Patients and clinical samples

This study, which encompassed the use of human tissues and related clinical data, was subject to review and obtained approval from the Institutional Review Board at the First Affiliated Hospital of Sun Yat-sen University (2022–056). Informed consent was provided by all subjects involved. All procedures adhered to the ethical guidelines established in the Declaration of Helsinki. For the cohort of HNSCC patients undergoing induction chemotherapy, the regimen included docetaxel (75 mg/m2 on day 1), cisplatin (75 mg/m2 on day 1), and 5-fluorouracil (750 mg/m2 by infusion over 120 h from day 1 to 5) [20,21,22]. This standard TPF regimen (docetaxel, cisplatin, and 5-fluorouracil) regimen was administered every three weeks for two cycles. The follow-ups were conducted at regular intervals: during the first two years, patients were monitored by physical examinations every three months. This was followed by biannual examinations during the third to fifth years. After the fifth year, examinations were conducted annually until either the patient’s death or the censoring of their data. The tumor's response to neoadjuvant chemotherapy was assessed using both clinical evaluation and imaging, based on the Response Evaluation Criteria in Solid Tumors (RECIST) guidelines [23, 24]. Clinicopathological parameters, encompassing age, sex, tumor location, smoking history and differentiation status, were extracted from electronic health records. Detailed characteristics of the HNSCC cohort are presented in Supplementary Table 1.

Cell culture

The HNSCC cell lines UM-SCC-1 (SCC-1) and UT-SCC-23 (SCC-23) were procured from the University of Michigan, while the CAL-27 cell line was purchased from ATCC. The cisplatin-resistant cell lines (SCC-1cisR, SCC-23cisR, and CAL-27cisR) were established following previously described methods [25]. All cell lines were cultivated in Dulbecco's modified Eagle's medium (DMEM), supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin. The cultures were maintained at 37 °C in a humidified atmosphere containing 5% CO2.

Plasmids, siRNAs, cell transfection, and lentivirus production

SiRNAs targeting c-Myc, MYH9, β-catenin or integrin β1 were synthesized by RiboBio (Guangzhou, China). Transient siRNA transfections were performed using Lipofectamine RNAiMAX (Thermo Fisher Scientific, Waltham, MA USA), while plasmid DNAs were transfected with Lipofectamine 3000 (Thermo Fisher Scientific), as per the manufacturer's instructions. Short hairpin RNA (shRNA) oligonucleotides targeting ACTN1 were inserted into the LV3-pGLV-h1-GFP-puro vector. Concurrently, the full-length human ACTN1 gene was integrated into the pGCL-GFP lentiviral expression vector. Lentiviral particles were produced by transfecting HEK293T cells with the recombinant lentiviral expression vectors and the necessary packaging plasmids. After incubation for 72 h, the lentivirus-containing supernatants were collected and concentrated. The HNSCC cells were transfected with the purified lentiviruses at a multiplicity of infection of 30. Lentiviral particles carrying MYH9 or β-catenin shRNA were sourced from GeneChem (Shanghai, China). The sequences of oligonucleotides used in this study are listed in Supplementary Table S2.

MTT assay

Cells with the indicated modifications were seeded at a density of 3000 cells per well in 96-well cell culture plates. After an incubation period ranging from one to four days, each well was supplemented with 20 μL of a 5 mg/mL MTT solution at the designated time points. The cells were subsequently incubated for 4 h at 37 °C in a humidified incubator. The supernatant was removed and the formazan product solubilized with 200 μL of dimethyl sulfoxide. Absorbance at 570 nm was measured using a Synergy HT multi-detection reader (Bio-Tek Instruments, Winooski, VT, USA).

Colony formation assay

Cells subjected to specified treatments were seeded in six-well plates. Following a two-week incubation, the cells were fixed with 4% paraformaldehyde and stained with 0.5% crystal violet.

Sphere formation assay

Cells were seeded in ultra-low-attachment plates and cultured in DMEM/F12 (Gibco, Grand Island, NY, USA) supplemented with 1% B27 (Invitrogen, Carlsbad, CA, USA), 1% N2 (Invitrogen), 20 ng/mL EGF, and 10 ng/mL bFGF. The culture medium was replaced every two days until tumor spheres developed.

