**Background:** Neutrophils within the tumor microenvironment (TME) can be polarized into anti-tumor N1 or pro-tumor N2 states, but the regulatory mechanisms are poorly defined. TGF-β signaling is known to promote tumor growth and can stimulate a pro-tumor N2 neutrophil phenotype, yet the role of the canonical Smad3 pathway in TAN polarization was unknown. This study investigates TGF-β/Smad3 signaling in tumor-associated neutrophils (TANs) in non-small cell lung carcinoma (NSCLC).
**Methods:** The study used a multi-pronged approach: (1) Human tissue analysis: 72 NSCLC patient tumor samples were analyzed by multiplex immunofluorescence for CD16b (neutrophil marker), iNOS (N1), CD206 (N2), and p-SMAD3. (2) Mouse models: Smad3-WT and Smad3-KO mice (C57BL/6J background) were inoculated subcutaneously with 2×10⁶ Lewis lung carcinoma (LLC) cells. Tumor growth was monitored, and TANs were analyzed by flow cytometry and immunofluorescence. (3) Neutrophil depletion: Anti-Ly6G antibody (clone 1A8, 50 µg/mouse) was given intraperitoneally starting 3 days before LLC inoculation and daily thereafter. (4) Adoptive transfer: 3×10⁶ untouched Smad3-WT or Smad3-KO bone marrow-derived neutrophils (BMDNs) were injected via tail vein on days 15 and 20 post-LLC inoculation. (5) In vitro assays: BMDNs were co-cultured with LLC cells at ratios of 1:5, 1:10, and 1:20 for cytotoxicity (LDH release) and phagocytosis (Dil-labeled LLC uptake) assays. (6) Single-cell RNA-seq: FACS-sorted Ly6G⁺CD11b⁺ TANs from LLC tumors (8 tumors/group pooled) underwent 10x scRNA-seq, yielding 3293 Smad3-WT and 1323 Smad3-KO TANs after filtering. (7) ChIP-seq: BMDNs from Smad3-WT mice were stimulated with LLC conditioned medium (LLC-CM) for 2 hours and assayed for Smad3 binding. (8) Human validation: SMAD3 was silenced in human peripheral blood-derived neutrophils (PBDNs) by siRNA, and these were co-cultured with A549 cells or adoptively transferred into A549-bearing NSG mice. (9) Pharmacological inhibition: Smad3 inhibitor SIS3 was administered daily at 2.5, 5, or 10 µg/g from day 7 post-LLC inoculation.
**Key Results:** In human NSCLC, N2 TANs (CD206⁺CD16b⁺) represented over 50% of total CD16b⁺ cells in tumors versus less than 10% in control lung tissue (P=0.004). SMAD3 activation (p-SMAD3) was prominent in N2 TANs and negatively correlated with N1 TAN proportion (Pearson correlation). Higher N1 TAN abundance (≥10% of area) was associated with better disease-free survival (log-rank test). In mice, Smad3-KO mice showed markedly less tumor growth on day 15 (P=0.0011) with increased neutrophil infiltration (P=0.0002 for TANs in TME) and a switch from predominant N2 to predominant N1 TAN phenotype (N1: P=0.0003; N2: P=0.0002 by immunofluorescence). Neutrophil depletion in Smad3-KO mice abrogated this protection (tumor weight P=0.0003 vs IgG-treated KO), while depletion in WT mice had no effect. Adoptive transfer of Smad3-KO BMDNs significantly inhibited tumor growth in WT mice (tumor weight P=0.0068 vs WT-BMDN transfer), whereas WT-BMDN transfer had no effect. Smad3-KO BMDNs showed enhanced binding to LLC cells (P=0.0021), increased cytotoxicity (P=0.0001 at 1:20 ratio), and increased phagocytosis (P=0.003). scRNA-seq revealed distinct transcriptome profiles: Smad3-KO TANs expressed N1 markers (Tnf, Icam1, Fas), while Smad3-WT TANs expressed N2 markers (Arg1, Ccl2, Vegf-b). RNA velocity and pseudotime analysis showed that N2 is an early TAN state that progressively develops toward N1, with Smad3 blocking this maturation. ChIP-seq identified 64 direct Smad3 target genes under LLC-CM stimulation, including neutrophil development genes (Afdn, Cdon, Elf4enif1, Limd1). In human cells, SMAD3 siRNA knockdown in PBDNs significantly enhanced killing of A549 cells (P=0.0077 at 1:10 ratio) and adoptive transfer of SMAD3-knockdown PBDNs into A549-bearing NSG mice significantly inhibited tumor growth (P=0.0001 vs control). Pharmacological Smad3 inhibition with SIS3 (10 µg/g) suppressed LLC tumor growth (P=0.0001 vs vehicle) and shifted TANs to a predominant N1 phenotype.
**Clinical Implications:** This study identifies Smad3 as a key regulator of TAN polarization in NSCLC, demonstrating that Smad3 signaling maintains TANs in an immature, pro-tumor N2 state, while Smad3 deletion or inhibition allows maturation to an anti-tumor N1 state with potent cytotoxic activity. The finding that SMAD3 activation in TANs negatively correlates with N1 abundance and patient survival in a 72-patient NSCLC cohort supports clinical relevance. Pharmacological Smad3 inhibition with SIS3 effectively suppressed tumor growth and promoted N1 polarization in mice, suggesting Smad3 as a druggable target. Given that TGF-β receptor inhibitors (e.g., Vactosertib) are already in clinical trials for NSCLC (NCT04515979), targeting Smad3 specifically may offer a more targeted approach with fewer side effects than broader TGF-β receptor inhibition. This work provides a rationale for developing neutrophil-based immunotherapies targeting the Smad3 pathway in NSCLC.