THBS1 upregulation by KLF7 in hypopharyngeal squamous cell carcinoma contributes to lung metastasis through the p38 MAPK signaling

Summary

This study aimed to analyze the specific role of thrombospondin-1 (THBS1) in hypopharyngeal squamous cell carcinoma (HPSCC) and its mechanism. The expression of histone deacetylase 6 (HDAC6), Kruppel-like factor 7 (KLF7), and THBS1 in the tumor and peritumor tissues of patients with HPSCC, HPSCC cells, and human oral keratinocytes was examined. The function of the HDAC6/KLF7/THBS1/p38 MAPK axis in HPSCC cells was explored using lentivirus-mediated genetic interventions in combination with EdU, colony formation, wound healing, Transwell assays, and Western blot assays. An in vivo lung metastasis model in nude mice was constructed by the tail vein injection of FaDu cells. HDAC6 expression was significantly downregulated in HPSCC, losing the removal capacity of H3K9ac marks at the KLF7 promoter, leading to the upregulation of KLF7, which in turn induced THBS1 transcription to activate p38 MAPK signaling and promote the epithelial-mesenchymal transition (EMT) and lung metastasis of HPSCC cells. These findings indicate that HDAC6 hinders KLF7/THBS1/p38 MAPK axis transduction and curtails HPSCC malignant progression by impairing EMT.

Graphical abstract

Highlights

• •KD-THBS1 curtails the malignant behaviors of HPSCC
• •The p38 activator LX-3 promotes HPSCC cell behaviors in the presence of KD-THBS1
• •KLF7 activates THBS1 transcription to promote lung metastasis of HPSCC cells
• •HDAC6 removes H3K9ac marks at the KLF7 promoter to inhibit KLF7 expression

Biological sciences; Microbiology

Introduction

Head and neck squamous cell carcinoma (HNSC) is a highly aggressive malignancy, often resulting in an unfavorable prognosis, mainly when diagnosed at advanced stages, and exhibiting a persistently low five-year survival rate due to high rates of recurrence and therapeutic resistance.¹ HNSC primarily affects the oral cavity, oropharynx, hypopharynx, and larynx.² Hypopharyngeal cancer accounts for 3–5 % of HNSC, and the most common pathological type of hypopharyngeal cancer is squamous cell carcinoma (HPSCC).³ Lymph node metastasis is a prevalent phenomenon, and this occurrence is significantly associated with a poor overall survival rate for patients with HPSCC.⁴ Furthermore, distal metastasis, most commonly to the lungs, contributes to unfavorable prognoses and restricts therapeutic alternatives.⁵ Collectively, lymphatic and hematogenous dissemination are significant contributors to patient survival, underscoring the necessity of experimental models that accurately reflect these metastatic routes. The identification of thrombospondin-1 (THBS1, also known as TSP1) as an angiogenesis inhibitor in 1990 generated interest in its role in cancer biology and its potential as a therapeutic target.⁶ THBS1 activates TGFβ, which can behave as an oncogenic factor in progressive stages, contributing to metastasis and aggressiveness.⁷ For instance, THBS1 overexpression rescued the inhibited proliferative and invasive abilities of glioblastoma cells after VSIG4 knockdown.⁸ It has been recently reported that upregulated THBS1 in oral keratinocytes activated p38 signaling to polarize M1-like tumor-associated macrophages.⁹ p38 mitogen-activated protein kinase (p38 MAPK) plays an important role in metastasis.¹⁰ To acquire the abilities required to form metastases, epithelial stem cells or differentiated epithelial cells must undergo what is referred to as the epithelial-mesenchymal transition (EMT).¹¹ Our previous study has highlighted the involvement of the p38 MAPK in the regulation of HPSCC cell growth and migration.¹² Herein, we aimed to investigate the role of THBS1-mediated p38 MAPK during metastatic cascades and clarify the underlying mechanisms governing the THBS1 upregulation in HPSCC.

Krüppel-like factors (KLFs) are a group of zinc-finger transcription factors (TFs), and KLF7, also termed ubiquitous KLF based on its ubiquitous expression in adult human tissues, is a conserved gene in animals.¹³ KLF7 was found to trigger the transcription of IGF2BP2 by directly binding to its promoter and super-enhancer regions, and patients with HNSC with the high expression of KLF7 showed poorer prognosis.¹⁴ Moreover, KLF7 has been reported to maintain the stemness of oral squamous cell carcinoma.¹⁵ However, its specific role in HPSCC remains unclear. Aberrant histone modifications have gained recognition as critical contributors to the development and progression of HNSC during which histone-modifying enzymes are actively engaged, including histone acetyltransferases, histone deacetylases (HDACs), and histone methyltransferases.¹⁶ The HDAC6 gene is located on chromosome Xp11.23 and encodes a protein of 1215 amino acids, the largest protein of the HDAC family.¹⁷ This study set out with the objective of conducting an investigation into the mechanisms by which HDAC6/KLF7/THBS1/p38 signaling pathways regulate metastasis in HPSCC.

Results

Thrombospondin-1 is upregulated in hypopharyngeal squamous cell carcinoma tissues and cell lines

We screened the differentially expressed genes (DEGs) between 14 HPSCC tissues and 4 normal control tissues in the GSE2379 dataset from the public GEO database (Figure 1A) with adj.p. Val <0.05, |LogFC| > 2. To obtain genes with more prognostic significance, we downloaded the prognostic markers for HNSC from UALCAN (https://ualcan.path.uab.edu/index.html) (Data S1) and obtained 42 intersections (Figure 1B) in Jvenn (https://jvenn.toulouse.inrae.fr/app/example.html). Thereafter, we used STRING (https://string-db.org/) to set the minimum required interaction score as high confidence (0.700) and hid the disconnected nodes in the work, constructing the PPI interaction network of 42 intersecting genes. As revealed in Figure 1C, IL6 and THBS1 were the two hub genes with the highest number of interaction evidence lines (n = 15). Since IL6 has been widely studied in HPSCC,^18,19 THBS1 was chosen for further study.

Figure 1.

THBS1 is highly expressed in HPSCC tissues and cell lines

(A) A volcano plot demonstrating DEGs between HPSCC and normal control tissues in the GSE2379 dataset (adj. p Val <0.05, |LogFC| > 2).

