay

The EdU assay (C10310-1, Ribobio, Guangdong, China) was carried out using the guide provided by the manufacturer as mentioned in previous study [32].

Cell invasion and migration assays

Cell invasion and migration assessment was carried out based on the previous studies [31]. For the invasion assay, Matrigel matrix (Corning, NY, USA) was used to coat the inserts. The cells were then imaged and counted under a microscope.

M6A RNA methylation quantification

According to the previously described method, total RNA was isolated [31]. Then, the m6A RNA methylation levels were quantified using the acquired RNA via the EpiQuik m6A RNA Methylation Quantitation Kit (Colorimetric) (#P-9005, Epigentek, NY, USA) per the kit’s instructions. The m6A levels were measured by assessing the absorbance at 450 nm.

M6A–mRNA and long non-coding RNA (lncRNA) epitranscriptomic microarray analysis

Whole RNA from 12 clinical samples (6 pairs of LUAD and adjacent non-cancerous tissue) was extracted. The protocol of Arraystar Human mRNA and lncRNA Epitranscriptomic Microarray (8 × 60,000, Arraystar) was utilized for the sample preparation and microarray hybridization. Agilent Scanner G2505C was utilized for scanning the arrays. Agilent Feature Extraction software (version 11.0.1.1) was employed to evaluate the obtained array images. Microarray experiments and data assessment were conducted by KangChen Biotech (Shanghai, China).

RNA immunoprecipitation (RIP)

Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (No. 17-700, Merck Millipore, MA, USA) was employed for RIP assay, according to the kit’s guide. Briefly, using the kit’s RIPA buffer, the cells were lysed and collected and incubated with RIP buffer comprising magnetic beads linked with the relevant antibody or control normal IgG. For sample digestion, proteinase K was employed to isolate the immunoprecipitated RNA, which was then subjected to qRT‒PCR to identify the presence of the binding targets.

M6A RIP (MeRIP)

For MeRIP analysis, a Magna MeRIP™ m6A Kit (No. 17-10499, Merck Millipore, MA, USA) was employed, per the kit’s instructions as previously mentioned [33]. The m6A-containing mRNA enrichment was assessed by qRT‒PCR and normalized to the input. The primers utilized for MeRIP PCR are provided in Table S4.

RNA stability

Actinomycin D (Act-D, 5 μg/ml) (cat. no. S8964, Selleck, TX, USA) was added to the LUAD cells for RNA stability. The cells were collected after incubation at the indicated times for RNA isolation via TRIzol reagent, and then reverse transcription was carried out to measure the remaining mRNA using qRT‒PCR.

Immunofluorescence staining

Immunofluorescent staining was performed as described previously [31]. The stained cells images were acquired via a confocal microscope (ZEISS, BW, Germany) under standardized conditions.

Luciferase assay

Fragments of SRPK1–3′ untranslated region (UTR) comprising the wild-type and mutant m6A motifs (C instead of m6A) were prepared at Azenta Life Science (Shanghai, China). The fragments of mutant and wild-type SRPK1–3′ UTR were inserted into the psiCHECK2 luciferase vector. Cells were propagated in 24-well plates and cotransfected with 500 ng wild-type/mutant luciferase reporter and 0.5 μg control/SRPK1-overexpressing plasmid. The relative luciferase activity, cell lysates were collected and Dual-Luciferase Reporter Assay kit (Promega, WI, USA) was utilized 24 h later.

Quantification of energy metabolites by liquid chromatography tandem mass spectrometry (LC–MS/MS)

MetWare (http://www.metware.cn/) were utilized to detect all the metabolites based on the AB Sciex QTRAP® 6500 LC–MS/MS platform. Then the pathways with substantially modulated metabolites mapped were imported in metabolite set enrichment analysis (MSEA), and their significance was assessed via the p values of the hypergeometric test.

Extracellular acidification rate (ECAR)

The ECAR was assessed with the help of a Seahorse Glycolysis Stress Test Kit (103017) on a Seahorse XFp Analyser (Agilent Technologies, CA, USA), per the kit’s guide. Briefly, 10,000 cells per well were propagated in a Seahorse XFp cell culture microtiter plate for 24 h and then used for ECAR measurement. After the baseline assessment, 10 mM of glucose, 1 μM of oligomycin (oxidative phosphorylation inhibitor), and the 50 mM of 2-DG (glycolysis inhibitor) were sequentially added to each well at a specified timepoint. The data were analysed by Seahorse Wave software. The cells were counted again after the measurement. The results were normalized to the number of cells.

Glucose uptake and lactate generation analysis

For the glucose uptake assay, cells with a density of 5 × 106/ml were collected and resuspended in 1 ml of distilled water. The suspension underwent ultrasonication, 10 min of boiling, and then centrifuged at 25 ℃ and 8000g for 10 min. A glucose detection kit (#BC2500, Solarbio Science & Technology, Beijing, China) was then used per the kit’s protocols. The absorbance at 505 nm was measured.

For the lactate production assay, cells with a density of 5 × 106/ml were collected and resuspended in extracting solution from the lactate detection kit (#BC2230, Solarbio Science & Technology, Beijing, China). The following steps were performed per the kit’s method. The absorbance was measured at 570 nm.

Co‑immunoprecipitation

Co-immunoprecipitation was carried out as mentioned previously [31]. IgG-, hnRNPA1-, SRPK1-, or Flag-bound proteins were isolated via SDS-PAGE and then subjected to western blotting assay.

In vivo xenograft model

BALB/c athymic nude mice (aged 4–6 weeks, female, with a weight of 16–20 g) were acquired from the Laboratory Animal Center of Suzhou Medical College, Soochow University and bred in pathogen-free environment. The first batch of 16 mice was randomly categorized into two cohorts (8 per group). For model establishment, control or METTL3-overexpressing A549 cells suspended in 100 μl 1640 RPMI medium with 50% Matrigel (serum-free) and injected subcutaneously into nude mice flanks. Tumour volumes (V) were identified by assessing the tumour length (L) and width (W) with a Vernier calliper via the following formula: V = (L × W2) × 0.5. The second batch of 18 mice was first randomly divided into two groups: the control (n = 6 mice) and the SRPK1-overexpressing (n = 12 mice), inoculated with control cells and SRPK1-overexpressing cells, respectively. At the tumour volume of 100–150 mm3, the overexpressing group was once again randomly divided into the vehicle group and STM2457 group. Then the mice were treated with vehicle or 50 mg/kg/day STM2457 (cat. no. S9870, Purity 99.95%, Selleck, TX, USA) via intraperitoneal injection until sacrifice. All in vivo analyses were carried out by following the Guide for the Care and Use of Experimental Animals Center of Soochow University.

Statistical measurements

All assays were conducted independently in triplicate. GraphPad Prism 9.0 (GraphPad, CA, USA) was utilized for all the statistical assessments and the data are illustrated as the mean ± SD. The intergroup significant differences were elucidated by a nonpaired Student’s t-test. Whereas for the significant differences between more than two groups, one-way or two-way ANOVA was utilized. All statistical tests were two-tailed and P < 0.05 was deemed statistically significant.