USP33 is an integrin α6 deubiquitinase and promotes esophageal squamous cell carcinoma cell migration and metastasis Authors: Qinglei Hang, Shiying Zuo, Yawen Yang, Yuanzhi Wang, Caimin Li, Wenqian Li, Jingya Guo, Sicong Hou, Haifeng Huang Published: 2024 ABSTRACT PurposeThe deubiquitinating enzymes (DUBs) have been linked to cancer initiation and progression. Although ubiquitin-specific protease 33 (USP33) represents a significant factor in regulating various tumor cell behaviors, its specific biological functions and precise mechanisms in esophageal squamous cell carcinoma (ESCC) progression remain unclear.MethodsThe expressions of USP33 mRNA in GEO databases, clinical ESCC samples, and USP33 protein were analyzed using bioinformatics, RT-PCR, and immunohistochemistry, respectively. Using Kaplan–Meier survival curves, the log-rank test was used to determine the cumulative survival rate. Western blotting was used to determine indicated protein expression. The cell biological functions were evaluated by cell growth assay, transwell, cell adhesion, and cell spreading assay, respectively. The interaction between USP33 and integrins was detected by immunoprecipitation, and the deubiquitination was performed by deubiquitination assay. The metastatic ability was checked by tail vein injection.ResultsA significant positive correlation was found between USP33 expression and clinical TNM stage, T classification, and poor prognosis in patients with ESCC. USP33 promoted laminin-dependent adhesion, spreading, and migration of ESCC cells but not their proliferation. Mechanistically, USP33 mediates cell migration through binding, deubiquinating, and stabilizing integrin α6. USP33 knockdown could inhibit ESCC cell migration and metastasis majorly through integrin α6.ConclusionThis study reveals a novel mechanism of USP33 in promoting laminin-dependent ESCC cell migration and metastasis through integrin α6, suggesting that USP33 may be a promising target for treating ESCC. Introduction Esophageal cancer (EC) ranks seventh among oncologic diseases in terms of mortality rate, of which more than half of the new cases and deaths in China (Bray et al. 2024). In the Chinese population, esophageal squamous cell carcinoma (ESCC) represents over 90% of esophageal cancer cases. Despite advances in surgery, immunotherapy, chemoradiotherapy, and targeted therapy, the survival rate for EC is only about 20%, and metastasis is the leading cause of poor prognosis (Li et al. 2021). Therefore, it is vital to uncover the molecular mechanism of metastasis to predict the survival prognosis of patients with ESCC, effectively intervening in tumor progression, thus improving their survival rate. Ubiquitination is one of the most pervasive and crucial post-translational modifications (PTM) in eukaryotic cells, and it is involved in modulating protein homeostasis and many physiological and pathological events (Lee et al. 2023). In the ubiquitination reaction, ubiquitin is covalently attached to the lysine residues of target proteins, and aberrant protein ubiquitination is correlated with cancer initiation and progression (Deng et al. 2020). Notably, the ubiquitination modification is a dynamic and reversible PTM, and the opposite reaction is tightly controlled by deubiquitinases (DUBs) (Lange et al. 2022). Among the largest subfamilies of DUBs, ubiquitin-specific proteases (USPs) are believed to play a critical role in tumor progression, regulating the stability or the functions of specific substrate proteins (Dewson et al. 2023; Harrigan et al. 2018). For instance, USP3 can bind and deubiquitinate COL9A3/COL6A5 to facilitate gastric cancer cell migration and progression (Wu et al. 2021). USP10 promotes colorectal cancer metastasis through the ZEB1 stabilization (Sun et al. 2023). These studies provide a theoretical foundation of USPs for targeted therapy. USP33 is also called the VDU1 (VHL-interacting deubiquitinating enzyme 1) (Li et al. 2002). The human USP33 gene is located on chromosome 1p31.1, which encodes a protein of 942 amino acids. The expression of USP33 has been altered in several types of cancer, including ovarian cancer (OC), pancreatic cancer (PC), glioma, and gastric adenocarcinoma (GAC), and was implicated in malignant behaviors by deubiquitinating and stabilizing the target proteins (Chen et al. 2018, 2024; Liu et al. 2023; Zhang et al. 2022a, b). However, the expression pattern, prognostic significance, biological functions, and regulatory mechanisms of USP33 in ESCC remain largely unexplored. In fact, the effects of USP33 on cancer cell functions are controversial. Here, we first checked USP33 expression via public database and tissue microarrays. The results showed that USP33 was remarkably upregulated in ESCC tissues compared with