USP14 promotes head and neck squamous cell carcinoma progression via deubiquitinating and stabilizing CFL2

Abstract

Aim

To identify novel substrates of USP14 and elucidate the molecular mechanisms by which USP14 promotes head and neck squamous cell carcinoma (HNSCC) progression.

Materials and methods

USP14 expression patterns were examined in HNSCC tissues and cell lines. Functional effects were assessed using genetic knockout and overexpression models in vitro and in vivo. USP14-CFL2 interactions were evaluated by co-immunoprecipitation and GST pull-down assays. Deubiquitination activity was measured in vitro. Transcriptomic analysis and bioinformatics were used to identify downstream pathways and clinical relevance.

Results

USP14 was significantly overexpressed in HNSCC and correlated with poor prognosis. Genetic knockout of USP14 markedly suppressed HNSCC cell proliferation, migration, and tumor growth, while USP14 overexpression exerted opposite effects. Mechanistically, we identified CFL2 as a novel substrate of USP14; USP14 directly interacted with and deubiquitinated CFL2, thereby enhancing its stability by preventing proteasomal degradation. Clinically, CFL2 was also overexpressed in HNSCC and its elevated levels correlated with reduced overall survival. Functionally, CFL2 overexpression significantly rescued the anti-tumor effects of USP14 knockout, including impaired cell proliferation and migration.

Conclusion

In summary, our findings identify a novel USP14-CFL2 regulatory axis and establish USP14 as a critical promoter of HNSCC progression, acting through CFL2 deubiquitination and stabilization.

ARTICLE HIGHLIGHTS

• USP14 is significantly overexpressed in HNSCC and correlates with advanced N stage and poor prognosis.

• CRISPR/Cas9-mediated USP14 knockout suppresses HNSCC proliferation, migration, and tumor growth in vitro and in vivo.

• USP14 directly binds to and deubiquitinates CFL2, protecting it from proteasomal degradation and enhancing its stability.

• CFL2 overexpression rescues the functional defects caused by USP14 knockout, establishing CFL2 as a key downstream effector.

• This study identifies the USP14–CFL2 axis as a novel therapeutic target and prognostic biomarker in HNSCC.

1. Introduction

Head and neck squamous cell carcinoma (HNSCC) is one of the most common cancer types [1], mainly occurring in the oral cavity, pharynx, and larynx. Unfortunately, more than 60% of new cases are diagnosed at an advanced stage, resulting in a poor prognosis [2]. Thus, it is crucial to study HNSCC pathogenesis to identify early detection biomarkers and new therapeutic targets. Growing evidence suggests that the ubiquitin-proteasome system (UPS) is critically involved in tumor progression [3,4]. The UPS is a canonical non-lysosomal protein degradation pathway. Ubiquitination, as a vital post-translational modification, plays a pivotal role in modulating a diverse array of cellular processes, including protein activation and inactivation, signal transduction, DNA damage response (DDR), apoptosis, and drug resistance [5]. The ubiquitination process is reversible via a mechanism known as deubiquitination, during which deubiquitinating enzymes (DUBs) act as key regulators by catalyzing the removal of covalently attached ubiquitin chains from substrate proteins [6]. Ubiquitin-specific proteases (USPs) constitute the largest DUB subclass. Dysregulation or dysfunction of the UPS has been implicated in a wide range of diseases, including cancer [7].

Numerous studies have shown that USP14 is highly expressed in a variety of tumors and is closely related to the occurrence and development of various cancer types, including breast cancer [8], lung cancer [9], and prostate cancer [10]. Notably, USP14 exerts its influence by modulating both the canonical and noncanonical NF-κB signaling pathways, thereby promoting autophagy and cytokine release [11].

