CTSZ-dependent anoikis resistance enhances malignant characters of glioblastoma via NF-κB signaling Authors: Yiyang Cao, Wei Wu, Yichang Wang, Beichen Zhang, Haoyu Zhou, Mingjing Zhou, Wei Chen, Maode Wang, Jia Wang Published: 2026 ABSTRACT SummaryGlioblastoma (GBM) remains one of the most aggressive and therapy-resistant malignancies of the central nervous system, with anoikis resistance emerging as a critical facilitator of tumor invasion and therapeutic failure. In this study, we identified cathepsin Z (CTSZ) as a pivotal regulator of anoikis resistance in GBM. Clinical data revealed significant elevation of CTSZ levels in GBM tissues, with its expression strongly correlating with adverse patient outcomes. Through comprehensive functional analysis, we determined that genetic knockdown of CTSZ substantially increased anoikis sensitivity in GBM cells while simultaneously attenuating multiple oncogenic characteristics. Mechanistically, we established that CTSZ-mediated anoikis resistance enhanced multiple malignant traits of GBM by activating nuclear factor kappa B (NF-κB) signaling, as evidenced by the impaired nuclear translocation of p65 following CTSZ-knockdown. Our findings collectively position the CTSZ/NF-κB axis as a crucial determinant of GBM pathogenesis and highlight CTSZ as a potential therapeutic target for this devastating disease.Subject areas: molecular biology, cancer Introduction Glioblastoma (GBM), which is well known as the most aggressive and lethal central nervous system tumor in adults, remains incurable, with a median survival of only 15 months and a 5-year survival rate of <5% despite multimodal therapy including maximal safe surgical resection, adjuvant radiotherapy (RT), and concomitant temozolomide treatment.1,2,3,4,5,6,7,8 Anoikis resistance is reportedly a critical factor contributing to the aggressiveness and poor prognosis of GBM.9 Therefore, it is imperative to investigate the molecular mechanisms underlying anoikis resistance in GBM to develop previously uncharacterized therapeutic strategies. Recent studies demonstrate that anoikis, a specialized form of programmed cell death induced by loss of cell-cell or cell-extracellular matrix (ECM) adhesion, serves as a vital homeostatic mechanism by eliminating displaced cells.10 Originally characterized in epithelial and endothelial systems, anoikis has emerged as a key regulator of malignant properties and metastasis among a wide range of solid tumors, including GBM and breast, lung, and pancreatic cancers.11,12,13 The restoration of anoikis sensitivity can help overcome GBM chemoresistance.14 Acquired anoikis resistance enables detached cells to evade apoptotic signaling and thus facilitates the survival of cells in hostile microenvironments.15 Hence, functional anoikis mechanisms represent a critical ECM-mediated barrier to metastatic dissemination.16 Targeting anoikis may offer a not previously reported therapeutic approach for GBM; however, the expression and functional role of anoikis-related genes in GBM remain unclear. Cathepsin Z (CTSZ), also called cathepsin X, is one of the cysteine cathepsins, which are a family of lysosomal proteases.17 Studies have shown that Arg-Gly-Asp(RGD)-dependent extracellular binding of CTSZ to integrin β3, which localizes in the ECM and binds to cell surfaces, mediates cell adhesion and migration.18 Furthermore, high levels of CTSZ are associated with multiple malignant cell characteristics, including proliferation, invasion, and tumor progression, among various types of cancers, including prostate, colorectal, gastric, and liver cancers, melanoma, pancreatic neuroendocrine tumors, and GBM.16,19,20 It has been reported that cathepsin B promotes cell death via nuclear factor kappa B (NF-κB) activation, which also contributes to the growth and migration of GBM.21,22 CTSG promotes focal adhesion and myofibril protein degradation, as well as cardiomyocyte anoikis.23 However, few studies have been conducted on CTSZ, especially on its role in the regulation of cell adhesion in GBM. Members of the NF-κB transcription factor family comprise five subunits—RelA (p65), RelB, c-Rel, p50, and p52—and five inhibitory proteins, namely, I-κBα, I-κBβ, I-κBε, p105, and p100.24,25 NF-κB signaling has been proven to modulate critical oncogenic processes, such as cell survival, proliferation, invasion, metastasis, angiogenesis, and chemoresistance, in multiple cancer subtypes, including skin, ovarian, breast, cervical, colorectal, and hepatic cancers and GBM.26,27 Several studies have clearly demonstrated that NF-κB activation is associated with poor prognosis in GBM, and that NF-κB-regulated genes are significantly overexpressed in GBM cells.28,29 Furthermore, the activation of NF-κB signaling endows cancer cells with resistance to anoikis and contributes to tumor progression.30,31 Accumulating evidence indicates that NF-κB represses E-cadherin and thereby promotes anoikis resistance in GBM.32 Moreover, pharmacological inhibition of CTSZ has been shown to suppress NF-κB activation by preventing I-κBα proteolysis, ultimately mitigating apoptosis of human neuroblastoma (SH-SY5Y) cells.33 However, the role of CTSZ in regulating the NF-κB signaling pathway under anoikis conditions, particularly in GBM, remains poorly characterized. In this study, CTSZ was identified as a potential biomarker of GBM. CTSZ suppressed anoikis in GBM by activating NF-κB signaling, thus promoting GBM tumor progression. Exogenous inhibition of CTSZ significantly attenuated multiple malignant traits of GBM cells. Collectively, these findings reveal that CTSZ is a reliable prognostic biomarker for GBM and that targeting CTSZ may offer a viable therapeutic strategy against GBM. Results CTSZ is the gene most correlated with anoikis in GBM Data from the TCGA GBM database were utilized to investigate the gene expression characteristics of GBM and establish a proportional hazards model. Univariate Cox regression analysis identified 21,430 prognosis-related genes, among which 534 genes had HRs >1 (Figure 1A). MEGENA clustered these genes into 91 distinct modules (Figure 1B). During correlation analysis of module eigengenes with an established anoikis gene set as the phenotypic reference, module c1-13 showed the most significant positive correlation with anoikis (Figure 1C). Within this key module, we further identified the top five high-confidence