By interrogating public cancer databases, we found that B3GNT3, a member of the β-1,3-N-acetylglucosaminyltransferase (β3GnT) family [8], is aberrantly expressed in multiple human malignancies, including lung cancer, pancreatic cancer, breast cancer, and cervical cancer. In this study, analysis of TCGA dataset revealed that B3GNT3 expression was elevated in tumor tissues compared with matched adjacent normal tissues in the majority of NSCLC cases. Specifically, among 104 paired LUAD and LUSC samples, 90 cases exhibited higher B3GNT3 expression in tumor tissues (Fig. 1A and B). To evaluate the clinical relevance of B3GNT3 expression, we performed Kaplan–Meier survival analyses using follow-up data from the TCGA-LUAD and TCGA-LUSC cohorts. Results showed that elevated B3GNT3 expression was significantly associated with worse overall survival (Fig. 1C) and disease-free survival (Fig. 1D) in NSCLC patients.
To further validate the above findings, 12 pairs of frozen LUAD tissues and matched adjacent normal tissues were obtained from the Tissue Bank. Quantitative real-time PCR (qRT-PCR) analysis showed that B3GNT3 expression was significantly upregulated in tumor tissues in 10 out of 12 cases, using a tumor-to-normal ratio greater than 1.5 as the cutoff for overexpression (Fig. 1E). Consistently, western blot analysis of six paired specimens revealed that B3GNT3 protein levels were markedly increased in tumor tissues in five cases compared with their corresponding normal counterparts (Fig. 1F). Immunohistochemical staining of paired tumor and adjacent normal tissue sections further confirmed that B3GNT3 expression was substantially higher in LUAD tissues (Fig. 1G).
Collectively, these data demonstrate that B3GNT3 is frequently upregulated in NSCLC at both the mRNA and protein levels, and its high expression is associated with poor clinical prognosis, suggesting that B3GNT3 may play a functional role in NSCLC progression.
3.2 B3GNT3 promotes LUAD cell growth and proliferationWe examined B3GNT3 expression in human lung cancer cell lines. Both qRT-PCR and western blot analyses showed that B3GNT3 expression was significantly higher in A549 and H2126 cells than in PC9 cells (Fig. 1H and I). To investigate the functional role of B3GNT3 in lung cancer cells, we ectopically overexpressed B3GNT3 in PC9 cells, and knock it out in A549 cells (Fig. 1J).
CCK-8 cell proliferation assays revealed that enforced expression of B3GNT3 markedly enhanced the proliferative capacity of PC9 cells (Fig. 2A). Consistently, soft agar colony formation assays demonstrated that B3GNT3 overexpression significantly increased anchorage-independent growth in PC9 cells (Fig. 2B). Flow cytometric analysis of cell cycle distribution further showed that B3GNT3 overexpression led to a pronounced accumulation of cells in the S phase (Fig. 2C). In line with this observation, immunoblotting analysis revealed that the expression levels of key cell cycle regulators, including Cyclin D1, CDK4, and E2F1, were substantially upregulated following B3GNT3 overexpression (Fig. 2D). In vivo subcutaneous xenograft experiments demonstrated that PC9 cells overexpressing B3GNT3 formed significantly larger tumors with increased tumor growth rates compared with control cells in nude mice (Fig. 2E-G).
Fig. 1
B3GNT3 is upregulated in NSCLC and predicts poor prognosis. (A, B) Analysis of TCGA-LUAD and TCGA-LUSC datasets showing B3GNT3 mRNA expression levels in paired NSCLC tissues and adjacent normal tissues (n = 104). (C, D) Overall survival (C) and disease-free (D) survival analysis of NSCLC patients stratified by high and low B3GNT3 expression in the TCGA-LUAD and TCGA-LUSC cohorts. (E) qRT-PCR analysis of B3GNT3 mRNA levels in 12 paired frozen LUAD tissues and matched adjacent normal tissues; a tumor-to-normal ratio > 1.5 was defined as overexpression. (F) Western blot analysis of B3GNT3 protein expression in 6 paired frozen LUAD and adjacent normal tissue samples. (G) Representative immunohistochemical staining images showing B3GNT3 expression in LUAD tissues and matched adjacent normal tissues. (H, I) qRT-PCR (H) and western blot (I) analyses of B3GNT3 expression in human lung cancer cell lines PC9, A549, and H2126. (J) Validation of B3GNT3 overexpression in PC9 cells and B3GNT3 knockout in A549 cells by western blot analysis
Conversely, B3GNT3 knockout in A549 cells markedly suppressed cell proliferation (Fig. 3A), and significantly impaired colony-forming ability in soft agar (Fig. 3B). Cell cycle analysis revealed a significant accumulation of cells in G0/G1-phase and a significant reduction of cells in G2/M-phase cells upon B3GNT3 knockout (Fig. 3C). Correspondingly, the protein levels of Cyclin D1, CDK4, and E2F1 were markedly decreased in B3GNT3-depleted cells (Fig. 3D). Consistent with in vitro findings, in vivo subcutaneous xenograft experiments showed that B3GNT3 knockout significantly attenuated tumor growth of A549 cells in nude mice (Fig. 3E-G).
