ESCO2 promotes hypopharyngeal carcinoma progression in a STAT1-dependent manner

To interrogate the relationship between ESCO2 and the clinical-pathological characteristics of the HPC, the clinical data and RNA sequencing data from 44 normal specimens and 458 HNSC deposited in the TCGA database were downloaded from UCSC Xena Data Hubs (http://​xena.​ucsc.​edu/​). Forty-three samples had paired adjacent normal tissue and pathology information among the samples. The data were normalized (TMM [Mean of M-Values]), and ESCO2 expression was analyzed against different clinical stages, gender, and age. Clinical data of patients with different ESCO2 expression levels were retrieved, analyzed, and plotted with tableone package using R. Additionally, the R package “survminer” was used to determine the cut-off point of ESCO2 expression levels based on the maximally selected log-rank statistics. The influences of tumor stages and baseline variables on the outcome by Cox proportional hazards analysis. The baseline variables in the univariable analyses were gender and age.

To evaluate the functional importance of ESCO2 in HPC pathogenesis in vitro, the FaDu cell line was acquired from the Cell Bank of Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (RRID: CVCL_1218, stock number TCHu132). The cells were incubated at 37 °C in a humidified atmosphere with 5% CO₂ and were cultured in Minimum Essential Medium (MEM) supplemented with 10% FBS.

The ESCO2-targeting or control shRNAs (shESCO2-1, shESCO2-2, shCtrl) were purchased from Genechem (Shanghai, China) (Table 1) and constructed into the GFP-expressing GV115 lentiviral vector (Genechem, China). For STAT1 overexpression, full-length STAT1 (NM_001384880) was built into the Ubi-MCS-SV40-Cherry-IRES-puromycin plasmid. Virus packaging and harvesting were conducted as previously described [12]. Harvested viruses were used to transduce FaDu cells (2 × 10⁵) seeded into a 6-well plate for further studies.

As previously reported, the relative expression of ESCO2 in cells was determined by qPCR [13]. In brief, RNA was isolated from FADU cells utilizing TRIzol solution (Pufei, China). Complementary DNA (cDNA) was synthesized using the ReverTra Ace™ qPCR RT Kit (Promega, USA) according to the manufacturer’s instructions. Gene amplification was conducted in Light Cycler 480 II Real-Time PCR System (Roche, Germany) using M-MLV reverse transcriptase (Progema, US). The primer sequences of ESCO2 and GAPDH are shown in Table 2. The ESCO2 expression level was quantified relative to GAPDH using the 2^−ΔΔCt method.

Western blot analysis was performed as previously described [14]. In general, cells were solubilized in RIPA buffer (Fisher Scientific) containing a protease inhibitor cocktail (Sigma), and protein concentration was determined with the Pierce BCA Protein Assay Kit (Fisher Scientific). Subsequently, an equal amount of total cell lysate was separated in SDS-PAGE gels and transferred to polyvinylidene difluoride membranes (Millipore, USA). The membrane was blocked in 5% goat serum for 1 h and was incubated with primary antibodies at 4ºC overnight, followed by secondary antibodies (Table 3). The membranes were developed by chemiluminescent substrate, and the immunoreactive bands were visualized on X-ray films. The intensity of bands using the densitometric scanning module on the ImageJ software reflected the relative protein expression. GAPDH was used as the internal control.

To evaluate the effect of ESCO2 silencing on cell viability, FaDu cells (2 × 10³ cells/well) with or without ESCO2 depletion were seeded into 96-well plates and cultured for the indicated periods. Then, 20 µl of MTT solution (5 mg/ml, Genview) was added to each well and incubated for 4 h at 37 °C. After removing the liquid, insoluble formazan crystals were dissolved in 100 µl DMSO (Sigma, USA) in an orbital shaker. Subsequently, the absorbance was measured at 570 nm in a spectrophotometer (Tecan Infinite, Switzerland). The optical density was used to reflect the number of viable cells.

To determine how ESCO2 and STAT1 regulate cell proliferation, their expressions were modulated via lentivirus-based gene overexpression or knockdown techniques. Lentivirus packaging and cell transduction were carried out by Genechem Col, Ltd. (Shanghai, China). Specifically, lentiviral particles for ESCO2 knockdown (KD) and their corresponding negative controls (NC) express GFP; while those for STAT1 overexpression (OE) and the corresponding negative controls (NC) express RFP. Cells were divided into three groups based on the gene expression modifications: negative controls (NC + NC), ESCO2 knockdown (KD + NC), and ESCO2 knockdown in combination with STAT1 overexpressing (KD + OE). Subsequently, cells from each group were seeded into 96-well plates (2 × 10³/well) in triplicates. After that, the fluorescence intensity was determined by Celigo® Image Cytometer (Nexcelom) for 5 consecutive days to reflect cell proliferation [15].