Matrigel invasion assay

This assay was performed using Matrigel-coated chambers (BD Biosciences, Bedford, MA, USA). Cancer cells with a density of 5.0 × 105 cells per 300 µL of DMEM were placed into the upper chamber, while DMEM supplemented with 10% FBS was added to the lower compartment. After a 24 h incubation, non-invading cells on the upper surface of the Transwell membrane were gently swabbed off. Invading cells were fixed with 4% paraformaldehyde and stained with 0.5% crystal violet. The average invasion area, indicative of cell invasion, was calculated using Image software (Bethesda, Maryland, USA).

Immunohistochemistry (IHC)

Formalin-fixed and paraffin-embedded tissues were deparaffinized in xylene, rehydrated through graded ethanol, and then washed in phosphate-buffered saline. The tissue sections were treated with 0.3% H2O2 in methanol to quench endogenous peroxidase activity. Following a 2 h blocking step with goat serum at room temperature, the slides were incubated overnight with primary antibodies at 4 °C, and then with horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. Staining signals were visualized using a DAB kit, and slides were counterstained with hematoxylin. After washing with distilled water, the slides were dehydrated with xylene and mounted permanently. Each slide was assessed for staining intensity and assigned an H-score using a specific formula. The H-score is the sum of the products of staining intensity categories and their respective percentages of area stained. Intensity categories are scored as follows: 0 for no staining, 1 for weak staining, 2 for moderate staining, and 3 for strong staining. The H-score is calculated by multiplying the percentage of the area with strong staining by 3, the percentage with moderate staining by 2, and the percentage with weak staining by 1. Areas without staining are multiplied by 0 and thus do not contribute to the score. The final H-score is a cumulative total of these calculations, providing a range from 0 to 300 [26, 27]. Two independent pathologists blinded to the clinical data performed the assessments.

Western blotting

Tissues and cells were lysed in RIPA buffer (Beyotime, Shanghai, China) containing a protease inhibitor cocktail. Equal amounts of protein were resolved on 4%-20% SDS–polyacrylamide gels, electrophoresed at a constant voltage (150 V) until the tracking dye reached the bottom of the gels. Proteins were transferred to polyvinylidene fluoride membranes using a Trans-Blot Turbo system (Bio-Rad, Hercules, CA, USA). Membranes were blocked in Protein-Free Rapid Blocking Buffer (EpiZyme, Shanghai, China) for 10 min at room temperature, followed by an overnight incubation with primary antibodies at 4 °C. After five washes in TBST, membranes were incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. Following three additional washes with TBST, antibody-antigen complexes were detected using the Amersham ECL Plus Western Blotting Detection Reagent (GE Healthcare, Chicago, IL, USA). Primary antibodies specific for the following proteins were used in this study: ACTN1 (Proteintech, Chicago, IL, USA), GAPDH (Proteintech), SNAI1 (Cell Signaling Technology, Danvers, MA, USA), SNAI2 (Proteintech), TWIST1 (Proteintech), ZEB1 (Proteintech), Vimentin (Proteintech), E-cadherin (Proteintech), N-cadherin (Proteintech), Cyclin D1 (Proteintech), β-catenin (Proteintech), c-Myc (Proteintech), MMP-7 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), PTCH1 (Proteintech), PTCH2 (Abcam, Cambridge, UK), GLI1 (Proteintech), SHH (Cell Signaling Technology), HEY1 (Proteintech), HES1 (Cell Signaling Technology), p-Smad2 (Cell Signaling Technology), Smad2 (Proteintech), p-Smad3 (Cell Signaling Technology), Smad3 (Proteintech), p-FAK (Thermo Fisher Scientific), FAK (Proteintech), p-PI3K (Cell Signaling Technology), PI3K (Cell Signaling Technology), p-AKT (Proteintech), AKT (Proteintech), Integrin β1 (Proteintech), CD44 (Proteintech), MYH9 (Proteintech), GSK-3β (Proteintech), BIRC5 (Proteintech), and epitope tag antibodies (Proteintech).

