SerpinB7 promotes c-Myc-mediated glycolysis by interacting with ANXA2 to facilitate the progression of non-small cell lung cancer Authors: Fulei Zhao, Fanlian Zeng, Pei Zhou, Huaping Zheng, Yue Hu, Jie Tang, Yawen Hu, Guolin Li, Yuting Feng, Xiaoyan Wang, Jiadong Yu, Ya Li, Chengcheng Yue, Xiao Liu, Tingting Zhang, Nongyu Huang, Xiaochi Sun, Xinmeng Wang, Shishi Huang, Mingxiang He, Xinyu Zeng, Wenling Wu, Xinai Cui, Kaijun Cui, Jiong Li Published: 2026 ABSTRACT Metabolic reprogramming, particularly aberrant glycolysis, is a hallmark of NSCLC progression; however, the regulatory mechanisms remain incompletely understood. Here, we identify SerpinB7, an endogenous serine protease inhibitor, as a novel oncogenic regulator of glycolytic reprogramming in NSCLC. SerpinB7 is significantly overexpressed in NSCLC tissues and correlates with poor prognosis. Mechanistically, SerpinB7 functions in a non-canonical, protease-independent manner by directly binding to Annexin A2 (ANXA2), protecting it from NEDD4L-mediated ubiquitination and degradation, thereby stabilizing ANXA2 protein levels.As an RNA-binding protein, ANXA2 interacts with c-Myc via its C-terminal domain, enhancing c-Myc driven transcription of glycolytic genes. This establishes a previously unrecognized SerpinB7-ANXA2-c-Myc regulatory axis that promotes glycolytic activation and malignant transformation in NSCLC cells. Importantly, immunohistochemical analysis of clinical specimens confirms the co-overexpression of SerpinB7 and ANXA2 in tumors, underscoring the clinical relevance of this pathway. Targeting SerpinB7 using siRNAs encapsulated in ferritin protein shells effectively suppresses NSCLC tumor growth and metastasis. These findings reveal SerpinB7 as a tumor-specific modulator of glycolysis and a potential metabolic target in NSCLC.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12943-026-02684-9. Introduction Lung cancer remains the leading cause of cancer-related deaths worldwide [1], with a five-year survival rate of less than 20% [2]. Non-small cell lung cancer (NSCLC) accounts for approximately 85% of all lung cancer cases, with its main subtypes including lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). Despite advances in targeted therapy and immunotherapy, lung cancer patients continue to face challenges such as poor prognosis, lack of therapeutic targets, drug resistance,and tumor heterogeneity [3–5]. Therefore, elucidating the molecular mechanisms underlying NSCLC progression and identifying novel potential therapeutic targets have become urgent goals. In recent years, emerging concepts have established metabolic reprogramming as a key hallmark of lung cancer [6]. Cancer cells primarily rely on glycolysis for glucose metabolism reprogramming, even under aerobic conditions. Aerobic glycolysis (the Warburg effect) has been shown to be a major consequence of oncogenic drivers [7]. The mechanisms underlying active glycolysis in tumor cells are multifaceted, involving the activation of oncogenes and signaling pathways, inactivation of tumor suppressors, altered expression and activity of metabolic enzymes, and influences from the tumor microenvironment [8]. Oncogenes such as c-Myc, mTOR and hypoxia-inducible factor 1α (HIF-1α) enhance glycolytic activity to meet the anabolic demands of highly proliferating cancer cells [9–11]. Under normoxic conditions, c-Myc acts as a “master regulator” of glycolysis and tumor proliferation [12], transcriptionally upregulating genes encoding glucose transporter 1 (GLUT1), lactate dehydrogenase A (LDHA), hexokinase 2 (HK2) and pyruvate kinase isozyme M2 (PKM2), thereby promoting glycolytic activity by increasing glucose uptake and accelerating the conversion of glucose to lactate [13, 14]. Previous strategies aimed at directly inhibiting key metabolic enzymes have often shown limited efficacy and are prone to systemic toxicity [15]. Therefore, the key to lung cancer treatment lies in identifying critical molecules (such as oncogenes, signaling proteins and transcription factors) that specifically regulate tumor metabolic reprogramming,rather than targeting metabolic enzymes themselves.The integration of sequencing projects like The Cancer Genome Atlas (TCGA) with multi-omics technologies has enabled the identification of oncogenes with profound and sustained impact through large-scale bioinformatics and omics-based screening. Serine proteases are a class of proteolytic enzymes characterized by a serine residue in their active site, which selectively or non-specifically hydrolyze peptide bonds in substrate proteins [16]. Endogenous serine protease inhibitors (Serpins) are natural protein molecules that inhibit the proteolytic activity of serine proteases by binding to their active sites [17, 18]. Some studies suggest that serpins may also regulate protease function through interactions outside the reactive center loop (RCL), influencing substrate binding and downstream signaling [19, 20]. As key endogenous regulatory molecules in biological processes, the study of serine proteases and their inhibitors not only deepens our understanding of proteolysis and protein interaction networks but also provides new molecular targets for the diagnosis and treatment of various diseases, holding significant theoretical value and broad clinical application prospects. SerpinB7 (Serine protease inhibitor clade B member 7) is a member of the serpin superfamily. A study published in PNAS reported that ALK fusions in NSCLC lead to increased SerpinB7 expression [21]. Additionally, SerpinB7 has been identified as a potential biomarker for intravenous vitamin C therapy in NSCLC [22]. However, the exact expression pattern, functional role, and underlying mechanisms of the endogenous serine protease inhibitor SerpinB7 in NSCLC remain unclear. Its downstream targets and contributions to malignant tumor phenotypes require further elucidation. Using TMT-based proteomics, GST pull-down combined with immunoprecipitation (IP), and mass spectrometry, we identified ANXA2 as a protein interacting with SerpinB7. ANXA2 is a 36 kDa calcium-dependent phospholipid-binding protein belonging to the annexin family. It is widely expressed in various tissues and interacts with multiple ligands. Its structure is characterized by a highly conserved “annexin core” composed of four homologous domains [23]. Previous studies have reported that high expression of ANXA2 in pancreatic ductal adenocarcinoma is associated with metabolic alterations such as DNA repair and glycolysis, as well as poor prognosis [24]. Furthermore, research has shown that ANXA2 undergoes post-translational modifications (PTMs), such as phosphorylation [25], acetylation [26], and ubiquitination [27], which regulate its stability and activity. However, the PTM mechanisms of ANXA2 in NSCLC remain poorly understood. Additionally, the exact mechanisms underlying ANXA2 dysregulation and its role in NSCLC glycolysis are not fully elucidated. This study reveals that the endogenous serine protease inhibitor SerpinB7 drives aerobic glycolysis and malignant progression in NSCLC through a non-canonical mechanism independent of its enzymatic activity. SerpinB7 interacts with ANXA2 via a non-RCL region, a process independent of its classical protease inhibitory activity; even when the enzyme-binding site responsible for its classical inhibitory function is mutated, SerpinB7 still significantly promotes tumor proliferation, metastasis, and glycolysis. Mechanistically, SerpinB7 stabilizes ANXA2 protein levels by competitively inhibiting NEDD4L-mediated ubiquitination of ANXA2. Subsequently, SerpinB7 reprograms glycolysis and promotes malignant progression in NSCLC through the ANXA2/c-Myc axis. Co-overexpression of SerpinB7 and ANXA2, underscoring the clinical significance of this pathway. Targeting SerpinB7 using siRNAs encapsulated in ferritin protein shells effectively suppresses NSCLC tumor growth and metastasis. In summary, our study demonstrates that SerpinB7 plays a critical role in promoting c-Myc signaling, enhancing downstream aerobic glycolysis, and driving NSCLC tumorigenesis by inhibiting proteasome-dependent ubiquitination and degradation of ANXA2, positioning it as a potential therapeutic target in NSCLC. Results SerpinB7 expression associates with prognosis of lung cancer patients Current research reports indicate that the SerpinB family plays important roles in NSCLC [21], glioma [20], melanoma [17], and other tumors, as well as in tumor brain metastasis [28]. To determine the potential correlation between the expression of SerpinB family members and NSCLC patients, we investigated the expression of SerpinB family members using data from the TCGA-LUAD and TCGA-LUSC databases. The LUAD dataset included 541 tumor samples and 59 adjacent normal samples, while the LUSC dataset included 502 tumor samples and 51 adjacent normal samples.As shown in volcano plots 1 A and 1B, SerpinB3, SerpinB4, SerpinB5 and SerpinB7 were significantly upregulated in tumor tissues compared to adjacent normal tissues in both LUAD and LUSC patients. Related studies have already reported the specific mechanism of SerpinB3, SerpinB4 and SerpinB5 in NSCLC [21, 29, 30]. Therefore, we selected SerpinB7 from the SerpinB family for further investigation, as its specific role and mechanism in NSCLC warrant in-depth research.From the TCGA database, we found that the mRNA levels of SerpinB7 were significantly upregulated in both LUAD and LUSC compared to adjacent normal lung tissue (Fig. 1C-F). The expression levels of SerpinB7 vary among different histological subtypes, with high