To evaluate if ESCO2 is involved in cell migration, GFP-positive FaDu cells (3 × 10⁴/well) with or without ESCO2 depletion were cultured in Oris^™ cell migration plates with cell seeding stoppers in the incubator overnight. Then, the stoppers were removed, and the culture medium was replaced with the starving medium (1% FBS). Images were captured at indicated periods (0, 24, and 48 h) using an inverted fluorescence microscope. Cells observed in the wound area, the cell-free zone, were considered migrated cells. The wound closure rate was calculated using the formula: [(wound area at 0 h – wound area at 24 or 48 h)/ wound area at 0 h] × 100%.

As previously reported, cell migration was assayed using the Transwell system [16]. Transduced FaDu cells in 100 µl of FBS-free medium were loaded in the upper chamber, and 600 µl of normal culture medium was added in the bottom chamber. The Transwell system was then returned to the incubator for 48 h. Subsequently, the cells in the upper chamber were removed, and the migrated cells were fixed with 4% paraformaldehyde and stained with crystal violet. Migrated cells from each well were counted and photographed in 9 randomly selected fields.

Cell cycle assay was conducted to determine if ESCO2 is involved in cancer cell cycle progression. In brief, FaDu cells were collected and washed with ice-cold PBS twice. Then, the cells were fixed with 75% ice-cold ethanol overnight and resuspended in PBS. Subsequently, the cells were stained with podium iodide (PI) (Sigma, USA) and screened by a flow cytometer (Millipore, USA). The cell cycle phases were analyzed using ModFit software (ModFit, USA).

According to the manufacturer’s instructions, cell Apoptosis was evaluated using an Annexin V-APC apoptosis detection kit (eBioscience, 88-8007). In brief, control or ESCO2-silenced cells were harvested, prepared into single-cell suspension in binding buffer, and incubated with annexin V-APC and PI at room temperature for 15 min in the dark. Subsequently, the cell apoptosis was determined by a flow cytometer (Millipore, USA), and the results were analyzed using FCS Express software (version 3.0; DeNovo).

To confirm the role of ESCO2 in HPC progression in vivo, a cell line-derived xenograft model was established in nude mice. GFP-positive control or ESCO2-silenced cells (shCtrl, shESCO2-1) subcutaneously inoculated in BALB/c nude mice (female, 4 weeks old) purchased from Lingchang Biotechnology (Shanghai, China). In brief, mice were randomly allocated to 2 groups (n = 10), and each mouse received 4.0 × 10⁶ transduced cells in 200 µl of PBS/Matrigel mixture (BD Biosciences) subcutaneously under the right armpit. After that, the mice underwent tumor volume measurement every three days, and tumor volumes were plotted against time. Tumor volume was calculated using the formula (L × W × W)×π/6, where L and W are the tumor length and width, respectively [17]. On day 20, the fluorescent signals of the tumors were collected by the IVIS Imaging System (Perkin Elmer, USA). After scanning, animals were sacrificed by an overdose of 2% pentobarbital sodium and confirmed dead by cervical dislocation. Then, tumors were harvested and weighed. Experiments were carried out according to the National Institute of Health Guide for Care and Use of Laboratory Animals and were approved by the Institutional Bioethics Committee of the Second Affiliated Hospital of Xi’an Jiaotong University.

To identify the interactome of ESCO2 in HPC cells, FaDu stably expressing 3-Flag-tagged ESCO2 or vehicle vectors were solubilized in IP lysis buffer (ThermoFisher Scientific) supplemented with protease inhibitor cocktail and PMSF (Sigma) for 1 h at 4 °C. Cleared lysates (16,000 × g, 20 min) were processed for Co-IP using protein A beads conjugated with anti-Flag antibody (Invitrogen) at 4 °C overnight. The immobilized proteins were eluted with 3-Flag peptide (Sigma) and subjected to immunoblotting or Coomassie blue staining followed by proteomic analysis (shotgun mass spectrometry). Antibodies used for immunoblotting are listed in Table 3.