(B) The intersection of DEGs in the GSE2379 dataset and prognostic markers in HNSC.

(C) The PPI protein interaction network of 42 intersecting genes by STRING (disconnected nodes in the work are not shown).

(D) p-value and LogFC value of the THBS1 gene in the GSE2379 dataset.

(E–G) Prognostic significance of THBS1 expression in patients with HNSC analyzed by UALCAN. RT-qPCR (F) and Western blot (G) analysis of THBS1 expression differences between tumors and paired peritumoral tissues in 22 patients with HPSCC. RT-qPCR (H) and Western blot (I) analysis of THBS1 expression differences between HOK, FaDu, and Detroit 562 cells. The results are expressed as the mean ± SD.

(H–I). These values are derived from three independent experiments. Statistical analysis was performed using paired t test (F-G) or one-way ANOVA (H-I), followed by Tukey’s multiple comparison test.

Analysis of the GSE2379 dataset showed that THBS1 (LogFC = 2.539451) expression was significantly higher in HPSCC tissues than in normal control tissues (Figure 1D). Furthermore, the prognostic results in UALCAN indicated that patients with HNSC with high THBS1 expression had a significantly lower survival probability (Figure 1E).

To confirm the above prediction results, we collected primary tumor biopsy tissues and paired peritumoral tissues from 22 patients with HPSCC and analyzed THBS1 expression using RT-qPCR and Western blot analysis (Figures 1F and 1G). The mRNA and protein of THBS1 were higher in tumor tissues than in peritumor tissues of patients with HPSCC. Not only that, THBS1 levels were significantly upregulated within FaDu and Detroit 562 cells compared to HOK (Figures 1H and 1I).

Exogenous inhibition of thrombospondin-1 curtails the proliferation, migration, and invasion of hypopharyngeal squamous cell carcinoma cells

To further explore the cellular functions regulated by THBS1 in HPSCC, we infected HPSCC cells with lentiviral vectors containing three shRNAs targeting THBS1. Knockdown of THBS1 significantly downregulated the mRNA (Figure 2A) and protein levels (Figure 2B) of THBS1 in HPSCC cells compared with the KD-Scramble group. To minimize potential off-target effects, we selected KD-THBS1 #1 and KD-THBS1 #2 with superior knockdown efficiencies for the subsequent experiments.

Figure 2.

Exogenous inhibition of THBS1 significantly curtails the malignant behavioral capacity of HPSCC cells. RT-qPCR

(A and B) Western blot (B) analysis of THBS1 expression in FaDu and Detroit 562 cells after infection with KD-THBS1 #1–3.

(C) HPSCC cell proliferative capacity was examined using EdU staining.

(D) The colony formation of HPSCC cells.

(E–I) The migratory capacity of HPSCC cells was examined using wound healing assays. Transwell assay detects changes in the migratory (F) and invasive (G) abilities of HPSCC cells. Expression of E-cadherin and N-cadherin in HPSCC cells was examined using Western blot analysis (H) and dual-labeling immunofluorescence (I). The results are expressed as the mean ± SD (A–I). These values are derived from three independent experiments. Statistical analysis was performed using a one-way ANOVA (A–I), followed by Tukey’s multiple comparison test.

EdU (Figure 2C) and colony formation assay (Figure 2D) showed that knockdown of THBS1 significantly inhibited the proliferation of HPSCC cells. Wound healing assay (Figure 2E) and Transwell assay (Figures 2F and 2G) also revealed that the exogenous inhibition of THBS1 repressed HPSCC cell migration and invasion. We analyzed the expression of E-cadherin and N-cadherin in HPSCC cells using Western blot analysis and dual-labeling immunofluorescence. Knockdown of THBS1 significantly upregulated the E-cadherin protein level and downregulated N-cadherin (Figures 2H and 2I).

The p38 MAPK activator LX-3 rescues the malignant behavior of hypopharyngeal squamous cell carcinoma cells repressed by KD-thrombospondin-1

We first employed Western blot analysis to examine the phosphorylation levels of p38 alongside its total protein expression (Total p38) in FaDu and Detroit 562 cells, as well as the levels of its downstream targets p-ATF2 and p-MK2. Knockdown of THBS1 significantly reduced the extent of p38, ATF2, and MK2 phosphorylation, while the total p38 expression remained stable (Figure 3A). We further explored whether THBS1 could affect HPSCC cell malignant behavior through the p38 MAPK pathway using the p38 MAPK activator LX-3. Unsurprisingly, LX-3 intervention significantly upregulated intracellular p-p38, p-ATF2, and p-MK2 levels (Figure 3B). EdU and Transwell assays showed that the proliferation (Figure 3C), migration (Figure 3D), and invasion (Figure 3E) abilities of HPSCC cells were significantly upregulated after LX-3 administration. In addition, LX-3 treatment diminished E-cadherin expression and enhanced N-cadherin expression in HPSCC cells (Figure 3F).

Figure 3.

The p38 MAPK activator LX-3 rescues the malignant behavior of HPSCC cells repressed by the silencing of THBS1

(A) Western blot analysis of the extent of p38, ATF2, and MK2 phosphorylation, as well as total p38 expression in KD-THBS1-treated HPSCC cells.

(B) Western blot analysis of the extent of p38, ATF2, and MK2 phosphorylation, as well as total p38 expression in HPSCC cells treated with 2 μM LX-3 after knockdown of THBS1.

(C–E) HPSCC cell proliferative capacity was examined using EdU staining. Transwell assay detects changes in the migratory (D) and invasive (E) abilities of HPSCC cells.

(F) Expression of E-cadherin and N-cadherin in HPSCC cells was examined using dual-labeling immunofluorescence. The results are expressed as the mean ± SD (A–F). These values are derived from three independent experiments. Statistical analysis was performed using an unpaired t test.

