A novel CREB5/TOP1MT axis confers cisplatin resistance through inhibiting mitochondrial apoptosis in head and neck squamous cell carcinoma

The Ethics Committee of Shanghai Jiao Tong University approved our study. All experimental methods comply with the Declaration of Helsinki. All animal studies were conducted in accordance with the “Guidelines for the Care and Use of Laboratory Animals of the National Institutes of Health” and were approved by the Animal Care and Use Committee of Shanghai Jiao Tong University. The experimental mice were kept in the Central Laboratory Animal Facility of the Ninth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine.

All the samples were collected from the Department of Oral and Maxillofacial-Head and Neck Oncology, Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine (Shanghai, China). After tissues were collected from the human body, they were quickly transferred to a tissue preservation solution for storage. RNA was extracted before tissues were frozen. Because the tissue was not repeatedly frozen and thawed, the experimental data accurately reflected the level of RNA in the body.

Our research involving human participants, human material, or human data has been performed in accordance with the Declaration of Helsinki and approved by the ethics committee of Ninth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (the reference number: SH9H-2021-TK57-1).

The human HNSCC cell lines HN4 and HN30 were kindly provided by the University of Maryland Dental School, USA. All of the above cells were maintained in DMEM with 10% FBS and cultured at 37°C with 5% CO₂. CR-HNSCC cells (HN4/DDP and HN30/DDP) were constructed from HN4 and HN30 through a gradient increasing the cisplatin dose and maintained in the culture medium containing 20 μM cisplatin (Howson, China).

The Agilent SurePrint G3 Human Gene Expression v3 8x60K Microarray (DesignID:072363) chip experiment and the data analysis of four samples were conducted at OE Biotechnology Co., Ltd., (Shanghai, China). The Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis was performed by DAVID (https://​david.​ncifcrf.​gov).

Cell viability was checked using an MTT assay. To determine the half maximal inhibitory concentration of cisplatin (IC50_cisplatin) on HNSCC cells, the tumor cells were seeded in triplicate into 96-well plates with 3000 cells per well. Cells were incubated with serial doses of cisplatin at 24 h post-seeding. MTT (Sigma-Aldrich, USA) was added to each well (5 mg/ml) at 72 h post-treatment, then 150 μL of dimethyl sulfoxide (DMSO) was added to each well to dissolve the formazan formed after 4-h incubation. The absorbance at 490 nm was measured using a plate reader (SpectraMax i3x, Molecular Devices, USA).

To evaluate the cell proliferation ability, tumor cells were seeded in triplicate into 96-well plates with 1000 cells per well. During the cultivation, the growth of tumor cells was monitored by measuring the absorbance at 490 nm by daily MTT detection.

Total proteins from the whole cell lysate were separated on 4–20% polyacrylamide gradient gel and transferred onto the PVDF membrane (Merck Millipore, USA). After being blocked with 5% skim milk, PVDF membranes were blotted with the primary antibody overnight at 4 °C. The GAPDH was used as a loading control. The signal was developed using an Odyssey Infrared Imaging System (Biosciences, USA) or with ECLUltra (New Cell and Molecular Biotech, Suzhou, China). All the antibodies are listed in Additional file 1: Table S1.

The total RNA from cultured cells or tissues was extracted using TRIzol reagent (Invitrogen, USA) and then reverse transcribed into cDNA using the Prime Script RT kit (Takara, Japan). Real-time PCR was performed using the Hieff UNICON® qPCR SYBR Green Master Mix kit (YEASEN, China) and the ABI StepOne real-time PCR system (Applied Biosystems, USA). β-actin was used as a housekeeping gene. The primer sequences are listed in Additional file 1: Table S2.

CREB5 and TOP1MT overexpression lentivirus and control lentivirus were purchased from Hanyin Biotechnology Limited Company (Shanghai, China). Lentiviral transduction was performed according to the manufacturer’s instructions with polybrene (final concentration of 10 μg/ml). Puromycin (10 μg/ml) was used to select stable transduced cells at 72 h post-transduction. After continuous cultivation for 1 month, the corresponding experiment was carried out. CREB5 and TOP1MT gene overexpression cells and control cells were named CREB5, TOP1MT, and Vector, respectively.