candidate genes using the Matthews Correlation Coefficient algorithm (Figure 1D). CTSZ is the most correlated gene to anoikis in GBM (A) The volcano plot is used to display genes associated with the prognosis of GBM (HR > 1 or HR < −1, with p value <0.01). (B) The hierarchical clustering dendrogram displays the results of gene module clustering from multiscale embedded gene network analysis (MEGENA) analysis of differentially expressed genes. (C) This chart displays the correlation coefficients and significance between the characteristic gene modules identified by MEGENA and the anoikis phenotype. (D) The gene network plot assesses the robustness of gene interactions using the Matthews correlation coefficient, selecting the top 5 high-confidence interacting genes. (E) The hierarchical clustering dendrogram visualizes the gene modules identified by WGCNA, with each module assigned a unique color. (F) The module-relationship plot shows the correlation between each gene module and the anoikis phenotype. (G) The module membership vs. gene significance plot is a scatterplot with the x axis representing module membership and the y axis representing gene significance, used to evaluate the strength of association between genes within the module and the anoikis phenotype (MM: gene centrality within a module; higher values indicate greater intramodular importance) (GS: gene-phenotype correlation strength; higher values suggest stronger phenotypic relevance). (H) The Venn diagram shows the overlap between core genes identified by weighted gene co-expression network analysis (WGCNA) and those identified by multiscale embedded gene network analysis (MEGENA). WGCNA re-clustered the identified prognosis-related genes into 20 modules, each represented by a distinct color (Figure 1E).34 Using the anoikis gene set as the phenotypic reference, the turquoise module, which exhibited a significant correlation with the anoikis phenotype, was selected (Figure 1F). Within this module, five core genes were selected based on predefined thresholds for MM (≥0.8) and GS (≥0.2) (Figure 1G). Finally, the intersection of the five core genes, each from WGCNA and MEGENA analyses, identified CTSZ as the target gene associated with malignancy and anoikis in GBM; hence, it was selected for further investigation (Figure 1H). CTSZ expression is significantly upregulated in GBM The expression profiles of CTSZ were collected from TCGA and CGGA datasets. CTSZ expression was found to be more enriched in tumor subtypes with higher WHO grades, particularly in GBM, than in non-tumor tissues (Figures 2A and 2B). Moreover, CTSZ expression was more enriched in human GBM cells than in NHA cells at both the mRNA and protein levels (Figures 2C and 2D). Similarly, CTSZ was significantly upregulated in tumor tissues compared to their paired adjacent non-tumor tissues from surgically resected patients with GBM (Figure 2E). IHC staining performed to further explore the clinical significance of CTSZ revealed that it was more highly expressed in GBM than in lower-grade gliomas (Figure 2F). Kaplan-Meier analysis indicated that elevated CTSZ expression was significantly associated with poorer overall survival in patients with GBM (Figure 2G). Kaplan-Meier survival analysis based on TCGA and CGGA databases indicated similar results (Figures 2H–2K). Altogether, CTSZ can be considered a prognostic biomarker for GBM. Our findings confirmed that U251 and U373 cells exhibit high endogenous CTSZ expression at both the RNA and protein levels. Therefore, we selected these two cell lines for subsequent experiments. The expression of CTSZ is significantly up-regulated in GBM (A) Gene expression analysis of GBM and non-tumor tissues using the TCGA database (∗∗p < 0.01, Student’s t test). (B) Gene expression analysis of gliomas with different WHO grades using the CGGA database (∗∗p < 0.01 and ∗∗∗p < 0.001, Student’s t test). (C and D) qRT-PCR and western blot were conducted to evaluate CTSZ expression in commercial human GBM cell lines and NHA, with GAPDH as an internal reference (n = 3, ∗∗p < 0.01 and ∗∗∗∗p < 0.0001, Student’s t test). (E) Detection of CTSZ expression in tumor and adjacent non-tumor tissues (ANT) from surgically resected patients with GBM by western blot. (F) Representative tissue microarray immunohistochemistry (TMA-IHC) images of CTSZ in glioma samples. (G) Kaplan-Meier analysis of 41 patients with GBM was performed based on the expression of CTSZ via TMA-IHC (p = 0.027). (H and I) Kaplan-Meier analysis of CTSZ expression using GBM patient samples and brain tumor patient samples from the TCGA (log rank test). (J and K) Kaplan-Meier survival analysis of CTSZ expression using GBM patient samples and brain tumor patient samples from the CGGA database (log rank test). CTSZ is functionally required for multiple malignant properties of GBM shCTSZ lentiviruses #1 and #2 were successfully transfected into U251 and U373 cells to eliminate CTSZ expression. RT-PCR and western blot analyses revealed that CTSZ expression was significantly lower in shCTSZ-transduced cells than in shNT controls (Figures 3A and 3B). Moreover, CCK-8 cell proliferation, colony formation, and transwell assays indicated that a range of malignant characteristics of GBM cells, including proliferation, self-renewal capability, and cell migration, were significantly attenuated by CTSZ knockdown (Figures 3C–3G). Furthermore, flow cytometry analysis showed that the percentage of apoptotic cells in the CTSZ knockdown group was higher than that in the control group (Figures 3H and 3I). To simulate the function of CTSZ in vivo tumorigenesis, U251 cells transfected with shCTSZ or shNT were used to establish an intracranial tumor model. Mice transplanted with shCTSZ U251 cells exhibited reduced tumor volumes and significantly prolonged survival times compared to the control group (Figures 3J and 3K). Altogether, these results indicate that CTSZ is functionally required for multiple malignant properties of GBM both in vitro and in vivo. CTSZ is functionally required for multiple malignancies of GBM (A and B) qRT-PCR and western blot were used to detect the expression of CTSZ in U251 and U373 cells transfected with lentiviral shCTSZ#1 and #2, as well as in the control group (n = 3). (C and D) Plate cloning assay was conducted to investigate the effect of CTSZ knockdown on tumor cell proliferation (n = 3). (E) CCK-8 