To validate the molecular function of THBS1, we performed lentiviral infection of FaDu and Detroit 562 cells with an overexpression construct for THBS1. Western blot analysis confirmed the stable overexpression of THBS1 within the cells (Figure S1A). Cells were then treated with the p38 MAPK inhibitor SB 203580. As expected, THBS1 overexpression significantly upregulated p-p38, p-ATF2, and p-MK2 levels, while SB 203580 intervention markedly downregulated these proteins; however, Total-p38 levels remained unchanged (Figure S1B). More importantly, forced activation of THBS1 significantly increased the EdU-positive staining rate (Figure S1C), migration (Figure S1D), and the number of invaded cells (Figure S1E) in HPSCC cells. However, SB 203580 substantially downregulated the proliferation, migration, and invasion capabilities of HPSCC cells. Finally, the level of E-cadherin expression in HPSCC cells was downregulated following THBS1 overexpression and upregulated after SB 203580 treatment; N-cadherin exhibited the opposite pattern (Figure S1F).

Kruppel-like factor 7 induces the transcription of thrombospondin-1 in hypopharyngeal squamous cell carcinoma cells

What are the upstream molecular mechanisms underlying the overexpression of THBS1 in HPSCC? To address this question, we obtained a set of TFs with the highest number of genes within the genomic interval between the gene transcription start site and the enhancer midpoint and with binding sites on the THBS1 promoter and enhancer. These TFs were intersected with the DEGs in the GSE2379 dataset and prognostic markers on Jvenn to the only gene: KLF7 (Figure S2A) (LogFC = 2.025231) (Figure S2B). Prognostic analysis revealed that patients with HNSC with high KLF7 expression represented a weaker likelihood of survival (Figure S2C). We found an enhanced binding peak of KLF7 at the THBS1 promoter (chr15: 39,580,792-39,581,141) by ChIP-seq prediction in UCSC (https://genome.ucsc.edu/cgi-bin/hgGateway) (Figure S2D). It was also predicted in JASPAR (https://jaspar.elixir.no/) that KLF7 has numerous binding sites on the THBS1 promoter (Figure S2E). Correlation analysis by GEPIA (http://gepia.cancer-pku.cn/index.html) also showed that KLF7 was significantly positively correlated with THBS1 expression in the HNSC (Figure S2F). Therefore, we hypothesized that the KLF7 transcriptional activation of THBS1 further activates p38 MAPK pathway signaling and enhances EMT in HPSCC.

We examined the expression differences of KLF7 between primary tumor biopsies from 22 patients with HPSCC and paired peritumoral tissue (Figures 4A and 4B), and between HPSCC cells and HOK cells (Figures 4C and 4D). The expression level of KLF7 was elevated in HPSCC tissues and HPSCC cells.

Figure 4.

KLF7 transcriptional activation of THBS1 in HPSCC cells. RT-qPCR

(A–D) Western blot (B) analysis of KLF7 expression differences between tumors and paired peritumoral tissues in 22 patients with HPSCC. RT-qPCR (C) and Western blot (D) analysis of KLF7 expression differences between HOK, FaDu, and Detroit 562 cells.

(E and G) The degree of enrichment of the THBS1 promoter by anti-KLF7 in HPSCC cells was examined using ChIP. RT-qPCR (F) and Western blot analysis (G) of the knockdown efficiency of KLF7 in HPSCC cells.

(H) The mRNA levels of THBS1 in knockdown KLF7-treated HPSCC cells were detected by RT-qPCR.

(I) The luciferase activity of the THBS1 promoter in knockdown KLF7-treated HPSCC cells was detected by the dual-luciferase assay.

(J) The luciferase activity of the THBS1 promoter in HPSCC cells transfected with WT or Mut vectors was detected by the dual-luciferase assay. The results are expressed as the mean ± SD (A–J). These values are derived from three independent experiments. Statistical analysis was performed using a paired t test (A–B), unpaired t test (E, H–I), one-way ANOVA (C–D, F-G), or two-way ANOVA (J), followed by Tukey’s multiple comparison test.

ChIP experiments showed that anti-KLF7, but not IgG, was significantly enriched for the THBS1 promoter (chr15: 39,580,792-39,581,141) fragment (Figure 4E). We also infected HPSCC cells using lentiviral vectors containing KLF7 shRNAs and selected the KD-KLF7 #1 group of cells with the best knockdown efficiency for subsequent exploration (Figures 4F and 4G). Upon RT-qPCR analysis, it was found that the knockdown of KLF7 also suppressed the mRNA expression of THBS1 (Figure 4H). More importantly, a dual-luciferase assay showed that luciferase activity was also downregulated with KLF7 knockdown in HPSCC cells transfected with a luciferase vector bearing the THBS1 promoter (Figure 4I). Moreover, KLF7 knockdown did not reduce luciferase activity in HPSCC cells with the binding site mutated (THBS1 promoter^Mut), whereas it did reduce activity in cells expressing the wild-type promoter (THBS1 promoter^WT) (Figure 4J).

Reactivation of thrombospondin-1 rescues the hypopharyngeal squamous cell carcinoma malignant biological behavior curtailed by the knockdown of Kruppel-like factor 7

We combined THBS1 overexpression lentiviral intervention on HPSCC cells in the KD-KLF7 #1 group. Knockdown of KLF7 downregulated both KLF7 and THBS1 protein expression, whereas combined overexpression of THBS1 only upregulated the protein level of THBS1 (Figure 5A). More importantly, EdU and Transwell assays revealed that the knockdown of KLF7 significantly suppressed the proliferative activity (Figure 5B), migration (Figure 5C), and invasion (Figure 5D) of HPSCC cells, whereas the proliferative, migratory, and invasive capacities of the cells were again increased upon the overexpression of THBS1. We used a dual-labeling immunofluorescence assay and similarly found that the overexpression of THBS1 reversed the upregulation of E-cadherin and downregulation of N-cadherin upon knockdown of KLF7 (Figure 5E). Finally, we also analyzed intracellular p-p38 expression. Knockdown of KLF7 significantly downregulated the extent of p38, ATF2, and MK2 phosphorylation, while overexpression of THBS1 upregulated p-p38, p-ATF2, and p-MK2 protein levels (Figure 5F).

Figure 5.

Reactivation of THBS1 rescues the HPSCC cell malignant biological behavior curtailed by knockdown of KLF7

(A) Western blot analysis of protein expression of KLF7 and THBS1 in HPSCC cells treated with knockdown of KLF7 combined with the overexpression of THBS1.

(B–D) HPSCC cell proliferative capacity was examined using EdU staining. Transwell assay detects changes in the migratory (C) and invasive (D) abilities of HPSCC cells.