Small-interfering RNA (siRNA, Sangon Biotech, China) was transfected using Lipofectamine 2000 reagent (Invitrogen, USA) at a final concentration of 50 nM. The siRNA sequences are listed in Additional file 1: Table S3.

Around 1000 cells were seeded into six-well plates and cultured for 24 h, then the medium was replaced with fresh DMEM with or without cisplatin. The cells were cultured in the incubator for another 8–15 days until visible colonies were formed. The colonies were fixed and stained with crystal violet (YEASEN, China). The colony-forming ability was calculated based on the colony size and number.

Cells were incubated with or without cisplatin for 24 h. Then, the cells were removed from the culture using trypsin and washed twice with ice-cold PBS. Apoptotic cells were stained using APC Annexin V Apoptosis Detection Kit (BD Pharmingen^TM, USA) and quantified by flow cytometry (BD Biosciences, USA).

Cellular oxygen consumption rate (OCR) was analyzed on an XF96 Extracellular Flux Analyzer (Seahorse Bioscience, USA). The day before the experiment, 1 × 10⁵ cells/well were seeded on Seahorse 96-well culture plates and incubated overnight. The seahorse XF calibration solution was used to hydrate the probe overnight in a carbon dioxide-free incubator at 37 °C. On the day of the experiment, the detection solution was prepared with seahorse XF Base Medium .Take out Seahorse 96-well culture plates and observe the cell state and density under a microscope. Change the cell culture medium into the detection solution and place it in a carbon dioxide-free incubator at 37 ° C for 1h. Prepare and dilute the drug to the required concentration, and add it into the four dosing holes of ABCD on the test board according to the experimental design. For mitochondrial fitness tests, OCR was measured sequentially at basal, and following the addition of 1 μM oligomycin (O), 0.5 μM luoro-carbonyl cyanide phenylhydrazone (F), and 1 μM antimycin (A).

Cells were seeded onto six-well plates (10–20% confluence). About 48 h later, Mito-Tracker Red CMXRos (200 μM stock solution; Beyotime, China) was added to the culture medium (the final concentration was 20–200 nM) and incubated at 37°C for 15 min. Then, the supernatant was removed, a fresh cell culture medium (pre-warmed to 37°C) was added to the wells, and Hoechst 33342 (Beyotime China) was used to stain nuclei. Quantitative analysis of mitochondrial activity was performed using the Cytation™ 5 cell imaging multifunctional detection system (BioTek, USA).

The Enhanced ATP Assay Kit (Beyotime, China) was used to determine the ATP level in cells. Cells were washed with ice-cold PBS and then lysed with 200 μl of lysis buffer for each well of the six-well plate. Scrap to lift the cells and the lysate was clarified by centrifugation (12,000×g at 4°C for 5 min). In a 96-well plate, 20 μl of lysate was added to each well, which contained 100 μl of ATP detection solution per well. The detection was performed by a multifunctional microplate reader (SpectraMax i3x, Molecular Devices, USA).

Mitochondrial-ROS (MT-ROS) was measured by the change in fluorescence of the MitoSOX Red Mitochondrial Superoxide Indicator (YEASEN, China). Briefly, cells were plated in six-well plates and cultured in a cell incubator for 24 h. MitoSOX Red Mitochondrial Superoxide Indicator was added at 5 μM after being washed in ice-PBS and incubated at 37 °C in the dark for 10 min. The detection was performed by a multifunctional microplate reader (SpectraMax i3x, Molecular Devices, USA) after being washed with preheating buffer for 3 times.

Mitochondrial-ATP (MT-ATP) fluorescent probe pCMV-Mito-AT1.03 (Beyotime, China) was transfected by Lipofectamine 2000 reagent (Invitrogen, USA) to detect the mitochondrial ATP level in cells. The detection was performed by a multifunctional microplate reader (SpectraMax i3x, Molecular Devices, USA).

An enhanced mitochondrial membrane potential assay kit with JC-1 (Beyotime, China) was used to analyze the mitochondrial membrane protein. Remove the culture supernatant of adherent cells and wash it with PBS. Then, 1×JC-1 working solution was added to the medium for 20 min at 37 °C in the dark to label the mitochondria. Wash twice with JC-1 dyeing buffer, and add 2 ml cell culture medium. The detection was performed by a multifunctional microplate reader (SpectraMax i3x, Molecular Devices, USA). Normal mitochondrial potential showed red fluorescence (aggregate JC-1), and damaged mitochondrial potential showed green fluorescence (monomeric JC-1).