USP14 has a dual function by regulating both the stability and degradation of substrate proteins. USP14 is not only a deubiquitinating enzyme, but also contains a proteasome component that can enhance protein degradation by activating the proteasome [8]. Specifically, in HNSCC cells, USP14 upregulation can promote cell proliferation, survival, and metastasis by deubiquitinating and stabilizing heat shock transcription factor 1 (HSF1) [9]. In addition, USP14 binds to RELA and IκBα, reducing the ubiquitination of IκBα K48, which thereby facilitates its degradation. IκBα is a key inhibitor of the canonical NF-κB signaling pathway in HNSCC [10]. Additionally, USP14 can enhance the cancer stem-like properties of oral squamous cell carcinoma cells by promoting the deubiquitination of SOX2 [11]. Collectively, these findings suggest that USP14 may be a potential therapeutic target in HNSCC. Although numerous studies have investigated the role of USP14 in HNSCC, the majority of these studies have relied on RNA interference or small-molecule inhibitors to knock down USP14 expression [9,12]. While these approaches have provided valuable insights, they carry inherent limitations, including potential off-target effects and an inability to achieve complete functional ablation. Moreover, the downstream substrates that mediate USP14’s oncogenic functions beyond the known regulators HSF1, IκBα, and SOX2 remain largely unexplored. To address these gaps and advance the current understanding, we employed CRISPR/Cas9 gene editing technology to achieve definitive loss-of-function analysis.

In this study, we aimed to identify novel substrates of USP14 and elucidate its regulatory mechanisms. Notably, our work provides the first evidence that USP14 directly targets CFL2, a key actin-remodeling factor, for deubiquitination and stabilization, thereby revealing a previously unrecognized link between USP14 and cytoskeletal dynamics in HNSCC progression.

2. Materials and methods

2.1. Patient tissue specimens

HNSCC tissues and matched adjacent normal epithelial tissues were obtained for western blot analysis from eight HNSCC patients who were treated at the Third Affiliated Hospital of Sun Yat-sen University from June 2020 to June 2022 (Table S1). The immunohistochemistry (IHC) samples were obtained using pathological sections from 79 patients with HNSCC who underwent surgical treatment at the Third Affiliated Hospital of Sun Yat-sen University from March 2015 to March 2018 (Table S2). Inclusion criteria were: patients with a confirmed pathological diagnosis of HNSCC who received primary surgical resection. Exclusion criteria were: patients who had received chemotherapy or neoadjuvant therapy before surgery. All included patients were newly diagnosed with HNSCC. Two experienced clinical pathologists diagnosed the tissue samples independently. The conduct of this study was approved by the Clinical Research Ethics Committee of the Third Affiliated Hospital of Sun Yat-sen University.

2.2. Cell lines and reagents

The human HNSCC cell line Cal27 was provided by Procell Life Science & Technology Co., Ltd. (Wuhan, China), the human Laryngeal Squamous Cell Carcinoma (LSCC) cell line Tu212 was provided by iCell Bioscience, Inc. (Shanghai, China), and the HEK293T cell line was obtained from China Center for Type Culture Collection (CCTCC). The cells were maintained under standard culture conditions (37 °C, 5% CO₂, humidified atmosphere) in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin.

2.3. qPCR and Western blot analyses

Total RNA was extracted from cells and tissues using the HiPure Total RNA Mini Kit (Magen, Guangzhou, China) according to the manufacturer’s instructions. Then, qPCR and western blot analyses were performed as described previously [13]. The primary antibodies used for western blot analysis included the following: anti-USP14 antibody (14517-1-AP, 103 Proteintech, Wuhan, China), anti-CFL2 antibody (D163489, BBI, Canada), anti-GFP antibody (AB290, Abcam, UK), anti-UB antibody (sc-8017, Santa Cruz Biotechnology, Dallas, TX, USA), and anti-cadherin antibody (sc-73614, Santa Cruz Biotechnology). The detailed procedures for protein extraction and western blot quantification, as well as the relevant qPCR cycling conditions, primer sequences (Table S3), and IHC scoring methods, are provided in the Supplementary Materials and Methods.