(E) Expression of E-cadherin and N-cadherin in HPSCC cells was examined using dual-labeling immunofluorescence.

(F) Western blot analysis of the extent of p38, ATF2, and MK2 phosphorylation, as well as total p38 expression in HPSCC cells. The results are expressed as the mean ± SD (A–F). These values are derived from three independent experiments. Statistical analysis was performed using a one-way ANOVA followed by Tukey’s multiple comparison test.

Activation of the thrombospondin-1/p38 MAPK axis by Kruppel-like factor 7 promotes lung metastasis in nude mice

To alleviate the suffering of nude mice, we chose the less invasive FaDu cells for in vivo studies. We constructed the metastasis model by injecting 1 × 10⁶ FaDu cells suspended in 100 μL of PBS into the tail vein of nude mice. HE staining was used to assess the number of metastatic nodules in the lung tissues of nude mice. KD-KLF7 or KD-THBS1 significantly reduced the number of metastatic nodules in the lungs, but combined OE-THBS1 increased the number of metastatic nodules (Figures 6A and 6B). The expression of EMT-related markers (Slug and Vimentin) was analyzed by IHC and Western blot analysis. KD-KLF7 or KD-THBS1 significantly decreased the IHC scores and expression of Slug and Vimentin within lung metastases, whereas the additional OE-THBS1 was able to reverse the benefits of KD-KLF7 (Figures 6C and 6D).

Figure 6.

Activation of the THBS1/p38 MAPK axis by KLF7 promotes the lung metastasis of HPSCC cells

(A) Lung metastasis in nude mice injected with FaDu cells in the tail vein was assessed by HE staining.

(B–D) Quantitative results of lung metastatic nodules. IHC scores (C) and protein expression (D) of Slug and Vimentin within lung metastases in nude mice.

(E) RT-qPCR detection of mRNA expression levels of KLF7 and THBS1 in lung metastases.

(F) IHC scores of KLF7 and THBS1 within lung metastases in nude mice.

(G) Western blot analysis of the extent of p38, ATF2, and MK2 phosphorylation, as well as total p38 expression in lung metastases. The results are expressed as the mean ± SD (A–G). These values are derived from five mice. Statistical analysis was performed using a one-way ANOVA followed by Tukey’s multiple comparison test.

The results of RT-qPCR and IHC showed that effective KD-KLF7 or KD-THBS1 both downregulated the expression of THBS1, which was enhanced by combined OE-THBS1 (Figures 6E and 6F). Finally, we also probed the activation of p38 MAPK within lung metastases in nude mice and found that p-p38, p-ATF2, and p-MK2 levels were downregulated by KLF7 or THBS1 knockdown, and again activated upon THBS1 overexpression (Figure 6G).

Histone deacetylase 6 removes H3K9ac modification from the Kruppel-like factor 7 promoter to inhibit Kruppel-like factor 7 expression

We found that HDAC6 protein expression was downregulated in HNSC tumors by UALCAN prediction (Figure S3A) and predicted that low HDAC6 expression represented significantly poorer prognostic significance (Figure S3B). Analysis in ChIP-seq of UCSC revealed that the KLF7 promoter is modified by H3K9ac (Figure S3C). Finally, correlation prediction in GEPIA revealed that HDAC6 was negatively correlated with KLF7 expression in HNSC (Figure S3D).

We first explored the differences in the protein expression of HDAC6 within primary tumor biopsy tissues versus paired peritumoral tissues in 22 patients with HPSCC (Figure 7A), as well as within HPSCC cells versus HOK cells (Figure 7B). HDAC6 protein levels were found to be significantly downregulated in tumors of patients with HPSCC and HPSCC cells.

Figure 7.

HDAC6 removes H3K9ac modification at the KLF7 promoter to inhibit KLF7 expression

(A and B) Western blot analysis of HDAC6 expression differences between tumors and paired peritumoral tissues in 22 patients with HPSCC (A) and between HOK, FaDu, and Detroit 562 cells (B).

(C and D) The degree of enrichment of the KLF7 promoter by anti-HDAC6 and anti-H3K9ac in HPSCC cells was examined using ChIP. Western blot analysis (D) of the overexpression efficiency of HDAC6 in HPSCC cells.

(E) The degree of enrichment of the KLF7 promoter by anti-HDAC6 and anti-H3K9ac in HPSCC cells overexpressing HDAC6 was examined using ChIP.

(F) KLF7 and THBS1 mRNA levels in HPSCC cells overexpressing HDAC6 were examined using RT-qPCR.

(G) The binding of anti-H3K14ac and anti-H4K8ac to the KLF7 promoter in HPSCC cells overexpressing HDAC6 was examined using ChIP. The results are expressed as the mean ± SD (A–G). These values are derived from three independent experiments. Statistical analysis was performed using a paired test (A), unpaired t test (D, F), one-way ANOVA (B–C), and two-way ANOVA (E, G), followed by Tukey’s multiple comparison test.

The detection of HPSCC cells by ChIP revealed that both anti-HDAC6 and anti-H3K9ac were significantly enriched for the KLF7 promoter (chr2: 207,167,132–207,167,464) fragment compared to IgG (Figure 7C). Next, we verified the overexpression of HDAC6 in HPSCC cells by Western blot analysis (Figure 7D) and performed the ChIP assay. The forced expression of HDAC6 significantly increased the abundance of the KLF7 promoter by anti-HDAC6, but this treatment dramatically downregulated the abundance of the KLF7 promoter by anti-H3K9ac (Figure 7E). More importantly, HDAC6 overexpression also downregulated the transcript levels of KLF7 and THBS1 (Figure 7F). ChIP results indicated that the overexpression of HDAC6 did not affect the levels of KLF7 promoter fragments enriched by anti-H3K14ac and anti-H4K8ac (Figure 7G). This suggests that HDAC6 modifies and represses KLF7 expression by removing H3K9ac at the KLF7 promoter.