Cells were fixed in 37% formaldehyde at room temperature for 15 min, and then, the DNA was sheared to 500–1000 bp by sonication. The chromatin was immunoprecipitated with anti-CREB5 antibody and IgG negative control (ChIP Assay Kit, Beyotime, China; DNA Purification Kit, Beyotime, China). ChIP-PCR primers, 2000 bp upstream of the TOP1MT promoter region, were designed and synthesized by RiboBio, China. The purified chromatin was analyzed and quantified by real-time PCR (primers are listed in Additional file 1: Table S4).

Luciferase reporter (200 ng/well) and Renilla luciferase vector (10 ng/well; pRL-CMV; Hanyin, China) were transfected into HEK293T or tumor cells, which had been transfected with CREB5 overexpressing plasmid, at 24 h post-seeding into a 24-well plate (3×10⁴ cells per well) using Lipofectamine^TM 2000 (Invitrogen, USA). The luciferase activity was measured using a Dual-Luciferase Reporter Assay kit (Beyotime, Shanghai, China) according to the manufacturer’s instructions at 24 h post-transfection.

The cells were treated with an AKT inhibitor (MK2206; final concentration 10 μM; Beyotime, China) for 24 h, and then, the nuclear protein and plasma protein were extracted using a Nuclear and Cytoplasmic Protein Extraction Kit (Beyotime, China).

The tumor xenograft experiment was used to identify the role of CREB5 in HNSCC cells in vivo. CREB5 overexpression plasmid transfected HN30 cells or parent control HN30 cells were injected into BALB/C athymic nude mice (4-week-old; 1×10⁷ cells plus Matrigel per injection). The mixture was injected subcutaneously into the buttocks of nude mice to establish a tumor-bearing model. The tumor volume and mouse body weight were measured every 3–4 days. The mice were intraperitoneally injected with cisplatin (4 mg/kg) every 6 days of four injections at 11 days post-cell injection. The mice were sacrificed and the tumors were removed for subsequent analysis at 31 days. The tumor growth curve was plotted using tumor volume.

To evaluate the effect of targeting CREB5 and TOP1MT individually or in combination in cisplatin resistance in vivo, the nude mice were divided into four groups (three mice per group): the siNC group, siTOP1MT group, siCREB5 group, and siTOP1MT+siCREB5 group. Thereafter, HN30/DDP cells (1 × 10⁷ tumor cells plus Matrigel per injection) were subcutaneously injected into the buttocks of mice. Beginning approximately 11 days later, the mice received four intraperitoneal injections once every 6 days, with each injection containing 4 mg/kg cisplatin. For the siTOP1MT group, siCREB5 group, and siTOP1MT+siCREB5 group, the corresponding cholesterol-modified siRNA (5 nmol per spot; RiboBio, China) was injected around the tumor at the indicated time. The mice were sacrificed on day 31 for further analysis.

The animal research was conducted in accordance with the Basel Declaration outlines fundamental principles and approved by the ethics committee of Ninth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (the reference number: SH9H-2021-A896-1).

The xenografts were fixed, dehydrated, and embedded in paraffin. Then, the sections were deparaffinized, rehydrated, immersed in a citric acid buffer for heat-induced antigen retrieval, and then immersed in 0.3% hydrogen peroxide to block endogenous peroxidase activity, using 10% goat serum blocking agent and incubation with primary antibody overnight at 4°C. DAKO ChemMate Envision kit/HRP (Dako-Cytomation, USA) was then used for the development of sections, and the sections were counterstained with hematoxylin, dehydrated, removed, and fixed. Finally, the sections were observed under a microscope, and the tissues with brown staining on the cytoplasm, nucleus, or cell membrane were considered positive. A tissue microarray (TMA) was prepared using 70 HNSCC and 18 normal oral mucosal tissues obtained from the Ninth People’s Hospital Affiliated to Shanghai Jiao Tong University between 2007 and 2008, from patients who underwent surgery and were diagnosed through pathological examinations.