non-tumor tissues, and increased USP33 was significantly correlated with TNM stage, T classification, and poor survival. Furthermore, USP33 overexpression dramatically promoted the adhesive, migratory, and invasive functions of ESCC cells in a laminin-dependent manner. Moreover, we demonstrated that USP33 could mediate ESCC cell migration and metastasis through integrin α6. Mechanistically, USP33 can interact with integrin α6 and subsequently promote its deubiquitination and stabilization. These findings are the first to describe that USP33 promoted laminin-dependent cell migration by stabilizing integrin α6, which provides an insight into the mechanistic roles of USP33-integrin α6 axis in ESCC metastasis. Materials and methods Antibodies and reagents The antibodies and reagents were used as follows: Rabbit polyclonal antibody (pAb) against USP33 for WB (1:2000, Catalog #A300-925 A) and IHC (1:100, Catalog #A300-924 A) were from Bethyl Laboratories (Montgomery, USA). Rabbit monoclonal antibodies (mAbs) against integrin β4 (1:1000, Catalog #14803S), integrin β1 (1:3000, Catalog #9699S), and integrin α6 (1:2000, Catalog #3750S) for WB were from Cell Signaling Technology (CST, Massachusetts, USA). Mouse mAbs against integrin α6 (1:500, Catalog #sc-374057) and protein-A/G PLUS beads for immunoprecipitation, GAPDH (1:500, Catalog #sc-365062), integrin α3 (1:500, Catalog #sc-374242), and β-actin (1:500, Catalog #sc-47778) for WB were from Santa Cruz Biotechnology (Texas, USA). Mouse mAb against integrin α5 (1:2000, Catalog #610633) for WB was from BD Biosciences (New Jersey, USA). Laminin, fibronectin, collagen, fibrinogen, and RIPA lysate (Catalog#20–188) were purchased from Merck KGaA (Darmstadt, Germany). IgG for immunoprecipitation was from Bio-Rad (California, USA). The protease and phosphatase inhibitors were from TargetMol (Boston, USA). The Matrigel Invasion Chamber was from Corning (Catalog #354480, New York, USA). Human tumor samples and immunohistochemistry (IHC) The 97 ESCC tissues (T) and paired adjacent tissues (N) were obtained from Affiliated Hospital of Yangzhou University (Yangzhou, China) from 2012 to 2016. None of the patients received any radiochemotherapy before the operation. IHC staining of USP33 was performed and then analyzed by using the immunoreactive score (IRS) system as described previously (Hou et al. 2022). Briefly, the percentage of positive cells was scored on a scale of 0 to 4: 0 if 0% of tumor cells were positive, 1 if 1–25% of cells were positive, 2 if 26–50% were positive, 3 if 51–75% were positive, and 4 if 76–100% were positive. Staining intensity was scored on a scale of 0 to 3 (3 is the strongest). Final IRS score = (score of the staining intensity) × (score of the percentage of positive cells). Score = 0–3 was considered low expression and score = 4–12 was considered high expression. Bioinformatics analysis The expression profiles of GSE32424, GSE164158, GSE130078, and GSE111011 were retrieved and downloaded from the Gene Expression Omnibus (GEO) database (You et al. 2019; Chen et al. 2020). Cell lines and cell culture The HEK293T, Eca-109, KYSE-150, and TE-1 cells were from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). KYSE-30, KYSE-520, and TE-10 cells were obtained from the KeyGEN BioTECH (Jiangsu, China). KYSE-70, KYSE-140, and KYSE-450 were from the Feng Hui Biotechnology (Hunan, China). KYSE-180, KYSE-410, and KYSE-510 cells were from the Procell Life Science & Technology Co., Ltd (Hubei, China). The HEK293T cell line was cultured in DMEM medium and the ESCC cell lines were cultured in RPMI 1640 medium, supplemented with 10% fetal bovine serum (FBS). Overexpression and knockdown vectors The cDNAs of the human USP33 and integrin α6 were amplified from KYSE-150 cells and then cloned into the pDONR201 entry vector as described previously (Hou et al. 2019). The pENTR1A-GFP vector was requested from Professor Jianguo Gu’s lab of Tohoku Medical and Pharmaceutical University (Sendai, Japan). By using the Gateway™ cloning kit (Thermo Fisher Scientific, Massachusetts, USA), the cDNAs of GFP, integrin α6, or USP33 from the entry vectors were switched into the pLenti-CMV-Puro (Addgene, Catalog #17452) or pLenti-CMV-Hygro (Addgene, Catalog #17454) destination vectors. Regarding the knockdown lentiviral vectors, the specific oligonucleotides (shRNA#1 against USP33, 5’-GTGGAATTTGTCAGCAGATAT-3’; shRNA#2 against USP33, 5’-GCAACAGTGATAGAGCAGAAA-3’) were inserted into the pLKO.1-puro vector. Lentiviral infection The production and infection of USP33- or integrin α6- overexpression and USP33 knockdown virus were performed as described previously (Hang et al. 2015). The stable cell lines were selected by puromycin (2 µg/mL, 2 days) or hygromycin B (200 µg/mL, 4 days) after virus