cell proliferation assay was also used to investigate the effect of CTSZ knockdown on tumor cell proliferation (n = 3). (F and G) Transwell migration assay was used to evaluate the effect of CTSZ knockdown on tumor cell migration ability (n = 3). (H and I) Flow cytometry analysis was performed using the Annexin V-PE/7-ADD apoptosis detection kit to measure the apoptosis rate of U251 and U373 cells transduced with lentiviral shCTSZ compared to the negative control group (n = 3). The quadrant definitions are as follows: Q1 (upper-left, 7-AAD+/Annexin V−): Cellular debris or mechanically damaged cells; Q2 (upper-right, 7-AAD+/Annexin V+): Late apoptotic cells; Q3 (lower-left, 7-AAD-/Annexin V−): Viable/normal cells; Q4 (lower-right, 7-AAD-/Annexin V+): Early apoptotic cells. (J) Kaplan-Meier analysis showing the survival of orthotopic xenograft nude mice in the control group and the CTSZ knockdown group (p = 0.033, n = 5, log rank test). (K) Representative H&E staining of nude mice brain tissue sections showing the results after intracranial transplantation. (p = 0.033, n = 5, log rank test). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. Differences between the two groups were compared using a two-tailed Student’s t test. For comparisons among multiple groups, one-way ANOVA with post hoc tests was used. All data are presented as the mean ± standard deviation (SD). CTSZ knockdown can reduce the malignancy of GBM cells under low-attachment conditions To investigate the impact of CTSZ on the malignancy of GBM under anoikis conditions, we implemented a standardized anoikis model by culturing GBM cells in an ultra-low-attachment plate, as described previously.35 It has been reported that cells tend to form spheroid structures in an anchorage-independent environment. Moreover, cells that form larger spheroids have a higher survival rate.36 To assess the functional role of CTSZ in spheroid formation by GBM cells, U251 and U373 cells were transfected with shCTSZ or shNT lentiviruses under anchorage-independent culture conditions. When cells were seeded at a density of 200 cells/well in ultra-low-attachment 96-well plates, the shNT group formed significantly larger spheroids, and the total number of spheroids formed (diameter >100 μm) was notably higher, whereas the shCTSZ group showed a trend toward forming smaller spheroids. Moreover, Calcein AM/EthD-1 staining revealed a higher cell death rate in the knockdown group (Figures 4A–4C). This observation indicates that under anchorage-independent conditions, CTSZ knockdown made GBM cells more prone to apoptosis. Flow cytometry also detected a significant increase in the apoptosis rate in the CTSZ-knockdown group (Figures 4D and 4E). Furthermore, CCK-8 cell proliferation and transwell migration assays conducted using similar plates also showed that CTSZ knockdown significantly inhibited tumor cell proliferation and migration (Figures 4F–4H). In conclusion, our results indicate that CTSZ knockdown can reduce the malignancy of GBM cells in an anchorage-independent environment, which is vital for tumorigenesis. CTSZ knockdown can reduce the malignancy of GBM cells under low-attachment conditions (A–C) Phase-contrast images (left), calcein-AM staining for live cells (middle), and EthD-1 staining for dead cells (right). Bar charts show the number of spheroids formed by each group of U251 and U373 cells after 48 h of culture under low-attachment conditions (only spheroids with a diameter greater than 100 μm were counted) and the average diameter of the largest spheroids. (D and E) Flow cytometry analysis was performed using the Annexin V-PE/7-AAD apoptosis detection kit to determine the apoptosis rate of cells in each group after 48 h of culture under low-attachment conditions. The quadrant definitions are as follows: Q1 (upper-left, 7-AAD+/Annexin V−): Cellular debris or mechanically damaged cells; Q2 (upper-right, 7-AAD+/Annexin V+): Late apoptotic cells; Q3 (lower-left, 7-AAD-/Annexin V−): Viable/normal cells; Q4 (lower-right, 7-AAD-/Annexin V+): Early apoptotic cells. (F) CCK-8 cell proliferation assay was conducted to investigate the effect of CTSZ knockdown on tumor cell proliferation under low-attachment conditions. (G and H) Transwell migration assay was used to evaluate the effect of CTSZ knockdown on tumor cell migration ability under low-attachment conditions. n = 3, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. Differences between the two groups were compared using a two-tailed Student’s t test. For comparisons among multiple groups, one-way ANOVA with post hoc tests was used. All data are presented as the mean ± standard deviation (SD) of three independent experiments. CTSZ is crucial for anoikis resistance and tumor formation Apoptosis in response to lack of adhesion or inappropriate adhesion to the ECM has been termed anoikis; however, despite its own definition, anoikis is essentially an apoptotic process.37 In this study, western blot analysis showed increased cleavage of caspase-3 and a reduced Bcl-2/Bax ratio in the CTSZ-knockdown group under both low- and normal-attachment conditions. These findings confirm that CTSZ suppression induced apoptosis in GBM cells (Figure 5A).38,39,40,41,42 CTSZ is crucial for anoikis resistance and tumor formation (A) Western blot was used to detect the expression of apoptosis-related proteins in U251 and U373 cells after 48 h of culture under Normal-attachment and low-attachment conditions. (B and C) Flow cytometry analysis was performed to compare the apoptosis rate of cells in each group after 48 h of culture under Normal-attachment and low-attachment conditions. The quadrant definitions are as follows: Q1 (upper-left, 7-AAD+/Annexin V−): Cellular debris or mechanically damaged cells; Q2 (upper-right, 7-AAD+/Annexin V+): Late apoptotic cells; Q3 (lower-left, 7-AAD-/Annexin V−): Viable/normal cells; Q4 (lower-right, 7-AAD-/Annexin V+): Early apoptotic cells. (D) CCK-8 cell proliferation assay was conducted to compare the effects of CTSZ knockdown on the proliferation of tumor cells cultured under Normal-attachment and low-attachment conditions. (E and F) Transwell migration assay was used to compare the effects of CTSZ knockdown on tumor cell migration ability under Normal-attachment and low-attachment conditions. n = 3, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. Differences between the two groups were