This experiment was approved by the laboratory animal care and use committee of the hospital. The PDX model was constructed as previously reported (Zhou et al., 2019). The patient’s tumor tissue was cut into 20–30-mm³ pieces and planted on the flank of BALB/C nude mice (6 weeks, male, about 20g). When the tumor volume reached 1500–2000 mm³, the mice were sacrificed. Tumor tissues were cut into pieces and seeded again for passaging. Third-passage mice with tumor volumes of 150–250 mm³ were used to evaluate the therapeutic value of the siCREB5 and siTOP1MT. Tumor volume (TV) was recorded every 3–5 days, and mice were sacrificed after therapy for 4 weeks. The tumor growth inhibition rate (TGI) was calculated as

The statistical analysis was performed using GraphPad Prism. The Student’s t test or one-way ANOVA was used to analyze the significance between two or more groups. The data were presented as the mean ± SEM of three independent experiments, where p < 0.05 was considered statistically significant (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

First, we constructed CR-HNSCC cell lines (HN4/DDP and HN30/DDP) by gradually increasing the concentration of cisplatin (Fig. 1A). To seek differential genes in the CR-HNSCC cells, RNA sequencing was performed. The results showed that 241 genes were upregulated in both HN4/DDP and HN30/DDP cells (Fig. 1B). KEGG was used to analyze 241 genes, and it was found that the PI3K-AKT signaling pathway ranked first (Fig. 1C). It was also observed that cisplatin treatment could significantly induce the phosphorylation of AKT at Ser 473 in a dose- and time-dependent manner (Fig. 1D, E and Additional file 1: Fig. S1A-C).

In addition, CREB5 expression had the greatest difference in CR-HNSCC cells in the PI3K-AKT signaling pathway (Fig. 1F, G). CREB5 was indeed highly expressed in CR-HNSCC cells in a cisplatin-concentration-dependent manner (Fig. 1H, I and Additional file 1: Fig. S1D, E). To clarify whether CREB5 was regulated by AKT phosphorylation and entered the nucleus as a transcription factor, we used the AKT inhibitor (MK2206) to detect the enrichment of nuclear CREB5 after inhibiting AKT activity. The results showed that the expression of phosphorylated AKT (ser 473) was upregulated and the nucleus translocation of CREB5 was increased in HN4/DDP and HN30/DDP cells. In addition, its nucleus translocation was significantly reduced after inhibiting the PI3K-AKT signaling pathway (Fig. 1J and Additional file 1: Fig. S1F, G) and downregulating AKT expression (Additional file 1: Fig. S2). The above data indicate that the transcription factor CREB5 is regulated by AKT phosphorylation and then translocated into the nucleus.

To further investigate the correlation between CREB5 expression and HNSCC cisplatin resistance, we used real-time PCR to quantify CREB5 mRNA levels in the extended HNSCC cohort. Consistent with the above data, CREB5 was upregulated in 40 CR-HNSCC samples compared with 32 non-CR-HNSCC samples (Fig. 1K). As shown in Table 1, high CREB5 expression was significantly correlated with larger tumor size (diameter greater than 4 cm) and advanced TNM (Tumor Node Metastasis) stage. Interestingly, patients with high CREB5 expression had poor responses to the TPF (docetaxel, cisplatin, 5-FU) strategy. In addition, COX regression analysis showed that CREB5 expression was an independent predictor for overall survival in HNSCC (Fig. 1L).

To explore the biological function of CREB5 in the cisplatin resistance mechanism of HNSCC, we established cell lines with CREB5 overexpression or knockout (Fig. 1M, N). Ectopic expression of CREB5 notably elevated the IC50_cisplatin in HN4 cells (3.635 ± 0.2522 μM vs 9.757 ± 0.278 μM), so it is in HN30 cells (3.977 ± 0.921 μM vs 11.51 ± 0.8776 μM, Fig. 1O). In HN4/DDP cells, CREB5 knockout reduced the IC50_cisplatin from 71.23 ± 3.394 μM to 9.848 ± 1.231 μM (Fig. 1P). In HN30/DDP cells, CREB5 silencing decreased the IC50_cisplatin from 52.15 ± 1.832 μM to 5.935 ± 1.467 μM (Fig. 1P).