Combined overexpression of Kruppel-like factor 7 rescues the proliferation, migration, and invasion of hypopharyngeal squamous cell carcinoma cells curtailed by overexpression-histone deacetylase 6

The HPSCC cells were infected with OE-HDAC6 alone or in combination with OE-KLF7. Exogenous overexpression of HDAC6 significantly upregulated the protein level of HDAC6 and downregulated the level of KLF7, whereas forced expression of KLF7 did not affect HDAC6 expression and only upregulated the protein level of KLF7 (Figure 8A). The exogenous expression of HDAC6 was found to significantly curb the proliferation, migration, and invasion of HPSCC cells by EdU proliferation assay (Figure 8B), Transwell migration (Figure 8C), and invasion (Figure 8D) assays. Dual-labeling immunofluorescence visualization of E-cadherin and N-cadherin expression (Figure 8E) showed that OE-HDAC6 upregulated E-cadherin and downregulated N-cadherin. Nevertheless, OE-KLF7 compromised these effects of OE-HDAC6.

Figure 8.

Overexpression of KLF7 rescues the proliferation, migration, and invasion of HPSCC cells curtailed by the forced expression of HDAC6

(A) WB detection of protein levels of HDAC6 and KLF7 in HPSCC cells infected with OE-HDAC6 alone or in combination with OE-KLF7.

(B–D) HPSCC cell proliferative capacity was examined using EdU staining. Transwell assay detects changes in the migratory (C) and invasive (D) abilities of HPSCC cells.

(E) Expression of E-cadherin and N-cadherin in HPSCC cells was examined using dual-labeling immunofluorescence. The results are expressed as the mean ± SD (A–E). These values are derived from three independent experiments. Statistical analysis was performed using a one-way ANOVA followed by Tukey’s multiple comparison test.

Inhibition of Kruppel-like factor 7/thrombospondin-1/p38 MAPK signaling by histone deacetylase 6 impedes lung metastasis of hypopharyngeal squamous cell carcinoma cells

We also injected the FaDu cells with the same infection schemes through the tail vein into nude mice to construct a lung metastasis model. Once the nude mice were euthanized, we analyzed the lungs of nude mice by HE staining and found that OE-KLF7 significantly increased the number of lung metastatic nodules downregulated by OE-HDAC6 (Figures 9A and 9B). The IHC scores and protein expression of Slug and Vimentin within lung metastatic tumors were downregulated after overexpression of OE-HDAC6 and upregulated again following OE-KLF7 (Figures 9C and 9D). Lastly, we examined the successful overexpression of HDAC6 and KLF7 in the lung metastatic nodules using both IHC (Figure 9E) and Western blot analysis (Figure 9F).

Figure 9.

Inhibition of KLF7/THBS1/p38 MAPK signaling by HDAC6 impedes the lung metastasis of HPSCC cells

(A) Lung metastasis in nude mice injected with FaDu cells in the tail vein was assessed by HE staining.

(B–D) Quantitative results of lung metastatic nodules. IHC scores (C) and protein expression (D) of Slug and Vimentin within lung metastases in nude mice.

(E) IHC scores of KLF7 and HDAC6 within lung metastases in nude mice.

(F) Western blot analysis of KLF7 and HDAC6 protein expression in lung metastases. The results are expressed as the mean ± SD (A–F). These values are derived from five mice. Statistical analysis was performed using a one-way ANOVA followed by Tukey’s multiple comparison test.

The mean protein expression of HDAC6 (mean = 0.448), KLF7 (mean = 0.685), and THBS1 (mean = 0.277) in tumor tissue samples of 22 patients with HPSCC was used to categorize patients into high- and low-expression groups, respectively. We then performed correlation analyses between gene expression levels and patients' clinicopathological parameters. As shown in Table 1, we found that patients with low HDAC6 expression or high KLF7 and THBS1 expression were associated with higher T stage, N stage, M stage, and TNM stage.

Table 1.

Correlation between clinicopathological characteristics of patients with HPSCC and HDAC6/KLF7/THBS1 protein expression

Characteristics	HDAC6 protein expression (total = 22)			KLF7 protein expression (total = 22)			THBS1 protein expression (total = 22)
	Low (n = 12)	High (n = 10)	p value	Low (n = 11)	High (n = 11)	p value	Low (n = 10)	High (n = 12)	p value
Age
≤60	5	2	0.3808	6	1	0.0635	4	3	0.6517
>60	7	8		5	10		6	9
Sex
Male	9	7	>0.9999	8	8	>0.9999	7	9	>0.9999
Female	3	3		3	3		3	3
T stage
T1-T2	3	8	0.0300∗	9	2	0.0089∗	9	2	0.0019∗
T3-T4	9	2		2	9		1	10
N stage
N0-N1	2	9	0.0019∗	10	1	0.0003∗	8	3	0.0300∗
N2-N3	10	1		1	10		2	9
M stage
M0	7	10	0.0396∗	11	6	0.0351∗	10	7	0.0396∗
M1	5	0		0	5		0	5
TNM Stage
I-II	2	8	0.0083∗	9	1	0.0019∗	8	2	0.0083∗
III-IV	10	2		2	10		2	10

The statistical significance of the contingency table was analyzed using Fisher’s exact test, with ∗p < 0.05 indicating a significant difference.

Discussion

This study described the regulatory mechanism of the THBS1 upregulation in HPSCC and found that the HDAC6 removed the H3K9ac marks of the KLF7 promoter to reduce THBS1 transcription, thereby halting the progression and metastasis of HPSCC through p38 MAPK pathway impairment.

Giraud et al. revealed that THBS1⁺ regulatory myeloid cells expanded in hepatocellular carcinoma, which populated fibrotic lesions and were associated with poor prognosis.²⁰ In this study, we found that the upregulation of THBS1 was related to the dismal prognosis of patients with HPSCC using online prediction tools as well, and linked it to the EMT event. Interestingly, THBS1 was identified as one of the ten genes that were part of the EMT pathway in Merkel cell carcinoma.²¹ It has been demonstrated that M1-like tumor-associated macrophages may play a pivotal role in the transformation of oral squamous cell carcinoma into a mesenchymal/stem-like phenotype, which was orchestrated by the IL-6/Stat3/THBS1 feedback loop.²² In addition, this group further found that macrophages were activated by taking up THBS1-containing exosomes released from oral squamous cell carcinoma cells through p38 signaling at the early phase.²³ Interestingly, livin has been found to enhance tumorigenesis in HPSCC by activating the p38 MAPK signaling.²⁴ Here, the gain of E-cadherin and loss of N-cadherin upon THBS1 knockdown in HPSCC cells was reversed by p38 activator LX-3, and the downregulation of E-cadherin and upregulation of N-cadherin upon THBS1 overexpression in HPSCC cells were reversed by p38 inhibitor SB 203580. These findings indicated that the p38 MAPK signaling is indeed the effector pathway of THBS1 in HPSCC.

