HSP90B1 regulates autophagy via PI3K/AKT/mTOR signaling, mediating HNSC biological behaviors

Data source

Differential gene expression analysis in head and neck squamous cell carcinoma was conducted using the GEPIA2.0 database (Gene Expression Profiling Interactive Analysis 2.0, http://gepia2.cancer-pku.cn/#index), which is built upon the TCGA database (The Cancer Genome Atlas, https://portal.gdc.cancer.gov/). The selected tissue samples comprised 519 tumor and 44 paraneoplastic specimens. The inclusion criteria for the samples necessitated comprehensive patient history documentation, clinicopathological evaluation, availability of sufficient tissue for analysis, and consent for genomic study participation. The marked disparity between the quantity of tumor and normal samples could be attributed to challenges in sample collection, the focus of research, and the accessibility of data, among other factors. For prognosis analysis based on molecular expression, we employed the Kaplan–Meier Plotter (https://kmplot.com/analysis/). Additionally, clinical data and gene expressions of HNSC patients were retrieved from TCGA to discern gene-clinical correlations. The potential biological functions of HSP90B1 were investigated by downloading coexpression gene data from GEPIA2.0 and cross-referencing it with positively or negatively correlated gene sets from the UALCAN database (http://ualcan.path.uab.edu/index.html). KEGG pathway (https://www.genome.jp/kegg/kegg1a.html) and Wikipathway (https://www.wikipathways.org/) was performed for function enrichment analysis using Sangerbox (http://Sangerbox.com). These databases allowed researchers to download and analyze public datasets for scientific purposes.

Tumor samples

Tissue specimens were obtained from 45 HNSC patients treated in the Department of Otorhinolaryngology Head and Neck Surgery at the First Hospital of Guangxi Medical University between June 2022 and August 2023. Among these patients, 10 were diagnosed with oral squamous cell carcinoma (OSCC), 28 with laryngeal squamous cell cancer (LSCC), and seven with hypopharyngeal squamous cell cancer (HSCC). A total of 45 pairs of HNSC tissues and adjacent non-tumor tissues were collected with informed consent and the approval of the Institutional Ethics Board of the First Affiliated Hospital of Guangxi Medical University (NO.2022-KY-E-(195)). Inclusion criteria were stringent: patients must have had complete medical records, not have received chemotherapy or radiotherapy prior to surgery, and received a perioperative diagnosis of squamous cell carcinoma. Ethical compliance was ensured as informed consent was obtained from all participants, who signed the informed consent form approved by the Ethics Committee of the First Affiliated Hospital of Guangxi Medical University.

Cell cultivation

Normal human oral keratinocytes (HOK), laryngeal squamous cell cancer cells (TU686), hypopharyngeal squamous cell cancer cells (FaDu), and tongue squamous cell carcinoma cells (SAS) were acquired from Shanghai WHELAB Bioscience Co., Ltd (Shanghai, China). TU686 cells were maintained in RPMI 1640 medium (Gibco, Invitrogen, Carlsbad, CA, USA), FaDu cells in MEM medium (Gibco, Invitrogen, Carlsbad, CA, USA), and SAS cells in DMEM medium (Gibco, Invitrogen, Carlsbad, CA, USA). All media were supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. The cell cultures were incubated at 37 °C in a 5% CO2 atmosphere.

RNA extraction and quantitative real-time fluorescence quantitative PCR

Total RNA was isolated from cells and specimen tissues using Trizol reagent (Beyotime Scientific, China). RNA concentration was measured with a NanoDrop2000 (Thermo Fish Scientific, USA). RNA is stored at −80 °C. The expression level of HSP90B1 was determined by quantitative real-time PCR (qRT-PCR) (QuantStudio 6 Flex, Thermo Fisher Scientific, Waltham, MA, USA) using a reverse transcription kit (TransGen Biotech, Peking, China). Primers used were as follows: HSP90B1, 5′-CGAAGTTGGACAGTGGTAAAGAG-3′(forward), 5′-TGCCCAATCATGGAGATGTCT-3′(reverse); GAPDH, 5′-AATCCCATCACCATCTTCCAG-3′(forward), 5′-GAGCCCCAGCCTTCTCCAT-3′(reverse).