To verify whether CREB5 conferred cisplatin resistance in vivo, we subcutaneously injected HN30 expressing with or without CREB5 into nude mice to establish xenograft tumor models. Overexpression of CREB5 promoted tumor growth and resulted in greater tumor weight (Fig. 1Q, R). Moreover, we observed the upregulation of CREB5 mRNA and protein level in the CREB5-expressing group (Fig. 1S). In addition, the percentage of Ki67 staining was significantly increased in the CREB5 overexpression group compared to the control (Additional file 1: Fig. S3). These results indicated that CREB5 could promote the cisplatin resistance of HNSCC in vitro and in vivo.

The above results show that CREB5 promotes HNSCC cell resistance to cisplatin. Therefore, we next evaluated whether CREB5 affected cell proliferation, colony formation, and apoptosis. First, we determined the effect of CREB5 on cell proliferation and found that CREB5 promoted HN4 and HN30 cell proliferation (Fig. 2A), while CREB5 deletion significantly reduced HN4/DDP and HN30/DDP cell proliferation (Fig. 2B). We also observed that CREB5 improved the colony formation ability of HN4 and HN30 cells (Fig. 2C), while the number of colonies in HN4/DDP and HN30/DDP silencing CREB5 was greatly reduced (Fig. 2D). These results show that CREB5 can enhance HNSCC cell proliferation and colony formation ability.

To determine the role of CREB5 in HNSCC cell survival, a cell apoptosis assay was performed by flow cytometry. Compared with the vector group, the basal and cisplatin-induced apoptosis percentage of the CREB5 overexpression group was significantly reduced in HN4 and HN30 cells (Fig. 2E). Similarly, the knockdown of CREB5 greatly increased the basal and cisplatin-induced apoptosis percentage of HN4/DDP and HN30/DDP cells (Fig. 2F). To explore the mechanism, we detected the expression of mitochondrial apoptosis-related proteins (Fig. 2G). We found that overexpression of CREB5 could promote the expression of Bcl-2 and Bcl-xL and inhibit the expression of Bax and the release of cytochrome c, which inhibit the mitochondrial pathway of apoptosis and vice versa (Fig. 2H and Additional file 1: Fig. S4). Hence, we concluded that upregulation of CREB5 promoted tumor progression and inhibited mitochondrial apoptosis in HNSCC.

CREB, a key transcription factor family, plays an important role in maintaining cell survival by regulating the genes required to maintain mitochondrial function, autophagy, ROS production, and antioxidant response [18, 19]. There are several critical genes involved in mitochondrial biosynthesis, including peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α), nuclear respiratory factors (NRFs), mitochondrial transcription factor A (TFAM), mitochondrial DNA topoisomerase I (TOP1MT), peroxisome proliferator-activated receptor gamma (PPAR-γ), peroxisome proliferator-activated receptor gamma coactivator 1-beta (PGC-1β), mitochondrial transcription factor B1 (TFBM1), mitochondrial transcription factor B2 (TFBM2), transient receptor potential cation channel subfamily M member 2 (TRPM), and brain-derived neurotrophic factor (BDNF) [19‐27]. To explore whether the above genes are transcriptional targets of CREB5, RT-PCR was performed and showed that only TOP1MT had consistent upregulation in HN4 and HN30 cells after CREB5 overexpression (Fig. 3A and Additional file 1: Fig. S5). TOP1MT plays an indispensable role in the modification of DNA topology, relieving tension and DNA supercoiling generated in the mitochondrial genome during replication and transcription [28]. Immunoblotting assay further confirmed that CREB5 overexpression could elevate the protein level of TOP1MT (Fig. 3B). Conversely, silencing CREB5 expression inhibited the mRNA and protein expression of TOP1MT in cisplatin-resistant cell lines (Fig. 3C, D). In addition, we choose CREB inhibitor (KG-501) and AKT inhibitor (MK2206) to detect further the relationship between CREB5, AKT, and TOP1MT. The result showed that the expression level of TOP1MT in cisplatin-resistant cells decreased after CREB5 was inhibited, while the combination with MK2206 resulted in a more significant decrease in TOP1MT expression (Additional file 1: Fig. S6). It is noteworthy that we found a positive correlation between CREB5 and TOP1MT expression in cisplatin-resistant HNSCC samples (Fig. 3E). Furthermore, TOP1MT overexpression promoted the expression of anti-apoptotic protein, including Bcl-2 and Bcl-xL, and inhibited the expression of pro-apoptotic protein, including Bax and cytochrome c (Fig. 3F and Additional file 1: Fig. S7A, B). In contrast, TOP1MT knockdown inhibited the anti-apoptotic protein and promoted pro-apoptotic protein in cisplatin-resistant cells (Fig. 3G and Additional file 1: Fig. S7C, D). This imbalance subsequently activated caspases 3 and 7 and PARP, initiating mitochondrial apoptosis. The results indicate that TOP1MT has a significant function in regulating the activation of the mitochondrial apoptosis pathway.