compared using a two-tailed Student’s t test. For comparisons among multiple groups, one-way ANOVA with post hoc tests was used. All data are presented as the mean ± standard deviation (SD) of three independent experiments. Anoikis resistance, the ability to evade cell death under low-attachment conditions in an environment devoid of adhesion to the ECM, is a prerequisite step for cancer metastasis.36 To specifically assess CTSZ-dependent anoikis resistance in GBM progression independent of its other oncogenic effects, we compared the GBM cell phenotypes under low- (to simulate anoikis conditions) and normal-attachment conditions. Flow cytometry indicated that, compared with cells under normal-attachment conditions, CTSZ-knockdown cells under low-attachment conditions exhibited a higher rate of apoptosis (Figures 5B and 5C). CCK-8 cell proliferation and transwell migration assays indicated that the proliferative and migratory capacities of CTSZ-knockdown cells were further diminished under low-attachment conditions compared with those under normal-attachment conditions (Figures 5D–5F). Collectively, these results demonstrate that the downregulation of CTSZ expression can inhibit GBM progression by enhancing anoikis. CTSZ modulates the NF-κB signaling pathway, thereby influencing the malignant progression of GBM To further elucidate the underlying mechanisms of CTSZ in GBM, genome-aligned RNA sequencing was performed on CTSZ-knockdown and control samples (three each) to identify DEGs. Comparative transcriptomic analysis identified 147 DEGs (|log2FC| > 2, adjusted p < 0.05) between CTSZ-knockdown U251 cells and controls, including 86 upregulated (red) and 61 downregulated (blue) genes (Figures 6A and 6B). KEGG pathway analysis of these DEGs revealed CTSZ-associated regulation of cell-adhesion molecules, which have been confirmed as critical mediators of anoikis (Figure 6C).43,44 GSEA demonstrated significant enrichment of NF-κB signaling pathway-related genes in CTSZ-expressing cells (Figure 6D).45 The NF-κB signaling pathway was downregulated in samples with CTSZ knockdown (Figure 6E). CTSZ modulates the NF-κB signaling pathway, thereby influencing the malignant progression of GBM (A) Volcano plot shows differentially expressed genes between CTSZ-knockdown U251 cells and control group cells (log2FC > 2 or < −2, adjusted p value < 0.05). (B) Heatmap displays significantly differentially expressed genes between experimental and control groups. (C) KEGG pathway analysis of up - and down - regulated genes in CTSZ-knockdown U251 cells versus control. (D) Gene set enrichment analysis (GSEA) shows a positive correlation between CTSZ and NF-κB signaling. (E) GSEA curve plot of the NF-κB pathway gene set shows a comparison of gene expression profiles between CTSZ-knockdown U251 cells and control group cells. NES: normalized enrichment score; p value indicates enrichment significance. (F) Western blot analysis was used to detect the levels of NF-κB signaling pathway proteins and apoptosis-related proteins in CTSZ-knockdown GBM cells cultured under Normal-attachment or low-attachment conditions, with or without TNF-α pretreatment. (G–I) Phase-contrast images (left), calcein-AM staining for live cells (middle), and EthD-1 staining for dead cells (right). Bar graphs show the effect of TNF-α on the number of spheroids (only spheroids with a diameter greater than 100 μm were counted) and the average maximum diameter of spheroids in CTSZ-knockdown GBM cells after 48 h of culture under low-attachment conditions. (J and K) Flow cytometry analysis was performed to assess the effect of TNF-α on the apoptosis rate of CTSZ-knockdown U251 cells under Normal-attachment or low-attachment conditions. The quadrant definitions are as follows: Q1 (upper-left, 7-AAD+/Annexin V−): Cellular debris or mechanically damaged cells; Q2 (upper-right, 7-AAD+/Annexin V+): Late apoptotic cells; Q3 (lower-left, 7-AAD-/Annexin V−): Viable/normal cells; Q4 (lower-right, 7-AAD-/Annexin V+): Early apoptotic cells. (L) CCK-8 cell proliferation assay was used to verify the effect of TNF-α on the proliferative capacity of CTSZ-knockdown U251 cells cultured under Normal-attachment or low-attachment conditions. (M and N) Plate cloning assay was performed to evaluate the effect of TNF-α on the proliferative capacity of CTSZ-knockdown U251 cells cultured under Normal-attachment conditions. (O and P) Transwell migration assay was conducted to investigate the effect of TNF-α on the migration ability of CTSZ-knockdown U251 cells cultured under Normal-attachment or low-attachment conditions. n = 3, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. For comparisons among multiple groups, one-way ANOVA with post hoc tests was used. All data are presented as the mean ± standard deviation (SD) of three independent experiments. It is well known that the TNF-α-dependent activation of NF-κB signaling regulates vital cellular activities such as proliferation, apoptosis, and inflammation.46,47,48 Therefore, we treated CTSZ-knockdown U251 cells with TNF-α. At the mechanistic level, western blot analysis confirmed that deficiency of CTSZ significantly attenuates IKKα phosphorylation and impedes IκBα degradation, thereby suppressing the activation of the NF-κB pathway. Collectively, CTSZ knockdown significantly reduced the levels of proteins in the NF-κB signaling pathway and apoptosis-related biomarkers; TNF-α treatment could partially reverse these effects (Figure 6F). Since it is well established that NF-κB activation depends on the IKK complex—primarily composed of the catalytic subunits IKKα and IKKβ and the regulatory subunit IKKγ/NEMO.49 We further investigated the relationship between CTSZ and the expression of IKKβ and IKKγ using western blot analysis. The results showed that regulating the expression of CTSZ did not significantly change the phosphorylation level of IKKβ (P-IKKβ) or the total protein amount of IKKγ (Figure S1A). Meanwhile, it has been reported in the literature that the NF-κB pathway can be classified into canonical and non-canonical branches.50 To investigate whether CTSZ can also induce the non-canonical NF-κB pathway, western blot analysis was performed again. The results demonstrated that there were no significant changes in p100/p52 processing or RelB expression upon the modulation of CTSZ levels (Figure S1A). Taken together, our findings suggest