Protein western blot analysis

Cells were harvested for protein extraction, and the protein concentration was quantified using the bicinchoninic acid (BCA) method (Biosharp Biotech, Shanghai, China). Proteins were separated by electrophoresis, transferred onto polyvinylidene difluoride (PVDF) membranes (GE Healthcare, Chicago, IL, USA), and then the membranes were blocked with 5% skimmed milk for 1.5 h at room temperature. Primary antibodies were applied and incubated overnight at 4 °C. Subsequently, membranes were incubated with the appropriate secondary antibodies for 1.5 h at room temperature, followed by three washes with TBST buffer. Protein detection was performed using an Odyssey infrared imaging system, and signal intensity was quantified using ImageJ software. The primary antibodies used included: HSP90B1 (1:1000, rabbit, Biolab Biotech, China), CASP3 (1:1000, rabbit, KleanAB Scientific, China), BCL2 (1:1000, rabbit, Zenbio Biotech, China), Bax (1:1000, rabbit, Zenbio Biotech, China), P62 (1:1000, rabbit, Proteintech Group, China), LC3B (1:1000, rabbit, Zenbio Biotech, China), and GAPDH (1:2000, mouse, Zenbio Biotech, China). Secondary antibodies included anti-rabbit (1:15000, Zenbio Biotech, China) and anti-mouse (1:15000, Zenbio Biotech, China).

Immunohistochemistry

A total of 45 pairs of tissue samples were prepared for immunohistochemistry (IHC). The tissues were sectioned into 3 µm slices, dewaxed, rehydrated, and treated with 3% hydrogen peroxide for one hour to block endogenous peroxidase activity. Following antigen retrieval, the sections were incubated with an HSP90B1 primary antibody (1:1000, rabbit, Biolab, China) overnight at 4 °C. This was succeeded by an incubation with a secondary antibody (ZB-2305, ZSGB-BIO, CN) for one hour at room temperature. For visualization, 3,3’-diaminobenzidine (DAB) reagent (ZLI-9018, ZSGB-BIO, Beijing, China) was applied to develop the peroxidase reaction, followed by hematoxylin counterstaining. Microscopic images were captured using an Olympus C-5050 camera. Independent and blind evaluations of the staining results were performed by two pathologists who categorized the immunohistochemical outcomes based on the intensity of staining in positive cells and the proportion of positive staining. The expression level of HSP90B1 was thus determined.

Transient transfection and establishment of stable cell lines

For the transient transfection of HNSC cell lines, si-HSP90B1 RNAs and control RNAs were synthesized (WZ Biosciences, Shandong, China) and introduced into cells using the Lipofectamine 2000 reagent (Thermo Fisher, Waltham, MA, USA) as per the provided guidelines. The shHSP90B1 lentiviral vectors were created by choosing an appropriate siRNA (small interfering RNA) sequence. Following puromycin selection, cells with stable transfection were employed in subsequent experiments. In a parallel approach, HSP90B1 overexpression was accomplished in cell lines using an HSP90B1 overexpression plasmid (HSP90B1-OE) and a corresponding empty vector control (HSP90B1-EV). Sequence of siHSP90B1: 5′-TCGCCTCAGTTTGAACATTGA-3′(palindromic sequence), 5′-AAGTTGATGTGGATGGTACAT-3′(antisense sequence). Overexpression sequence:pLV[Exp]-CMV>LPCAT1[NM_024830.5 ] -CBH>GFP-2A-Puro (Cas9X™, Haixing Biosciences, China).