expression observed in certain subtypes (Supplementary Fig. 1A-B). Analysis using the Kaplan–Meier plotter database indicated that high expression of SerpinB7 mRNA in lung cancer patients was associated with reduced overall survival (OS) (Fig. 1G). Immunofluorescence staining of LUAD and LUSC tissue samples and lung cancer cells revealed that SerpinB7 protein was predominantly localized in the cytoplasm of lung cancer cells (Fig. 1H and Supplementary Fig. 1 C). Additionally, compared to adjacent normal tissues, SerpinB7 protein levels were elevated in LUAD and LUSC tissues(Fig. 1I-J). Consistently, immunohistochemical staining demonstrated that SerpinB7 expression was significantly higher in LUAD and LUSC tissues compared to paired adjacent normal lung tissues (Fig. 1 K-L). The mRNA expression levels of SerpinB7 in LUAD tissues were also elevated compared to adjacent normal tissues (Fig. 1 M). Furthermore, SerpinB7 was highly expressed in multiple tumor cell lines of NSCLC compared to normal pulmonary epithelial cells (Fig. 1N-O). Analysis of the publicly available microarray dataset GSE297013 (from the Gene Expression Omnibus) showed that SerpinB7 mRNA expression was significantly higher in CL1-5 cells than in CL1-0 cells (Supplementary Fig. 1D). Collectively, these results suggest that SerpinB7 is upregulated in NSCLC and predicts poor prognosis of NSCLC patients.Fig. 1SerpinB7 expression associates with prognosis of lung cancer patients. A-B The volcano plot displays the expression patterns of differentially expressed genes in the SerpinB family from the LUAD and LUSC datasets within the TCGA database. C-F Quantitative comparison of paired and unpaired SerpnB7 expression in LUAD, LUSC and adjacent normal tissues in the TCGA-LUAD and TCGA-LUSC database; G Analysis of overall survival of 2166 patients with lung cancer. According to SerpinB7 expression in lung cancer tissue, the patients were divided into two groups according to the median (high expression group: n=1068, low expression group: n=1098). H-J Representative micrographs of LUAD, LUSC patient tissue and adjacent normal tissue stained with anti-SerpinB7 antibodies; The figure showed the quantitative average expression of SerpinB7 per high-magnification field of view in the tissue. scale bar, 50 μm. n=6. K Representative images from the tissue microarray analysis of SerpinB7 expression in paired LUAD specimens. scale bar, 200 μm. n=69. L Representative images from the tissue microarray analysis of SerpinB7 expression in paired LUSC specimens. scale bar, 200 μm. n=86. M mRNA expression levels of SerpinB7 in 15 pairs of LUAD tissues (T) and corresponding adjacent tissues (N). n=15. N-O Protein expression levels of SerpinB7 in normal human lung epithelial BEAS-2B and NSCLC cell lines. BEAS-2B was used as the normal counterpart to the NSCLC cell lines. n=3. SerpinB7 expression associates with prognosis of lung cancer patients. A-B The volcano plot displays the expression patterns of differentially expressed genes in the SerpinB family from the LUAD and LUSC datasets within the TCGA database. C-F Quantitative comparison of paired and unpaired SerpnB7 expression in LUAD, LUSC and adjacent normal tissues in the TCGA-LUAD and TCGA-LUSC database; G Analysis of overall survival of 2166 patients with lung cancer. According to SerpinB7 expression in lung cancer tissue, the patients were divided into two groups according to the median (high expression group: n=1068, low expression group: n=1098). H-J Representative micrographs of LUAD, LUSC patient tissue and adjacent normal tissue stained with anti-SerpinB7 antibodies; The figure showed the quantitative average expression of SerpinB7 per high-magnification field of view in the tissue. scale bar, 50 μm. n=6. K Representative images from the tissue microarray analysis of SerpinB7 expression in paired LUAD specimens. scale bar, 200 μm. n=69. L Representative images from the tissue microarray analysis of SerpinB7 expression in paired LUSC specimens. scale bar, 200 μm. n=86. M mRNA expression levels of SerpinB7 in 15 pairs of LUAD tissues (T) and corresponding adjacent tissues (N). n=15. N-O Protein expression levels of SerpinB7 in normal human lung epithelial BEAS-2B and NSCLC cell lines. BEAS-2B was used as the normal counterpart to the NSCLC cell lines. n=3. SerpinB7 promotes tumor growth and metastasis both in vivo and in vitro To reveal the biological functions of SerpinB7 in NSCLC, we knocked down SerpinB7 in H1299, HCC827, SK-MES-1 and H520 cells using short hairpin (sh) RNAs (sh1, sh2, and sh3) and overexpressed SerpinB7 in A549, PC9, SK-MES-1 and H520 cells using a SerpinB7-encoding lentiviral vector (Lv-SerpinB7). The mRNA and protein levels of SerpinB7 in the transfected cells were validated (Supplementary Fig. 2A-D). The transfection efficiencies of sh1 and sh2 were satisfactory for subsequent experiments. The results of CCK-8, 5-ethynyl-2 -deoxyuridine (EdU) and colony formation assays demonstrated that the knockdown and upregulation of SerpinB7 substantially reduced and increased cell proliferation potential, respectively (Supplementary Fig. 3A-H). On cell cycle regulation, SerpinB7 knockdown resulted in cell cycle arrest while overexpression of SerpinB7 promoted cell cycle progression in NSCLC cells (Supplementary Fig. 3I-N). Moreover, Transwell experiments demonstrated that the down-regulation of SerpinB7 substantially limited NSCLC cell migration and invasion, whereas SerpinB7 upregulation facilitated these abilities (Supplementary Fig. 4A-L). Collectively, these results strongly indicated that SerpinB7 facilitates the proliferation, migration, and invasion of NSCLC cells in vitro. A subcutaneous xenograft mouse model was established to evaluate the proliferative capacity of NSCLC cells in vivo (Fig. 2A). SerpinB7-knockdown H1299 cells, SerpinB7-overexpressing A549 cells, and their corresponding control NSCLC cells were inoculated into nude mice. Decreased tumor volumes and weights were observed in the SerpinB7 knockdown group, whereas SerpinB7 overexpression yielded the opposite effect (Fig. 2B-F). Xenograft tumors were harvested for IHC analysis (SerpinB7 and Ki-67). The results showed that SerpinB7 knockdown dampened the proliferative ability of NSCLC cells in vivo. In contrast, up- regulation of SerpinB7 promoted NSCLC cell proliferation (Supplementary Fig. 5 A). To identify the role of SerpinB7 in affecting NSCLC metastasis, in vivo metastasis experiment was carried out by injecting SerpinB7 stably transfected A549 or SerpinB7-knockdown H1299 cells and control cells in the tail vein of nude mice (Fig. 2G). The colonization of cancer cells was also significantly changed, as we found that quantification of lung bioluminescence intensity in mice receiving SerpinB7-transfected cells on day 50 showed more lung cancer cell metastasis in the SerpinB7-transfected cells than in the control group; In contrast, the lung cancer cell metastasis in the SerpinB7 knockdown group was significantly lower than that in the control group (Fig. 2H-K). As shown in Supplementary Fig. 5B, histological analysis showed that mice receiving the SerpinB7-OE subgroup developed more lung metastases compared to the control group. Consistently, overexpression of SerpinB7 resulted in a stronger Ki67 staining in lung metastatic tumors in comparison to the control group (Supplementary Fig. 5 C). A549 cells overexpressing SerpinB7 and H1299 cells with SerpinB7 knockdown were injected into mouse lungs to establish orthotopic lung cancer models [31, 32]. Observation via IVIS fluorescence imaging showed that SerpinB7 knockdown inhibited lung tumor growth, while SerpinB7 overexpression significantly promoted lung tumor growth (Fig. 2L-N). Furthermore, hematoxylin–eosin (H&E) staining also demonstrated that SerpinB7 knockdown inhibited tumor formation and progression, whereas SerpinB7 overexpression promoted tumor formation and progression (Supplementary Fig. 5D-E). Collectively, these data indicate that SerpinB7 promotes the malignant phenotype of NSCLC cells in vivo.Fig. 2SerpinB7 Promotes NSCLC Progression In Vivo and Upregulates Glycolytic Gene Expression. A Schematic illustration of the subcutaneous tumorigenesis model established in nude mice using cells with knockdown or overexpression of SerpinB7. B Photographs of xenograft tumors induced by the subcutaneous inoculation of nude mice with A549-transfected and H1299-transfected cells. n=8. C Growth curves of xenograft tumor volumes in A549 cells. n=8. D Graphs of A549 cell xenograft tumor weights. n=8. E Growth curves of xenograft tumor volumes in H1299 cells. n=8. F Graphs of H1299 cell xenograft tumor weights. n=8. G Schematic illustration of the lung metastatic model established in nude mice using cells with knockdown or overexpression of SerpinB7. H-K Representative images showing luciferase activity in lungs from BALB/c nude mice. Quantification of total lung bioluminescence from BALB/c nude mice. n=6. L Schematic illustration of the Orthotopic mouse model established in nude mice using cells with knockdown or overexpression of SerpinB7. M-N In vivo IVIS images and quantitative fluorescence analysis results of orthotopic lung cancer models in each group at 15 days post-inoculation. n=6. O Scheme displaying the procedure used for identifying the GO and KEGG pathways regulated by SerpinB7 through transcriptome. P The functional interpretation of differentially expressed genes was subjected to KEGG enrichment analysis. Q Heatmap of differentially expressed genes related to proliferation, metastasis and glycolysis from RNA-seq of A549 transfected