that CTSZ promotes progression and tumorigenesis of GBM via the canonical NF-κB pathway in an IKKα-independent manner. To further validate the role of CTSZ in activating the NF-κB signaling pathway in GBM cells, confocal immunofluorescence analysis and multiplex immunohistochemical staining assays were employed. The results showed that deficiency of CTSZ significantly inhibited the nuclear translocation of p65 (Figures S1B and S1C). Consistent with this observation, the absence of CTSZ markedly suppressed the RNA expression levels of downstream target genes of the NF-κB pathway, including the p65 target genes IL-6 and IL-1β (Figure S1D). These findings support a mechanism by which CTSZ promotes NF-κB pathway activation through enhancing p65 nuclear translocation. To functionally validate the regulatory role of CTSZ in the NF-κB signaling pathway, we conducted spheroid formation and flow cytometry analyses, and CCK-8 cell proliferation, plate cloning, and transwell migration assays. As expected, CTSZ knockdown decreased the proliferative and migratory capacities of GBM cells and promoted anoikis, whereas TNF-α treatment partially reversed these effects (Figures 6G–6P). Collectively, these findings demonstrate that CTSZ promotes the malignancy of GBM by activating the NF-κB signaling pathway. Discussion This study elucidates that CTSZ directly drives anoikis resistance in GBM cells by activating the NF-κB signaling pathway, independent of the known ITGA5/PI3K/AKT or EGFR/Integrin β1/FAK pathways.14,51 Differing from previous studies that focused on integrin (such as ITGA5) or growth factor receptor (such as EGFR) pathways, this study found that CTSZ regulates the core transcriptional hub, NF-κB, affecting the expression of a series of downstream genes related to survival and invasion, thus providing a mechanistic explanation for anoikis resistance at the level of protein homeostasis. This discovery not only enriches the regulatory network of anoikis resistance in GBM but also highlights the value of CTSZ as a target for therapeutic intervention. CTSZ expression was significantly higher in GBM than in non-tumor tissues and low-grade gliomas, and its high expression was associated with poor patient prognosis.16 Functional validation demonstrated that CTSZ promotes GBM cell proliferation, migration, and invasion while inhibiting apoptosis, establishing its role as a key oncogenic driver in GBM progression. Notably, evidence indicates that CTSG (a member of the cathepsin protease family) downregulates focal adhesion kinase to induce anoikis.52 As a special member of the cathepsin family,53 CTSZ modulates cell adhesion molecules and focal adhesion formation, thereby influencing ECM protein attachment.54 Although its carboxypeptidase activity may not directly degrade ECM components, CTSZ can modulate cell adhesion/migration through integrin binding or proteoglycan-mediated integrin activation, thereby indirectly influencing ECM interactions to promote tumor migration, invasion, and proliferation.55 These findings suggest that CTSZ may also induce anoikis resistance. Verifying this, our experimental results demonstrate that compared with control cells, CTSZ-knockdown cells exhibited significantly reduced malignant behaviors under low-attachment conditions, confirming that CTSZ deficiency indeed markedly decreases anoikis resistance in GBM cells. Accumulating evidence is establishing a close association between NF-κB and anoikis resistance.56 Notably, emerging studies are revealing that NF-κB induces anoikis resistance in GBM by suppressing E-cadherin expression. The present study demonstrates that CTSZ deficiency leads to the downregulation of the NF-κB signaling pathway. Importantly, restoring the activity of this pathway significantly enhanced both tumor malignancy and anoikis resistance, which is consistent with previous findings. Thus, the NF-κB signaling pathway was identified as a critical downstream mediator of CTSZ function. This places the CTSZ/NF-κB axis within a growing network of signaling pathways implicated in GBM therapy resistance and aggressiveness. For instance, Liu et al. demonstrated that a three-dimensional (3D)-bioprinted glioma model exhibited enhanced radioresistance mediated through the ITGA2/p-AKT pathway,57 highlighting the importance of tumor microenvironment interactions in activating pro-survival signals. Similarly, Qiao et al. identified that TMED3 promotes GBM malignancy via the ZBTB7A signaling axis.58 Although ITGA2, TMED3, and CTSZ represent distinct molecular initiators, their convergence on pathways underscores the redundancy and complexity of survival signaling in GBM. Our work adds CTSZ as an upstream regulator of NF-κB specifically in the context of anoikis resistance, a phenotype essential for metastatic spread. Unlike the focus on RT resistance in 3D models57 or general proliferation/migration,58 this study links CTSZ to the evasion of detachment-induced death, offering a complementary perspective on how GBM cells overcome physiological barriers. Notably, given the marked tumor heterogeneity in GBM, therapeutic strategies targeting single biomarkers such as CTSZ may trigger unpredictable compensatory responses. These limitations underscore the necessity of establishing an integrated framework incorporating multi-omics analysis and personalized therapeutic evaluation before clinical intervention. Limitations of the study Although this study has preliminarily elucidated the potential role of CTSZ in GBM and related pathways, further mechanistic investigations are required to fully delineate the molecular basis by which CTSZ contributes to GBM pathogenesis through enhanced anoikis resistance, and to validate its clinical significance. First, the clinical sample size of this study is relatively limited, primarily owing to the difficulty in obtaining high-quality paired surgical specimens. A smaller sample size may affect statistical power, increase the uncertainty of the conclusions, and make it difficult to comprehensively cover the heterogeneity of GBM, which limits the generalizability of the research findings to some extent. In the future, it is necessary to validate the expression pattern and prognostic value of CTSZ using a larger-scale, multi-center clinical cohort to enhance the reliability of the conclusions. Second, complementary studies on CTSZ overexpression phenotypes