Identification of differentially m6A-modified lncRNAs in NSCLC patients

Three paired human lung adenocarcinoma tissues and adjacent noncancerous tissues were collected at Harbin Medical University Cancer Hospital, with detailed clinical information provided in Supplementary Table S1. Arraystar m6A lncRNA epitranscriptomic microarrays were used to screen differentially m6A-methylated lncRNAs in NSCLC, and array figures were shown in Supplementary Fig. S1. A total of 39 differential metylated candidates were identified with 24 upregulated genes and 15 downregulated genes (Table 1). The differentially m6A-modified lncRNAs were shown in Fig. 1a. Functional analyses were subsequently conducted, and lncRNA-related mRNAs were predicted using Pearson’s correlation analysis (|r|>0.3, P < 0.05) based on TCGA lung adenocarcinoma data. These mRNAs underwent Gene Ontology (GO) (Fig. 1b) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (Fig. 1c) pathway enrichment analysis. The results demonstrated that differentially methylated lncRNAs were found to be involved in various tumor-related pathways, including cellular senescence, ubiquitin-mediated proteasome degradation, DNA replication, and homologous recombination.

Fig. 1

SCIRT m6A modification and expression levels were elevated in NSCLC. (a) Heatmap of differentially m6A-methylated lncRNAs. (b, c) Differentially m6A-methylated lncRNA-related mRNAs were analyzed by GO and the KEGG using TCGA data. (d) SCIRT, RP11-385J1.2, and SNHG9 m6A methylation levels in 10 paired NSCLC and adjacent noncancerous tissues. (e) Relative expression levels of SCIRT in 10 paired NSCLC and adjacent noncancerous tissues. (f) SCIRT expression levels of NSCLC (A549, H1299, and H358) and bronchial epithelial cell lines (BEAS-2B). Data are presented as mean ± SD from at least three independent experiments. **P < 0.01, ***P < 0.001

Table 1 lncRNAs with different m6A methylation levelsSCIRT m6A modification and expression levels are elevated in NSCLC

By integrating lncRNA expression data, three lncRNAs were identified – two upregulated and one downregulated. MeRIP and qRT‒PCR assays were used to detect m6A methylation in 10 pairs of patient samples to validate whether the m6A methylation levels aligned with microarrays result of these three lncRNAs. Notably, SCIRT m6A methylation and expression levels were markedly higher in LUAD tissue samples compared to adjacent noncancerous tissue samples (P < 0.001) (Fig. 1d and e). Additionally, SCIRT expression levels were analyzed in NSCLC cell lines. SCIRT expression was significantly elevated in A549 and H358 cells compared to BEAS-2B cells, whereas H1299 cells showed no significant difference in expression levels relative to BEAS-2B (Fig. 1f).

Aberrant METTL3 inhibition in NSCLC decreases the abundance of m6 A modifications and the stability of SCIRT

Microarray analysis identified the m6A modification site within SCIRT sequence (Fig. 2a). The METTL3 motif, GGAC, corresponds to specific m6A modification sites on SCIRT. Subsequent m6A-specific RIP-qPCR analyses elucidated significantly reduced m6A levels of SCIRT in A549 and H1299 cells expressing the m6A-WT SCIRT respectively (Fig. 2b and c), suggesting that the 180 A site in exon 2 was the m6A modification stie of SCIRT in NSCLC.