Cell Counting Kit-8 assays

Post-transfection, cells were centrifuged and counted. In a 96-well plate, each well received 1,000 cells in 100 µL of medium. The experiment was divided into four groups: the siHSP90B1 group(si) and its control group (siNC), overexpression group (OE) and empty plasmid vector group (EV), wild group,with each group having 5 replicate wells per day. Daily, 10 µL of CCK-8 reagent (Invitrogen™, Thermo Fisher, Waltham, MA, USA) was added to each well at a consistent time and incubated for 2 h in the same environment. The optical density (OD) of the cells was then measured at 450 nm.

Clonogenic assay

In these colony formation assays, cells were plated in six-well plates at a density of 1,000 cells per well and cultured for two weeks. Following the cultivation period, the medium was discarded, and the established colonies were fixed with a 4% paraformaldehyde solution. The fixed colonies were then stained with a 0.1% crystal violet solution. Photographs of the stained six-well plates were taken to document the colony formation, and the number of colonies was quantitatively assessed.

Transwell assay

Dilute the matrix gel with serum-free medium at a 1:29 ratio. Take 100 µL of the diluted gel, spread it on the upper chamber of the Transwell (Corning, NY, USA), and incubate for 1.5 h. Following transfection, 5  × 10⁴ cells were placed in the upper chamber, and 600 µL of complete medium (containing 10% serum DMEM) was added to the lower chamber. The cells were cultured for 24 h, fixed with 4% paraformaldehyde, stained with 1% crystal violet, photographed under the microscope, and counted.

Wound healing experiment

Following transfection, the cells were pelleted by centrifugation and subsequently enumerated. Scratch inserts by ibidi GmbH, Germany, were affixed to twelve-well plates, and three replicates per group were established for si, siNC, wild, EV, and OE treatments. A consistent volume seeding 30,000 cells was placed into the insert’s designated compartment. Post-attainment of full confluence within the chambers, and robust adherence to the walls, the scratch inserts were carefully extracted. The resultant cellular monolayers were then transferred to six-well plates containing a serum-free growth medium for further culture. At the outset and after twenty-four hours, the wounds induced by the insert removal were examined using microscopy, and corresponding images were captured. The rate of wound closure, indicative of the healing rate, was quantified at the conclusion of the observation period.

Flow cytometric assessment

The supernatant was saved and the cells were treated with EDTA-free trypsin. The trypsin was neutralized with the saved supernatant and the cell suspension was collected. It was then washed twice with PBS and centrifuged at 1,000 rpm for 5 min. Cells were suspended in buffer containing Annexin V-Alexa Flour647/7-AAD Kit (4A Biotech, Beijing, China), and stained for apoptosis detection using flow cytometry.

In vivo experiments

For constructing a model of hematogenous metastases, TU686 cells (1 × 10⁷) were subcutaneously injected into the caudal vein of the mice. The animal experiments conducted in this study were approved by the Ethics Committee of Guangxi Medical University (approval number: 202302003). We obtained male BALB/c nude mice of SPF grade from the school’s animal facility, which were 5 weeks old, to create subcutaneous tumorigenic and hematogenous metastatic tumor models. TU686 cells were transfected with lentivirus containing shHSP90B1 or shNC. For constructing a model of Subcutaneous tumor, TU686 cells (5  × 10⁶) were subcutaneously injected into the armpits of the mice.For constructing a model of hematogenous metastases, TU686 cells (1 × 10⁷) were subcutaneously injected into the caudal vein of the mice. These procedures are performed without anesthesia because of the small injection volume of cell fluid (100–150 ul), short injection duration, and skilled injection technique. All mice were kept in animal facilities without pathogens. The animal room was kept at 18–23 °C, the humidity was 40%–60%, and the light and shade cycle was 10–14 h. In order to get the most scientific experimental results and according to the principle of 3R reduction, there are six mice in each group, a total of 12 mice were randomly divided into two groups. All food, cages, water and other items that came into contact with rats were sterile, and they were treated in certified biosafety cabinets using aseptic technology. Tumor size were measured every 7 days. On day 28, the maximum tumor diameter was about 15 mm. Thus, all mices of ubcutaneous tumorigenic groups were euthanized. Then, tumor size and weight were measured,and tumor tissues were harvested for immunohistochemistry. On day 48 (when fluorescence was detected in all mice of shNC groups), detecting visceral hematogenous metastases by animal in vivo imaging. During the experiment, the mices showed no cachexia and abnormal pain, all mices were killed by cervical dislocation after anesthesia, while those with tumors injected into the tail vein were observed for 4 months. Cervical dislocation euthanasia after anethetized; after anesthetizing the mice with isoflurane, the experimenters grabbed the base of the tail and lifted it up, placed it on the cage cover or other rough surface, and pressed down hard on the head and neck with the thumb and index finger of the other hand, the right hand grabbed the base of the tail and pulled it backward and upwards, causing the dislocation of the cervical spine and the separation of the spinal cord from the brain stem. The experimental animals died immediately.