should be conducted. Both in vitro and in vivo experiments are needed to systematically evaluate the consequences of CTSZ overexpression for more comprehensive conclusions. In addition, the precise molecular interactions among CTSZ function, NF-κB pathway regulation, and the anoikis microenvironment remain unclear. In particular, as a lysosomal cysteine protease, the specific molecular pattern by which CTSZ activates NF-κB has not been elucidated. Whether its effect is through direct intracellular signal transduction or indirect regulation by altering secreted factors remains to be identified. So future investigations employing conditioned medium experiments, co-immunoprecipitation, or proximity ligation assays will be necessary to determine whether CTSZ regulates NF-κB directly or through indirect mechanisms. Meanwhile, this study has not systematically screened whether other survival pathways (such as PI3K-AKT and MAPK) are activated as compensatory mechanisms after CTSZ inhibition. Moreover, additional validation is required to confirm whether NF-κB signaling serves as the sole or predominant downstream effector pathway mediating CTSZ-induced anoikis resistance; this may be addressed through co-immunoprecipitation experiments. Furthermore, there are significant knowledge gaps regarding the upstream regulatory mechanisms controlling CTSZ expression in GBM. Promoter analysis coupled with chromatin immunoprecipitation assays may provide preliminary insights to plug these gaps. Finally, while the ultra-low-attachment culture system is a well-established model for studying anoikis, it represents a simplification of the complex tumor microenvironment in vivo. This model does not fully recapitulate the three-dimensional architecture, heterotypic cell-cell interactions, or dynamic ECM remodeling that occur in actual GBM tumors. Future studies employing brain organoid-based anoikis models of GBM, once technically feasible, may provide more physiologically relevant insights. Resource availability Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Dr. Jia Wang (jiawang_xjtu@163.com). Materials availability This study did not generate new unique reagents. Data and code availability •Data: RNA-seq data have been deposited in the SRA database under accession code SRP619480. Data: RNA-seq data have been deposited in the SRA database under accession code SRP619480. •Code: This paper does not report original code. Code: This paper does not report original code. •Other items: Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Other items: Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Acknowledgments This study was supported by the National Natural Science Foundation of China (82173285) and Innovation Capacity Support Plan of Shaanxi Province (2024ZC-KJXX-091). Author contributions Y.C., J.W., and W.W. designed the research. M.W. and J.W. guided the experiments. Y.C. wrote the paper. W.W and J.W. supervised the study and revised the manuscript. Y.C., B.Z., Y.W, H.Z., and M.Z. performed the experiments and analyzed data. W.C. provided patient samples and performed IHC staining. All authors reviewed and approved the manuscript. J.W. acquired the funding. Declaration of interests The authors declare no competing interests. STAR★Methods Key resources table REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesanti-CTSZAbcamCat# ab180580anti-NF-κB p65Cell Signaling TechnologyCat# 8242; RRID: AB_10859369anti-p-IKKαAbcamCat# ab38515; RRID: AB_881450anti-IKKαProteintechCat# 84372-5-RR; RRID: AB_3671908anti-IκBαZenbioCat# R23322anti-caspase-3PTM BIOCat# 5752anti-cleaved caspase-3AbcamCat# ab32042; RRID: AB_725947anti-BCL-2Cell Signaling TechnologyCat# 4223Tanti-BAXZenbioCat# 380709anti-GAPDHAbcamCat# ab8245; RRID: AB_2107448anti-rabbit IgGCell Signaling TechnologyCat# 7074; RRID: AB_2099233anti-mouse IgGCell Signaling TechnologyCat# 7076; RRID: AB_330924Anti-CTSZServicebioGB112171Immunohistochemistry kit (Anti - rabbit secondary antibody)ServicebioG1215anti-NF-κB p65ServicebioGB15997Bacterial and virus strainspsPAX2Addgene#12260pMD2.GAddgene#12259Biological sampleshuman tissue samplesthe First Affiliated Hospital of Xi’an Jiaotong UniversityApproval No. 2021-695Chemicals, peptides, and recombinant proteinsDMEMGibcoC11995500BTFBSVazymeF103-01penicillin/streptomycinBiosharpBL505ALipo8000Beyotime BiotechnologyC0533puromycinThermo Fisher ScientificA1113803RevertAid Master MixThermo Fisher ScientificM163252× Universal SYBR Green Fast qPCR MixMedChemExpressHY-K0501ARIPA bufferServicebio BiologicalG2002protease and phosphatase inhibitorsMedChemExpressHY-K0010/HY-K0021/HY-K00220.1% crystal violetServicebioG1014HematoxylinBeyotime BiotechnologyST2067Critical commercial assaysCell RNA Extraction KitVazymeRC102BCA Protein Assay KitBeyotime BiotechnologyP0012SCalcein AM/EthD-1 Cell Viability Assay KitProteintech GroupCat No. PF00008CCK8 kitBeyotime BiotechnologyC0038mIHC Immunohistochemical Double Staining KitServicebioG1267DAB Chromogenic KitServicebioG1212Annexin V-PE/7-AAD Apoptosis Detection KitVazymeA213Deposited dataRNA-seq datasetSRA databaseSRA: SRP619480Experimental models: Cell linesHuman GBM (U87, U251, U373, and LN229)Chinese Academy of Sciences Cell Bank and maintainedN/AHEK293TChinese Academy of Sciences Cell Bank and maintainedN/ANHAChinese Academy of Sciences Cell Bank and maintainedN/AExperimental models: Organisms/strainsBALB/c-nuShanghai Model Organisms Center, Inc.SM-014OligonucleotidesPrimersThis paperMETHOD DETAILS (Quantitative Real-Time PCR)shRNAThis paperMETHOD DETAILS (Cell Culture and Stable Cell Line Construction)Software and algorithmsImageJNational Institutes of Health(NIH)https://imagej.nih.gov/ij/GraphPad Prism 8.0GraphPad Softwarehttps://www.graphpad.com/featuresSlideViewer3DHISTECHhttps://www.3dhistech.com/software/slide-viewer/The R ProjectBell Laboratoryhttps://www.r-project.org/Other12% precast SDS-PAGE gelsACE BiotechnologyET15420LGelPVDFThermo Fisher Scientific88518ultra-low-attachment 96-well platesCorningCLS3474ultra-low-attachment 6-well platesCorningCLS3471transwell plateCorning3422 Experimental model and study participant details Anoikis model Cells were seeded into ultra-low-attachment 6- and 96-well plates (CLS3471/CLS3474, Corning, USA) and cultured for 48 h. All suspended cells were harvested for subsequent experiments. In vivo intracranial xenograft tumor model Six-week-old male BALB/c-nu