Collectively, these gain- and loss-of-function studies demonstrate that B3GNT3 plays a critical role in promoting LUAD proliferation and tumor growth both in vitro and in vivo.
Fig. 2
B3GNT3 promotes LUAD proliferation in vitro and tumor growth in vivo. (A) Cell proliferation of PC9 cells overexpressing B3GNT3 or vector control assessed by CCK-8 assays. (B) Anchorage-independent growth of PC9 cells evaluated by soft agar colony formation assays. (C) Flow cytometric analysis of cell cycle distribution showing increased S-phase population upon B3GNT3 overexpression. (D) Western blot analysis of cell cycle-related proteins (Cyclin D1, CDK4, and E2F1) in PC9 cells with B3GNT3 overexpression. (E–G) Subcutaneous xenograft assays showing representative tumor images (E), tumor growth curves (F), and tumor weights (G) from nude mice injected with PC9 cells overexpressing B3GNT3 or control. Data are presented as mean ± SEM
Fig. 3
Knockout of B3GNT3 suppressed LUAD proliferation in vitro and tumor growth in vivo. (A) CCK-8 assays showing significantly reduced proliferation of A549 cells following B3GNT3 knockout. (B) Soft agar colony formation assays demonstrating decreased anchorage-independent growth upon B3GNT3 depletion. (C) Flow cytometric analysis of cell cycle distribution indicating a significant accumulation of cells in G0/G1-phase and a significant reduction of cells in G2/M-phase upon B3GNT3 knockout. (D) Western blot analysis of Cyclin D1, CDK4, and E2F1 expression in control and B3GNT3- knockout cells. (E–G) Subcutaneous xenograft experiments showing representative tumor images (E), tumor growth curves (F), and tumor weights (G) from nude mice injected with control or B3GNT3-knockout A549 cells. Data are presented as mean ± SEM
3.3 B3GNT3 enhances the invasive and metastatic potential of LUADTo evaluate the role of B3GNT3 in LUAD invasion and metastasis, Matrigel-coated Transwell assays were performed. As shown in Fig. 4A, ectopic expression of B3GNT3 significantly increased the invasive capacity of PC9 cells. In contrast, genetic ablation of B3GNT3 markedly impaired the invasive ability of A549 cells (Fig. 4B).
Given that resistance to anoikis is a critical prerequisite for metastatic dissemination, we further assessed the effect of B3GNT3 on anoikis sensitivity. Anchorage-independent culture assays revealed that B3GNT3 overexpression significantly enhanced the resistance of PC9 cells to anoikis-induced apoptosis (Fig. 4C). Conversely, loss of B3GNT3 rendered A549 cells more susceptible to anoikis under non-adherent conditions (Fig. 4D).
Collectively, these findings indicate that elevated B3GNT3 expression promotes an aggressive phenotype in LUAD by enhancing invasive capacity and conferring resistance to anoikis, thereby facilitating metastatic potential.