2.4. Plasmids and lentivirus infection

To establish stable cell lines, the CRISPR/Cas9 technique was employed. Two single-guide RNAs (sgRNAs) targeting human USP14 (sgRNA1: 5′-TTTCACCATAACTTTCTGTCTGG-3′, sgRNA2: 5′-TCATGTAACAAGTGTTACCAAGG-3′) were cloned into the LentiCRISPR V2 vector. The USP14 and CFL2 gene sequences were cloned in their full-length forms into the pCDH-Neo-Venus/DEST vector, with the specific procedures described in a previous publication [14]. HEK293T cells were used to package the lentivirus, then lentiviral particles were collected from the cell culture supernatant. Cells were infected to establish stable cell lines and were subsequently selected using puromycin or G418. Single-cell clones were selected by limiting dilution, with the knockout efficiency verified by western blot analysis and genomic sequencing. The detailed procedures for sgRNA design, oligonucleotide synthesis, and lentivirus production, titration, and infection are provided in the Supplementary Materials and Methods.

2.5. Cell counting kit-8 (CCK-8) assay

Twenty-four hours after transfection, cells were seeded into 96-well plates at a density of 3 × 10³ cells per well. On days 0, 1, 2, 3, 4, and 5, the medium in each well was replaced with 100 μL of fresh complete medium and 10 μL of CCK-8 reagent (Shanghai Yesen Biotechnology Co., Ltd.), then incubated for 1 hour at 37 °C with 5% CO₂. Subsequently, a multifunctional microplate detection system was used to measure the absorbance value of each well. Each experiment was performed with three independent biological replicates, with each biological replicate including three technical replicates.

2.6. Wound healing, Transwell, colony formation, and IHC assays

The wound healing, Transwell, colony formation, and IHC assays were performed as described previously [13,15]. For wound healing, Transwell, and colony formation assays, the experiments were conducted using three independent biological replicates with no technical replicates. Each condition used one well per replicate, with two wounds per well averaged for wound healing assays, one chamber per condition used for Transwell assays, and one well per condition used for colony formation assays.

2.7. Co-immunoprecipitation (Co-IP)

First, untreated CAL27 and Tu212 cells were collected using a CelLytic/Immunoprecipitation buffer supplemented with PMSF and a mixture of protease and phosphatase inhibitors. IP was performed using antibodies against CFL2 (D163489, BBI) or USP14 (sc-398009, Santa Cruz Biotechnology). In brief, the lysates were first pre-cleared by incubating with Protein A/G magnetic beads at 4 °C for 1 hour. The Sepharose agarose beads were then removed by centrifugation. Subsequently, the pre-cleared lysates were incubated with the primary antibody overnight at 4 °C with gentle shaking. Rabbit IgG was used as a negative control. Next, the magnetic beads were washed with CelLytic/Immunoprecipitation buffer, resuspended in SDS-PAGE loading buffer, and loaded onto a 10% gradient gel after heating. Mass spectrometry analysis was then performed by Guangzhou FITGENE Company.

To explore the interaction between CFL2 and USP14, 293 T cells were co-transfected with tagged pCDH-venus-USP14/pCDH-venus and pCMV-GST-CFL2. Pull-down experiments were performed using anti-Flag magnetic beads and detected using western blot analysis. We further conducted a reverse co-IP experiment, in which Flag-tagged pCDH-venus-USP14 was co-transfected with pCMV-GST-CFL2/pCMV-GST in 293 T cells. Pull-down experiments were performed using GST-coupled magnetic beads and detected using western blot analysis.

2.8. Protein half-life assay

TU212 cells stably overexpressing USP14 were treated with cycloheximide (CHX; final concentration of 20 μg/mL) for various amounts of time (0, 4, and 8 hours). At each time point, the medium was removed, then the cells were lysed to extract proteins. The protein samples were then analyzed by western blot analysis to determine the half-life of CFL2.

2.9. Ubiquitination analysis

In parallel, we examined the ubiquitination status of CFL2 in Cal27 and TU212 cells with stable USP14 knockdown. These cells were treated with MG132, a well-known proteasome inhibitor, for 6 hours before lysis to extract proteins. Additionally, 293 T cells were co-transfected with pCDH-Venus, pCDH-Venus-USP14, pCMV-GST-CFL2, and HA-UB plasmids. After 48 hours, the culture medium was discarded, then the cells were treated with MG132 (final concentration of 20 μM) for 6–8 hours. The cells were gently washed twice with PBS at room temperature and lysed. GST magnetic beads were used to pull down CFL2 protein, then western blot analysis was performed to detect the expression levels of ubiquitin (UB).