infection. Western blotting and immunoprecipitation The western blotting (WB) was performed as described previously (Hang et al. 2015). The immunoprecipitation (IP) was performed as described previously (Hou et al. 2023). Cell growth analysis The cell proliferation ability under normal culture or on laminin coating conditions was performed as described previously by our group (Hou et al. 2022). Migration and invasion assays Migration assay under normal culture or on laminin coating conditions and the invasion assay were performed similarly to those previously described by our group (Hou et al. 2022). Cell adhesion and spreading assays The cell adhesion and spreading assays were performed as described previously by our group (Hou et al. 2022). Deubiquitination assay The denature IP for the deubiquitination assay was performed as described previously (Hang et al. 2021). RNA isolation and quantitative PCR Total RNA from tissues was extracted as described previously (Hou et al. 2022), followed by quantitative PCR assay. The primer sequences were listed as follows: forward, 5’-AAAATCCCTTGGTACTTGTCAGG-3’ for USP33 and 5’-GGAGTCAACGGATTTGGT-3’ for GAPDH; reverse, 5’-TCGAAGAGTGGTAAGGTTCACA-3’ for USP33 and 5’-GTGATGGGATTTCCATTGAT-3’ for GAPDH. Animal study The Con, USP33-sh#1, USP33-sh#1 + USP33, and USP33-sh#1 + Integrin α6 KYSE-150 cells (1 × 106) were injected intravenously into the tail vein of 5-week-old BALB/c-nude mice. The lungs were removed at the endpoint, followed by HE staining. The number of lung metastatic nodules was quantified. Statistical analysis The quantifications were performed via GraphPad Prism10 and SPSS 26.0. The χ2 test was used to analyze the association between USP33 expression and clinicopathological features. The log-rank test was used to compare Kaplan-Meier survival curves. Other data were analyzed using a paired t-test, two-tailed test, one-way ANOVA, or two-way ANOVA depending on the indicated experiment situation. The error bars are shown as the mean ± s.d. P < 0.05 was defined as statistically significant (n.s., P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****, P < 0.0001). Results USP33 is upregulated in ESCC tissues and correlated with poor survival First, we compared the mRNA expression of USP33 in ESCC tissues with adjacent normal tissues using GEO databases and found it was significantly overexpressed in ESCC tissues (Fig. 1A-D). Significantly, this phenotype can also be validated in fresh paired ESCC and normal tissues on our side through a real-time PCR test (Fig. 1E). To verify if USP33 protein levels align with transcriptional levels, immunohistochemistry (IHC) was performed in ESCC specimens from 97 patients. IHC results showed that the protein expression level of USP33 was higher in ESCC tissues compared to adjacent normal esophageal epithelium (Fig. 1F, results from two representative ESCC patients). Based on comprehensive database analysis and experimental verification, we speculated that USP33 may play a key role in ESCC occurrence and development. Fig. 1High USP33 level is correlated with poor survival in ESCC. (A-D) USP33 mRNA expression in paired ESCC (T) and normal (N) tissues from GSE32424 (A), GSE111011 (B), GSE164158 (C), and GSE130078 (D) datasets (paired t-test). (E) The fold change of mRNA levels of USP33 in paired T and N tissues was determined by real-time PCR (n = 21, paired t-test). (F) Immunohistochemical (IHC) staining of USP33 in representative ESCC (c, d, g, and h) and adjacent specimens (a, b, e, and f). Scale bar, 100 μm. (G) Kaplan-Meier curves of overall survival of ESCC patients (n = 97), stratified by the expression level of USP33 protein (P = 0.042, log-rank test) High USP33 level is correlated with poor survival in ESCC. (A-D) USP33 mRNA expression in paired ESCC (T) and normal (N) tissues from GSE32424 (A), GSE111011 (B), GSE164158 (C), and GSE130078 (D) datasets (paired t-test). (E) The fold change of mRNA levels of USP33 in paired T and N tissues was determined by real-time PCR (n = 21, paired t-test). (F) Immunohistochemical (IHC) staining of USP33 in representative ESCC (c, d, g, and h) and adjacent specimens (a, b, e, and f). Scale bar, 100 μm. (G) Kaplan-Meier curves of overall survival of ESCC patients (n = 97), stratified by the expression level of USP33 protein (P = 0.042, log-rank test) Furthermore, we examined the relationship between clinical and pathologic characteristics of ESCC and USP33 expressions. The baseline information is shown in Table 1; we found that USP33 expression was significantly correlated with T classification (P = 0.015) and TNM stage (P = 0.016). In addition, based on a multivariate Cox proportional hazards model, we found that TNM stage (P = 0.00817) and USP33 expression (P = 0.0460) were independent