Fig. 2

Aberrant METTL3 inhibition in NSCLC decreases the abundance of m6 A modification and the stability of SCIRT. (a) The motif of METTL3 had matched sites on the m6A modification of SCIRT. (b) Schematic diagram of human SCIRT exons and its mutated forms used in m6 A RIP assay. (c) The m6A RIP assay showed that the SCIRT 180 A is its m6A site in NSCLC cells. (d, e) The expression levels of METTL3 were assessed in 10 paired NSCLC tissues by qRT‒PCR (d) and western blotting. (e). (f) The expression levels of METTL3 were measured in NSCLC cell lines (A549, H1299 and H358) and bronchial epithelial cell line (BEAS-2B). (g) Pearson’s correlation analysis between SCIRT and METTL3 expression levels in NSCLC tissues. (h) The expression levels of METTL3 in METTL3-knockdown A549 and H1299 cells were verified by western blotting. (i) SCIRT m6A modifications in METTL3 knockdown A549 and H1299 cells. (j) SCIRT expression in the presence of the transcriptional inhibitor ActD treatment after 0, 4, 8, 12, and 24 h. All results were shown with one representative image from three independent experiments. Data are mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001

To investigate the regulatory role of SCIRT m6A modification, the expression levels of METTL3, an important methyltransferase, were measured in 10 paired NSCLC tissues using qRT‒PCR and Western blot assays. METTL3 expression was significantly upregulated in NSCLC tissues (Fig. 2d and e) and showed a strong positive correlation with SCIRT expression (P < 0.001) (Fig. 2g). Consistently, METTL3 expression levels were higher in A549, H1299 and H358 cells compared to the bronchial epithelial cell line BEAS-2B, mirroring the SCIRT expression patterns (Fig. 2f).

To investigate the role of METTL3 in regulating SCIRT m6A modification in NSCLC cells, we analyzed changes in SCIRT m6A modification in METTL3-knockdown A549 and H1299 cells. As shown in Fig. 2h, METTL3 expression levels were significantly reduced in both the A549 shMETTL3 and H1299 shMETTL3 groups compared to the shNC groups (P < 0.05), indicating successful METTL3 knockdown. MeRIP-qPCR analysis revealed that downregulation of METTL3 inhibited SCIRT m6A modification in A549 and H1299 cell lines (Fig. 2i).

SCIRT expression was then analyzed following treatment with the transcriptional inhibitor actinomycin D (ActD) after 0, 4, 8, 12, and 24 h. The results revealed that SCIRT degraded more rapidly in METTL3-knockdown A549 and H1299 cells compared to A549 and H1299 shNC cells (Fig. 2j). These results indicated that METTL3 inhibition in NSCLC decreased SCIRT m6A modification and compromised its stability.

SCIRT exhibits oncogenic activity in NSCLC cells

The function of SCIRT in NSCLC cell lines was studied via loss- and gain-of-function experiments. A549 cells were infected with sh-SCIRT#1 and sh-SCIRT#2 plasmid-packaged viruses, and stable cell lines were selected with puromycin. qRT‒PCR analysis confirmed successful knockdown of SCIRT in both sh-SCIRT#1 and the sh-SCIRT#2 A549 cell lines, with sh-SCIRT#1 exhibiting higher knockdown efficiency (Fig. 3a). Additionally, the pcDNA3.1(+)-SCIRT overexpression (OE-SCIRT) plasmid was transfected into H1299 cells using Lipofectamine 3000, with an empty vector serving as control (vector group). Stable cell lines were subsequently selected with G418. As shown in Fig. 3b, qRT‒PCR results indicated dramatic upregulation of SCIRT expression in OE-SCIRT H1299 cells (Fig. 3b). CCK-8 assay demonstrated that SCIRT knockdown inhibited the proliferation rate of A549 cells, whereas SCIRT overexpression significantly enhanced the proliferation rate of H1299 cells (P < 0.05) (Fig. 3c and d).