Real-time PCR and cytoskeletal remodeling array

Total RNA from cell samples was extracted using the Quick-RNA™ kit (Zymo Research Corp, Irvine, CA, USA), followed by its reverse transcription into cDNA via SuperScript III Reverse Transcriptase (Invitrogen). Real-time PCR was subsequently carried out on a Bio-Rad CFX96 system using Light Cycler 480@ SYBR Green I MasterMix (Roche, Applied Science, Indianapolis, IN, USA). Relative gene expression alterations were calculated using the 2−ΔΔCT method, employing GAPDH as the internal control. The primers used in this study are listed in Supplementary Table 2. For qPCR array analysis, cDNA from the experimental samples was loaded onto a GeneQuery™ Human Cytoskeletal Remodeling qPCR Array (ScienCell Research Laboratories, Carlsbad, CA, USA), and subsequent procedures were performed in accordance with the manufacturer's instructions.

Apoptosis assay

The cells with indicated treatments were collected and subjected to Annexin V/propidium iodide (PI) staining (Invitrogen) according to the manufacturer’s instructions. Briefly, cells were washed with cold PBS and resuspended in binding buffer. After being stained with Annexin V-APC and PI, cells were then analyzed by flow cytometry.

Flow cytometry

Stained cells were subjected to analysis utilizing a DxFLEX flow cytometer (Beckman Coulter, Brea, CA, USA). The APC fluorescence signal was detected in the FL6 channel, whereas the PI signal was captured in the FL2 channel. Analytical gating strategies were employed to exclude cellular debris and doublet events. A forward scatter versus side scatter plot was generated to identify the cellular population. Subsequently, an Annexin V-APC versus PI plot was created to categorize cells into viable (Annexin V-negative, PI-negative), early apoptotic (Annexin V-positive, PI-negative), late apoptotic (Annexin V-positive, PI-positive), and necrotic (Annexin V-negative, PI-positive) statuses. The proportion of cells undergoing apoptosis was quantitatively assessed based on these criteria.

Luciferase reporter assay

To assess the activity of β-catenin signaling, TOPFlash or FOPFlash luciferase reporter vectors were co-transfected with the relevant plasmids into the cancer cells. The TOPFlash and FOPFlash luciferase reporter systems are commonly used for evaluating the canonical Wnt/β-catenin signaling pathway's activity [28]. The TOPFlash reporter plasmid comprises two repeats of three optimal copies of wild-type TCF binding sites, positioned upstream of the thymidine kinase minimal promoter and the luciferase open reading frame. Conversely, the FOPFlash reporter plasmid harbors the same thymidine kinase promoter but with mutated TCF binding sites, along with the same luciferase open reading frame as the TOPFlash reporter plasmid, serving as a negative control for TOPFlash activity. In the context of the c-Myc related reporter assay, the TOPFlash or FOPFlash plasmid was replaced by either the wild-type (pGL3-ACTN1-wt) or the mutant ACTN1 (pGL3-ACTN1-mut) luciferase reporter plasmid. Cell transfection was conducted using Lipofectamine 3000 transfection reagent (Invitrogen) as per the manufacturer's guidelines. Relative luciferase activity was measured 24 h post-transfection utilizing the Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA).

Chromatin immunoprecipitation-quantitative polymerase chain reaction (ChIP-qPCR)

The ChIP assay was conducted using the EZ ChIP Chromatin Immunoprecipitation Kit (Millipore, Billerica, MA, USA), adhering to previously outlined methods [25]. Briefly, DNA–protein cross-links were established by treating cell samples with 1% formaldehyde for 10 min at room temperature. Subsequently, samples were washed thrice with PBS, followed by lysis in RIPA buffer. Sonication was then employed to fragment the genomic DNA. The resulting chromatin fragments were immunoprecipitated with an anti-c-Myc antibody at 4 °C overnight. Associated DNA sequences were identified through qPCR. Primer sequences used for ChIP-qPCR are available upon request.

Clinical relevance of ACTN1 in HNSCC determined by public databases

Datasets GSE127165, GSE143224, GSE26549, GSE6631, GSE37991, GSE58911, GSE25099, GSE34105, GSE55550, GSE23558, GSE30784, GSE65858, GSE41613, GSE30788, GSE117973, GSE27020, GSE23036, GSE40774, GSE136037 and GSE85195 were retrieved from the Gene Expression Omnibus (GEO) repository of the NCBI (https://www.ncbi.nlm.nih.gov/geo/). For the TCGA HNSCC cohort, RNA-seq data along with corresponding clinical details were procured from the National Cancer Institute Genomic Data Commons portal (https://gdc.cancer.gov/). Survival analysis was performed by employing X-tile software (https://medicine.yale.edu/lab/rimm/-research/software/) to define the optimal cutoff for separating HNSCC patients into high and low ACTN1 expression cohorts. Gene Set Enrichment Analysis (GSEA) was performed after segregating patients into ACTN1-high and ACTN1-low groups based on the median value of ACTN1 expression.