2.10. Xenograft tumorigenesis

Nude mice (6–8 weeks old) were obtained from Guangdong Yaokang Biotechnology Co., Ltd. (Guangzhou, China). A total of 2 × 10⁶ Cal27 cells were collected in 100 µL of serum-free DMEM and injected subcutaneously into the left axilla of each mouse. Tumor volumes were measured after cultivation for seven days. The tumor volume was calculated as follows: Volume = (length × width²)/2. At day 28 post-inoculation, the mice were sacrificed, then the tumors were weighed and processed for histological analysis, which was performed as described previously [13]. This animal experiment was conducted in accordance with the ARRIVE reporting guidelines [16]. For comprehensive methodological details, please refer to the Supplementary Materials and Methods.

2.11. Statistical analysis

Data analysis was performed using the statistical software SPSS (version 22) and GraphPad Prism (version 8). The student’s t-test and one-way ANOVA were used to compare two groups and multiple groups, respectively. Fisher’s exact test or the chi-square (χ²) test was applied for categorical data. Significant differences between groups: *P < 0.05; highly significant differences between groups: **P < 0.01; no significant difference: ns.

3. Results

3.1. USP14 is overexpressed in HNSCC tissues

Western blot analysis of eight fresh frozen HNSCC specimens and matched adjacent normal mucosa revealed that the USP14 protein expression levels were consistently higher in the tumor tissues than in the corresponding normal samples (Figure 1(A,B)). To confirm this observation, IHC was performed on a tissue microarray containing 79 HNSCC specimens together with their paired normal mucosal tissues. Representative images are shown in Figure 1(C). Positive USP14 staining was predominantly cytoplasmic and displayed moderate-to-strong intensity in tumor cells, whereas the adjacent normal epithelium exhibited weak or negative staining. Quantitative analysis demonstrated that the mean USP14 IHC score was significantly higher in tumors than in matched normal tissues (P < 0.001, paired t-test). High USP14 expression patterns (moderate-to-strong staining) were detected in 58 of the 79 (73.42%) tumor samples, while only 32.91% of the paired normal tissues (26 cases) exhibited high USP14 expression (Figure 1(D)). The observed correlations between the USP14 protein expression levels and clinicopathological variables are summarized in Table 1. High USP14 expression levels were positively associated with advanced N stage (P = 0.008, χ² test), but were not related to patient age, sex, T stage, or histological differentiation grade.

Figure 1.

USP14 expression in HNSCC patient clinical samples. (A–B) Western blot analysis of eight paired samples of non-tumorous region (N) and tumor region (T) from the same patient. (C–D) USP14 immunohistochemical staining was performed in 79 paired clinical HNSCC tumor patients. (E–F) Schematic of the designed USP14 knockout primers and western blot analysis of USP14 protein expression in TU212 and Cal27 cells infected with lentivirus carrying either the vector control or USP14-KO constructs (sg1 or sg2).

Table 1.

Associations between USP14 expression and clinicopathological characteristics HNSCC patients.

Characteristics	USP14 expression		P value
	High(58)	Low(21)
Age			0.96
>60	28	10
≤60	30	11
Gender			0.654
Male	56	19
Female	2	2
T classification			0.472
T1	13	5
T2	17	4
T3	21	11
T4	7	1
N classification			0.041*
N0	28	10
N1	10	8
N2	19	3
N3	1	0
Tumor differentiation			0.591
Well	14	4
moderate	20	10
Poor	24	7