Fig. 3

SCIRT exhibits oncogenic activity in NSCLC cells. (a, b) qRT‒PCR verified SCIRT expression in SCIRT-knockdown A549 and SCIRT-overexpressed H1299 cells. (c) CCK-8 assay of A549 cells transfected with shSCIRT#1, shSCIRT#2, or sh-NC. (d) CCK-8 assay of H1299 cells transfected with OE-SCIRT or empty vector. (e) Representative images and quantification of the colony formation assay in A549 cells transfected with shSCIRT#1, shSCIRT#2, or sh-NC. (f) Representative images and quantification of colony formation assay in H1299 cells transfected with OE-SCIRT or empty vector. (g) Representative images and quantification of EdU assays in A549 cells transfected with shSCIRT#1, shSCIRT#2, or sh-NC (Scale bar, 100 μm). (h) Representative images and quantification of the EdU assays in H1299 cells transfected with OE-SCIRT or empty vector (Scale bar, 100 μm). (i) Representative images and quantification of wound healing assays in A549 cells transfected with shSCIRT#1, shSCIRT#2, or sh-NC (Scale bar, 100 μm). (j) Representative images and quantification of wound healing assays in H1299 cells transfected with OE-SCIRT or empty vector (Scale bar, 100 μm). (k) Representative images and quantification of migration and invasion assays in A549 cells transfected with shSCIRT#1, shSCIRT#2, or sh-NC (Scale bar, 100 μm). (l) Representative images and quantification of migration and invasion assays in H1299 cells transfected with OE-SCIRT or empty vector (Scale bar, 100 μm). (m, n) The effect of SCIRT on the EMT formation of A549 and H1299 cells was detected in western blot. All results were shown with one representative image from three independent experiments. Data are mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001

Furthermore, EdU staining and colony formation assays revealed that SCIRT knockdown conspicuously suppressed cell proliferation, as well as colony number and size, whereas SCIRT overexpression produced the opposite effect (Fig. 3e-h). In summary, these findings confirmed that SCIRT promotes the proliferative capacity of NSCLC cells.

The impact of SCIRT on A549 cell migration, invasion, and EMT were assessed using wound healing, Transwell, and Western blot assays. The wound healing assay results revealed a smaller wound healing area in the shSCIRT group compared to the sh-NC group (P < 0.05). Similarly, the Transwell assay revealed that A549 cells transfected with sh-SCIRT#1 or sh-SCIRT#2 exhibited significantly reduced migration and invasion compared to the sh-NC group, whereas SCIRT overexpression markedly enhanced the migration and invasion of A549 cells (Fig. 3i-l). Givern that in vitro migration and invasion are often associated with EMT, we further investigated the role of SCIRT in modulating EMT.

We used Western blotting to detect two important metastasis-associated proteins, E-cadherin and N-cadherin. Compared to the sh-NC group, the sh-SCIRT#1 group exhibited a diminished ability to induce EMT (P < 0.05) (Fig. 3m). Although the sh-SCIRT#2 group also demonstrated a reduced capacity to induce EMT, the difference was not significant. Conversely, SCIRT overexpression exerted the opposite effect on H1299 cells, enhancing EMT (Fig. 3n). In summary, these findings indicate that SCIRT facilitates the migration, invasion, and EMT, functioning as a potential oncogene in NSCLC cells.

SCIRT interacts with the splicing factor SFPQ

To further investigate the role of SCIRT in NSCLC progression, FISH analysis was performed to detect the cellular distribution of A549 and H1299 cells using specific cy3-labeled SCIRT probes. U6 and 18S rRNA served as makers for the nucleus and cytoplasm, respectively. The results showed that SCIRT was predominantly localized in the nucleus of NSCLC cells (Fig. 4a). Notably, many nuclear lncRNAs exert their functions through interactions with proteins. Using CatRAPID lncRNA–protein prediction, we identified potential SCIRT-interacting proteins. Among these candidate proteins, SFPQ, an RNA-binding protein crucial for DNA double-strand break repair, emerged as a notable candidate. RNA immunoprecipitation and RNA pull-down assays in A549 cells demonstrated the interaction between SCIRT and SFPQ (Fig. 4b and c). To determine the region of SCIRT that binds to SFPQ, a series of truncated SCIRT constructs were generated. Mapping experiments revealed that the residues 801–1600 of SCIRT were required for its interaction with SFPQ (Fig. 4d). Moreover, SFPQ was found to be highly expressed in NSCLC cells (Fig. 4e).