Co-immunoprecipitation (Co-IP) assay

Cells subjected to indicated modifications were lysed with RIPA buffer. Post-centrifugation at 14,000 × g for 20 min at 4 °C, the supernatant was incubated overnight with primary antibodies at 4 °C. The resulting immunocomplexes were then coupled to prewashed magnetic beads (EpiZyme) for 6 h at 4 °C to form antigen–antibody-bead complexes. Following three washes with elution buffer, the complexes were denatured at 100 °C for 10 min and subsequently analyzed via western blotting.

Liquid chromatography with tandem mass spectrometry (LC–MS/MS)

Proteins interacting with ACTN1 were separated via SDS-PAGE. Subsequently, candidate bands were digested with sequencing-grade trypsin (Promega) and submitted to MS/MS analysis for protein identification.

Cycloheximide (CHX) chase assay

The indicated cells were treated with 100 μg/mL CHX to block de novo protein synthesis. At designated time points post-treatment, total protein lysates were collected and subjected to western blotting to assess GSK-3β degradation rate.

Animal experiments

All animal procedures complied with the guidelines of the Institutional Animal Care and Use Committee (IACUC) at Southern Medical University. Six-week-old BALB/c nude male mice served as the in vivo model for the subcutaneous tumor study. A volume of 100 μL containing a single-cell suspension at a density of 2 × 107 cells/mL was subcutaneously injected into the dorsal skin of each mouse. The mice were subsequently euthanized four weeks post-injection, and the weight and volume of the resulting tumors were recorded. Following this, tumor tissues were fixed, embedded in paraffin, and prepared for subsequent IHC analysis. To establish patient-derived xenograft (PDX) mouse models, surgical specimens from the primary tumors of HNSCC patients were dissected into small fragments (2–3 mm3) and subcutaneously transplanted into NSG mice within 4 h post-resection. Routine monitoring of body weight, tumor growth, and overall health status was performed. Upon reaching a volume of approximately 1 cm3, the tumors were excised, and the animals were euthanized. These tumors were serially transplanted into fresh NSG mice, initiating the creation of passage 1 (P1) PDX tumors. This procedure was repeated for the generation of passage 2 (P2) PDX tumors. We successfully established three distinct PDX models (PDX-1, PDX-2, and PDX-3), each derived from a unique HNSCC patient, for subsequent analyses. To evaluate the effects of both ACTN1 depletion and cisplatin on PDX growth, primary tumor cells from PDX-1 were transduced with the corresponding lentiviral vectors (shCTRL, shACTN1 #1). An equal number of these transduced cells were subsequently implanted subcutaneously into NSG mice. The mice were then categorized into four groups (n = 6): shCTRL, shACTN1 #1, cisplatin (5 mg/kg, administered intraperitoneally), and shACTN1 #1 plus cisplatin. The treatment period lasted for four weeks. To assess the potential of shACTN1, alone or in combination with cisplatin, to influence the tumorigenicity of HNSCC cells, primary tumor cells were isolated from PDX1 tissues collected from mice in different treatment groups. These cells were then implanted into mice at varying cell counts (1 × 104, 1 × 105). The number of mice that developed tumors was evaluated, and the CSC frequency was determined using the Extreme Limiting Dilution Analysis software (https://bioinf.wehi.edu.au/software/elda/).

Statistical analysis

Statistical analyses were performed with GraphPad Prism 9.0 (GraphPad Software, San Diego, CA, USA). One-way analysis of variance and Student's t-tests were applied to analyze group differences. Data are reported as means ± standard deviations unless otherwise indicated. Kaplan–Meier method was utilized to estimate survival probabilities, and survival distributions were compared using the log-rank test. The degree of association between two variables was quantified using the Pearson correlation coefficient. P values less than 0.05 were considered statistically significant.