Fig. 4
B3GNT3 enhances invasion and anoikis resistance in LUAD. (A) Matrigel-coated Transwell invasion assays showing increased invasive capacity of PC9 cells overexpressing B3GNT3. (B) Matrigel-coated Transwell invasion assays showing reduced invasion of A549 cells following genetic ablation of B3GNT3. (C) Anchorage-independent culture assays showing enhanced anoikis resistance in PC9 cells with B3GNT3 overexpression. (D) Anchorage-independent culture assays showing increased sensitivity to anoikis in B3GNT3-deficient A549 cells. Quantitative data are shown as mean ± SEM
3.4 B3GNT3 is associated with EMT- and NF-κB signaling-related transcriptional programs in LUADTo elucidate the signaling pathways regulated by B3GNT3 that contribute to LUAD progression, we performed RNA sequencing in PC9 cells ectopically expressing B3GNT3 and corresponding control cells (Online Resource 2). GSEA analysis revealed that elevated B3GNT3 expression was significantly associated with EMT-related gene signatures (Fig. 5A), as well as activation of NF-κB signaling pathways (Fig. 5B).
To further characterize the transcriptional programs induced by B3GNT3, Gene Ontology (GO) analysis was conducted on genes significantly upregulated upon B3GNT3 overexpression (fold change ≥ 2, P ≤ 0.05). This analysis demonstrated that B3GNT3-regulated genes were predominantly enriched in biological processes related to extracellular matrix organization, cell-cell adhesion, and cell-matrix interactions (Fig. 5C), pathways that are closely linked to EMT regulation. Subsequently, qRT-PCR was performed to validate representative EMT-associated genes identified from the RNA sequencing dataset, including COL6A3, LOX, WNT5A, VIM, and MMP2. Notably, enforced expression of B3GNT3 resulted in a marked upregulation in the expression levels of these EMT-related genes (Fig. 5D).
Consistent with these transcriptional changes, morphological examination revealed that enforced expression of B3GNT3 induced a pronounced phenotypic shift in PC9 cells. Whereas control cells exhibited a typical epithelial morphology characterized by a rounded shape and tight intercellular contacts, B3GNT3-overexpressing cells displayed a scattered, spindle-like appearance indicative of mesenchymal transition (Fig. 5E).
Immunofluorescence analysis further demonstrated that overexpression of B3GNT3 significantly reduced the expression of the epithelial marker E-cadherin in PC9 cells (Fig. 5F). In agreement with these observations, western blot analysis confirmed a decrease in E-cadherin protein levels accompanied by a concomitant increase in the mesenchymal markers N-cadherin and vimentin following B3GNT3 overexpression (Fig. 5G).
Collectively, these results demonstrate that elevated B3GNT3 expression induces EMT-associated transcriptional programs and phenotypic changes in LUAD, thereby promoting a more invasive and aggressive cellular state.
3.5 B3GNT3 interacts with NFKB2 and activates non-canonical NF-κB signaling in LUADTo elucidate the molecular mechanisms underlying B3GNT3-mediated regulation of LUAD proliferation and EMT, we performed immunoprecipitation followed by LC-MS/MS analysis to identify candidate B3GNT3-interacting proteins (Online Resource 3). The top ten candidate interactors were subsequently subjected to STRING network analysis to explore potential functional associations (Fig. 6A). Gene Ontology enrichment analysis further demonstrated significant enrichment in pathways related to non-canonical NF-κB signaling, inflammatory responses, and transcriptional regulation (Fig. 6B), thereby providing functional context linking the B3GNT3 interactome to the biological processes implicated in our phenotypic observations. In addition, we performed single-gene GSEA analysis in the TCGA-LUAD cohort to evaluate the correlation between B3GNT3 expression and established gene expression signatures [17] of canonical NF-κB1 (Fig. 6C) and non-canonical NFKB2 (Fig. 6D) pathways. The results demonstrated that elevated B3GNT3 mRNA levels were significantly associated with enrichment of both canonical and non-canonical NF-κB target gene sets, with a particularly strong correlation observed for the NFKB2-dependent non-canonical pathway. These findings provide further support for the functional cooperation between B3GNT3 and NFKB2 signaling in NSCLC.
The interaction between B3GNT3 and NFKB2 was subsequently validated by co-immunoprecipitation assays. In PC9 cells ectopically expressing Flag-tagged B3GNT3, endogenous p100, the precursor form encoded by NFKB2, was readily detected in B3GNT3 immunoprecipitates (Fig. 6E). Conversely, immunoprecipitation of HA-tagged p100 efficiently pulled down B3GNT3 (Fig. 6F). These results confirm that B3GNT3 physically interacts with NFKB2 in PC9.