Fig. 4

SCIRT interacts with splicing factor SFPQ. (a) FISH analysis was performed to detect the cellular distribution of SCIRT in A549 and H1299 cells (Scale bar, 50 μm). (b, c) RNA immunoprecipitation and RNA pull-down assays in A549 cells confirmed the interaction between SCIRT and SFPQ; IgG was used as control. (d) RNA pull-down assay and western blotting were used to assess the interaction between SFPQ and SCIRT truncations. Input and antisense were used as control. (e) The expression levels of SFPQ were measured in NSCLC cells (A549, H1299, and H358) and the bronchial epithelial cell line (BEAS-2B). All results were shown with one representative image from three independent experiments. Data are mean ± SD. **P < 0.01

SCIRT is involved in homologous recombination repair

Numerous genes and proteins are directly or indirectly implicated in DNA damage and repair pathways, contributing to either pro- or anti-tumor effects in NSCLC. SFPQ, a DNA damage response RNA-binding protein (DRBP), plays pro-apoptotic and other critical roles in cancer. We aimed to explore the correlations between SCIRT and HR repair. In this context, we examined γH2AX, a well-known marker for DSBs and Rad51, a widely recognized marker for HR repair. A549 cells were treated with CPT for 12 h and allowed to recover for the indicated durations. γH2AX expression was assessed via Western blotting (Fig. 5a and b). At the 1-hour recovery point, γH2AX expression was similar between Vector, OESCIRT, shNC and shSCIRT A549 cells. However, after 12 h of recovery, γH2AX expression elevated in OESCIRT A549 cells compared with Vector group, while it was significantly reduced in the shSCIRT group compared with shNC group. Immunofluorescence experiments revealed that SCIRT overexpression increased the number of γH2AX foci compared to the control, while SCIRT knockdown markedly reduced them (Supplementary Fig. S2). The Rad51 foci formation dropped significantly after 12 h recovery in OESCIRT group compared to Vector group, and SCIRT knockdown markedly increased them (Fig. 5c).

The sustained activation of γH2AX and decrease Rad51 foci formation in OESCIRT cells 12 h post-CPT recovery strongly suggested that the HR repair process was impaired after SCIRT overexpression. In addition, the presence of a high number of spontaneous γH2AX foci and reduced Rad51 foci formation in OESCIRT A549 cells supported a positive correlation between SCIRT and HR repair.

Fig. 5

Prolonged γH2AX activation and decreased Rad51 foci formation in OESCIRT A549 cells indicated the impairment of homologous recombination repair. (a) A549 cells in the Vector and OESCIRT groups were treated with 18 µmol/L CPT for 12 h and recovered for the indicated times, and the expression of γH2AX was analyzed by western blotting. (b) A549 cells in shNC and shSCIRT groups were treated with 18 µmol/L CPT for 12 h and recovered for the indicated times, and the expression of γH2AX was analyzed by western blotting. (c) Immunofluorescence experiments were used to measure the Rad51 foci in shNC, shSCIRT, Vector, and OESCRT A549 cells (Scale bar, 5 μm). All results were shown with one representative image from three independent experiments. Data are mean ± SD. **P < 0.01, ***P < 0.001

SFPQ reverses the influence of SCIRT on A549 cell proliferation, migration, and invasion after DNA damage

To elucidate the functional significance of their interaction, specific shRNAs targeting SFPQ (shSFPQ) and SFPQ overexpression vectors (pcDNA3.1(+)-SFPQ) were transfected into OESCIRT A549 cells and shSCIRT A549 cells, respectively. Western blot analysis confirmed the successful knockdown and overexpressed of SFPQ (Fig. 6a). qRT‒PCR analysis revealed that the knockdown or overexpression of SFPQ did not alter SCIRT expression levels (data not shown). However, Western blot results indicated that SFPQ expression levels could be effectively restored by its knockdown or overexpression in SCIRT-overexpressing or SCIRT-knockdown A549 cells.

