Journal: Non-coding RNA Research

Article Title: CircSMAD4 shapes matrix-remodeling TAMs in lung adenocarcinoma

doi: 10.1016/j.ncrna.2026.03.003

Figure Lengend Snippet: circSMAD4 is enriched in LUAD TAMs and is associated with advanced disease and poor prognosis. (A) Workflow for isolating human LUAD TAMs and paired normal tissue-resident macrophages (NTRMs) for circRNA profiling. (B) Heatmap of the top differentially expressed circRNAs between TAMs and NTRMs (z-score logCPM). (C) Volcano plot of circRNA-seq (TAMs vs NTRMs) showing differentially expressed circRNAs. Differential-expression categories were defined as follows: Up (red), log2FC ≥ 1 and FDR <0.05; Down (blue), log2FC ≤ −1 and FDR <0.05; all remaining circRNAs were classified as Normal (gray). (D) RT–qPCR validation of selected circRNA candidates in TAMs versus NTRMs. (E) Independent cohort validation showing higher circSMAD4 expression in TAMs than in NTRMs. (F) Representative images of combined CD68 immunofluorescence (green) and circSMAD4 ISH (red) in LUAD tumor and adjacent normal tissues; nuclei were counterstained with DAPI (blue). Scale bar, 50 μm. (G) Kaplan–Meier analysis of overall survival stratified by circSMAD4 expression in TAMs (high vs low). (H) Schematic of circSMAD4 genomic origin and back-splice junction validation by Sanger sequencing. (I) Divergent-primer PCR showing circSMAD4 detection in cDNA but not gDNA; GAPDH serves as a linear control. (J, L) RNase R digestion assay showing resistance of circSMAD4 relative to linear SMAD4 mRNA in patient-derived TAMs (J) and TC-BMDMs (L). (K, M) Actinomycin D chase assay showing greater stability of circSMAD4 than SMAD4 mRNA in patient-derived TAMs (K) and TC-BMDMs (M). Half-life estimated by one-phase decay (Y0 = 1, Plateau = 0). ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001; ns, not significant.

Article Snippet: Reverse transcription for mRNA/circRNA was performed with HiScript III (Vazyme, Cat# R312-01), and miRNA reverse transcription was performed using the Vazyme miRNA RT system. qPCR was carried out using ChamQ SYBR qPCR Master Mix (Vazyme, Cat# Q331) on a qTOWER3 84 G Real-Time PCR Thermal Cycler. circSMAD4 was quantified using divergent primers spanning the back-splice junction, whereas linear transcripts were detected using convergent primers.

Techniques: Quantitative Proteomics, Quantitative RT-PCR, Biomarker Discovery, Expressing, Immunofluorescence, Sequencing, Control, Derivative Assay

Journal: Non-coding RNA Research

Article Title: CircSMAD4 shapes matrix-remodeling TAMs in lung adenocarcinoma

doi: 10.1016/j.ncrna.2026.03.003

Figure Lengend Snippet: circSMAD4 drives tumor-educated M2-like polarization of macrophages and promotes tumor-cell aggressiveness. (A) Workflow for generating TC-hMDMs and TC-BMDMs, circSMAD4 knockdown, and downstream functional assays. (B) RT–qPCR analysis of M1-associated markers (MHC-II [HLA-DRA in TC-hMDMs; H2-Ab1 in TC-BMDMs], NOS2, and CD86) and M2-associated markers (CD163, CD206, and ARG1) in TC-hMDMs and TC-BMDMs. (C) Representative flow-cytometry histograms for HLA-DR, iNOS, CD86, CD163, CD206, and ARG1 in TC-hMDMs. Gating strategy and marker thresholds were defined based on FMO controls (see ). (D) Flow-cytometry quantification of marker-positive cells in TC-hMDMs and TC-BMDMs. (E) ELISA of IL-10, TGF-β, and iNOS in culture supernatants. (F) CCK-8 assays of A549 and LLC cells. (G) Colony-formation assays of A549 and LLC cells with quantification. (H) Bioluminescence-based growth readouts of patient-derived LUAD organoids (PDO #1 and PDO #2) after co-culture with TC-hMDMs. (I) Immunoblot analysis of EMT-related proteins (E-cadherin, N-cadherin, Vimentin) in A549 and LLC cells. (J) Transwell migration and invasion assays of A549 and LLC cells with quantification. Scale bar, 50 μm. ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001; ns, not significant.

Article Snippet: Reverse transcription for mRNA/circRNA was performed with HiScript III (Vazyme, Cat# R312-01), and miRNA reverse transcription was performed using the Vazyme miRNA RT system. qPCR was carried out using ChamQ SYBR qPCR Master Mix (Vazyme, Cat# Q331) on a qTOWER3 84 G Real-Time PCR Thermal Cycler. circSMAD4 was quantified using divergent primers spanning the back-splice junction, whereas linear transcripts were detected using convergent primers.