Given that CREB5 is a transcription factor, we speculate that CREB5 might promote the transcription of TOP1MT. We designed eight primers based on the transcription factor binding motif of CREB5 in the TOP1MT promoter region. ChIP-PCR results showed that only primer 2 could amplify the TOP1MT promoter fragment in HN4/DDP and HN30/DDP cells (Fig. 3H). The binding sites of CREB are divided into two types: the canonical motif (TGACG) and the non-canonical motif (TGGCG) [29]. There are two canonical motifs and three non-canonical motifs in the TOP1MT promoter region. To clarify the specific binding position of CREB5 in the TOP1MT promoter, we constructed five TOP1MT deletion mutation vectors (pro-Mut 1-5) based on known binding sites in the canonical and non-canonical motifs (Fig. 3I). A dual-luciferase reporter assay revealed that only the pro-Mut 1 vector had significantly lower luciferase activity than the wild-type (WT) in HN4 and HN30 cells (Fig. 3J). Moreover, TOP1MT promoter activation was significantly increased in cisplatin-resistant cells, which could be reduced by the deletion of −253 to −257 bp (Fig. 3K). Therefore, we concluded that the binding motif between CREB5 and the TOP1MT promoter was TGACG, which was located at ch8:143361725-143361721. To sum up, we conclude that TOP1MT participating in mitochondrial apoptosis was a novel transcriptional target of CREB5 in cisplatin-resistant cells.

To explore the role of TOP1MT in CREB5 regulating mitochondrial activity, we examined the effects of CREB5 and TOP1MT on oxidative phosphorylation (OXPHOS). The results showed that both CREB5 or TOP1MT overexpression can promote the process of cellular OXPHOS. Further analysis found that basal respiration, ATP production, maximal respiration, and respiratory capacity were all increased after CREB5 and TOP1MT overexpression (Fig. 4A–D). Cellular oxygen consumption rate (OCR) gives a deeper idea of the mitochondrial activity and bioenergetic. These results demonstrated that both CREB5 and TOP1MT enhanced mitochondrial activity. In addition, Mito-Tracker Red CMXRos also was used to detect mitochondrial activity. The results show that overexpression of CREB5 enhances mitochondrial activity with an increasing red fluorescence intensity in HN4 and HN30 cells, which is reduced by TOP1MT knockdown. Conversely, CREB5 silencing inhibited mitochondrial activity, while the overexpression of TOP1MT reversed it in cisplatin-resistant cells (Fig. 4E–H).

Mitochondrial dysfunction leads to abnormal respiratory chain function and decreased ATP production, triggering mitochondrial apoptosis. We found that overexpression of CREB5 increased cellular ATP levels, which was reduced by TOP1MT knockdown in HN4 and HN30 cells (Fig. 4I). Furthermore, CREB5 silencing decreased ATP levels, while the overexpression of TOP1MT reversed that effect in cisplatin-resistant cells (Fig. 4J). The increasing red fluorescence intensity and ATP levels in cisplatin-resistant cells also proved that mitochondrial activity upregulation was a key event during cisplatin resistance (Additional file 1: Fig. S8). Besides, we detected the intracellular levels of mitochondrial superoxide (MT-SOX indicator), mitochondrial ATP (MT-ATP), and JC-1. The results showed that the levels of MT-ROS and MT-ATP were significantly increased after CREB5 expression upregulation, and the level of monomeric JC-1 was significantly decreased. On this basis, we found that the levels of MT-ROS and MT-ATP were correspondingly decreased after TOP1MT knockdown and the level of monomeric JC-1 was also increased (Additional file 1: Fig. S9). This is also consistent with our previous conclusion that CREB5 regulated mitochondrial activity through TOP1MT.