3.2. USP14 promotes HNSCC cell proliferation and migration in vitro and in vivo

To investigate the functional role of USP14 in HNSCC, we generated stable USP14 knockout cell lines using CRISPR/Cas9-mediated gene editing in TU212 and Cal27 cells. Two specific sgRNAs (sg1 and sg2) targeting USP14 were designed (Figure 1(E) and Figure S1). Western blot analysis confirmed the efficient depletion of USP14 protein expression in both cell lines transduced with lentivirus carrying the USP14-KO-sg1 or USP14-KO-sg2 constructs compared with the expression in cell lines transduced with the vector control (Figure 1(F)). The impact of USP14 knockout on cell proliferation was evaluated using CCK-8 assays. Depletion of USP14 significantly reduced the proliferation rate of both TU212 (Figure 2(A)) and Cal27 (Figure 2(B)) cells over time compared with no depletion. This anti-proliferative effect was consistently observed using both sgRNAs. Consistent with the CCK-8 results, colony formation assays demonstrated that USP14 knockout dramatically reduced the clonogenic survival of both TU212 and Cal27 cells. Cells lacking USP14 formed significantly fewer and smaller colonies than the vector control cells (Figure 2(C,D)). We next assessed the effect of USP14 knockout on cell migration using Transwell assays. The migratory ability of both TU212 and Cal27 cells was significantly impaired upon USP14 depletion (Figure 2(E,F)). This finding was corroborated by wound healing assays, which showed markedly delayed wound closure in both the USP14-KO-sg1 and USP14-KO-sg2 groups compared with that in the vector controls in TU212 (Figure 2(G,H)) and Cal27 (Figure 2(I,J)) cells. We further established a subcutaneous xenograft model to assess the in vivo tumor suppressive role of USP14 knockout. Mice implanted with USP14-sg1-Cal27 cells exhibited markedly reduced tumor formation. Representative images revealed smaller tumors in the USP14-sg1-Cal27 groups compared with the findings in the vector control group (Figure 2(K)). Tumor growth curves confirmed a sustained reduction of tumor volume throughout the experimental period (Figure 2(L)). Consistent with this, excised tumors from the USP14-sg1-Cal27 groups showed significantly decreased weight (Figure 2(M)). Collectively, these data demonstrate that the genetic ablation of USP14 using CRISPR/Cas9-mediated gene editing could effectively inhibit HNSCC cell proliferation, clonogenicity, and migration.

Figure 2.

Knocking out USP14 suppresses HNSCC cell proliferation and migration in vitro and tumor growth in vivo. Cell proliferation was assessed by CCK-8 assays in TU212 cells (A) and Cal27 cells (B) infected with lentivirus vector, USP14-KO-sg1, or USP14-KO-sg2. (C–D) Colony formation assays were performed in TU212 and Cal27 cells infected with lentivirus vector, USP14-KO-sg1, or USP14-KO-sg2. (E–F) Transwell migration assays were used to evaluate the cell migration ability following USP14 knockout in TU212 and Cal27 cells. (G–J) Wound healing assays were used to assess the migration ability of HNSCC cells after USP14 knockout. Panels (G) and (H) show the results in TU212 cells, while panels (I) and (J) show the results in Cal27 cells. (K) Subcutaneous xenograft tumors (representative images and morphology). (L) Tumor growth curves. (M) Tumor weight comparison.

3.3. USP14 binds directly to CFL2

Co-IP assays were performed to identify the proteins that interact with USP14. The proteins that were pulled down were sent for silver staining (Figure 3(A,B)). We analyzed the specific proteins identified through the silver-stained specific band molecular weights, their corresponding mRNA expression levels in the TCGA database, and their correlations with HNSCC survival rates, as detailed in Table S4. The TCGA database analysis suggested that CFL2 overexpression in HNSCC tissues is significantly associated with the patient overall survival rate (Figure S2(A,B)). Reciprocal co-IP assays in HNSCC cells (TU212) confirmed an endogenous interaction between USP14 and CFL2. IP using an anti-USP14 antibody pulled down CFL2 (Figure 3(C)), while IP with an anti-CFL2 antibody pulled down USP14 (Figure 3(D)). To validate this interaction exogenously, we performed reciprocal pull-down assays in HEK293T cells. GST pull-down of CFL2 co-precipitated Venus-tagged USP14 (Figure 3(E)), while Flag pull-down of USP14 co-precipitated GST-tagged CFL2 (Figure 3(F)). These results consistently indicate a specific interaction between USP14 and CFL2, both endogenously in HNSCC cells and exogenously in a heterologous system.

Figure 3.