Chaim et al. showed that the brain sections of mice bearing glioblastoma stem cells with THBS1 KD revealed less invasive tumors, and their transcription was regulated by GNA12.²⁵ In this study, KLF7 was predicted to be the only TF that was differentially expressed in HPSCC and might be both a prognostic biomarker and THBS1 regulator. KLF7 was the most significant prognostic gene among the 17 family members in high-grade serous ovarian cancer, and the mechanistic targets of KLF7 included genes involved in EMT.²⁶ In a cohort of patients diagnosed with breast cancer, KLF7 expression demonstrated a correlation with the aggressiveness of the intrinsic breast cancer subtype and tumor grading.²⁷ KLF7 was highly expressed in human hepatocellular carcinoma samples and correlated with patients’ differentiation and metastasis status, and KLF7 overexpression contributed to cell proliferation and invasion of hepatocellular carcinoma cells.²⁸ The metastasis-promoting effects of KLF7 overexpression have been validated in hepatocellular carcinoma and colon adenocarcinoma through different targets.^29,30,31 Our in vitro and in vivo evidence revealed that the anti-metastatic properties of KLF7 knockdown were reversed by THBS1 overexpression.

As for the upstream modifier of KLF7, it has been reported to be modulated by a microRNA in non-small cell lung cancer.³² In this study, we mainly focused on HDAC6-mediated acetylation. Zamperla et al. determined that the inactivation of HDAC6 using ITF3756, an HDAC6 inhibitor, siRNAs, or CRISPR/Cas9 gene editing led to substantial alterations in chromatin accessibility (particularly increased acetylation of histone H3 lysines 9, 14, and 27).³³ HDAC6-mediated H3K9ac modification has been implicated in thyroid tumorigenesis.³⁴ H3K9ac is generally related to transcription initiation and unfolded chromatin, thus positively influencing gene expression.³⁵ In this study, we predicted the downregulation of HDAC6 protein expression in H NSC and the presence of H3K9ac modification in the KLF7 promoter, which led to the hypothesis that the KLF7 upregulation in HPSCC was related to the loss of HDAC6 and the strengthened H3K9ac modification. Even though HDAC6 has been reported as a tumor suppressor in liver cancers,^36,37 its involvement in HPSCC has not been revealed. Our findings here are that the impairment of EMT and lung metastasis in HPSCC by the overexpression of HDAC6 was compromised by KLF7 overexpression.

The present study revealed that HDAC6 loss led to the transcription of KLF7 and THBS1, thus facilitating the EMT and lung metastasis of HPSCC cells via the p38 MAPK axis (Figure 10). Beyond the mechanistic insights, the findings of this study may have important clinical and translational implications. The HDAC6/KLF7/THBS1/p38 signaling axis is a promising therapeutic target in HPSCC. Modulation of KLF7 activity through pharmacological means, namely inhibitors, has the potential to serve as a mechanism for the precise regulation of downstream transcriptional programs.²⁶ Regulatory interventions targeting the p38 MAPK pathway, a prominent focus in oncology drug development,³⁸ hold considerable potential for enhancing therapeutic efficacy when integrated with these strategies.

Figure 10.

A schematic diagram

HDAC6 removes the H3K9ac modification of the KLF7 promoter, leading to the downregulation of KLF7 expression, which in turn impairs the transcriptional activation of KLF7 on THBS1, further curtailing p38 MAPK signaling and impeding HPSCC progression.

Limitations of the study

However, it is imperative to acknowledge the presence of certain limitations that require further elucidation. First, in this study, the primary focus was on p38 MAPK as the principal downstream effector of THBS1. However, given the multifunctional nature of THBS1, the involvement of other signaling pathways, such as PI3K/AKT signaling, is a plausible consideration,³⁹ and future studies will be needed to evaluate these additional cascades. Secondly, the focus was directed toward HDAC6 due to its poor prognostic features and negative correlation with KLF7 expression. This provided a robust rationale for examining its effect on H3K9ac modification at the KLF7 promoter. Nonetheless, additional HDAC family members, such as HDAC1, might also play a role in the regulation of KLF members.⁴⁰ Additional studies must be designed to examine the potential contributions of other HDACs to achieve a more profound comprehension of the epigenetic regulation of KLF7. Additionally, the present study was unable to evaluate the influence of sex on HPSCC, as the clinical sample numbers were limited and unbalanced. Finally, although FaDu cells are a well-established HPSCC model with reproducible tumorigenicity and relevance to the disease context, reliance on a single cell line raises the possibility of cell line-specific effects. Subsequent research, encompassing further cellular models or patient-specific xenografts, is imperative to substantiate the generalizability of our observations.

Resource availability

Lead contact

Further information and requests concerning resources and reagents should be directed to and will be answered by the lead contact, Weiwei Wang (wangwwent@163.com).

Materials availability

This study did not generate any unique new reagents. All reagents used in this study are commercially available.

Data and code availability

• •All data reported in this article will be shared by the lead contact upon request.
• •This article does not report original code.
• •Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

We thank the Youth Program of the National Natural Science Foundation of China (No. 62203117) for the funding support.