Fig. 6

SFPQ reversed the influence of SCIRT on A549 cell proliferation, migration and invasion after DSBs through PI3K/Akt pathway. (a) The expression levels of SFPQ were verified in indicated A549 cells with or without OESFPQ transfected with shSCIRT or vector and with or without shSFPQ transfected with OESCIRT or vector. (b) CCK-8 assay of shSCIRT, shSCIRT + OESFPQ and Vector A549 cells. (c) CCK-8 assay of OESCIRT, OESCIRT + shSFPQ and Vector A549 cells. (d) Representative images and quantification of the colony formation assay in shSCIRT, shSCIRT + OESFPQ and Vector A549 cells. (e) Representative images and quantification of the colony formation assay in OESCIRT, OESCIRT + shSFPQ and Vector A549 cells. (f) Representative images and quantification of wound-healing assays in shSCIRT, shSCIRT + OESFPQ and Vector A549 cells (Scale bar, 100 μm). (g) Representative images and quantification of wound-healing assays in OESCIRT, OESCIRT + shSFPQ and Vector A549 cells (Scale bar, 100 μm). (h) Representative images and quantification of transwell migration and invasion assays in shSCIRT, shSCIRT + OESFPQ and Vector A549 cells (Scale bar, 100 μm). (i) Representative images and quantification of transwell migration and invasion assays in OESCIRT, OESCIRT + shSFPQ and Vector A549 cells (Scale bar, 100 μm). (j) Expression levels of SFPQ, phosphorylation levels of PI3K (p-PI3K) and Akt (p-Akt) in shSCIRT, shSCIRT + OESFPQ and Vector A549 cells. (k) Expression levels of SFPQ, phosphorylation levels of PI3K (p-PI3K) and Akt (p-Akt) in OESCIRT, OESCIRT + shSFPQ and Vector A549 cells. All results were shown with one representative image from three independent experiments. Data are mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001

We further investigated the mechanism through which SFPQ and SCIRT contribute to the promotion of NSCLC growth, proliferation, invasion, and migration under DNA damage conditions. Following treatment with CPT (18 µmol/L) for 12 h, CCK-8 and colony formation assays indicated that the promotion of A549 cell growth and proliferation by SCIRT overexpression was attenuated after SFPQ knockdown (Fig. 6c and e). Conversely, the inhibitory effect of SCIRT knockdown on A549 cell proliferation were alleviated by SFPQ overexpression (Fig. 6b and d). Wound healing and Transwell assays further confirmed that SCIRT knockdown suppressed A549 cell migration and invasion, the effects were reversed upon SFPQ overexpression (Fig. 6f and h). Conversely, SCIRT overexpression combined with SFPQ knockdown yielded the opposite results (Fig. 6g and i). Overall, our data suggested that the oncogenic functions of SCIRT in NSCLC cells were critically dependent on the expression of SFPQ following DNA damage.

SCIRT responds to DNA double-strand breaks through the PI3K/Akt pathway and is required for SFPQ function

The findings above indicated that SCIRT was involved in HR repair. However, the mechanism by which SCIRT responds to DSBs remains unclear. To explore the pathway by which SCIRT mediated its response to DSBs, pcDNA3.1(+), shSCIRT, and shSCIRT + OESFPQ A549 cells were subjected to CPT treatment for 12 h. As shown in Fig. 6j and k, SCIRT overexpression led to increased phosphorylation of Akt (p-Akt) and PI3K (p-PI3K) compared to the empty vector group, but this effect was attenuated following SFPQ knockdown. Conversely, SCIRT knockdown reduced Akt and PI3K phosphorylation, which was reversed upon SFPQ overexpression. Collectively, these results suggested that SCIRT responds to DSBs through the PI3K/Akt pathway and relied on its interaction with SFPQ to exert its function.