risk factors for ESCC prognosis (Table 2). Moreover, Kaplan-Meier survival analysis showed that ESCC patients with high expression levels of USP33 had shorter overall survival than patients with low levels of USP33 (P = 0.042) (Fig. 1G). Altogether, these results suggest that high expression of USP33 may be positively associated with ESCC progression and may serve as a potential indicator for predicting the prognosis of patients with ESCC. Table 1Clinicopathological parameters and expression of USP33 in 97 cases of ESCCCharacteristics n USP33 expressionχ2 P High (%)Low (%)Specimen Adjacent normal tissue9748 (49.5%)49 (50.5%)9.4970.002* ESCC9769 (71.1%)28 (28.9%)Gender Female2016 (80.0%)4 (20.0%)0.9640.326 Male7753 (68.8%)24 (31.2%)Age ≤ 602721 (77.8%)6 (22.2%)0.8040.37 > 607048 (68.6%)22 (31.4%)Smoking history None smoking3226 (81.3%)6 (18.7%)2.380.123 Smoking6543 (66.2%)22 (33.8%)TNM stage I-II5836 (62.1%)22 (37.9%)5.7730.016* III3933 (84.6%)6 (15.4%)T classification T1 + T24728 (59.6%)19 (40.4%)5.9340.015* T3 + T45041 (82.0%)9 (18.0%)N classification No5535 (63.6%)20 (36.4%)3.4780.062 Yes4234 (81.0%)8 (19.0%)Differentiation Well2217 (77.3%)5 (22.7%)0.5220.47 Moderately/poorly7552 (69.3%)23 (30.7%)Tumor size < 43625 (69.4%)11 (30.6%)0.080.778 ≥ 46144 (72.1%)17 (27.9%)Statistical analysis was performed by the Pearson χ2 test*, P < 0.05 was considered significant Table 1Clinicopathological parameters and expression of USP33 in 97 cases of ESCCCharacteristics n USP33 expressionχ2 P High (%)Low (%)Specimen Adjacent normal tissue9748 (49.5%)49 (50.5%)9.4970.002* ESCC9769 (71.1%)28 (28.9%)Gender Female2016 (80.0%)4 (20.0%)0.9640.326 Male7753 (68.8%)24 (31.2%)Age ≤ 602721 (77.8%)6 (22.2%)0.8040.37 > 607048 (68.6%)22 (31.4%)Smoking history None smoking3226 (81.3%)6 (18.7%)2.380.123 Smoking6543 (66.2%)22 (33.8%)TNM stage I-II5836 (62.1%)22 (37.9%)5.7730.016* III3933 (84.6%)6 (15.4%)T classification T1 + T24728 (59.6%)19 (40.4%)5.9340.015* T3 + T45041 (82.0%)9 (18.0%)N classification No5535 (63.6%)20 (36.4%)3.4780.062 Yes4234 (81.0%)8 (19.0%)Differentiation Well2217 (77.3%)5 (22.7%)0.5220.47 Moderately/poorly7552 (69.3%)23 (30.7%)Tumor size < 43625 (69.4%)11 (30.6%)0.080.778 ≥ 46144 (72.1%)17 (27.9%)Statistical analysis was performed by the Pearson χ2 test*, P < 0.05 was considered significant Clinicopathological parameters and expression of USP33 in 97 cases of ESCC Statistical analysis was performed by the Pearson χ2 test *, P < 0.05 was considered significant Table 2Contribution of various potential prognostic factors to survival by Cox regression analysis in 97 ESCC specimensHazard ratio95% Confidence Interval P Age (years)1.6650.893–3.1020.108Gender0.6650.348–1.2710.217Tumor size1.0020.582–1.7250.994TNM stage3.0171.326–6.8660.008*Smoking history1.2400.685–2.2450.477T classification1.5130.848-2.7000.161N classification1.5120.741–3.0830.256Differentiation1.1060.577–2.1210.761USP331.8161.010–3.2630.046*Statistical analysis was performed using the log-rank test*, P < 0.05 was considered significant Table 2Contribution of various potential prognostic factors to survival by Cox regression analysis in 97 ESCC specimensHazard ratio95% Confidence Interval P Age (years)1.6650.893–3.1020.108Gender0.6650.348–1.2710.217Tumor size1.0020.582–1.7250.994TNM stage3.0171.326–6.8660.008*Smoking history1.2400.685–2.2450.477T classification1.5130.848-2.7000.161N classification1.5120.741–3.0830.256Differentiation1.1060.577–2.1210.761USP331.8161.010–3.2630.046*Statistical analysis was performed using the log-rank test*, P < 0.05 was considered significant Contribution of various potential prognostic factors to survival by Cox regression analysis in 97 ESCC specimens Statistical analysis was performed using the log-rank test *, P < 0.05 was considered significant USP33 promotes cell invasion, but not growth and migration, in normal cell culture conditions Next, we wonder about the specific cell biological function(s) affected by USP33. We first detected the protein expression of USP33 in multiple ESCC cells. We found that the USP33 was lowly expressed in KYSE-30 and KYSE-140 cells while highly expressed in KYSE-150 and KYSE-450 cells (Fig. 2A). We then constructed USP33 stable overexpression KYSE-30 cells (Fig. 2B) and USP33-knockdown KYSE-150 cells (Fig. 2E) to demonstrate the effect of USP33 on multiple cell biological behaviors. However, the Con and USP33-overexpressed KYSE-30 cells showed comparable cell proliferation ability (Fig. 2C). Interestingly, as shown by transwell assays, the cell invasion, but not migration, ability was increased in USP33-overexpressed cells compared to Con cells (Fig. 2D). As expected, knockdown of USP33 only attenuated KYSE-150 cell invasion (Fig. 2F). Together, our results indicate that USP33 can promote