cells. Rows are centered for each gene with unit variance scaling applied, gene clustering by correlation distance and average linkage. R Illustration of glycolysis process and related rate-limiting enzymes. S TCGA database analysis showed that SerpinB7 expression was positively correlated with the expressions of LDHA, ENO1, PKM2 and TPI1 in LUAD and LUSC, respectively. LUAD: n=513 , LUSC: n=501 SerpinB7 Promotes NSCLC Progression In Vivo and Upregulates Glycolytic Gene Expression. A Schematic illustration of the subcutaneous tumorigenesis model established in nude mice using cells with knockdown or overexpression of SerpinB7. B Photographs of xenograft tumors induced by the subcutaneous inoculation of nude mice with A549-transfected and H1299-transfected cells. n=8. C Growth curves of xenograft tumor volumes in A549 cells. n=8. D Graphs of A549 cell xenograft tumor weights. n=8. E Growth curves of xenograft tumor volumes in H1299 cells. n=8. F Graphs of H1299 cell xenograft tumor weights. n=8. G Schematic illustration of the lung metastatic model established in nude mice using cells with knockdown or overexpression of SerpinB7. H-K Representative images showing luciferase activity in lungs from BALB/c nude mice. Quantification of total lung bioluminescence from BALB/c nude mice. n=6. L Schematic illustration of the Orthotopic mouse model established in nude mice using cells with knockdown or overexpression of SerpinB7. M-N In vivo IVIS images and quantitative fluorescence analysis results of orthotopic lung cancer models in each group at 15 days post-inoculation. n=6. O Scheme displaying the procedure used for identifying the GO and KEGG pathways regulated by SerpinB7 through transcriptome. P The functional interpretation of differentially expressed genes was subjected to KEGG enrichment analysis. Q Heatmap of differentially expressed genes related to proliferation, metastasis and glycolysis from RNA-seq of A549 transfected cells. Rows are centered for each gene with unit variance scaling applied, gene clustering by correlation distance and average linkage. R Illustration of glycolysis process and related rate-limiting enzymes. S TCGA database analysis showed that SerpinB7 expression was positively correlated with the expressions of LDHA, ENO1, PKM2 and TPI1 in LUAD and LUSC, respectively. LUAD: n=513 , LUSC: n=501 SerpinB7 regulates the expression of key glycolysis genes clinically relevant to NSCLC To further clarify the biological processes associated with the function of SerpinB7 in NSCLC progression, we performed RNA sequencing of the A549-stably transfected cell line (Fig. 2O). Hierarchical clustering and principal component analysis (PCA) of A549 overexpressed cells showed that the gene expression profiles of Ev and SerpinB7-OE groups were different (Supplementary Fig. 6A-B). Compared with the control group, 833 genes showed increased expression and 1608 genes showed decreased expression after SerpinB7 overexpression (Supplementary Fig. 6 C). GO enrichment analysis (GO-BP, GO-MF, GO-CC) revealed that SerpinB7 mainly affected the proliferation, metastasis, adhesion and other biological functions of tumor cells (Supplementary Fig. 6D-F). Heatmaps demonstrate global, yet distinct, differences in expression of genes related to proliferation, migration and adhesion comparing SerpinB7-OE versus control A549 cells (Supplementary Fig. 6G). Furthermore, GSEA also found that after overexpression of SerpinB7, genes related to cell adhesion were significantly reduced, and genes related to cell cycle pathways and cell migration were increased (Supplementary Fig. 6H). KEGG pathway analysis revealed significant enrichment not only in proliferation-related pathways but also in energy metabolism pathways (Fig. 2P). In addition, it was further found that overexpression of SerpinB7 increased the expression of cell cycle, cell Metastasis and other genes, as well as the expression of genes regulating glycolysis (Fig. 2Q). Among them, the expression of genes regulating cell cycle, including CCND1, CDKN1B, CCNB1; Metastasis-related genes, such as CDH1, PHLDA2, SNAI2, SerpinE1 and CDH17; The main genes related to glycolysis are PKM2, ENO1, LDHA and TPI1. Subsequent Western blotting analysis demonstrated that in PC9 and A549 cells overexpressing SerpinB7 (SerpinB7-OE group), the protein levels of c-Myc and its downstream glycolytic targets (including LDHA, PKM2, ENO1, and TPI1), cell cycle regulators (Cyclin A, Cyclin E), as well as epithelial-mesenchymal transition (EMT) marker proteins (N-cadherin, Vimentin) were significantly elevated compared to the empty vector control group (Ev group). Conversely, knockdown of SerpinB7 markedly reduced the expression levels of these molecules. Notably, the protein level of the EMT marker E-cadherin was significantly decreased in the SerpinB7-OE group relative to the Ev control, whereas SerpinB7 knockdown resulted in a substantial increase in its expression (Supplementary Fig. 4 M). Previous studies have shown that the upregulation of PKM2 [33], ENO1 [34], LDHA [35] and TPI1 [36] can promote tumorigenesis. Glycolytic enzymes are key enzyme classes that catalyze a series of biochemical reactions in the glycolysis pathway, progressively degrading glucose into pyruvate to generate energy (ATP) and metabolic precursors (Fig. 2R). Glycolytic enzymes are considered potential therapeutic targets for cancer treatment [37, 38]. Consistent with our RNA-seq results, analysis of human LUAD and LUSC datasets revealed a positive correlation between the mRNA levels of PKM2, ENO1, LDHA and TPI1 and the mRNA levels of SerpinB7 in LUAD and LUSC samples (Fig. 2S). Furthermore, high mRNA levels of PKM2, ENO1, LDHA and TPI1 were associated with poor survival prognosis in lung cancer as well as the two subtypes, LUAD and LUSC (Supplementary Fig. 6I). These results suggest that the mechanism by which SerpinB7 overexpression upregulates glycolysis gene expression is relevant in human NSCLC. SerpinB7 potentiates glycolytic metabolism in NSCLC To further clarify the metabolism-related biological processes associated with SerpinB7 in the progression of NSCLC. We conducted untargeted metabolomics to further explore which metabolic pathway is crucial for the function of SerpinB7. Thus, we carried out untargeted metabolomics to further explore which metabolic pathway is critical for SerpinB7 function (Fig. 3A). Based on PLS-DA and OPLS-DA, the metabolic profiles were clearly segregated between the SerpinB7-OE and Lv-con groups, which indicated a consistent and major metabolic shift following the alteration of SerpinB7 expression level (Fig. 3B). The volcano plot and heatmap display the specific details of the differentially affected metabolites following the overexpression of SerpinB7 (Fig. 3C, D). A Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis further highlighted central carbon metabolism in cancer (CCM) as the most changed pathway based on the number of changed metabolites and statistical significance (Fig. 3E).Fig. 3SerpinB7 potentiates glycolytic metabolism in NSCLC. A Scheme displaying the procedure used for identifying the specific metabolic process regulated by SerpinB7. B PLS-DA and OPLS-DA of metabolomics. n=6. C Metabolomics volcano plots of A549 transfected cells (blue indicates downregulated metabolites; red indicates upregulated metabolites). n=6. D Heatmap showing dysregulated metabolites (red, upregulated metabolites; blue, downregulated metabolites) identified by metabolomics assays. n=6. E KEGG pathway enrichment analysis highlighted central carbon metabolism in cancer as the most changed pathway in terms of the P-value of changed metabolites. F-H Bar charts represent the relative abundance of metabolites between A549-transfected cells. n=6. I Glucose consumption, Lactate production and ATP levels in H1299 and A549 cells was determined. n=3. J-L The uptake of 2-NBDG in A549 cells overexpressing SerpinB7 and H1299 cells with knockdown of SerpinB7. n=3. M-Q In A549 cells overexpressing SerpinB7 and H1299 cells with knockdown of SerpinB7, the Extra Cellular Acidification Rate (ECAR), glycoPER, basal glycolysis rate, and compensatory glycolysis rate were measured. n=3. R Schematic diagram represents intracellular fluxes in A549 cells quantified using U-13C6 glucose tracer. Blue circle represents 12 C and red circles 13 C atoms. S-T Bar charts represent the relative abundance of metabolites between A549-transfected cells. n=3 SerpinB7 potentiates glycolytic metabolism in NSCLC. A Scheme displaying the procedure used for identifying the specific metabolic process regulated by SerpinB7. B PLS-DA and OPLS-DA of metabolomics. n=6. C Metabolomics volcano plots of A549 transfected cells (blue indicates downregulated metabolites; red indicates upregulated metabolites). n=6. D Heatmap showing dysregulated metabolites (red, upregulated metabolites; blue, downregulated metabolites) identified by metabolomics assays. n=6. E KEGG pathway enrichment analysis highlighted central carbon metabolism in cancer as the most changed pathway in terms of the P-value of changed metabolites. F-H Bar charts represent the relative abundance of metabolites between A549-transfected cells. n=6. I Glucose consumption, Lactate production and ATP levels in H1299 and A549 cells was determined. n=3. J-L The uptake of 2-NBDG in A549 cells overexpressing SerpinB7 and H1299 cells with knockdown of SerpinB7. n=3. M-Q In A549 cells overexpressing SerpinB7 and H1299 cells with knockdown of SerpinB7, the Extra Cellular Acidification Rate (ECAR), glycoPER, basal glycolysis rate, and compensatory glycolysis rate