For MS, the separated protein in the Coomassie blue-stained gel was dehydrated, vacuum-dried, and digested as previously described [18]. The generated polypeptides were identified through liquid chromatography-tandem mass spectrometry on the Q-Exactive HF-X quadrupole-Orbitrap mass spectrometer (ThermoFisher Scientific). Protein identification was achieved in the MASCOT 2.6 database and analyzed by Proteome Discoverer 2.1 (ThermoFisher Scientific).

Statistical analyses were performed with SPSS 19.0 (SPSS Inc., USA). Data are presented as means ± standard deviations (SD) of at least three independent experiments. Comparisons between the two groups were performed with the independent samples using the student’s t-test or Mann-Whitney non-parametric test. Multiple-group comparisons were performed by one-way analysis of variance with Bonferroni post-tests. The independence of categorical variables was analyzed using Fisher’s exact test. Asterisks were used to annotate statistical significance (*P < 0.05, **P < 0.01, ***P < 0.001).

First, we interrogated its expression level in tumor and normal tissues using a dataset from the TCGA database. The results showed elevated ESCO2 levels in HPC tissues compared with normal tissues (Fig. 1A, P < 0.001). Furthermore, the ESCO2 mRNA expression was compared in 43 paired specimens, and the results revealed significantly higher ESCO2 transcription levels in tumors than in paired-adjacent normal tissues (Fig. 1B, P < 0.001).

We then queried whether the ESCO2 expression level is involved in HPC patients’ tumor progression. We first allocated the patients into two groups based on their ESCO2 expression levels (ESCO2 high and ESCO2 low). Then, we retrieved the patients’ clinical information and checked whether the ESCO2 level was associated with the patients’ TNM stage, clinical stage, gender, and age (Table 4). Our data revealed that tumors with high ESCO2 expression are inclined to develop distant metastasis (M stage) (HR = 3.51, 95% CI = 1.08, 11.38, P = 0.037) and lymph node metastasis (N stage) (HR = 1.43, 95% CI = 1.01, 2.01, P = 0.04) (Fig. 1C). Together, these findings indicated that ESCO2 participates in the HNSC metastasis.

To further explore the role of ESCO2 in HPC development, ESCO2 was silenced in FaDu cells by shRNAs (shESCO2-1 and shESCO2-2), and the ESCO2 knockdown was validated by qPCR and Western blotting analyses (Fig. 2A, B, P < 0.01). Subsequently, we interrogated the function of ESCO2 in cell proliferation; both shRNAs significantly suppressed the tumor cell growth since day 2 (Fig. 2C, D). Surprisingly, introducing shESCO2-1 completely abrogated the FaDu cell growth (Fig. 2C, D). In parallel, we evaluated the cell viability in those cells and found that HPC viability was significantly impaired by ESCO2 depletion (Fig. 2E). Moreover, the inhibitory effect on cell viability of ESCO2 depletion was observed as early as 2 days after the experiment’s initiation. Therefore, these data indicated that ESCO2 relates to the HPC proliferation and viability.

Tumorigenesis is characterized by a disorganized cell cycle that leads to uncontrolled cellular proliferation [19]. Meanwhile, we observed that ESCO2 is closely related to cell growth and viability in HPC cells. Consequently, we sought to determine whether ESCO2 is involved in cell cycle progression. Thus, we compared the cell population in control or ESCO2-depleted cells at different cell cycle stages. Our results revealed that ESCO2 depletion led to a marginal increase in the G1 population (Fig. 2F, P < 0.01) but showed no significant difference in the S population.

Furthermore, we queried whether ESCO2 will affect apoptosis in HPC cells. The results demonstrated that ESCO2-targeting shRNA, especially ESCO2-1, significantly promoted cellular apoptosis in FaDu cells (Fig. 2G, P < 0.01). Specifically, in the control group, 4.23% of cells underwent apoptosis. Conversely, in cells expressing shESCO2-1 or shESCO2-2, the average apoptotic cell portion was 20.83% and 8.7%, respectively. Together, these findings suggested that ESCO2 appears to affect apoptosis in HPC cells significantly.

We then evaluated the importance of ESCO2 in cell migration. The wound-healing assay revealed that ESCO2 depletion significantly inhibited wound closure. As shown in Fig. 3A, the wound closing rates in the control group were approximately 50% and 75% at 24 and 48 h, respectively. However, the wound closure was significantly inhibited by ESCO2 depletion at the corresponding time points (Fig. 3A, C, P < 0.01). The Transwell migration assay was also conducted to confirm this result. We found that silencing ESCO2 inhibited the HPC cell migratory potential significantly compared with the cells transduced with control non-targeting shRNA (Fig. 3B, D, P < 0.01). More importantly, shESCO2-1 almost completely blocked the cancer cell migration (Fig. 3B, D, P < 0.01). Together, these data suggest that ESCO2 is involved in tumor cell migration.