Statistical analysis

The same type of experiment was repeated three times. Each data that needs manual measurement will be repeated three times, and the average value will be taken as the result. Data were analyzed using SPSS 23.0 software, and measurements were presented as mean ± standard deviation (X + SD). The correlation of clinicopathological characteristics was analyzed by chi-square test and Log-rank statistical analysis, and all experimental and control groups were statistically analyzed by independent samples T-test, and the difference was considered statistically significant at p < 0.05.0.01 <*p < 0.05, 0.001 < **p < 0.01, ***p < 0.001.

HSP90B1 is overexpressed in HNSC and correlated with poor prognosis in HNSC patients

To examine the differential expression of HSP90B1, we utilized the GEPIA database (http://gepia.cancer-pku.cn/) and performed analyses on both cell lines and tissue specimens. Data from GEPIA indicated that HSP90B1 expression was elevated in cancerous tissues relative to their normal counterparts (Fig. 1A). Furthermore, PCR analysis of cell lines revealed pronounced upregulation of HSP90B1 in head and neck squamous cell carcinoma (HNSC) cells (TU686 and Fadu, SAS) when contrasted with normal human oral keratinocytes (HOK) (Fig. 1B). Similarly, HSP90B1 displayed increased expression in HNSC tissue samples compared to adjacent non-tumorous tissues from our patient cohort (Fig. 1C). Western blot assays corroborated the heightened expression of HSP90B1 in HNSC cell lines (Fig. 1D). Immunohistochemistry (IHC) assays further verified the upregulation of HSP90B1 in cancer tissues versus normal tissues (Fig. 1E). Prognostic evaluation using the Kaplan–Meier plotter database revealed an inverse relationship between elevated HSP90B1 expression in HNSC patients and their overall survival (OS), suggesting an adverse prognostic impact (Fig. 1F). Collectively, these findings denote that HSP90B1 is overexpressed in HNSC and may serve as a biomarker for unfavorable prognosis.

To delve deeper into the clinical relevance of HSP90B1 in HNSC, we sourced expression profiles of HSP90B1 and clinical data from the TCGA database for comprehensive analysis. Our analysis revealed that patients exhibiting elevated HSP90B1 levels experienced significantly reduced survival times. Notably, heightened HSP90B1 expression correlated with advanced clinical T Stage, the presence of distant metastasis, and higher tumor grade (Table 1). Further, by correlating immunohistochemistry (IHC) results with clinicopathological features, a significant association between HSP90B1 expression and clinical stage was established (Table 2).