were measured. n=3. R Schematic diagram represents intracellular fluxes in A549 cells quantified using U-13C6 glucose tracer. Blue circle represents 12 C and red circles 13 C atoms. S-T Bar charts represent the relative abundance of metabolites between A549-transfected cells. n=3 CCM involves three main pathways: glycolysis, pentose phosphate pathway, and citric acid cycle [39–42]. Glycolysis is the most characterized metabolic pathway in tumor cells [43]. Through analysis of differential metabolites, we observed an increased abundance of 3-phosphoglycerate (Fig. 3F). In glycolysis, 3-phosphoglycerate can be further converted to 2-phosphoglycerate under the action of phosphoglycerate mutase [44]. The tricarboxylic acid (TCA) cycle serves as the central hub of carbon and energy metabolism, and is also interconnected with glycolysis, amino acid, and lipid metabolism [45]. The TCA cycle is recognized as one of the most crucial metabolic pathways present in nearly all living organisms, involving key intermediates including α-ketoglutarate, citrate, fumarate, succinate, oxaloacetate, and L-malate [46]. In our untargeted metabolomics study, we detected decreased abundance of TCA cycle-related metabolites such as L-malate and citrate derivatives (Fig. 3G-H). We wondered whether alterations in the NSCLC cells glycolysis pathway underlie the SerpinB7-regulated NSCLC progression. We found the glucose uptake, lactate release, ATP production and 2-NBDG uptake were diminished by the knockdown of SerpinB7 in H1299 cells, whereas the opposite effects were observed in A549 cells overexpressing SerpinB7 (Fig. 3I-L). The results of Seahorse glycolytic rate analysis indicated that glycolytic proton efflux rate (glycoPER), basal glycolysis and compensatory glycolysis in NSCLC cells over-expressing SerpinB7 were higher than those in control group cells, knockdown of SerpinB7 in NSCLC cells resulted in the opposite effects (Fig. 3 M-Q and Supplementary Fig. 7A-H). To examine the direct effect of SerpinB7 on glycolysis, we performed a metabolic flux analysis in A549 cells using [13C] glucose as a substrate. Metabolic tracking revealed a significantly increased abundance of glycolysis metabolite M3 lactate in the SerpinB7-overexpressing group compared with the Ev group, but decreased levels of TCA-circulating metabolites such as M2 citrate, M2 malate, M2 succinate and M2 a-ketoglutarate (Fig. 3R-T). Combined with the results from the previous transcriptome analysis, where the overexpression of SerpinB7 led to increased expression of PKM2, ENO1, LDHA, and TPI1, all of which are key enzymes in glycolysis, we hypothesize that SerpinB7 may enhance aerobic glycolysis in NSCLC cells and reduce the flux of glucose-derived carbon into the TCA cycle. Taken together, these results indicate that SerpinB7 contributes to glycolysis. The regulation of SerpinB7 on tumors depends on glycolysis As described above, Ectopic overexpression of SerpinB7 upregulates the glycolytic program and alters transcription of key glycolytic genes. This led us to investigate the mechanism by which SerpinB7 regulates the malignant phenotype of lung cancer cells through the glycolytic pathway. To inhibit glycolysis, we used the following inhibitors: 2-deoxy-D-glucose (2-DG), a hexokinase inhibitor and glucose analog [37, 47]. After treating PC9, A549, HCC827, and H1299 cells with 2-DG, the proliferation, migration, invasion, basal glycolysis and compensatory glycolysis of lung cancer cells were all inhibited (Supplementary Fig. 8A-O). 2-DG treatment significantly restored the enhanced proliferation ability induced by SerpinB7 overexpression and also notably diminished the original difference in proliferation ability between the control and knockdown groups. (Supplementary Fig. 9A-B). We further evaluated the effects of the same treatment on the performance of stable cell lines, and the results showed that 2-DG treatment significantly restored the enhanced migration and invasion abilities caused by SerpinB7 overexpression, while the inhibition of glycolysis in the knockdown and control groups also significantly reduced the original differences in migration and invasion abilities between the two groups. (Supplementary Fig. 9C-F). Furthermore, the data indicated that 2-DG treatment significantly restored the elevated levels of glycoPER, basal glycolysis, and compensatory glycolysis caused by SerpinB7 overexpression, while also significantly reducing the original differences in glycolytic levels between the control group and the knockdown group (Supplementary Fig. 10A-J). Knockdown of the key glycolytic enzyme LDHA and treatment with the oxidative phosphorylation modulator ROT/AA (0.5 μM) significantly restored the enhanced proliferation, migration, and invasion abilities induced by SerpinB7 overexpression, and also notably diminished the original differences in these abilities between the control and knockdown groups (Supplementary Fig. 11A-J). Therefore, 2-DG, knockdown of the key glycolytic enzyme LDHA, and ROT/AA can restore the changes in glycolytic levels and malignant phenotypes of lung cancer cells caused by alterations in SerpinB7 expression levels, indicating that the influence of SerpinB7 on the malignant phenotype of tumor cells is dependent on glycolysis. SerpinB7 contributes to malignancy and glycolysis in NSCLC in an enzyme-independent manner It is known that overexpression of SerpinB7 induces mesangial cell proliferation and expansion, leading to the accumulation of type IV collagen and laminin, confirming its activity as a protease inhibitor [48]. Therefore, we sought to investigate whether the typical protease inhibitor activity of SerpinB7 is required for its role in promoting malignancy and glycolysis in NSCLC. As previously reported, substitution of Arg340 at the P1′ site in the RCL of SerpinB5 with alanine abolished its serine protease inhibitory function [49]. The RCL mutant of SerpinB1 (F343A/C344A) lost its binding capacity to neutrophil serine proteases [19]. Like most Serpins [50, 51], the RCL region of SerpinB7 contains a scissile bond (P1-P1') located near the C-terminus of the protein sequence (Fig. 4A). Accordingly, we constructed wild-type (WT) and protease inhibitor-inactive mutant (MUT#1: K347A/Q348A) versions of SerpinB7 and reintroduced SerpinB7-WT and SerpinB7-MUT#1: K347A/Q348A into H1299 and HCC827 cells with endogenous SerpinB7 knockdown (Fig. 4B-D).Fig. 4SerpinB7 promotes malignant progression and glycolysis in NSCLC in a non-enzymatic manner. A Cleavage sites in the reactive site loops of human clade B serpins with overlapping reactive centers. B The schematic diagram of a SerpinB7 mutant (K347A/Q348A). C Schematic illustration of WT or MUT#1 version of SerpinB7. D SerpinB7 expression in NSCLC cells after reconstituting SerpinB7-WT and SerpinB7-MUT#1 into H1299 and HCC827 cells with endogenous SerpinB7 knockdown. E Protease activity were detected in H1299 and HCC827 cells. n=3. F-I Transwell assays were performed to assess the migratory and invasive capacities of NSCLC cells after reconstituting SerpinB7-WT and SerpinB7-MUT#1 into H1299 and HCC827 cells with endogenous SerpinB7 knockdown. scale bar, 200 μm. n=3. J-K Colony formation assays were employed to assess the proliferation capacity of NSCLC cells after reconstituting SerpinB7-WT and SerpinB7-MUT#1 into H1299 and HCC827. n=3. L-N Glucose consumption, ATP levels and Lactate production in NSCLC cells after reconstituting SerpinB7-WT and SerpinB7-MUT#1 into H1299 and HCC827 cells with endogenous SerpinB7 knockdown was determined. n=3. O-P The ECAR, glycoPER, basal glycolysis, and compensatory glycolysis of NSCLC cells after reconstituting SerpinB7-WT and SerpinB7-MUT#1 into H1299 and HCC827 cells with endogenous SerpinB7 knockdown were determined. n=3 SerpinB7 promotes malignant progression and glycolysis in NSCLC in a non-enzymatic manner. A Cleavage sites in the reactive site loops of human clade B serpins with overlapping reactive centers. B The schematic diagram of a SerpinB7 mutant (K347A/Q348A). C Schematic illustration of WT or MUT#1 version of SerpinB7. D SerpinB7 expression in NSCLC cells after reconstituting SerpinB7-WT and SerpinB7-MUT#1 into H1299 and HCC827 cells with endogenous SerpinB7 knockdown. E Protease activity were detected in H1299 and HCC827 cells. n=3. F-I Transwell assays were performed to assess the migratory and invasive capacities of NSCLC cells after reconstituting SerpinB7-WT and SerpinB7-MUT#1 into H1299 and HCC827 cells with endogenous SerpinB7 knockdown. scale bar, 200 μm. n=3. J-K Colony formation assays were employed to assess the proliferation capacity of NSCLC cells after reconstituting SerpinB7-WT and SerpinB7-MUT#1 into H1299 and HCC827. n=3. L-N Glucose consumption, ATP levels and Lactate production in NSCLC cells after reconstituting SerpinB7-WT and SerpinB7-MUT#1 into H1299 and HCC827 cells with endogenous SerpinB7 knockdown was determined. n=3. O-P The ECAR, glycoPER, basal glycolysis, and compensatory glycolysis of NSCLC cells after reconstituting SerpinB7-WT and SerpinB7-MUT#1 into H1299 and HCC827 cells with endogenous SerpinB7 knockdown were determined. n=3 To verify the effect of SerpinB7-MUT#1: K347A/Q348A on serine protease activity, we conducted enzymatic activity assays in SerpinB7-downregulated NSCLC cells with or without transfection of SerpinB7-WT or SerpinB7-MUT#1: K347A/Q348A. In SerpinB7-downregulated NSCLC cells transfected with SerpinB7-WT, serine protease activity was restored. However, in cells transfected with SerpinB7-MUT#1: K347A/Q348A, no significant change in serine protease