cell invasion but has no remarkable effect on cell proliferation and migration in normal cell culture conditions. Fig. 2The effects of USP33 on indicated ESCC cell biological functions. (A and B) Cell lysates from multiple ESCC cell lines (A), control (Con), and USP33-overexpressed (USP33) KYSE-30 cells (B) were harvested, then immunoblotted with antibodies against USP33 and GAPDH. (C) Growth curve of Con and USP33-overexpressed KYSE-30 cells (n = 3 biological replicates, n.s., not statistically significant, two-way ANOVA). (D) Transwell migration and invasion assays of Con and USP33-overexpressed KYSE-30 cells (Scale bar, 250 μm, n = 3 biological replicates, two-tailed t-test). (E) Cell lysates from Con, USP33-sh#1, and USP33-sh#2 KYSE-150 cells were harvested and immunoblotted with antibodies against USP33 and GAPDH. (F) Transwell migration and invasion assays of USP33-knockdown KYSE-150 (n = 3 biological replicates, one-way ANOVA) The effects of USP33 on indicated ESCC cell biological functions. (A and B) Cell lysates from multiple ESCC cell lines (A), control (Con), and USP33-overexpressed (USP33) KYSE-30 cells (B) were harvested, then immunoblotted with antibodies against USP33 and GAPDH. (C) Growth curve of Con and USP33-overexpressed KYSE-30 cells (n = 3 biological replicates, n.s., not statistically significant, two-way ANOVA). (D) Transwell migration and invasion assays of Con and USP33-overexpressed KYSE-30 cells (Scale bar, 250 μm, n = 3 biological replicates, two-tailed t-test). (E) Cell lysates from Con, USP33-sh#1, and USP33-sh#2 KYSE-150 cells were harvested and immunoblotted with antibodies against USP33 and GAPDH. (F) Transwell migration and invasion assays of USP33-knockdown KYSE-150 (n = 3 biological replicates, one-way ANOVA) USP33 promotes laminin-dependent cell adhesion and migration Since cell-matrix adhesion is closely correlated with cell migration and invasive potentials in tumors (Hamidi and Ivaska 2018; Chastney et al. 2024), we wondered whether USP33-mediated invasion is specific ECM component(s) dependent. First, we detected the binding abilities of Con and USP33-overexpressed KYSE-30 cells on different ECMs. As shown in Fig. 3A, compared with other ECM components, the binding ability of USP33-overexpressed cells was significantly increased in the laminin (which is also the primary component of Matrigel) group, indicating that USP33-mediated invasion may be related to the inducement of the adhesiveness to laminin. Fig. 3USP33 promotes laminin-mediated adhesion, spreading, and migration of ESCC cells. (A) Con and USP33-overexpressed KYSE-30 cells’ adhesion abilities on different ECM components (n = 3 biological replicates, two-tailed t-test). (B) Growth curve of Con and USP33-overexpressed KYSE-30 cells cultured on laminin (n = 3 biological replicates, two-way ANOVA). (C and D) Transwell migration assay of Con and USP33-overexpressed KYSE-30 (C) or USP33-knockdown KYSE-150 (D) cells cultured on laminin. (n = 3 biological replicates, two-tailed t-test in C, and one-way ANOVA test in D). (E) Cell spreading assay of Con and USP33-overexpressed KYSE-30 cells on laminin (Scale bar, 50 μm; n = 3 biological replicates, two-tailed t-test) USP33 promotes laminin-mediated adhesion, spreading, and migration of ESCC cells. (A) Con and USP33-overexpressed KYSE-30 cells’ adhesion abilities on different ECM components (n = 3 biological replicates, two-tailed t-test). (B) Growth curve of Con and USP33-overexpressed KYSE-30 cells cultured on laminin (n = 3 biological replicates, two-way ANOVA). (C and D) Transwell migration assay of Con and USP33-overexpressed KYSE-30 (C) or USP33-knockdown KYSE-150 (D) cells cultured on laminin. (n = 3 biological replicates, two-tailed t-test in C, and one-way ANOVA test in D). (E) Cell spreading assay of Con and USP33-overexpressed KYSE-30 cells on laminin (Scale bar, 50 μm; n = 3 biological replicates, two-tailed t-test) Next, we questioned laminin affect the effect of USP33 on cell proliferation and migration. Interestingly, overexpression of USP33 significantly increased the cell migration (Fig. 3C) but not the proliferation (Fig. 3B) ability upon laminin coating. As expected, the migration of USP33-knockdown cells was largely inhibited on laminin, compared to Con cells (Fig. 3D). In addition, cell spreading assay result showed that USP33-overexpressed cells exhibit increased spreading ability on laminin compared with Con cells (Fig. 3E). Taken together, these results demonstrate that USP33 can significantly promote laminin-mediated adhesion, spreading, and migration in ESCC cells. USP33 binds and stabilizes integrin α6 through deubiquitination Considering that integrins are major ECM receptors that mediate cell adhesion and movement (Chastney et al. 2024), we