Fig. 5
B3GNT3 is associated with EMT- and NF-κB signaling-related transcriptional programs in LUAD. (A, B) GSEA showing significant enrichment of EMT-related (A) and NF-κB signaling-related (B) gene signatures in PC9 cells overexpressing B3GNT3. (C) GO analysis of genes upregulated upon B3GNT3 overexpression, highlighting enrichment in extracellular matrix organization and cell adhesion-related processes. (D) qRT-PCR validation of representative EMT-associated genes (COL6A3, LOX, WNT5A, VIM, and MMP2) following B3GNT3 overexpression. (E) Representative phase-contrast images showing morphological changes consistent with EMT in PC9 cells overexpressing B3GNT3. (F) Immunofluorescence staining of E-cadherin in PC9 cells with B3GNT3 overexpression. (G) Western blot analysis of epithelial and mesenchymal markers (E-cadherin, N-cadherin, and vimentin) in PC9 cells. Quantitative data are presented as mean ± SEM
To determine whether the interaction between B3GNT3 and p100 depends on its glycosyltransferase activity, we generated a catalytic-dead mutant of B3GNT3 (D297A) targeting a conserved residue within the predicted catalytic domain. Using the UMP/CMP-Glo Glycosyltransferase Assay, we confirmed that the D297A mutant was enzymatically inactive (Fig. 6G). Co-immunoprecipitation assays showed that the B3GNT3 (D297A) mutant retained the ability to interact with p100 with similar binding affinity to wild-type B3GNT3 (Fig. 6H).
To further evaluate whether B3GNT3 enzymatic activity is required for activation of non-canonical NF-κB signaling, we examined p100 phosphorylation and processing in PC9 cells overexpressing either wild-type B3GNT3 or the D297A mutant. Western blot analysis revealed that enforced expression of both constructs markedly increased p100 phosphorylation and enhanced the generation of the processed active form p52 (Fig. 6I). Collectively, these results indicate that B3GNT3 promotes NFKB2 processing and activation of the non-canonical NF-κB pathway independently of its glycosyltransferase activity.
Consistent with enhanced NFKB2 processing, nuclear-cytoplasmic fractionation analysis demonstrated increased nuclear accumulation of p52 in cells overexpressing either wild-type B3GNT3 or the catalytic-dead mutant B3GNT3 (D297A) (Fig. 6J and K). These findings were further supported by immunofluorescence staining, which revealed a pronounced increase in nuclear localization of p52 following overexpression of both constructs (Online Resource 5 Fig. S1A, S1B). Functionally, EMT marker analysis showed that B3GNT3 (D297A) promoted EMT-associated protein expression in PC9 cells to a similar extent as wild-type B3GNT3 (Online Resource 5 Fig. S1C). Consistently, Transwell invasion assays demonstrated that the D297A mutant enhanced the invasive capacity of PC9 cells comparably to wild-type B3GNT3 (Online Resource 5 Fig. S1D).
Taken together, these results indicate that B3GNT3 facilitates p100 phosphorylation, processing to p52, and activation of the non-canonical NF-κB pathway in LUAD independently of its enzymatic activity.