activity levels was observed (Fig. 4E). The study found that overexpression of both SerpinB7-WT and SerpinB7-MUT#1 significantly enhanced tumor cell proliferation, migration, invasion and glycolytic capacity (Fig. 4F-P). More importantly, apart from partial loss of function in promoting cell proliferation, SerpinB7-MUT#1: K347A/Q348A exhibited effects comparable to SerpinB7-WT overexpression in NSCLC progression and glycolytic phenotype regulation (Fig. 4F-P). These data suggest that SerpinB7 exerts its oncogenic effects in NSCLC not only through its protease inhibitor activity but also via a previously unknown enzyme-independent regulatory mechanism that promotes glycolytic function. SerpinB7 can directly interact with ANXA2 To elucidate the specific molecular target responsible for the enzyme inhibitor-independent function of SerpinB7, we performed immunoprecipitation (IP) assays using anti-Flag beads to successfully pull down GST-tagged SerpinB7, followed by LC–MS/MS analysis and integration with upregulated proteins identified through proteomic profiling to characterize its interacting proteins (Fig. 5A). Functional enrichment analysis of potential interactors was conducted using the KOBAS online platform (http://kobas.cbi.pku.edu.cn/), revealing that SerpinB7-interacting proteins are primarily involved in metabolic pathways (Supplementary Fig. 12A-B). This approach identified 113 potential Flag-SerpinB7-interacting proteins through IP assays and 32 potential GST-SerpinB7-interacting proteins through GST pull-down assays. SerpinB7-interacting proteins are listed in Fig. 5B-C according to the number of unique peptides identified. By integrating results from IP-MS, GST pull-down, and SerpinB7 overexpression-upregulated protein analyses, we identified ANXA2 as a key protein that both interacts with and is regulated by SerpinB7 (Fig. 5D and Supplementary Fig. 12C-D).Fig. 5SerpinB7 interacts with ANXA2. A Scheme displaying the procedure used for identifying the specific target of SerpinB7. B, C The top Seven SerpinB7-interacting proteins according to the number of identified unique peptides. D Intersection of the results from the proteomics, GST-pulldown/MS and IP-MS analyses. E-G IP analysis of the interaction between SerpinB7 and ANXA2 in HEK293T cells. HEK293T cells transfected with the indicated expression vectors were subjected to IP assays and followed by immunoblotting with indicated antibodies. H, I IP analysis of the interaction between SerpinB7 and ANXA2 in NSCLC cells. A549 and H1299 cell lines expressing pcDNA3.1 and SerpinB7-Flag were subjected to IP and immunoblotting assays. J, K A549 and H1299 cell lines expressing pcDNA3.1 and ANXA2-Myc were subjected to IP and immunoblotting assays. L Immunofluorescent images showing the colocalization of SerpinB7 (red) and ANXA2 (green) in A549 and H1299 cell lines in combination. The nuclear was counterstained with DAPI (blue). M Plots of pixel intensity along the white line region, with colors consistent with the merged images. N Design of SerpinB7 truncations. O Immunoprecipitation and Western blotting analyses showing the interactions between Flag-tagged truncated SerpinB7 and Myc-tagged ANXA2 proteins in HEK293T cells. P Docking model of the SerpinB7/ANXA2 complex. Red dashed lines indicate hydrogen bonds, respectively SerpinB7 interacts with ANXA2. A Scheme displaying the procedure used for identifying the specific target of SerpinB7. B, C The top Seven SerpinB7-interacting proteins according to the number of identified unique peptides. D Intersection of the results from the proteomics, GST-pulldown/MS and IP-MS analyses. E-G IP analysis of the interaction between SerpinB7 and ANXA2 in HEK293T cells. HEK293T cells transfected with the indicated expression vectors were subjected to IP assays and followed by immunoblotting with indicated antibodies. H, I IP analysis of the interaction between SerpinB7 and ANXA2 in NSCLC cells. A549 and H1299 cell lines expressing pcDNA3.1 and SerpinB7-Flag were subjected to IP and immunoblotting assays. J, K A549 and H1299 cell lines expressing pcDNA3.1 and ANXA2-Myc were subjected to IP and immunoblotting assays. L Immunofluorescent images showing the colocalization of SerpinB7 (red) and ANXA2 (green) in A549 and H1299 cell lines in combination. The nuclear was counterstained with DAPI (blue). M Plots of pixel intensity along the white line region, with colors consistent with the merged images. N Design of SerpinB7 truncations. O Immunoprecipitation and Western blotting analyses showing the interactions between Flag-tagged truncated SerpinB7 and Myc-tagged ANXA2 proteins in HEK293T cells. P Docking model of the SerpinB7/ANXA2 complex. Red dashed lines indicate hydrogen bonds, respectively Consistent with previous studies demonstrating that high ANXA2 expression in pancreatic ductal adenocarcinoma is associated with metabolic alterations such as DNA repair and glycolysis, as well as poor prognosis [24], ANXA2 was selected as a candidate target of SerpinB7. Reciprocal IP assays demonstrated that ANXA2 indeed interacted with SerpinB7 in multiple cell lines (Fig. 5E-K). Immunofluorescent staining revealed that SerpinB7 partially colocalized with ANXA2 in both A549 and H1299 cells (Fig. 5L-M, yellow color in merge images). Similar to other serpins, SerpinB7 folds into a metastable structure consisting of nine α-helices and three β-sheets with a RCL [52]. Additionally, it has been reported that the carboxyl-terminal of Serpin inhibits the activation of pro-caspase-1/−4/−5/−11 [19]. To verify whether the interaction between SerpinB7 and ANXA2 depends on the RCL and carboxy-terminal, we constructed truncations of SerpinB7 (Fig. 5N). IP analysis confirmed that their interaction is not affected by the RCL and carboxy-termina (Fig. 5O). To elucidate the structural basis of the SerpinB7-ANXA2 interaction, we employed AlphaFold 3 to predict the three-dimensional structure of their complex (Fig. 5P). The results demonstrate that the N-terminal helix of ANXA2 is embedded in the CD-loop and helix E region of SerpinB7, where the side chain of Lys10 and the backbone of Val4 from ANXA2 form hydrogen bonds with the side chains of Ser74 and Cys117 residues in SerpinB7, respectively. Taken together,these results suggested that the interaction between SerpinB7 and ANXA2 affects NSCLC cancer growth, metastasis and glycolysis. SerpinB7 stabilizes ANXA2 by interrupting NEDD4L-mediated ubiquitination degradation We next determine whether there is a regulatory relationship between SerpinB7 and ANXA2. As shown in Fig. 6A, ectopic expression of SerpinB7 resulted in an increase in ANXA2 protein levels. Conversely, knockdown of SerpinB7 using shRNAs reduced ANXA2 protein levels.Furthermore, SerpinB7 maintained the stability of ANXA2 in the presence of the protein synthesis inhibitor, cycloheximide (CHX) (Fig. 6B-D). In addition, the proteasome inhibitor, MG132, reversed the SerpinB7-knockdown-induced decrease in ANXA2 levels, suggesting that SerpinB7 regulated ANXA2 stability in a proteasome-mediated manner (Fig. 6E). Consistent with these observations, ANXA2 ubiquitination was enhanced by SerpinB7 downregulation and inhibited by SerpinB7 upregulation (Fig. 6F-G).Fig. 6SerpinB7 stabilizes ANXA2 by interrupting NEDD4L-mediated ubiquitination degradation of ANXA2. A ANXA2 expression levels in NSCLC cell lines with SerpinB7 knockdown or upregulation. B-D ANXA2 expression levels in NSCLC cells treated with CHX for the indicated times and relative ANXA2 protein levels. n=3. E ANXA2 expression levels in SerpinB7-knockdown or SerpinB7-overexpression NSCLC cells treated with or without MG132. F-G Ubiquitination of ANXA2 in H1299 and A549 transfected cells. H The UbiBrowser database was used to analyze the upstream E3 ubiquitinligase of ANXA2. I The STRING database confirmed that NEDD4L is the E3 ubiquitin ligase in the SerpinB7 binding gene. J The interaction of NEDD4L and ANXA2 was detected by immunoprecipitation and Western blotting analyses in HEK293T cells with SerpinB7 knockdown or overexpression. K Ubiquitination of ANXA2 in HEK293T cells with SerpinB7 knockdown or overexpression SerpinB7 stabilizes ANXA2 by interrupting NEDD4L-mediated ubiquitination degradation of ANXA2. A ANXA2 expression levels in NSCLC cell lines with SerpinB7 knockdown or upregulation. B-D ANXA2 expression levels in NSCLC cells treated with CHX for the indicated times and relative ANXA2 protein levels. n=3. E ANXA2 expression levels in SerpinB7-knockdown or SerpinB7-overexpression NSCLC cells treated with or without MG132. F-G Ubiquitination of ANXA2 in H1299 and A549 transfected cells. H The UbiBrowser database was used to analyze the upstream E3 ubiquitinligase of ANXA2. I The STRING database confirmed that NEDD4L is the E3 ubiquitin ligase in the SerpinB7 binding gene. J The interaction of NEDD4L and ANXA2 was detected by immunoprecipitation and Western blotting analyses in HEK293T cells with SerpinB7 knockdown or overexpression. K Ubiquitination of ANXA2 in HEK293T cells with SerpinB7 knockdown or overexpression Considering the inability of SerpinB7 to directly influence protein ubiquitination, we speculated that an E3 ligase or deubiquitinase (DUB) mediates SerpinB7-regulated ANXA2 expression. The UbiBrowser database was used to explore the related upstream E3 ubiquitin ligase of ANXA2, and 20 previously reported E3 ubiquitin ligases, including NEDD4L were identified (Fig. 6H). NEDD4L is an E3 ligase involved in ubiquitin-mediated protein degradation, bound to the Annexin 1 domain of ANXA2 and promoted its degradation [53]. Through integrated analysis