USP14 binds directly to CFL2 and stabilizes CFL2 protein through deubiquitination. (A) Silver staining of USP14 RNA pull-down assay products revealed differential bands in the probe group compared with the findings in the control group. (B) Venn diagram depicting the proteins identified by mass spectrometry analysis of the USP14 pull-down products. USP14 specifically bound to 98 proteins compared with the control group. (C) Endogenous co-immunoprecipitation (co-IP) using an anti-USP14 antibody demonstrated CFL2 co-precipitation. (D) Reciprocal co-IP using an anti-CFL2 antibody confirmed USP14 co-precipitation, indicating endogenous binding between USP14 and CFL2. (E) Flag pull-down assays in 293T cells co-transfected with pCDH-venus-USP14 and pCMV-GST-CFL2 detected GST-tagged CFL2. (F) GST pull-down assays in 293T cells co-transfected with pCDH-venus-USP14 and pCMV-GST-CFL2 detected Venus-tagged USP14. (G) CFL2 protein expression levels were increased in Cal27 and TU212 cells overexpressing USP14 (OE). (H–I) Cycloheximide (CHX) chase assay in TU212 and Cal27 cells. Western blot analysis revealed accelerated CFL2 degradation upon USP14 knockout at 0, 4, and 8 hours post-CHX treatment. (J) CFL2 protein expression levels were decreased in USP14-knockout (KO) Cal27 and TU212 cells. (K–L) MG132 (6-hour treatment) rescued CFL2 protein expression levels in USP14-KO Cal27 and TU212 cells. (M) Co-IP assays demonstrated reduced CFL2 polyubiquitination in cells overexpressing USP14 compared with that in the controls.

3.4. USP14 enhances CFL2 stability via deubiquitination

We next investigated the regulatory relationship between USP14 and CFL2 in HNSCC cells. This regulation was determined to occur post-transcriptionally because USP14 knockout did not alter CFL2 mRNA expression levels (Figure S3(A,B)). USP14 overexpression conversely increased CFL2 protein abundance, while genetic knockout of USP14 significantly reduced the endogenous CFL2 protein expression levels in Cal27 and TU212 cells (Figure 3(J,G)). CHX chase assays revealed that USP14 overexpression significantly slowed the CFL2 protein degradation rate in both TU212 and Cal27 cells (Figure 3(H,I)). Furthermore, treatment with the proteasome inhibitor MG132 rescued the decreased CFL2 protein expression levels resulting from USP14 knockout (Figure 3(K,L)), indicating that USP14 can stabilize CFL2 by inhibiting its proteasome-mediated degradation. Consistent with the deubiquitinase function of USP14, co-IP assays demonstrated that USP14 overexpression markedly reduced the polyubiquitination levels of co-expressed GST-CFL2 in 293 T cells (Figure 3(M)). Collectively, these results suggest that USP14 can stabilize the CFL2 protein by deubiquitinating it, thereby protecting it from proteasomal degradation in HNSCC cells.

3.5. CFL2 promotes HNSCC progression and rescues USP14 knockout‑induced defects

We conducted a series of cellular functional assays, including CCK-8, wound healing, colony formation, and Transwell assays, in HNSCC cells. The results demonstrated that knocking out CFL2 significantly reduced the HNSCC cell proliferation and migration rates (Figure 4(A–I)). Overall, this indicates that CFL2 contributes to HNSCC progression.

Figure 4.

CFL2 overexpression restores the proliferation and migration defects caused by USP14 knockout in HNSCC cells; correlation analysis of USP14 and CFL2 expression patterns. (A) Western blot analysis of CFL2 protein expression levels in TU212-sg1 and Cal27-sg1 cells with CFL2 overexpression. (B–C) CCK-8 assays were used to evaluate the proliferative capacity of Venus- and Venus-CFL2-overexpressing TU212-USP14-sg1 and Cal27-USP14-sg1 cells. (D–E) Colony formation assays were used to assess the clonogenic potential of Venus- and Venus-CFL2-overexpressing TU212-USP14-sg1 and Cal27-USP14-sg1 cells. (F–G) Transwell migration assays were used to examine the effect of CFL2 overexpression on the migratory ability of TU212 and Cal27 cells with stable USP14 knockout. (H–I) Wound healing assays demonstrated that overexpression of CFL2 significantly enhanced the migratory ability of both TU212 and Cal27 cells following USP14 knockout. (J) Immunohistochemical analysis of CFL2 and USP14 protein expression patterns. (K) Correlation analysis between USP14 and CFL2 expression levels.