Fig. 6
B3GNT3 interacts with NFKB2 and regulates cell cycle and EMT gene expression. (A, B) STRING interaction network (A) and functional enrichment analysis (B) of B3GNT3-interacting proteins identified by LC-MS/MS, highlighting a prominent association with the non-canonical NF-κB signaling pathway. (C, D) GSEA analysis showing significant enrichment of NF-κB 1(C) and NF-κB2 (D) signaling-related gene signatures in PC9 cells overexpressing B3GNT3. (E) Co-immunoprecipitation showing interaction between Flag-tagged B3GNT3 and endogenous p100 (NFKB2) in PC9 cells. (F) Reciprocal co-immunoprecipitation demonstrating association between HA-tagged p100 and B3GNT3. (G) Glycosyltransferase activity assay of wild-type B3GNT3 and the catalytic-dead mutant B3GNT3 (D297A) using the UMP/CMP-Glo Glycosyltransferase Assay, confirming loss of enzymatic activity in the D297A mutant. (H) Co-immunoprecipitation analysis showing that the catalytic-dead mutant B3GNT3 (D297A) retains the ability to interact with p100 (NFKB2), with comparable binding to wild-type B3GNT3 in PC9 cells. (I) Western blot analysis of p100 phosphorylation and p52 generation in PC9 cells overexpressing B3GNT3 or B3GNT3 (D297A). (J) Nuclear-cytoplasmic fractionation analysis showing increased nuclear accumulation of p52 following B3GNT3 or B3GNT3 (D297A) overexpression. Wcl, whole cell lysate. Nuc, nucleus. Cyto, cytoplasm. (K) Quantification of nuclear and cytoplasmic p52 levels shown in (J). Band intensities were first normalized to Wcl input and subsequently calculated as the ratio of nuclear to cytoplasmic fractions
3.6 NFKB2 is required for the tumor-promoting effects of B3GNT3 in LUADTo determine whether B3GNT3 exerts its tumor-promoting effects through NFKB2, NFKB2 was genetically ablated in PC9 cells stably overexpressing B3GNT3 (Fig. 7A). We first examined key cell cycle regulators, including CCND1, MYC, and E2F1. As expected, these genes were markedly upregulated upon B3GNT3 overexpression; however, their elevated expression was significantly attenuated following NFKB2 depletion (Fig. 7B), indicating that B3GNT3-driven cell cycle progression depends on NFKB2. Similarly, the induction of EMT-associated genes by B3GNT3 overexpression was substantially reduced upon NFKB2 knockout (Fig. 7C), suggesting that NFKB2 is required for B3GNT3-mediated EMT-related transcriptional programs.
Conversely, to further validate the functional linkage between B3GNT3 and NFKB2, p52 was ectopically expressed in B3GNT3-deficient A549 cells (gB3#2) (Online Resource. 6 Fig. S2A). Restoration of p52 partially rescued the suppressed expression of cell cycle and EMT-associated markers induced by B3GNT3 loss (Online Resource 6 Fig. S2B, S2C), supporting the notion that NFKB2/p52 functions downstream of B3GNT3.
To assess the in vivo relevance of this axis, subcutaneous xenograft assays were performed using PC9 cells overexpressing B3GNT3 with or without NFKB2 knockout. Consistent with the in vitro findings, NFKB2 depletion significantly reversed the accelerated tumor growth driven by B3GNT3 overexpression, as reflected by reduced tumor volume and tumor weight (Fig. 7D-F).
Functionally, Matrigel-coated Transwell invasion assays demonstrated that the enhanced invasive capacity conferred by B3GNT3 overexpression was markedly impaired upon NFKB2 ablation in PC9 cells (Fig. 7G). In contrary, re-expression of p52 in B3GNT3-deficient A549 cells restored invasive capacity (Online Resource 6 Fig. S2D). Anoikis resistance assays further revealed that NFKB2 loss significantly increased sensitivity to detachment-induced apoptosis in B3GNT3-overexpressing PC9 cells (Fig. 7H).
Collectively, these results establish NFKB2 as a critical downstream effector of B3GNT3 and demonstrate that B3GNT3 promotes LUAD proliferation, invasion, and anoikis resistance through NFKB2-dependent activation of non-canonical NF-κB signaling.
Fig. 7
NFKB2 mediates the tumor-promoting effects of B3GNT3 in vivo and in vitro. (A) Validation of NFKB2 knockout in PC9 cells overexpressing B3GNT3. (B) qRT-PCR analysis of cell cycle-related genes (CCND1, MYC, and E2F1) following NFKB2 depletion. (C) qRT-PCR analysis of EMT-associated genes in PC9 cells with B3GNT3 overexpression and NFKB2 knockout. (D–F) Subcutaneous xenograft assays showing representative tumor images (D), tumor growth curves (E), and tumor weights (F), indicating reversal of B3GNT3-induced tumor growth upon NFKB2 ablation. (G) Matrigel-coated Transwell invasion assays showing that invasive capacity which was enhanced by B3GNT3 overexpression was significantly impaired following NFKB2 knockout. (H) Anoikis resistance assays demonstrating restored sensitivity to detachment-induced apoptosis in NFKB2-deficient, B3GNT3-overexpressing PC9 cells. Data are presented as mean ± SEM
Comments (0)