with the STRING database, we identified 10 potential SerpinB7-interacting proteins, including NEDD4L (Fig. 6I). Thus, we hypothesized that SerpinB7 stabilizes ANXA2 by interrupting NEDD4L-mediated ubiquitination. As shown in Fig. 6 J-K, SerpinB7 knockdown significantly promoted the interaction between NEDD4L and SerpinB7 as well as NEDD4L-induced ANXA2 ubiquitination. In contrast, NEDD4L-ANXA2 interaction and NEDD4L-mediated ANXA2 ubiquitination were markedly weakened when SerpinB7 was overexpressed. Thus, SerpinB7 stabilizes ANXA2 by interrupting NEDD4L-mediated ubiquitination and degradation of ANXA2. SerpinB7 enhances glycolysis and accelerates malignant phenotype in NSCLC by up-regulating ANXA2 Subsequently, to investigate whether ANXA2 promoted glycolysis and accelerates malignant phenotype in NSCLC, ANXA2 protein was significantly expressed in both LUAD and LUSC, with increased expression in both cancerous tissues compared to adjacent normal tissues (Supplementary Fig. 13A-C). Kaplan–Meier Plotter database showed reduced long-term survival in lung cancer patients with high ANXA2 expression (Supplementary Fig. 13D-E). To further validate the above results, we constructed stable A549 cells overexpressing ANXA2 and H1299 cells with stable ANXA2 knockdown using lentiviral infection methods (Supplementary Fig. 13F-G). knockdown of ANXA2 restrained NSCLC cell proliferation, migration, and invasion, whereas upregulation of ANXA2 exerted the opposite effects (Supplementary Fig. 13H-N). Flow cytometry assays indicated that the ANXA2 knockdown groups contained more apoptotic cells, whereas upregulation of ANXA2 inhibited apoptosis (Supplementary Fig. 14A-B). Moreover, Glucose uptake, lactate release, ATP production, 2-NBDG uptake, glycoPER, basal glycolysis and compensatory glycolysis decreased in H1299 cells with ANXA2 knockdown, and the reverse effects were observed in A549 cells overexpressing ANXA2 (Fig. 7A-H). Therefore, we inferred that SerpinB7 facilitates glycolysis and glycolysis-related malignancy in NSCLC cells by upregulating ANXA2.Fig. 7SerpinB7 enhances glycolysis and accelerates malignancy in NSCLC by up-regulating ANXA2. A The Glucose consumption, Lactate production and ATP levels in H1299 and A549 cells was determined. n=3. B, C The uptake of 2-NBDG in A549 cells overexpressing ANXA2 and H1299 cells with knockdown of ANXA2. n=3. D-H In A549 cells overexpressing ANXA2 and H1299 cells with knockdown of ANXA2, the ECAR, glycoPER, basal glycolysis rate and compensatory glycolysis rate were measured. n=3. I-J The ECAR, glycoPER, basal glycolysis, and compensatory glycolysis were measured in A549 cells (sh-con or shANXA2) with or without SerpinB7 up-regulation. n=3. K-L The ECAR, glycoPER, basal glycolysis, and compensatory glycolysis were measured in H1299 cells (sh-con or sh1) with or without ANXA2 up-regulation. n=3. M Schematic illustration of the orthotopic tumor model established in nude mice using cells with knockdown or overexpression of SerpinB7 and ANXA2. N-O In vivo IVIS images and quantitative fluorescence analysis results of orthotopic lung cancer models in each group at 15 days post-inoculation. n=6. SerpinB7 enhances glycolysis and accelerates malignancy in NSCLC by up-regulating ANXA2. A The Glucose consumption, Lactate production and ATP levels in H1299 and A549 cells was determined. n=3. B, C The uptake of 2-NBDG in A549 cells overexpressing ANXA2 and H1299 cells with knockdown of ANXA2. n=3. D-H In A549 cells overexpressing ANXA2 and H1299 cells with knockdown of ANXA2, the ECAR, glycoPER, basal glycolysis rate and compensatory glycolysis rate were measured. n=3. I-J The ECAR, glycoPER, basal glycolysis, and compensatory glycolysis were measured in A549 cells (sh-con or shANXA2) with or without SerpinB7 up-regulation. n=3. K-L The ECAR, glycoPER, basal glycolysis, and compensatory glycolysis were measured in H1299 cells (sh-con or sh1) with or without ANXA2 up-regulation. n=3. M Schematic illustration of the orthotopic tumor model established in nude mice using cells with knockdown or overexpression of SerpinB7 and ANXA2. N-O In vivo IVIS images and quantitative fluorescence analysis results of orthotopic lung cancer models in each group at 15 days post-inoculation. n=6. To confirm this conclusion, H1299 cells were transfected with sh-SerpinB7 (sh1) and Lv-ANXA2, A549 cells were transfected with sh-ANXA2 and Lv-SerpinB7, and ANXA2 expression level was evaluated by western blotting (Supplementary Fig. 15 A). These results showed that overexpression of ANXA2 reversed the inhibitory effects of SerpinB7 knockdown on the proliferation, migration and invasion of H1299 cells; whereas in A549 cells, overexpression of SerpinB7 reversed the inhibitory effects of ANXA2 knockdown on cell proliferation, migration and invasion (Supplementary Fig. 15B-I). Further results showed that overexpression of ANXA2 restored the inhibitory effects of SerpinB7 knockdown on glucose uptake, lactate release, ATP production, glycoPER, basal glycolysis, and compensatory glycolysis in H1299 cells. However, in A549 cells, upregulation of SerpinB7 reversed the inhibitory effects of ANXA2 knockdown on these glycolytic parameters (Fig. 7I-L and Supplementary Fig. 16 A). The transfected A549 and H1299 cells were orthotopically implanted into nude mice. In vivo experiments demonstrated that upregulation of ANXA2 reversed the inhibitory effects of SerpinB7 knockdown on tumor growth, while upregulation of SerpinB7 reversed the inhibitory effects of ANXA2 knockdown on tumor growth (Fig. 7 M-O and Supplementary Fig. 15 J-L). In summary, ANXA2 plays a critical role in glycolysis in NSCLC. Meanwhile, SerpinB7 promotes glycolysis and accelerates malignant progression of NSCLC by upregulating ANXA2. To further investigate the potential association between ANXA2 and key glycolysis-related genes in NSCLC tumors. Analysis of the TCGA database revealed a positive correlation between ANXA2 expression and glycolytic enzymes PKM2, ENO1, LDHA, and TPI1 in lung cancer patients (Supplementary Fig. 17A-C), suggesting a potential regulatory role of ANXA2 in key glycolytic enzymes. As an mRNA-binding protein, ANXA2 binds to the localization signal in the 3' untranslated region (UTR) of c-Myc mRNA, promoting its perinuclear localization and transcriptional activation [54–56]. The mRNA-binding site is located within its domain IV (Supplementary Fig. 17D). Given the characteristics of ANXA2 as an RNA-binding protein, we evaluated the effects of SerpinB7 on c-Myc mRNA stability and translational efficiency. Polysome profiling revealed that SerpinB7 overexpression led to a significant shift of c-Myc mRNA from light to heavy polysome fractions, indicating enhanced translational efficiency (Supplementary Fig. 17E). Concurrently, Actinomycin D chase assays demonstrated that SerpinB7 overexpression extended the half-life of c-Myc mRNA, thereby increasing its stability (Supplementary Fig. 17 F). This dual post-transcriptional regulatory mechanism is consistent with the established role of ANXA2 as an RNA-binding protein that scaffolds c-Myc mRNA to the perinuclear cytoskeleton, thereby facilitating efficient translation and protecting the transcript from degradation [56]. Furthermore, our observation of increased c-Myc protein levels (Supplementary Fig. 17G–H) aligns with previous reports that phosphorylated ANXA2 can inhibit the proteasomal degradation of c-Myc [57]. This suggests that in NSCLC, SerpinB7/ANXA2 exerts multi-level control over the protein homeostasis of c-Myc. Given that c-Myc is a key regulator of tumor glycolysis [58]. Combined with our RNA-seq observations of increased MYC mRNA levels in SerpinB7-overexpressing cells (Supplementary Fig. 17I), and as shown in Supplementary Fig. 4 M, c-Myc protein levels were significantly elevated in PC9 and A549 cells overexpressing SerpinB7, whereas this trend was reversed in H1299 and HCC827 cells with SerpinB7 knockdown. we hypothesize that SerpinB7-mediated stabilization of ANXA2 may play an important role in regulating NSCLC malignancy through c-Myc-mediated glucose metabolism. To demonstrate that ANXA2 induces glycolytic metabolic changes through c-Myc, we first analyzed the TCGA database, which revealed a positive correlation between ANXA2 and MYC expression (Supplementary Fig. 17 J). Further reporter gene assays showed that ANXA2 binds to the localization signal in the 3' UTR of c-Myc mRNA and regulates the transcriptional activity of c-Myc (Supplementary Fig. 17 K). Overexpression of ANXA2 enhanced the direct binding of c-Myc to the promoters of its downstream glycolytic genes, including LDHA, ENO1, PKM2, and TPI1 (Supplementary Fig. 17L). We subsequently established cells with c-Myc overexpression and knockdown (Supplementary Fig. 18 A). The results demonstrated that overexpression of c-Myc significantly reversed the inhibitory effects of ANXA2 knockdown on proliferation, migration, invasion and glycolysis in H1299 cells. Conversely, knockdown of c-Myc abrogated the ANXA2 overexpression-induced exacerbation of NSCLC progression and glycolytic enhancement (Supplementary Fig. 18B-P). Thus, the ANXA2/c-Myc signaling axis mediates SerpinB7-regulated glycolysis. SerpinB7 was associated with the expression levels of ANXA2 and glycolytic related molecules in NSCLC tumor tissues To further validate the above findings and explore the clinical significance of SerpinB7, immunofluorescence staining of lung cancer tissues revealed partial co-localization of SerpinB7 and ANXA2 in lung cancer tissues. The expression levels of both proteins were higher in LUAD and LUSC