To investigate the functional interplay between USP14 and CFL2 in HNSCC progression, we first established USP14-knockout models in TU212 and Cal27 cells (TU212-USP14-sg1, Cal27-USP14-sg1). Western blot analysis confirmed successful depletion of USP14 protein, with concomitant downregulation of CFL2 protein expression levels in both cell lines (Figure 3(J)). Functional rescue experiments revealed that CFL2 overexpression significantly restored the proliferative capacity of USP14-deficient cells, as evidenced by both CCK-8 assays (Figure 4(B,C)) and colony formation assays (Figure 4(D,E)). Notably, Transwell migration assays demonstrated that CFL2 reconstitution completely reversed the migration defects caused by USP14 knockout (Figure 4(F,G)). Further validation through wound healing assays corroborated these results. USP14 knockout markedly suppressed cell migration, while CFL2 overexpression robustly enhanced the migratory ability in both cell models (Figure 4(H,I)). To assess the clinical relevance of these findings, IHC analysis of patient tissues showed co-expression patterns of USP14 and CFL2 (Figure 4(J)). Subsequent correlation analysis revealed a significant positive association between USP14 and CFL2 expression levels (Pearson’s r = 0.72, P < 0.001; Figure 4(K)), supporting their functional synergy in HNSCC pathogenesis.

4. Discussion

USP14 is a well-established carcinogenic gene in HNSCC. Inhibition of USP14 has been shown to overcome radioresistance in oral squamous cell carcinoma by activating autophagy-dependent apoptosis, which holds promise as a novel strategy for radiosensitization [17]. Moreover, USP14 overexpression in HNSCC stabilizes IκBα to activate NF-κB signaling, driving TNF-α expression and resistance to radiotherapy [10]. In HNSCC cells, depleting USP14 could enhance the degradation of HSF1 [9], but enhance the stability of SOX2 [11]. HSF1 expression in cancer-associated fibroblasts (CAFs) drives HNSCC progression by inducing the epithelial-mesenchymal transition (EMT) and promoting tumor cell proliferation, migration, and invasion, serving as a potent prognostic biomarker [18]. Additionally, HSF1 promotes HNSCC tumor growth by driving FOXO3a-dependent transcription of ΔNp63α, leading to CDK4 upregulation, with ablation or pharmacological inhibition of HSF1 suppressing this pathway and tumorigenesis. Therefore, HSF1 drives HNSCC progression through dual mechanisms: in CAFs it can induce EMT and promote tumor cell proliferation/migration/invasion, while in tumor cells it can support growth via the FOXO3a-ΔNp63α-CDK4 axis. These previous data establish HSF1 as both a key prognostic biomarker and therapeutic target in HNSCC [18,19]. SOX2 is expressed in cancer stem cells (CSCs) and has been used widely as a marker to identify these cells in HNSCC tissues [20,21]. SOX2 binds LIF-SE, activating LIF transcription to promote CSC maintenance and tumor development in HNSCC [22]. Overall, these prior findings demonstrate that USP14 plays a critical role in HNSCC pathogenesis by modulating key signaling pathways.