Techniques: Knockdown, Functional Assay, Quantitative RT-PCR, Flow Cytometry, Marker, Enzyme-linked Immunosorbent Assay, CCK-8 Assay, Derivative Assay, Co-Culture Assay, Western Blot, Migration

Journal: Non-coding RNA Research

Article Title: CircSMAD4 shapes matrix-remodeling TAMs in lung adenocarcinoma

doi: 10.1016/j.ncrna.2026.03.003

Figure Lengend Snippet: circSMAD4 physically associates with IGF2BP2 in macrophages. (A) LC–MS/MS summary of proteins enriched by circSMAD4 RNA pull-down. (B) Western blot validation of IGF2BP2 in circSMAD4 sense (vs antisense) pull-down from TC-hMDMs. (C) IGF2BP2 RIP–qPCR showing circSMAD4 enrichment over IgG in TC-hMDMs. (D–E) catRAPID prediction and ViennaRNA RNAfold secondary-structure modeling indicating multiple candidate IGF2BP2-binding regions on circSMAD4. (F) Western blot of IGF2BP2 after pull-down with circSMAD4 fragments (1#–3#). (G) Schematic of IGF2BP2 domain architecture and the Flag-tagged truncation/deletion constructs used for mapping circSMAD4 interaction (designed based on catRAPID prediction and annotated RRM/KH domain boundaries). (H) Anti-Flag RIP–qPCR showing circSMAD4 enrichment precipitated by the indicated Flag-tagged IGF2BP2 truncation/deletion constructs (presented as % input and normalized to IgG). (I) Nuclear–cytoplasmic fractionation followed by RT–qPCR showing circSMAD4 distribution and the effect of IGF2BP2 knockdown on the nuclear-to-cytoplasmic ratio of circSMAD4 in TC-hMDMs. Fractionation quality was validated using nuclear/cytoplasmic marker transcripts/proteins. (J) Representative immunofluorescence/ISH images showing circSMAD4 signals and IGF2BP2 staining in macrophages (CD163) with nuclear counterstaining (DAPI). Scale bar, 50 μm. (K–N) qPCR and Western blot showing no reciprocal change in expression between circSMAD4 and IGF2BP2 upon knockdown/overexpression. ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001; ns, not significant.

Article Snippet: Reverse transcription for mRNA/circRNA was performed with HiScript III (Vazyme, Cat# R312-01), and miRNA reverse transcription was performed using the Vazyme miRNA RT system. qPCR was carried out using ChamQ SYBR qPCR Master Mix (Vazyme, Cat# Q331) on a qTOWER3 84 G Real-Time PCR Thermal Cycler. circSMAD4 was quantified using divergent primers spanning the back-splice junction, whereas linear transcripts were detected using convergent primers.

Techniques: Liquid Chromatography with Mass Spectroscopy, Western Blot, Biomarker Discovery, Binding Assay, Construct, Fractionation, Quantitative RT-PCR, Knockdown, Marker, Immunofluorescence, Staining, Expressing, Over Expression

Journal: Non-coding RNA Research

Article Title: CircSMAD4 shapes matrix-remodeling TAMs in lung adenocarcinoma

doi: 10.1016/j.ncrna.2026.03.003

Figure Lengend Snippet: circSMAD4 facilitates IGF2BP2-dependent stabilization of m6A-marked transcripts. (A) Venn diagram intersecting ENCORI-predicted IGF2BP2 targets with DEGs from shIGF2BP2 versus shNC and shcircSMAD4 versus shNC mRNA-seq, identifying shared candidates. (B) MeRIP–qPCR showing m6A enrichment on COL4A1, SPI1, and ACTA2 candidate regions (CRDs) in shNC and shIGF2BP2 cells. (C) IGF2BP2-RIP–qPCR showing IGF2BP2 binding to COL4A1, SPI1, and ACTA2 CRDs in shNC + Vector, shcircSMAD4 + Vector, shNC + IGF2BP2, and shcircSMAD4 + IGF2BP2 groups. (D) Biotin-circSMAD4 pull-down followed by qPCR showing enrichment of COL4A1, SPI1, and ACTA2 CRDs in Vector + shNC, circSMAD4 + shNC, Vector + shIGF2BP2, and circSMAD4 + shIGF2BP2 groups. (E–G) Schematics of m6A-site mutations introduced into COL4A1, SPI1, and ACTA2 reporters. (H–J) Dual-luciferase assays for CRD reporters (WT and m6A-mutant) in Vector, circSMAD4, and IGF2BP2 groups. (K–M) MeRIP–qPCR for WT and m6A-mutant CRD reporters in Vector, circSMAD4, and IGF2BP2 groups. (N–P) mRNA decay assays of endogenous COL4A1, SPI1, and ACTA2 following circSMAD4 knockdown with Vector or IGF2BP2 overexpression. Half-life estimated by one-phase decay (Y0 = 1, Plateau = 0). (Q–S) mRNA decay assays of endogenous COL4A1, SPI1, and ACTA2 following circSMAD4 overexpression with shNC or shIGF2BP2. Half-life estimated by one-phase decay (Y0 = 1, Plateau = 0). ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001; ns, not significant.

Article Snippet: Reverse transcription for mRNA/circRNA was performed with HiScript III (Vazyme, Cat# R312-01), and miRNA reverse transcription was performed using the Vazyme miRNA RT system. qPCR was carried out using ChamQ SYBR qPCR Master Mix (Vazyme, Cat# Q331) on a qTOWER3 84 G Real-Time PCR Thermal Cycler. circSMAD4 was quantified using divergent primers spanning the back-splice junction, whereas linear transcripts were detected using convergent primers.

Techniques: Binding Assay, Plasmid Preparation, Luciferase, Mutagenesis, Knockdown, Over Expression