The overexpression of CREB5 inhibited mitochondrial apoptosis by increasing Bcl-2 and Bcl-xL as well as decreasing Bax, cytochrome c, cleaved caspase, and PARP. This effect was reversed by TOP1MT knockdown (Fig. 4K and Additional file 1: Fig. S10A, B). In contrast, in cisplatin-resistant cells, CREB5 silencing induced the mitochondrial apoptosis pathway, which was rescued by TOP1MT overexpression (Fig. 4L and Additional file 1: Fig. S10C, D). These results proved that CREB5 regulated mitochondrial apoptosis through TOP1MT, especially in cisplatin-resistant cells.

According to the above findings, in which CREB5 regulated mitochondrial apoptosis through TOP1MT in cisplatin-resistant cells, we speculated that CREB5 might confer resistance to cisplatin through TOP1MT in HNSCC. Flow cytometry and an IC50 assay revealed that ectopic expression of CREB5 reduced the percentage of apoptotic cells and elevated the IC50_cisplatin, while silencing TOP1MT increased apoptosis and decreased the IC50_cisplatin (Fig. 5A, B). We observed similar results in HN30 cell lines (Fig. 5C, D). In the resistant cell lines, CREB5 knockdown significantly increased the percentage of apoptotic cells and reduced the IC50_cisplatin, while the overexpression of TOP1MT reversed that trend in HN4/DDP and HN30/DDP cell lines (Fig. 5E, H). The above results show that CREB5 confers cisplatin resistance through TOP1MT via apoptosis regulation in HNSCC.

To determine whether double-targeting of CREB5 and TOP1MT overcomes HNSCC resistance to cisplatin, we established a cisplatin-resistant xenograft model of HNSCC by subcutaneously injecting HN30/DDP cells into the buttocks of nude mice. Cholesterol-modified siCREB5 and siTOP1MT were used to silence the expression of CREB5 and TOP1MT in the xenograft tumor models. The results showed that compared with that in the control group, tumor weights in the gene silencing groups were significantly reduced. Among them, the double-targeting group (siCREB5 and siTOP1MT) resulted in the most obvious inhibition of tumor weight (Fig. 6A). Similarly, tumor growth rates in the siTOP1MT group, siCREB5 group, and siCREB5+siTOP1MT group were significantly reduced, while the combination group had the best response to cisplatin treatment (Fig. 6B).

Moreover, the synergistic inhibition of TOP1MT and CREB5 could dramatically decrease the positive percentage of Ki67 (Fig. 6C). These results showed that double-targeting of TOP1MT and CREB5 had a potent antitumor effect in the cisplatin-resistant model. To confirm the gene silencing effect in vivo, the tumor tissues were analyzed by RT-PCR and immunoblotting assay. The mRNA and protein levels of TOP1MT and CREB5 were significantly suppressed, especially in the double-targeting group (Fig. 6D, E).

To evaluate the therapeutic value of siCREB5 and siTOP1MT in HNSCC in vivo, we constructed the PDX model with cisplatin-resistant patient samples (Fig. 6F, G). The results showed that siCREB5 and siTOP1MT significantly inhibited the growth of HNSCC PDXs in vivo (Fig. 6H, I). The TGI was 82.35%. The tumor weight was significantly decreased in the treatment group compared with the control group (Fig. 6J). These results indicated that double-targeting of CREB5 and TOP1MT had a significant effect in overcoming the cisplatin resistance of HNSCC.

To further explore the correlation between CREB5 and TOP1MT, immunohistochemistry was performed in an HNSCC tissue microarray. First, we found that both CREB5 and TOP1MT were significantly higher in tumor tissues than in the normal control (Fig. 7A–C). Moreover, there was a positive correlation between the expressions of CREB5 and TOP1MT (r = 0.438, Fig. 7D). In addition, we analyzed the correlation between protein expression and clinical characteristics and found that higher expression of CREB5 and TOP1MT correlated with the advanced TNM stage, while not with others (Fig. 7E–H).

Overall, this study demonstrated that CREB5 nucleus translocation mediated by AKT phosphorylation induced by cisplatin resistance transcriptionally activated TOP1MT, which inhibited mitochondrial apoptosis through Bcl-2 upregulation and Bax downregulation in HNSCC (Fig. 7I).