wondered whether some integrin(s) were the potential substrate(s) for USP33. We first detected the expression of various integrins, especially the laminin receptor integrins (including the integrin α3, integrin α6, integrin β1, and integrin β4), in Con and USP33-knockdown KYSE-150 cells. As shown in Fig. 4A, compared with Con cells, only the expression of integrin α6, but not other integrins, was decreased in USP33-knockdown KYSE-150 cells. Thus, we speculated that integrin α6 may be the potential substrate protein of USP33. Fig. 4USP33 regulates integrin α6 stability via deubiquitination. (A) Cell lysates from USP33-knockdown KYSE-150 cells were harvested and then immunoblotted with antibodies against integrin β1, integrin β4, integrin α3, integrin α5, integrin α6, USP33, and GAPDH. (B) KYSE-150 cell lysates were collected and immunoprecipitated using an anti-USP33 antibody; the immunoprecipitants were immunoblotted with antibodies against integrin β4, integrin α6, and USP33. (C) KYSE-150 cells were co-transfected with HA-ubiquitin (Ub) and USP33 or GFP, followed by denature immunoprecipitation through anti-integrin α6 antibody and then immunoblotting with antibodies against integrin α6 and HA USP33 regulates integrin α6 stability via deubiquitination. (A) Cell lysates from USP33-knockdown KYSE-150 cells were harvested and then immunoblotted with antibodies against integrin β1, integrin β4, integrin α3, integrin α5, integrin α6, USP33, and GAPDH. (B) KYSE-150 cell lysates were collected and immunoprecipitated using an anti-USP33 antibody; the immunoprecipitants were immunoblotted with antibodies against integrin β4, integrin α6, and USP33. (C) KYSE-150 cells were co-transfected with HA-ubiquitin (Ub) and USP33 or GFP, followed by denature immunoprecipitation through anti-integrin α6 antibody and then immunoblotting with antibodies against integrin α6 and HA To further demonstrate the effect of USP33 on integrin α6, we did an endogenous co-IP assay in KYSE-150 cells. The result showed that USP33 could interact with integrin α6 and β4 (Fig. 4B). It is reasonable to speculate that USP33 maintained the stability of integrin α6 via deubiquitination. To test this, we co-transfected HA-ubiquitin, GFP or USP33 plasmids into KYSE-150 cells, followed by denature-IP. As results shown in Fig. 4C, overexpression of USP33 largely decreased integrin α6 polyubiquitination in KYSE-150 cells. Altogether, these results indicate that USP33 could bind and stabilize integrin α6 through deubiquitination. USP33 regulates laminin-dependent ESCC cell migration and metastasis through integrin α6 As described above, USP33 could bind and deubiquitinate integrin α6 to stabilize its expression; it is reasonable to infer that the USP33-integrin α6 axis plays an essential role in ESCC cell migration on laminin. To validate this presumption, firstly, we overexpressed integrin α6 or rescued USP33 expression in both USP33-knockdown KYSE-150 and KYSE-450 cells, respectively (Fig. 5A and B). We found that integrin α6 and USP33 overexpression in USP33-knockdown cells can largely rescue cell migration abilities on laminin in both KYSE-150 and KYSE-450 cells compared to indicated knockdown cells (Fig. 5C), which highlights the importance of integrin α6 in USP33-mediated ESCC cell migration on laminin. Fig. 5USP33 promotes laminin-mediated cell migration and metastasis through integrin α6. (A and B) Cell lysates were harvested from the Con, USP33-sh#1, USP33-sh#1 + USP33, USP33-sh#1 + Integrin α6 KYSE-150 (A) and KYSE-450 (B) cell groups, and then immunoblotted with antibodies against USP33, integrin α6, and β-actin. (C) Transwell migration assay of indicated cells described in (A and B) on laminin-coated condition. (n = 3 biological replicates, one-way ANOVA test). (D) H&E staining images and quantification of the number of metastatic nodules in the lungs from BALB/c-nude mice following tail vein injection of Con, USP33-sh#1, USP33-sh#1 + USP33, USP33-sh#1 + Integrin α6 KYSE-150 cells (Scale bar, 200 μm; n = 5, one-way ANOVA) USP33 promotes laminin-mediated cell migration and metastasis through integrin α6. (A and B) Cell lysates were harvested from the Con, USP33-sh#1, USP33-sh#1 + USP33, USP33-sh#1 + Integrin α6 KYSE-150 (A) and KYSE-450 (B) cell groups, and then immunoblotted with antibodies against USP33, integrin α6, and β-actin. (C) Transwell migration assay of indicated cells described in (A and B) on laminin-coated condition. (n = 3 biological replicates, one-way ANOVA test). (D) H&E staining images and quantification of the number of metastatic nodules in the lungs from BALB/c-nude mice following tail vein injection of Con, USP33-sh#1, USP33-sh#1 + USP33, USP33-sh#1 + Integrin α6 KYSE-150 cells (Scale bar, 200 μm; n = 5, one-way ANOVA) In