In this study, we demonstrated that USP14 interacts with CFL2 and promotes its deubiquitination, which stabilizes CFL2 protein expression by suppressing its polyubiquitination in HNSCC cells (Figure 5). CFL2, a member of the ADF/cofilins family, is an essential regulator of actin remodeling, severing F-actin to modulate cytoskeletal mechanical stress [23]. Actin cytoskeleton dynamics modulate YAP activity in the Hippo signaling pathway, which can thereby regulate cancer cell proliferation and differentiation [24,25]. CFL2 expression levels are reportedly upregulated in prostate and gastric cancers, with this abnormal pattern associated with the disease pathogenesis [26,27]. Elevated CFL2 expression serves as a predictive biomarker for radiotherapy resistance and represents a potential therapeutic target for enhancing radiosensitivity in nasopharyngeal carcinoma [28]. CFL2 elevation, mediated by circ_0008673 via miR-153-3p sponging, also drives breast cancer progression [29]. Similarly, CFL2 activation via the SOX2-OT/miR-369-3p axis accelerates prostate cancer malignancy [30]. Our current results identify CFL2 as an unfavorable prognostic factor in HNSCC. Furthermore, CFL2 overexpression could rescue USP14 knockout-induced impairments in cell proliferation and migration, demonstrating its critical role in promoting HNSCC progression. Previously identified USP14 substrates, such as HSF1 [9], IκBα [10], and SOX2 [11], are primarily involved in stress signaling, inflammation, and stemness. However, CFL2 links USP14 to the regulation of actin cytoskeleton dynamics and cell migration, a previously unrecognized function of USP14 in HNSCC progression. We also acknowledge that while CFL2 is a critical downstream effector, other USP14 substrates (HSF1, IκBα, SOX2, and others) may cooperatively contribute to the cell proliferation and migration phenotypes observed upon USP14 knockout.

Figure 5.

Schematic diagram illustrating the regulatory mechanism of the USP14-CFL2 axis in HNSCC malignant progression.

Our findings also have implications for biomarker development in HNSCC. The consistent upregulation of USP14 and CFL2 in tumor tissues and their correlation with advanced N stage and reduced overall survival position them as promising candidate prognostic biomarkers for this malignancy.

This study has several limitations. First, our IHC cohort included only 79 HNSCC patients from a single institution, with the survival analysis derived from the TCGA database rather than from our own cohort. Therefore, larger multi-center studies are needed to validate the clinical significance of our results. Second, although USP14 stabilizes CFL2, its potential link to Hippo signaling remains unexplored. We did not directly analyze actin dynamics, such as the F-actin/G-actin ratio, or downstream pathways beyond cell proliferation and migration. Third, future in vitro assays with purified proteins are required to determine if USP14 directly deubiquitinates CFL2 or acts in an indirect manner, such as via other DUBs or E3 ligases. Fourth, although we used two independent sgRNAs with consistent phenotypes, we did not perform rescue experiments re-expressing USP14 in knockout cells or conduct a genome-wide off-target analysis. Nevertheless, the observed consistency between the sgRNAs and rescue experiments using CFL2 overexpression partially support specificity. Fifth, our xenograft study only measured tumor volume and weight, lacking IHC validation of USP14/CFL2 knockdown, Ki-67, or TUNEL staining, which limits the in vivo mechanistic depth. Despite these limitations, this study provides the first evidence that USP14 can promote HNSCC, at least in part, via CFL2 stabilization, establishing a foundation for future investigations.

5. Conclusion

In conclusion, our study demonstrates that USP14 is overexpressed in HNSCC and promotes tumor progression by deubiquitinating and stabilizing CFL2. These findings identify the USP14-CFL2 axis as a novel oncogenic mechanism and highlight USP14 and CFL2 as potential prognostic biomarkers and therapeutic targets in HNSCC.

Funding Statement

This work was supported by funds from the Science and Technology Program of Guangzhou (201903010024 and 202206010081), National Natural Science Foundation Cultivation Special Project Funding of The Third Affiliated Hospital of Sun Yat-sen University (2026GZRPYMS08), National Science Foundation of China (81902771, 81772540 and 82272996), and the Fundamental Research Funds for the Central Universities (21625105).

Ethics approval and consent to participate

This study was approved by the Research Ethics Committee of the Third Affiliated Hospital of Sun Yat-sen University (20220216501). Animal procedures were authorized under the approval of the Institutional Animal Care and Use Committee of Jinan University (IACUC-20220901-12).

Consent for publication

No details of the identified patients are contained within this manuscript.

Disclosure statement

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Availability of data and materials

The datasets supporting the conclusions of this article are included within the article (and its additional files). The TCGA gene expression and survival data analyzed in this study were accessed via the UALCAN platform (https://ualcan.path.uab.edu/analysis.html) [31].