Fig 2: RelB and IL‐8 reciprocity in promoting PCa expansibility. (A) RelB and IL‐8 were ectopically expressed in LNCaP cells, and no transfection served as a control. (B) The concentrations of IL‐8 in established cell lines were measured using a relative ELISA kit (n = 4). (C, D) Cell mineralization in RelB‐ or IL‐8‐overexpressed LNCaP cells was quantified with equal cell numbers (upper panel). C4‐2B cell line, an LNCaP cell‐derived bone metastatic lineage, serves as a positive control (n = 6). (E) Cell migration in the established LNCaP cell lines (n = 20). (F, G) Cell invasion in the established LNCaP cell lines (n = 6). (H) Expression of EMT‐associated proteins in the established LNCaP cell lines. (I) Scheme of orthotopic implantation of RelB‐ or IL‐8‐overexpressed LNCaP cells into nude mice prostates. (J, K) Formed tumours in mouse prostates were imaged by IVIS (n = 4). Data are shown as the mean ± SD. *p < .05, **p < .01, ***p < .001, ns, no significance; one‐way ANOVA.

Fig 3: The mRNA expression profiles in RelB‐silenced PC‐3 cells. (A, B) After silencing RelB in PC‐3 cells (A), the mRNA expression profiles were examined by RNA sequencing. IL‐8 (CXCL8) and S100A4 are indicated by green arrows (B). (C–E) The mRNA profiles were analyzed by KEGG pathway enrichment, assembling in cytokine (C), cell adhesion (D), and calcium signalling (E). (F) The correlation between S100A4 and RelB or RelA in PCa was analyzed using the Oncomine database. (G, H) S100A4 mRNA and protein levels in PCa cell lines vs. a normal cell line (PZ‐HPV‐7). (I, J) S100A4 mRNA and protein levels in RelB‐ or IL‐8‐overexpressed LNCaP cells. Data are shown as the mean ± SD. n = 4, *p < .05, ***p < .001, ns, no significance; one‐way ANOVA).

Fig 4: Sp1‐RelB cooperation in regulating SNAL1 and TWIST1 genes. (A) Sp1 and NF‐κB binding sites were identified in a 5′ flanking region of the human Snall 1 and Twist 1 genes. PCR primer sets were designed to amplify DNA fragments containing the binding sites, the sequence numbers are according to the transcriptional start site as +1. (B) PLT suppressed the Sp1 and NF‐κB binding activity in PC‐3 cells. (C–E) The Sp1 and NF‐κB interaction in PC‐3 cells was analyzed using Duolink PLA assay. The interaction of p50:RelA and p52:RelB dimers served as positive controls, and PLT was used as an inhibiter (C). RelA‐Sp1 and RelB‐Sp1 interactions were determined at 24 and 48 h (D). PLT inhibited the RelB‐Sp1 interaction (E). (F‐I) The Sp1 and NF‐κB binding sites were precipitated from chromatins derived from RelB‐overexpressed LNCaP cells by ChIP using a RelB antibody. Chromatins without pulldown served as input controls, pulldown with IgG served as negative antibody controls, and PLT served as an inhibiter. Sp1 and RelB proteins in the precipitated complexes were quantified by immunoblotting using specific antibodies, and IgG served as a loading control (F). Precipitated Sp1 and NF‐κB binding sites were amplified by regular PCR (G) and qPCR (H, I) with specific primers shown in (A) and Table S1. Data are shown as the mean ± SD. n = 4, **p < .01, ***p < .001; t‐text (B, D, E), one‐way ANOVA (H, I). Scale bar, 20 µm in (C).

Fig 5: RelB‐upregulated S100A4 in promoting PCa aggressiveness. (A, B) The correlation of S100A4 with Gleason score (GS) of PCa patients (n = 6). (C–E) TNF‐α increased the levels of nuclear RelB and cytosolic S100A4 in LNCaP cells. (F) The effect of S100A4 on LNCaP cell migration (n = 4). (G, H) The effect of S100A4 on LNCaP cell invasion (n = 4). (I, J) The S100A4‐overexpressed LNCaP cells were orthotopically implanted into nude mice prostates to examine the effect of S100A4 on tumour formation (n = 4). (K) The S100A4 enhancer was linked to the luciferase reporter. (L‐N) The reporter responses in RelB‐ and IL‐8‐overexpressed LNCaP cells (L), RelB‐ and IL‐8‐silenced PC‐3 cells (M), and TNF‐α‐ and PTL‐treated PC‐3 cells (N) (n = 4). (O, P) The RelB‐binding site was mutated to eliminate the reporter response in PC‐3 cells (n = 4). Data are shown as the mean ± SD. **p < .01, ***p < .001, ns, no significance; t‐test (E, J), one‐way ANOVA (B, F, H, L–N, P). Scale bar, 50 µm in (A) and 20 µm in (D).