We established a subcutaneous xenograft in nude mice to understand further the functional importance of ESCO2 in HPC growth in vivo. FaDu cells transduced with lentivirus expressing GFP-ESCO2-targeting shRNA (KD) or GFP-control shRNA (NC) were injected subcutaneously into nude mice. Tumor volume was measured with a caliper and preclinical imaging system. As reflected by the marginal fluorescence signals, ESCO2-depleted FaDu cells barely grow into the detectable tumor in all tested mice (Fig. 4A upper panel, n = 10).

In addition, at the end of this study, we harvested the tumors from all mice and compared the volumes. Consistent with our preclinical imaging results, ESCO2-depleted cells hardly formed sizable tumors, while those transduced with control shRNA grew into large tumors (Fig. 4B-C, n = 10, P < 0.01). In parallel, the average tumor weight at the end of the study also revealed significant differences between these two groups (Fig. 4D, n = 10, P < 0.01). Together, these data suggest that ESCO2 is critical in regulating tumor growth in vivo.

Notably, the reconstitution of STAT1 attenuated the inhibitory effects on cell growth, viability, and migration achieved by ESCO2 silencing. Thus, we speculated that ESCO2 might contribute to the tumorigenesis of HPC by influencing STAT1. Furthermore, it has been demonstrated that ESCO2 can promote the methylation of histone H3 at lysine 9 (H3K9), which drives tumor growth in vivo [20] by forming macromolecular complexes [21]. Therefore, we speculated that ESCO2 might similarly promote HPC progression. As our data indicated that ESCO2-silenced cells showed significantly inhibited tumor cell growth, we sought to identify ESCO2’s interactome with a focus on cell migration and/or proliferation-related pathways. So, we performed Co-IP/MS analysis and identified the proteins that potentially interact with ESCO2. As a result, we identified 2349 peptides in the control group and 5590 peptides in the ESCO2 overexpressing group, which matched to 757 and 1270 proteins, respectively. From the candidates, we tested the interaction between ESCO2 and cell migration and/or proliferation-related proteins, including calcyclin-binding protein (CACYBP), cullin 1 (CUL1), cadherin 1 (CDH1), histone deacetylase 2 (HDAC2), catenin beta 1 (CTNNB1), martin 3 (MATR3), galectin 3 (LGALS3), and signal transducer and activator of transcription 1 (STAT1). The data indicated that ESCO2 interacts with STAT1 (Fig. 5). Notably, STAT1 depletion promotes HNSC progression and metastasis [22]. Therefore, it is reasonable to speculate that ESCO2 may interact with STAT1 to inhibit and impair its tumor suppressor activity. However, the molecular activity needs further investigation.

So far, our data have indicated that ESCO2 depletion significantly suppressed tumor growth in vitro and in vivo. In addition, a recent bioinformatics study demonstrated the potential significance of the STAT1/ESCO2 pathway in breast cancer [23]. It is reasonable to speculate that STAT1 may also be involved in ESCO2-associated tumorigenesis in HPC. Therefore, we investigated whether STAT1 plays a role in ESCO2-related HPC biology. Thus, we overexpressed STAT1 in ESCO2-depleted cells and evaluated how this modulation affects ESCO2-mediated cellular activities. First, we performed a proliferation assay and found that when ESCO2 was silenced, the cell proliferation was inhibited compared with that of the control group (Fig. 6A, B, P < 0.05).

Moreover, when we further overexpressed STAT1 in those proliferation-inhibited cells, the ESCO2 depletion-mediated cell growth suppression was significantly impaired (Fig. 6A, B, P < 0.05). Meanwhile, the cell viability assay also demonstrated that ESCO2 silencing significantly suppressed cell viability, but this effect was compromised by STAT1 overexpression (Fig. 6C, P < 0.05). We also evaluated whether STAT1 overexpression could counteract the ESCO2-depletion-mediated suppressive effect on HPC mobility. Consistent with our results, ESCO2-silenced cells exhibited reduced migratory potency compared with the control group. However, we also noticed that the ESCO2-deficiency-inhibited cell mobility was recovered when STAT1 was overexpressed (Fig. 6D, E, P < 0.05). Together, these results suggested a promotive role of STAT1 in HPC progression. More importantly, combined with observing the physical interaction between STAT1 and ESCO2, these findings indicated the involvement of STAT1 in ESCO2-mediated HPC progression.