tissues compared to adjacent normal tissues (Fig. 8A). The expression of SerpinB7 and ANXA2 showed a positive correlation (Fig. 8B–C). Furthermore, analysis of clinical data from the TCGA database indicated that the expression of SerpinB7 and ANXA2 was highly and positively correlated in NSCLC patient tissues (Fig. 8D–E). Patients with high co-expression of both SerpinB7 and ANXA2 exhibited the worst overall survival (Fig. 8F). We used immunohistochemistry (IHC) to detect the expression levels of SerpinB7, ANXA2, LDHA, ENO1, PKM2 and TPI1 in six pairs of clinical NSCLC tissues and their corresponding adjacent normal tissues. Compared to adjacent normal tissues, the expression of SerpinB7, ANXA2, LDHA, ENO1, PKM2 and TPI1 was significantly elevated in NSCLC tissues (Fig. 8G).Fig. 8Levels of SerpinB7 and ANXA2 are clinically relevant in patients with NSCLC. A-C Representative micrographs of LUAD, LUSC tissue and adjacent normal tissue stained with anti-ANXA2 and anti-SerpinB7 antibodies; A quantitative correlation analysis of average ANXA2 and SerpinB7 expression per high-magnification field in LUAD (n=20) and LUSC (n=23). scale bar, 50 μm. D-E Scatter plots for correlation analysis of SerpnB7 and ANXA2 expression in LUAD and LUSC by TCGA-LUAD and TCGA-LUSC database. LUAD: n=513, LUSC: n=501. F Kaplan-Meier analysis of overall survival in patients with lung cancer depending on the number of upregulated molecular markers. G H&E staining and immunohistochemical detection of SerpinB7, ANXA2, PKM2, ENO1, LDHA and TPI1 in LUAD, LUSC and adjacent normal tissues. This figure shows quantitative analysis of the mean optical density values of SerpinB7, ANXA2, PKM2, ENO1, LDHA and TPI1 per high-magnification field of view in LUAD and LUSC. scale bar, 100 μm. n=6. H In vivo IVIS images and quantitative fluorescence analysis results of orthotopic lung cancer models in each group after inoculation. The BALB/c nude mice were treated with PBS, tHFn(+)1-159, sicon@tHFn(+)1-159 or siSerpinB7@tHFn(+)1–159. n=6. I Representative H&E staining of lung tissues from each group and statistical analysis of tumor nodules in the orthotopic lung cancer model. The BALB/c nude mice were treated with PBS, tHFn(+)1-159, sicon@tHFn(+)1–159 or siSerpinB7@tHFn(+)1–159. n=6. J Schematic diagram summarizing our working model: SerpinB7 stabilizes ANXA2 by interrupting NEDD4L-mediated ubiquitination degradation, thereby further promoting the c-Myc signaling pathway, which subsequently enhances glycolytic activity and facilitates NSCLC tumor promotion. Targeting endogenously overexpressed SerpinB7 in tumors may block disease progression in NSCLC patients Levels of SerpinB7 and ANXA2 are clinically relevant in patients with NSCLC. A-C Representative micrographs of LUAD, LUSC tissue and adjacent normal tissue stained with anti-ANXA2 and anti-SerpinB7 antibodies; A quantitative correlation analysis of average ANXA2 and SerpinB7 expression per high-magnification field in LUAD (n=20) and LUSC (n=23). scale bar, 50 μm. D-E Scatter plots for correlation analysis of SerpnB7 and ANXA2 expression in LUAD and LUSC by TCGA-LUAD and TCGA-LUSC database. LUAD: n=513, LUSC: n=501. F Kaplan-Meier analysis of overall survival in patients with lung cancer depending on the number of upregulated molecular markers. G H&E staining and immunohistochemical detection of SerpinB7, ANXA2, PKM2, ENO1, LDHA and TPI1 in LUAD, LUSC and adjacent normal tissues. This figure shows quantitative analysis of the mean optical density values of SerpinB7, ANXA2, PKM2, ENO1, LDHA and TPI1 per high-magnification field of view in LUAD and LUSC. scale bar, 100 μm. n=6. H In vivo IVIS images and quantitative fluorescence analysis results of orthotopic lung cancer models in each group after inoculation. The BALB/c nude mice were treated with PBS, tHFn(+)1-159, sicon@tHFn(+)1-159 or siSerpinB7@tHFn(+)1–159. n=6. I Representative H&E staining of lung tissues from each group and statistical analysis of tumor nodules in the orthotopic lung cancer model. The BALB/c nude mice were treated with PBS, tHFn(+)1-159, sicon@tHFn(+)1–159 or siSerpinB7@tHFn(+)1–159. n=6. J Schematic diagram summarizing our working model: SerpinB7 stabilizes ANXA2 by interrupting NEDD4L-mediated ubiquitination degradation, thereby further promoting the c-Myc signaling pathway, which subsequently enhances glycolytic activity and facilitates NSCLC tumor promotion. Targeting endogenously overexpressed SerpinB7 in tumors may block disease progression in NSCLC patients Targeting SerpinB7 suppresses NSCLC tumor growth and metastasis Based on the above findings, we hypothesized that targeting SerpinB7 could represent a unique therapeutic strategy for NSCLC. To explore this possibility, we first analyzed the expression of SerpinB7 in normal tissues using the GEPIA3 database (Supplementary Fig. 19 A). SerpinB7 was negative in most normal tissues and showed low expression levels in the lung and kidney (Supplementary Fig. 19 A). SerpinB7 was highly expressed in various tumors, especially in NSCLC (Supplementary Fig. 19B). We subsequently developed a specific siRNA targeting SerpinB7 and encapsulated it within a ferritin shell by adjusting the pH (Supplementary Fig. 19C-D) [59]. Ferritin, as an important iron storage protein, possesses a hollow interior cavity structure suitable for drug delivery. Ferritin nanocages offer significant advantages due to their human origin, resulting in low toxicity and low immunogenicity. tHFn(+)1–159 is an optimized ferritin nanocarrier that is more suitable for siRNA encapsulation and enables lysosomal escape, demonstrating excellent delivery efficiency and therapeutic efficacy in tumor treatment [59]. Furthermore, ferritin has a high affinity for transferrin receptor 1 (TfR1), which is often highly expressed in cancer cells [60]. We validated that TfR1 is indeed highly expressed in cancer cells (Supplementary Fig. 19E), and that 50 μg/mL of SerpinB7 siRNA encapsulated within the ferritin shell (siSerpinB7@tHFn(+)1–159) reduced SerpinB7 expression (Supplementary Fig. 19 F). Next, we investigated the biodistribution and biosafety of siSerpinB7@tHFn(+)1–159. PBS, tHFn(+)1–159, cy5-sicon@tHFn(+)1–159, and cy5-siSerpinB7@tHFn(+)1–159 were injected into mice bearing orthotopic lung cancer models via the tail vein. Ex vivo fluorescence imaging revealed significant fluorescent signals in the lung tissues of mice treated with cy5-sicon@tHFn(+)1–159 and cy5-siSerpinB7@tHFn(+)1–159, indicating that siSerpinB7@tHFn(+)1–159 accumulated at the tumor site in the lungs (Supplementary Fig. 19G). Additionally, we evaluated the potential toxicity of siSerpinB7@tHFn(+)1–159 to normal tissues in healthy BALB/c nude mice. The major organs of mice treated with siNC@tHFn(+)1–159 and siSerpinB7@tHFn(+)1–159 maintained typical physiological morphology, with no obvious histopathological abnormalities or damage observed (Supplementary Fig. 19H). In conclusion, the siSerpinB7@tHFn(+)1–159 nanocarrier can accumulate at tumor sites and exhibits excellent biosafety. Orthotopic mouse models demonstrated that siSerpinB7@tHFn(+)1–159 inhibits NSCLC tumor growth and metastasis in vivo (Fig. 8H-I). Taken together, these results further elucidate the molecular mechanism by which SerpinB7 promotes NSCLC tumorigenesis: through enhancing ANXA2-mediated c-Myc signaling pathway, thereby regulating tumor proliferation, metastasis and glycolytic process. Targeting SerpinB7 via siSerpinB7@tHFn(+)1–159 may represent a unique therapeutic strategy for NSCLC (Fig. 8 J). Discussion A substantial body of evidence demonstrates that the Serpin protein family plays a widespread and critical role in tumorigenesis, making it a promising biomarker for early disease diagnosis, prognosis assessment, and potential therapeutic targeting [28]. The functional complexity of Serpins in cancer stems from their intricate interactions with other proteins. Notably, elevated expression of neuroserpin and SerpinB2 in brain-metastasizing cells of lung and breast cancer promotes metastatic progression by inhibiting plasminogen activation [28]. SerpinB9 expression in cancer cells confers resistance to CAR T-cell therapy [61], whereas its loss enhances tumor cell susceptibility to T cell-mediated killing [62]. Although Serpins are endogenous regulators of hemostasis [63], their oncogenic roles remain paradoxical. Recent studies have linked SerpinB7 to the pathogenesis of lung cancer [21, 22], but its precise mechanisms remain unclear. Our bioinformatics analysis revealed significant overexpression of SerpinB7 in NSCLC tissues, suggesting its potential as a key regulator in NSCLC development. Interestingly, despite comparable SerpinB7 protein levels between CL1-0 and CL1-5 cells, the GSE297013 dataset revealed elevated SerpinB7 mRNA expression in CL1-5 cells, implying that post-transcriptional or post-translational mechanisms may regulate SerpinB7 function in tumor progression. Further investigations using paired cell lines with well-defined progression stages, tumor tissue sections, and gain- or loss-of-function experiments are needed to clarify the specific role of SerpinB7 in tumor progression. Preliminary quantitative comparison showed that SerpinB7 is highly expressed in SCLC compared to adjacent normal tissues. In SCLC, SerpinB7 expression is predominantly localized in the cytoplasm of tumor cells (Supplementary Fig. 20 A). This expression pattern is consistent with that observed in NSCLC as reported in our paper. Future studies will aim to validate these findings in SCLC models and explore the common and specific mechanisms across different