Author contributions

Huijuan Cheng: conceptualization, methodology, and writing – original draft preparation. Dongfang Tang: validation and writing – review and editing. Shousen Hu: visualization, investigation, and writing – review and editing. Zizi Zhang: formal analysis, validation, and supervision. Weiwei Wang: validation, data curation, funding acquisition, and writing – review and editing.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
E-cadherin	Cell Signaling Technologies	Cat# 14472; RRID: AB_2728770
N-cadherin	Invitrogen	Cat# PA5-19486; RRID: AB_10979609
Goat anti-Mouse IgG Secondary Antibody, Alexa Fluor Plus 488	Invitrogen	Cat# A32723TR; RRID: AB_2866489
Goat anti-Rabbit IgG Secondary Antibody, Alexa Fluor 647	Invitrogen	Cat# A-21245; RRID: AB_2535813
KLF7	Cusabio	Cat# CSB-PA126129
THBS1	BioAb	Cat# A2125; RRID: AB_2764144
Slug	ProteinTech Group	Cat# 12129-1-AP; RRID: AB_2191889
Vimentin	ProteinTech Group	Cat# 10366-1-AP; RRID: AB_2273020
HDAC6	ProteinTech Group	Cat# CSB-12834-1-AP; RRID: AB_10597094
goat anti-rabbit IgG secondary antibody, HRP	Invitrogen	Cat# 65–6120; RRID: AB_2533967
E-cadherin	Cell Signaling Technologies	Cat# 3195; RRID: AB_2291471
Phospho-p38 MAPK (Thr180/Tyr182)	Cell Signaling Technologies	Cat# 4511; RRID: AB_2139682
total p38	Cell Signaling Technologies	Cat# 9212; RRID: AB_330713
p-ATF2 (Thr71)	Invitrogen	Cat# 44-295G; RRID: AB_2533624
p-MK2 (Thr334)	Cell Signaling Technologies	Cat# 3007; RRID: AB_490936
GAPDH	BioAb	Cat# AC001; RRID: AB_2619673
KLF7	MyBioSource	Cat# MBS9129225
H3K9ac	Abcam	Cat# ab4441; RRID: AB_2118292
H3K14ac	Invitrogen	Cat# MA5-32814; RRID: AB_2810090
H4K8ac	Invitrogen	Cat# 701796; RRID: AB_2532510
Biological samples
HPSCC primary biopsy tissues and paired peritumoral tissues	First Hospital Affiliated to Zhengzhou University	N/A
Chemicals, peptides, and recombinant proteins
MEM	Procell	Cat# PM150410
penicillin-streptomycin	MedChemExpress	Cat# HY-K1006
LX-3	MedChemExpress	Cat# HY-125980
SB 203580	MedChemExpress	Cat# HY-10256
TRIzol reagent	Invitrogen	Cat# 15596018CN
Matrix-Gel	Beyotime	Cat# C0371
Lipofectamine 3000 transfection reagent	Invitrogen	Cat# L3000001
Critical commercial assays
RevertAid First-Strand cDNA Synthesis Kit	Thermo Fisher Scientific	Cat# K1622
Power SYBR Green PCR premix	Applied Biosystems	Cat# 4367659
E-Click EdU Cell Proliferation Imaging Assay Kit	Elabscience	Cat# E-CK-A377
HE staining kit	Sangon	Cat# E607318
Pierce BCA Protein Assay Kit	Thermo Fisher Scientific	Cat# 23227
SimpleChIP Enzymatic ChIP Kit	Cell Signaling Technologies	Cat# 9003
Renilla-Firefly luciferase dual assay kit	Thermo Fisher Scientific	Cat# 16186
Experimental models: Cell lines
Human oral keratinocytes	ScienCell	Cat# 2610
FaDu	Procell	Cat# CL-0083
Detroit 562	Procell	Cat# CL-0330
Experimental models: Organisms/strains
BALB/c nude mice	Vital River	Cat# 401
Oligonucleotides
KD-THBS1 #1: AGACATCTTCCAAGCATATAA	This paper	N/A
KD-THBS1 #2: GTAGGTTATGATGAGTTTAAT	This paper	N/A
KD-THBS1 #3: CGTGACTGTAAGATTGTAAAT	This paper	N/A
KD-KLF7 #1: ACGGGTGCCGGAAAGTTTATA	This paper	N/A
KD-KLF7 #2: ACTGTCATGCACTCAACTATA	This paper	N/A
KD-KLF7 #3: TCAACGCAGTGACCTCATTAA	This paper	N/A
THBS1 forward primer 5′-GCTGGAAATGTGGTGCTTGTCC-3′ and reverse primer 5′-CTCCATTGTGGTTGAAGCAGGC-3′	This paper	N/A
KLF7 forward primer 5′-CTCACGAGGCACTACAGGAAAC-3′ and reverse primer 5′-TGGCAACTCTGGCCTTTCGGTT-3′	This paper	N/A
GAPDH forward primer 5′-GTCTCCTCTGACTTCAACAGCG-3′ and reverse primer 5′-ACCACCCTGTTGCTGTAGCCAA-3′	This paper	N/A
Recombinant DNA
pGL3 Basic Vector	Addgene	Cat# 212936
pRL-TK vector	Youbio	Cat# VT1568
Software and algorithms
ImageJ software	Version 1.8.0	RRID: SCR_003070
Prism	Version 10.3.0	RRID: SCR_002798

Experimental model and study participant details

Patients

This study was approved by the Clinical Research Ethics Committee of the First Hospital Affiliated to Zhengzhou University (approval number: 2020-KY-0177-001), and all subjects provided consent for research participation. All clinical samples (HPSCC primary biopsy tissues and paired peritumoral tissues) were obtained from patients diagnosed with HPSCC at the First Hospital Affiliated to Zhengzhou University between February 2020 and April 2025, and the samples were collected during surgery. A total of 22 patients aged 33–72 years were included, comprising 16 males and 6 females. All patients were of Han ethnicity, and ancestral origin information was not collected. Additionally, for correlation analysis, the 22 HPSCC patients were divided into high-expression and low-expression groups based on the mean protein expression of HDAC6 (Mean = 0.448), KLF7 (Mean = 0.685), or THBS1 (Mean = 0.277). None of the patients enrolled had received radiotherapy or chemotherapy before surgery.

Cell culture and reagents

Human oral keratinocytes (HOK, 2610, ScienCell, Carlsbad, CA, USA) were cultured using oral keratinocyte medium (2611, ScienCell). The HPSCC cell lines FaDu (CL-0083) and Detroit 562 (CL-0330) cells were purchased from Procell (Wuhan, Hubei, China). FaDu and Detroit 562 were maintained in MEM (PM150410, Procell) containing 10% FBS, penicillin-streptomycin (HY-K1006, MedChemExpress, Monmouth Junction, NJ, USA). All the cells were placed at 37°C in a 5% CO₂ humidified environment. All cell lines were identified by short tandem repeat and routinely tested for mycoplasma contamination, with all results negative.