addition, we detected the effect of USP33-integrin α6 regulation on cell metastasis in vivo. As the tail vein injection result shown in Fig. 5D, the number of metastatic nodules in the lungs was significantly decreased in the USP33-knockdown group compared to the Con ones, which was largely rescued with restoration of USP33. Significantly, overexpression of integrin α6 in USP33-knockdown cells could also increase the metastatic nodule numbers compared to the USP33-knockdown group (Fig. 5D). Altogether, these findings demonstrate that integrin α6 plays a critical role in USP33-mediated ESCC cell metastasis. Discussion Deubiquitination has been considered a critical pathway for target protein degradation, which plays an essential role in maintaining multiple cell biological functions (Dewson et al. 2023). USPs are the most prominent DUB family and have long been linked to multiple cancer occurrences and progression (Harrigan et al. 2018). Among the USPs, altered expression of USP33 has been first described in B-cell acute lymphoblastic leukemia (B-ALL) (Pitta et al. 2005), then successively confirmed in carcinomas of the breast, lung, colon, stomach, liver, etc. (Chen et al. 2018; Yuasa-Kawada et al. 2009; Wen et al. 2014; Huang et al. 2015; Gan et al. 2020). Given the critical role of USP33 in multiple malignant characteristics of human cancers, we considered that USP33 might be involved in ESCC progression. The effects of USP33 on tumor behavior are controversial. Several groups showed that USP33 functions as a tumor promoter by deubiquitinating multiple oncoproteins. For example, a recent study revealed that enhanced USP33 expression was associated with reduced survival rates in OC patients; mechanistically, USP33 promotes OC cell growth and metastasis via deubiquitinating and stabilizing CBX2 (Chen et al. 2024). Liu et al. reported that USP33 expression was increased among PC samples, which was correlated with poor survival of patients with PC; they found that USP33 could promote PC cell growth, migration, and invasion by triggering the deubiquitination of TGFBR2 (Liu et al. 2023). However, paradoxically, some other evidence shows that USP33 exerts tumor-suppressive effects in several cancer types. As examples, USP33 was down-regulated in lung cancer patients and correlated with poor prognosis, inhibiting lung cancer cell migration through stabilization of the Robo1 protein (Wen et al. 2014). Additionally, USP33 could also inhibit BC cell migration via binding Robo1 receptor (Yuasa-Kawada et al. 2009). These findings implicate the complicated and multifaced roles of USP33 in different cancer types. Here, for the first time, our results revealed that USP33 expression was increased in ESCC samples, and increased USP33 was highly correlated with T classification, TNM stage, and shorter survival time in ESCC patients, indicating the USP33 might act as a marker of poor prognosis and control the malignant behaviors of ESCC. Given the complexity of USP33 in regulating tumor cell behaviors, a further detailed investigation is needed to elucidate its-mediated cell signaling in ESCC. ESCC is the most predominant histological subtype of EC and accounts for nearly 5% of all cancer-related mortality globally (Bray et al. 2024). Several factors are responsible for the poor prognosis of ESCC, including late-stage diagnosis, frequent metastasis, and limited treatment response, among which invasion and metastasis are the main reasons for high mortality (Yang et al. 2020). Many studies indicated that USPs hold the potential to affect almost every step during cancer metastasis (Dewson et al. 2023). For example, the aberrant expression of USP3, USP13, and USP29 in patients with gastric cancer was positively correlated with high invasiveness, advanced TNM stage, and unsatisfactory survival. Further evidence showed that EMT might be the primary mechanism for the enhanced metastatic property in GC cells: USP13 and USP29 promoted TGF-β1-induced EMT through interacting with and deubiquitinating Snail, while USP3 facilitated this process by deubiquitinating SUZ12 (Wu et al. 2021; Zhang et al. 2022a, b; Qian et al. 2020). In fact, the pro-migratory role of USP33 has also been reported in OC, PC, and HCC (Chen et al. 2024; Liu et al. 2023; Gan et al. 2020). Consistent with the observations in these types of cancer, we found that USP33 promoted cell adhesion and migration under laminin-coating conditions, indicating candidate mediator proteins in laminin receptors are involved in USP33-mediate cell migratory ability. As we know, laminins are significant components of ECM, which is central to modulating cell behaviors by binding to integrins (Arimori et al. 2021). It has been shown that laminin-10/11 and laminin-8 located in bone marrow and