histological subtypes. EGFR, KRAS, and ALK are the three most important and mutually exclusive driver genes in NSCLC [64]. Research indicates that Serpin genes (including SerpinB4 and SerpinB7) are markedly upregulated following ALK fusion induction in NSCLC, with STAT3, NF-κB, and AP-1 identified as the primary transcription factors downstream of ALK fusion regulating SerpinB4 [21]. Our analysis suggests that SerpinB7 likely acts as a downstream “effector” of these oncogenic mutations rather than a direct target. This “convergent” upregulation implies that SerpinB7 may function as a universal metabolic rheostat in NSCLC, with its expression widely activated by various upstream driver mutations. This potentially explains its broad prognostic significance across diverse patient cohorts. To elucidate whether SerpinB7 exerts its prognostic role independently of other co-regulated family members, we performed detailed transcriptional correlation analyses. In both LUAD and LUSC, the mRNA expression levels of SerpinB7 showed no significant correlation with those of ALK (Supplementary Fig. 21 A). Although SerpinB1, B2, B3, B4, B5, B8, B10, B11, B12 and B13 displayed statistical correlations with SerpinB7 (Supplementary Fig. 21B), among these correlated members, SerpinB4 has been reported in NSCLC or ALK-positive tumors [21], with mechanisms distinct from those of SerpinB7. Therefore, while some Serpin family members may be modestly modulated by shared upstream factors, the expression pattern of SerpinB7 and its association with poor prognosis are unique and cannot be explained by other known family members with established oncogenic roles, such as ALK, or by most other family members. Functionally, KRAS is a key driver of glycolysis [65], and our study found that SerpinB7 promotes glycolysis, suggesting a potential synergistic effect between the two. Meanwhile, EGFR mutation can also activate c-Myc [66]. This study has not yet conducted a detailed stratified analysis based on ALK/EGFR/KRAS mutation status. Investigating the interplay between the SerpinB7/ANXA2/c-Myc axis and these classical driver gene mutations represents a logical and important next step.Although the exact cause of SerpinB7 overexpression remains to be elucidated, our in vitro and in vivo experiments confirm that SerpinB7 robustly drives malignant phenotypes in NSCLC, underscoring its pivotal role in disease progression. Traditionally, research on the mechanistic roles of Serpin proteins in tumors has focused on their canonical inhibitory functions through substrate protease binding. For example, HAI-2, a novel plasmin inhibitor, effectively suppresses the invasion and metastasis of lung cancer cells, highlighting its critical role as a serine protease inhibitor in lung cancer [18]. Similarly, SerpinB9, a physiological inhibitor of granzyme B, promotes tumor cell death via granzyme B-dependent mechanisms when genetically deleted [17]. However, accumulating evidence suggests that Serpin proteins possess non-canonical functions beyond classical protease inhibition [19]. Recent studies have revealed that SerpinE2 promotes hepatocellular carcinoma metastasis by preventing c-Cbl-mediated EGFR ubiquitination and degradation, demonstrating a non-canonical mechanism [67]. Meanwhile, SerpinE1 directly interacts with the low-density lipoprotein receptor-related protein (LRP1), activating the AKT/ERK1/2 signaling pathway to induce endothelial-mesenchymal transition (EndoMT) in lymphatic endothelial cells (LECs) [68]. Although SerpinB7 has been classically defined as a serine protease inhibitor targeting plasmin [48], its non-enzymatic regulatory roles in cell signaling have not been thoroughly explored. To elucidate the molecular mechanisms underlying SerpinB7-mediated enhancement of glycolysis in NSCLC, we employed an integrated approach combining proteomics, GST-pull-down with co-immunoprecipitation (Co-IP) and LC–MS/MS analysis, identifying ANXA2 as the key effector molecule responsible for SerpinB7-regulated cell proliferation, metastasis, and glycolysis. Subsequent Co-IP and co-localization experiments further validated this interaction. Notably, through comprehensive Co-IP assays and AlphaFold 3 structural predictions, we successfully characterized the specific binding domain (non-RCL region) between SerpinB7 and ANXA2. Given the absence of a resolved crystal structure for human SerpinB7, our study relied on computational modeling. Strikingly, Although the RCL mutation reduces the enzymatic inhibitory activity of SerpinB7, the introduction of either wild-type SerpinB7 or the SerpinB7 mutant (MUT#1) into knockout cells significantly enhances the proliferation, metastasis, and glycolytic potential of NSCLC cells. However, the introduction of SerpinB7-MUT#1 partially impairs the pro-proliferative capacity of SerpinB7. These findings collectively support the novel concept that SerpinB7 exerts non-enzymatic functions through its interaction with ANXA2. Our study provides direct evidence for the enzyme activity-independent oncogenic signaling of the Serpins, revealing its pleiotropic roles beyond classical protease inhibition. These critical findings suggest that targeting the non-canonical mechanisms of the Serpins may offer innovative therapeutic strategies for NSCLC intervention. Although aerobic glycolysis is well-documented in NSCLC, its regulatory mechanisms remain incompletely understood. Surprisingly, we found that SerpinB7 overexpression upregulates key glycolytic enzymes (PKM2, LDHA, ENO1, TPI1), which collectively drive malignant progression[69, 70]. These findings support the conclusion that SerpinB7 plays a critical role in tumor metabolism. This study elucidates the impact of SerpinB7-mediated metabolic reprogramming on malignant phenotypes in NSCLC. Given the clinical success of metabolic normalization strategies[71, 72],targeting SerpinB7 as a glycolytic “gatekeeper” may globally suppress tumor cell metabolism. Our study confirmed that ANXA2 is degraded via the ubiquitin–proteasome system (UPS), a target of anticancer therapies[73]. Through predictive analysis combining the UbiBrowser and String databases, along with experimental validation, we unexpectedly discovered that SerpinB7 competitively inhibits NEDD4L[74] -mediated ubiquitination of ANXA2, thereby stabilizing ANXA2 protein levels. Thus, this work elucidates how SerpinB7 regulates ANXA2 ubiquitination, enabling its critical role in glycolytic reprogramming and malignant progression in NSCLC. Our findings further emphasize the necessity of studying protein ubiquitination to deepen the understanding of NSCLC oncogenesis and advance therapeutic interventions. ANXA2 is a well-established cancer biomarker [75] and is positively regulated by SerpinB7. Although the metabolic role of ANXA2 in NSCLC tumors remains unclear, recent studies have indicated that ANXA2 overexpression is associated with DNA repair defects, metabolic dysregulation, and poor prognosis in pancreatic ductal adenocarcinoma [24]. This study demonstrates the critical role of ANXA2 in glycolytic regulation in NSCLC, as evidenced by clinical correlation and energy metabolism analysis. Furthermore, we provide multifaceted evidence demonstrating that ectopic ANXA2 expression enhances the glycolytic capacity of NSCLC cells. The mechanisms underlying enhanced glycolysis in tumor cells are multifaceted, involving oncogene and signaling pathway activation, tumor suppressor inactivation, altered expression and activity of metabolic enzymes, and influences from the tumor microenvironment. Extensive research has shown that cancer cell metabolism is regulated by diverse factors, including proto-oncogenes (e.g., Myc), transcription factors (e.g., HIF-1), signaling pathways (e.g., PI3K/Akt/mTOR), and tumor suppressors(e.g.,p53) [13, 76, 77]. These regulators can upregulate multiple glycolytic genes (such as GLUTs, HK2, PFK, ENO1, LDHA and PDK1), thereby enhancing glycolysis and promoting tumor growth [78, 79]. Intriguingly, most candidate glycolytic genes regulated by ANXA2 are also transcriptional targets of MYC or HIF-1α. In NSCLC, SerpinB7 interacts with ANXA2 to synergistically enhance the stability and translational efficiency of c-Myc mRNA at the post-transcriptional level, which is consistent with previous reports [56]. In esophageal cancer, phosphorylated ANXA2 binds to c-Myc, thereby inhibiting the proteasomal degradation of c-Myc [57]. Consistent with our findings, ectopic SerpinB7 overexpression increases c-Myc protein stability, suggesting that ANXA2/c-Myc axis-mediated glycolysis plays an important regulatory role in SerpinB7-driven NSCLC malignancy. This study reveals that SerpinB7 upregulates ANXA2 expression through their interaction, thereby driving malignant progression in NSCLC via ANXA2/c-Myc axis-dependent glycolytic reprogramming. Thus, our work establishes ANXA2 as an indispensable mediator in SerpinB7-driven metabolic reprogramming. In summary, this study demonstrates that upregulation of SerpinB7 in NSCLC promotes tumor glycolysis and malignant progression. Mechanistically, SerpinB7 stabilizes ANXA2 by interfering with NEDD4L-mediated ubiquitination and degradation, and the ANXA2/c-Myc axis mediates SerpinB7-regulated glycolysis. Our findings suggest that targeting endogenously overexpressed SerpinB7 in tumors may block disease progression in NSCLC patients. In this study, the long-term biosafety of siSerpinB7@tHFn(+)1–159 still requires further investigation. Supplementary Information Supplementary Material 1. Supplementary Material 2. Supplementary Material 1. Supplementary Material 2.