mice in good health condition (five per group) were used to establish intracranial xenograft tumor models. The mice were randomly assigned to experimental groups. Humane endpoints for euthanasia included the following criteria: gait instability, hunched posture, >10% body weight loss, or hindlimb paralysis. The mice were euthanized via transcardial perfusion with ice-cold PBS followed by 4% (w/v) paraformaldehyde (PFA). Brains were dissected, fixed in 4% PFA for 24 h, transferred to 10% formaldehyde, and processed for sectioning. Cell lines Human GBM (U87, U251, U373, and LN229), HEK293T, and normal human astrocyte (NHA) cell lines were obtained from the Chinese Academy of Sciences Cell Bank and maintained in our laboratory. All cell lines were authenticated by short tandem repeat (STR) analysis and confirmed to be free of mycoplasma contamination. Ethics statement and human tissue samples All human tissue samples used in this study were obtained with approval from the Scientific Ethics Committee of the First Affiliated Hospital of Xi’an Jiaotong University (Approval No. 2021-695). Written informed consent was acquired from all participating patients. Glioma and non-tumorous tissue samples were collected from patients undergoing surgical resection. Based on pathological examination results, the samples were grouped for subsequent analyses. Animal ethics All animal experiments were approved by the Ethics Committee of Xi’an Jiaotong University School of Medicine (Approval No. 2016-18). Method details Gene expression profiles Gene expression profiles with clinical annotations were systematically retrieved from two authoritative glioma databases: The Cancer Genome Atlas (TCGA, https://www.cancer.gov/ccg/research/genome) and the Chinese Glioma Genome Atlas (CGGA, http://www.cgga.org.cn/).59 Identification of prognosis-related risk genes Cox regression analysis was performed on raw data files using the survivalROC package in R. The results were grouped based on the hazard ratio (HR; either > 1 or < −1), with a pvalue < 0.01. Subsequently, a volcano plot was generated using the ggplot2 package. Weighted gene Co-expression network analysis (WGCNA) The top 25% of the most variable genes were selected using the WGCNA package (v1.72) in R to construct co-expression networks. Pairwise Pearson correlation coefficients were calculated, and soft threshold (β = 8) selection and dynamic tree cutting were performed for module identification. Modules associated with anoikis were determined by correlating module eigengenes with anoikis phenotypes. Gene significance (GS) and module membership (MM) correlations were computed to identify hub genes. Multiscale embedded gene Co-expression network analysis (MEGENA) The MEGENA package in R was applied to differentially expressed genes (DEGs). A planar filtered network was constructed using absolute Pearson correlation coefficients while retaining significant correlations (FDR (false discovery rate) < 0.005). Multiscale clustering analysis was performed to identify hierarchical gene modules. Gene Set Enrichment Analysis (GSEA) and Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis GSEA was conducted using the clusterProfiler package (v4.8.1; Bioconductor) with gene sets from the Molecular Signatures Database (https://www.gsea-msigdb.org/gsea/msigdb). KEGG pathway analysis of DEGs was performed using the DAVID database (https://david.ncifcrf.gov/). Module–phenotype correlations were evaluated using normalized enrichment scores calculated with the fgsea R package to assess shared expression patterns across phenotypes. Cell culture and stable cell line construction The cells were cultured at 37°C with 5% CO2 in Dulbecco's Modified Eagle Medium (DMEM) (C11995500BT, Gibco, USA) containing 4.5 g/L glucose and supplemented with 10% fetal bovine serum (FBS) (F103-01, Vazyme, China) and 1% penicillin/streptomycin (BL505A, Biosharp, China). CTSZ-targeting shRNAs (shCTSZ#1: 5′-CCGGCGAATACCAGGACACCACATACTCGAGTATGTGGTGTCCTGGTATTCGTTTTTG-3′; shCTSZ#2: 5′-CCGGCCAGGACACCACATATATAAACTCGAGTTTATATATGTGGTGTCCTGGTTTTTG-3′) were designed and synthesized by Xi’an Qinke Biotechnology Co., Ltd. (Xi’an, China).Viral particles were packaged in HEK293T cells using the plasmids psPAX2 and pMD2.G (#12260 and #12259, Addgene,USA) with the Lipo8000 transfection reagent (C0533, Beyotime Biotechnology,China). Viral supernatants were collected and used to transduce U251 and U373 cells. The transduced cells were selected with 1 μg/mL puromycin (A1113803, Thermo Fisher Scientific,USA) for 48 h. Quantitative real-time PCR (qRT-PCR) Total RNA was isolated using a Cell RNA Extraction Kit (RC102, Vazyme, China). RNA concentration was quantified with a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA), and cDNA was synthesized using the RevertAid Master Mix (M16325, Thermo Fisher Scientific, USA). qRT-PCR60 was conducted with 2× Universal SYBR Green Fast qPCR Mix (HY-K0501A, MedChemExpress, USA) to measure CTSZ mRNA levels, which were normalized to those of GAPDH mRNA, the internal control. Relative mRNA expression was calculated using the 2−ΔΔCt method. Primer sequences: CTSZ-F: 5′-GTGCAGAACGTCATCGACTG-3′, CTSZ-R: 5′-TTGCAGGTCTCGTCAGGGA-3′; GAPDH-F: 5′-ACCCAGAAGACTGTGGATGG-3′, GAPDH-R: 5′-TTCAGCTCAGGGATGACCTT-3′. Western Blot Samples were lysed in RIPA buffer (G2002, Servicebio Biological,China) containing protease and phosphatase inhibitors (HY-K0010/HY-K0021/HY-K0022, MedChemExpress, USA) for protein extraction. Protein concentration was quantified using a BCA Protein Assay Kit (P0012S, Beyotime Biotechnology, China). Equal amounts of protein lysates (20 μg per lane) were resolved in 12% precast SDS-PAGE gels (ET15420LGel, ACE Biotechnology, China) and then transferred to PVDF membranes (88518, Thermo Fisher Scientific, USA). The membranes were blocked with 5% skim milk at room temperature for 1 h and then incubated overnight with primary antibodies at 4°C. After three 10 min washes with TBST (TBS containing 0.1% Tween 20), the membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies at room temperature for 1 h. Following additional TBST washes, protein signals were visualized using the ChemiScope Capture Imaging System (Clinx Science Instruments, China). The antibodies used were anti-CTSZ (Cat# ab180580, Abcam, UK; 1:1,000), anti-NF-κB p65 (Cat# 8242, Cell Signaling Technology, USA; 1:1,000), anti-p-IKKα (Cat# ab38515, Abcam; 1:1,000), anti-IKKα (Cat# 84372-5-RR, Proteintech, USA; 1:1,000), anti-IκBα (Cat# R23322, Zenbio, China; 1:1,000), anti-caspase-3 (Cat# 5752, PTM BIO, China; 1:1,000), anti-cleaved caspase-3 (Cat# ab32042, Abcam; 1:1,000), anti-BCL-2 (Cat# 4223T, Cell Signaling Technology, USA; 1:1,000), anti-BAX (Cat# 380709, Zenbio; 1:1,000), anti-GAPDH (Cat# ab8245, Abcam; 1:1,000), anti-rabbit IgG (Cat# 7074, Cell Signaling Technology; 1:10,000), and anti-mouse IgG (Cat# 7076, Cell Signaling Technology; 1:10,000). Spheroid formation assay Suspended cells cultured in ultra-low-attachment 96-well plates (CLS3474, Corning, USA) for 48 h were stained using the Calcein AM/EthD-1 Cell Viability Assay Kit (Cat No. PF00008, Proteintech Group, China). Calcein AM (green fluorescence) labeled viable cells, whereas EthD-1 (red fluorescence) identified dead cells. Images were acquired using an inverted fluorescence microscope. The maximum spheroid diameter per well was measured, and spheroids exceeding 100 μm in diameter were quantified. Cell proliferation assay Cells were seeded at a density of 2,000 cells/well in a 96-well plate. The CCK-8 kit (C0038, Beyotime Biotechnology, China) was used according to the manufacturer’s instructions to study cell proliferation. The optical density values were measured at 450 nm using a microplate reader 24, 48, 72, and 96 h after seeding. Colony formation assay Cells were seeded at a density of 2,000 cells/well in a six-well plate. After 2 weeks of culture, the cells were fixed with methanol and stained with 0.1% crystal violet (G1014, Servicebio, China). Colonies were observed under a light microscope, and those with diameter ≥ 2 mm were counted. Cell migration assay A total of 20,000 cells were seeded in the upper chamber of a transwell plate (3422, Corning, USA) containing 200 μL of serum-free culture medium, while the lower chamber was supplemented with 750 μL of complete culture medium containing 10% FBS as a chemoattractant. After incubation for 12 h, the migrated cells on the lower surface of the membrane were fixed with methanol for 20 min and then stained with 0.1% crystal violet at room temperature for 30 min. After washing twice with 1× Phosphate Buffered Saline (PBS), non-migrated cells on the upper surface of the membrane were removed using a cotton swab, and images were captured under an inverted microscope. Flow cytometry analysis of apoptosis Cells cultured for 48 h were collected and processed according to the instructions of the Annexin V-PE/7-AAD Apoptosis Detection Kit (A213, Vazyme, China). The stained cells were then analyzed using a NovoCyte flow cytometer (Agilent, USA). The nuclear translocation of p65 was observed via confocal microscopy Fix, permeabilize and block the processed cells. Subsequently, they underwent overnight incubation at 4°C with primary antibodies targeting NF-κB p65 (Cell Signaling Technology, Cat# 8242, USA), diluted at 1:500. After thorough triple washing with PBST, the cells were exposed to secondary antibodies labeled with FITC, diluted at 1:500. Nuclear counterstaining was then performed in a dark environment. Finally, the cells were examined using a laser confocal microscope. TMA-IHC Microcylindrical tissue cores were obtained via needle biopsy, arranged in paraffin blocks, and sectioned into 4μm-thick slices. Immunohistochemical (IHC) staining was performed following standardized protocols by Servebio Biotechnology Co., Ltd. (Wuhan, China). Anti-CTSZ primary antibodies (GB112171, Servicebio, China) and horseradish peroxidase (HRP)-conjugated secondary antibodies with DAB chromogen (G1212, Servicebio, China) were employed. Following staining, sections were counterstained with hematoxylin (ST2067, Beyotime Biotechnology, China) in-house. Digital whole-slide images were acquired using a high-throughput scanner and analyzed with SlideViewer software (v2.5.0.143918). mIHC Multiplex immunohistochemistry (mIHC) staining was performed following standardized protocols by Servebio Biotechnology Co., Ltd. (Wuhan, China). Double staining for CTSZ and p65 was performed using the mIHC Immunohistochemical Double Staining Kit (G1267, Servicebio, China) with specific primary antibodies against CTSZ(Cat# ab180580, Abcam, UK)and p65 (GB15997, Servicebio, China). Following staining, sections were counterstained with hematoxylin (ST2067, Beyotime Biotechnology, China) in-house. Digital whole-slide images were acquired using a high-throughput scanner and analyzed with SlideViewer software (v2.5.0.143918). Hematoxylin and eosin (H&E) staining Tumor tissues were fixed in 4% PFA, paraffin-embedded, and sectioned into 4 μm thick slices. Brain sections were stained with hematoxylin for 5 min, differentiated in 1% acid alcohol for 30 s, and rinsed with distilled water. Counterstaining was performed using eosin for 3 min. Finally, the sections were mounted with neutral resin and imaged under an optical microscope. Quantification and statistical analysis All statistical analyses were performed using GraphPad Prism 8.0 software. No samples were excluded from the analyses. Homogeneity of variance between compared groups was confirmed. Differences between two groups were assessed using the two-tailed Student’s t-test. For multi-group comparisons, one-way analysis of variance followed by Dunnett’s post hoc test was applied. Kaplan–Meier survival curves were analyzed via the log-rank test. A threshold of p < 0.05 was considered statistically significant. All results were expressed as means ± standard deviations. (Statistical details of experiments can be found in the respective figure legends.). Published: ▪▪▪ Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.isci.2026.115992. Contributor Information Wei Chen, Email: wscw7982@126.com. Maode Wang, Email: maodewang@163.com. Jia Wang, Email: jiawang_xjtu@163.com. Supplemental information Document S1. Figure S1 References Associated Data Supplementary Materials Document S1. Figure S1 Data Availability Statement •Data: RNA-seq data have been deposited in the SRA database under accession code SRP619480. Data: RNA-seq data have been deposited in the SRA database under accession code SRP619480. •Code: This paper does not report original code. Code: This paper does not report original code. •Other items: Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Other items: Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.