blood vessels enhanced stem and progenitor cell migration through binding to integrin α6 (Gu et al. 2003). Our group recently also reported that integrin α6 activated focal adhesion kinase (FAK) signaling, thus promoting laminin-mediated migration in ESCC cells. Given the similar findings here, further investigation is needed to elucidate the specific laminin isoform implicated in the stimulation of USP33-mediated ESCC cell metastasis. Considering that accumulating evidence highlights the crucial role of USP33 in multiple cancer development and progression, effects have been undertaken to study the possibility of targeting USPs for cancer therapeutics (Dewson et al. 2023). To date, no specific USP33 inhibitor has been reported. GSK2643943A, the inhibitor of USP20, has been reported to affect USP33 expression because USP33 and USP20 share approximately 59% of their identities (Li et al. 2022). Furthermore, Ward et al. developed a novel probe USP inhibitor that functionally inhibit USP33 activity; however, several other USPs could also be inhibited, including USP4 and USP16 (Ward et al. 2016). Aside from USP inhibitors, USP33 expression can also be controlled by the upstream factors. For instance, USP33 has been shown to be negatively regulated by p97, the p97 is responsible for HERC2-mediated polyubiquitination of USP33 (Chan et al. 2014), highlighting the HERC2-p97-axis-mediated ubiquitin-proteasome system (UPS) may be a potential target for USP33 degradation, which needs further investigation. Evidence also shows that USP33 expression could be post-transcriptionally regulated by several miRNAs, including miR-3591-5p, miR-365a-3p, and miR-206 (Lu et al. 2018; Wang et al. 2018; Ding et al. 2021). Given that miRNAs have the potential for clinical applications as therapeutic targets, targeting the USP33 expression through the specific miRNAs indirectly may also be a promising choice in the near future. However, due to the various substrates that miRNAs could also target, the specificity of the miRNAs-USP33 axis should be considered. Developing safe, specific, and effective inhibitors is the mainstream direction for targeting UPS33. Integrin α6 is involved in cell adhesion, growth, migration, and survival by mediating the interaction between cells and ECM. The aberrant expression or activation of integrin α6 has been reported in BC, HCC, and PC, affecting malignant biological behaviors such as invasion, angiogenesis, metastasis, immune evasion, and drug resistance (Campbell and Humphries 2011). For instance, integrin·α6β4 can induce the migration and invasion of endothelial cells through accumulating the NF-κB and ERK in nuclear (Nikolopoulos et al. 2005). As we know, the abnormal expression of integrin α6 in tumors is transcriptionally and post-transcriptionally regulated. On the one hand, transcription factors like MYC and RUNX1 can foster the transcription of integrin α6, while NFATs and KLF9 act an inverse role at the transcriptional level (Ying et al. 2014; Jauliac et al. 2002; Khademi et al. 2023). On the one hand, the integrin α6 expression level is mainly regulated by mRNA splicing, microRNAs, and N6-methyladenosine modification at the post-transcriptional level (Khademi et al. 2023). Our study firstly proposed that integrin α6 can be modified by USP33 at the post-translational level, which updated regulation approaches of integrin α6 expression in tumors. Previous studies have shown that USP33 regulates protein stability mainly by recognizing and binding to its substrate proteins, then removes specific K27- or/and K48- linked ubiquitin chains. For example, as a result of removing K27- and K48-linked ubiquitin chains from CBX2, USP33 promoted ovarian cancer progression (Chen et al. 2024). Moreover, Guo et al. reported that the K48-linked polyubiquitin chain of DUSP1 can be removed by USP33, and deubiquitinated DUSP1 further inhibits the JNK pathway, which ultimately restrains docetaxel-induced apoptosis in prostate cancer (Guo et al. 2020). Here, our study has demonstrated that the integrin α6 is a substrate of USP33, which can be deubiquitinated and stabilized by interacting with USP33. However, the specific ubiquitin linkage type(s) and site(s) of integrin α6 regulated by USP33 remains unknown and deserves further investigation. In summary, our study indicated that USP33 acts as an oncogene, which was upregulated in ESCC tissue and played a prominent role in ESCC prognosis. We observed that USP33 could stabilize integrin α6 via deubiquitination, promoting laminin-dependent adhesion and migration in ESCC cells. Our finding suggested that USP33 may be a promising marker for metastasis and prognosis of patients with ESCC, providing a theoretical basis for further development in related target therapies for ESCC.