HSP90B1 enhance proliferation and inhibits apoptosis in HNSC cell

To elucidate the role of HSP90B1 in head and neck squamous cell carcinoma, we conducted knockdown and overexpression experiments on TU686 and SAS cell lines (Fig. 2C). The CCK8 assay indicated that cells with HSP90B1 knockdown (siHSP90B1 group, si) exhibited a lower proliferation rate compared to the siRNA negative control group (siNC), while cells with HSP90B1 overexpression (OE group) demonstrated higher proliferation than the empty vector control group (EV) in both TU686 and SAS cells (Figs. 2A, 2B). WB assays and flow cytometry were employed to investigate HSP90B1’s role in apoptosis inhibition in HNSC cells. The WB assays showed increased expression of pro-apoptotic markers CASP3 and BAX and decreased expression of anti-apoptotic BCL2 in the siHSP90B1 group. In contrast, the overexpression group exhibited the reverse pattern. Correspondingly, flow cytometry results revealed an increase in apoptotic cells in the siHSP90B1 group and a decrease in the overexpression group (Fig. 2D). These findings suggest that high expression of HSP90B1 promotes proliferation and inhibits apoptosis in HNSC cells.

HSP90B1 promotes invasion, metastasis of HNSC cells

Cell migration and invasion are critical factors in the progression of HNSC. Wound healing assays demonstrated that HSP90B1 knockdown significantly increased the scratch width in the si group compared to the siNC at 12 h post-wounding. Conversely, cells with HSP90B1 overexpression exhibited decreasing trends (Fig. 3A). Furthermore, transwell assays indicated a decrease in the number of invasive cells in the si group compared to the siNC group, with an increase observed in OE group relative to the EV group (Fig. 3B). These findings indicate that elevated HSP90B1 expression enhances HNSC cell migration and invasion, contributing to the malignancy’s aggressive biological behavior.

HSP90B1 activates the AKT/mTOR pathway to inhibit autophagy

To explore the potential mechanisms of HSP90B1 in HNSC, we analyzed the top 200 genes closely related to HSP90B1 (similar genes) using the GEPIA database, identifying 1,613 positively correlated genes and 80 negatively correlated genes with a correlation coefficient greater than 0.3 from the UALCAN database. We intersected these datasets and pinpointed 75 genes consistently positively associated with HSP90B1. These genes underwent pathways enrichment analysis via Sangerbox (Fig. 4A). The analysis revealed significant enrichment of HSP90B1 such as focal adhesion, human papillomavirus infection, the PI3K/AKT signaling pathway, and ECM-receptor interactions (Fig. 4A). Notably, the PI3K/AKT signaling pathway emerged as the most prominent. Wikipath analysis also suggests that HSP90B1 participates in the PI3K/AKT/mTOR signaling pathway by regulating AKT phosphorylation (see Supplemental Information). Given that the PI3K/AKT/mTOR pathway is pivotal in regulating autophagy, we hypothesized that HSP90B1 influences autophagy through this pathway (Fig. 4B). Western blot analysis demonstrated that HSP90B1 knockdown reduced phosphorylated AKT (p-AKT), suggesting an inhibition of PI3K/AKT phosphorylation signaling (Fig. 4C). Meanwhile, P62 and LC3 are vital markers for autophagy process(Yu, Chen & Tooze, 2018), P62 levels decreased while LC3II/LC3I levels increased, indicating enhanced autophagy (Fig. 4C). In contrast, HSP90B1 overexpression resulted in inverse trends in these markers. These results indicate that HSP90B1 may suppress autophagy through the PI3K/AKT signaling pathway.

HSP90B1 promotes HNSC cell proliferation and metastasis in vivo

To further understand the in vivo functional role of HSP90B1 in head and neck squamous cell carcinoma (HNSC), we generated a cell line with stable HSP90B1 knockdown (shHSP90B1) and a control cell line (shNC). We established both a subcutaneous tumorigenic model (Fig. 5A) and a tail vein hematogenous metastasis model in BALB/C nude mice (Fig. 5C). The study revealed that tumors from HSP90B1-depleted cells were notably smaller and exhibited slower growth compared to the control (Fig. 5A). Immunohistochemical analysis confirmed reduced HSP90B1 expression in the sh-HSP90B1 group (Fig. 5B). In the hematogenous metastasis model, the sh-HSP90B1 group showed significantly lower compositional density (Fig. 5C), indicating that HSP90B1 plays a role in promoting proliferation and metastasis of HNSC in vivo.