Interfering lentiviral vectors containing human THBS1 shRNA (KD-THBS1 #1: AGACATCTTCCAAGCATATAA; KD-THBS1 #2: GTAGGTTATGATGAGTTTAAT; KD-THBS1 #3: CGTGACTGTAAGATTGTAAAT), human KLF7 shRNA (KD-KLF7 #1: ACGGGTGCCGGAAAGTTTATA; KD-KLF7 #2: ACTGTCATGCACTCAACTATA; KD-KLF7 #3: TCAACGCAGTGACCTCATTAA), and human THBS1 (vector name: pLV[Exp]-Puro-EF1A > hTHBS1 [NM_003246.4]), human KLF7 (vector name: pLV[Exp]-Puro-EF1A > hKLF7 [NM_003709.4]), human HDAC6 (vector name: pLV[Exp]-Puro-EF1A > hHDAC6 [NM_001321225.2]) gene expression lentiviruses, as well as control lentiviral vectors were purchased from VectorBuilder (Guangzhou, Guangdong, China). FaDu and Detroit 562 cells (1 × 10⁶) were seeded in 6-well plates and infected with the above lentiviruses or control lentiviruses (MOI = 50),⁴¹ respectively, once reaching 30%–40% density. The medium was refreshed with a conventional complete medium, followed by screening with 1.5 μg/mL puromycin for 14 days.

After stable infection of HPSCC cells with lentiviruses, HPSCC cells were treated with 2 μM LX-3 (HY-125980, MedChemExpress) to activate P38 MAPK signaling or with 20 μM SB 203580 (HY-10256, MedChemExpress) to inhibit p38 MAPK signaling. DMSO was then used as a control treatment. The concentration of DMSO was less than 0.005% in all assays.

Animal studies

All animal experiments were approved by the Institutional Animal Care and Use Committee of the First Hospital Affiliated to Zhengzhou University (approval number: 2025013) and were performed following the NIH Guide for the Care and Use of Experimental Animals. Male BALB/c nude mice (Genotype: CAnN.Cg-Foxn1^nu/Crl, 5–6 weeks old, weighing 15–21 g) were obtained from Vital River (Beijing, China). The mice were maintained in specific-pathogen-free grade cages with constant temperature (22 ± 1°C) in a 12/12-h dark/light cycle.

Nude mice were randomly divided into 9 groups, namely KD-Scramble, KD-THBS1 #1, KD-KLF7 #1, KD-KLF7 #1 + overexpression (OE)-negative control (NC), KD-KLF7 #1 + OE-THBS1, OE-NC, OE-HDAC6, OE-HDAC6 + OE-NC, OE-HDAC6 + OE-KLF7 (5 mice per group). In short, 1 × 10⁶ differently treated FaDu cells were suspended in 100 μL PBS and injected into nude mice via the tail vein. After 9 weeks of rearing under the above standard conditions, all nude mice were euthanized by intraperitoneal injection of 150 mg/kg sodium pentobarbital. The lung tissues of nude mice were collected, weighed, and cleaned. One part of the lungs was frozen and preserved, and the other part was fixed, embedded in paraffin, and sectioned for the subsequent experiments.

Method details

RNA extraction and reverse transcription (RT)-qPCR

Total RNA was extracted from HPSCC patient tumors, paired peritumoral normal tissues, HPSCC cells, and HOK using TRIzol reagent (15596018CN, Invitrogen Inc., Carlsbad, CA, USA) according to the manufacturer’s protocols, and cDNA was obtained by RT using the RevertAid First-Strand cDNA Synthesis Kit (K1622, Thermo Fisher Scientific Inc., Waltham, MA, USA). RNA expression levels were quantified using Power SYBR Green PCR premix (4367659, Applied Biosystems, Inc., Foster City, CA, USA) and on a CFX96 Touch deep-well real-time fluorescence qPCR detection system (Bio-Rad Laboratories, Hercules, CA, USA). THBS1 forward primer 5′-GCTGGAAATGTGGTGCTTGTCC-3′ and reverse primer 5′-CTCCATTGTGGTTGAAGCAGGC-3’; KLF7 forward primer 5′-CTCACGAGGCACTACAGGAAAC-3′ and reverse primer 5′-TGGCAACTCTGGCCTTTCGGTT-3’. The 2^−ΔΔCT quantification method using GAPDH (forward primer 5′-GTCTCCTCTGACTTCAACAGCG-3′ and reverse primer 5′-ACCACCCTGTTGCTGTAGCCAA-3′) primers for normalization was used to calculate the average mRNA expression.

Dual-labeling immunofluorescence

HPSCC cells were fixed using 4% paraformaldehyde (PFA), permeabilized with 1% Triton, and sealed with goat serum for 30 min. Primary antibodies to E-cadherin (1:200, 14472, Cell Signaling Technologies, Beverly, MA, USA) and N-cadherin (1:50, PA5-19486, Invitrogen) were incubated at 4°C overnight. The secondary antibodies, goat anti-mouse IgG, Alexa Fluor Plus 488 (1:100, A32723TR, Invitrogen) or goat anti-rabbit IgG Alexa Fluor 647 (1:500, A-21245, Invitrogen) were incubated at 37°C for 1 h. After staining the nuclei using DAPI (HY-D0814, MedChemExpress), images were captured by fluorescence microscopy, and the mean fluorescence intensity was analyzed using ImageJ software.

Cell proliferation assay

Cell proliferation capacity was assayed using the E-Click EdU Cell Proliferation Imaging Assay Kit (Red, Elab Fluor 594) (E-CK-A377, Elabscience, Wuhan, Hubei, China) according to the manufacturer’s instructions. Specifically, HPSCC cells were seeded into 6-well plates, labeled using 10 μM EdU staining solution, and incubated with PBS containing 4% PFA for 15 min at room temperature. The cells were incubated with 1 mL of PBS containing 0.3% Triton X-100 for 20 min, with 500 μL of the configured Click reaction solution for 30 min in the dark, and with 500 μL of DAPI working solution for 10 min in the dark (all at room temperature). The results were observed under the fluorescence microscope.

Colony formation assay

HPSCC cells were plated in 6-well plates at 1000 cells/well, incubated at 37°C for 2 weeks,⁴² and fixed using PBS containing 4% PFA. The colonies formed were counted after 0.1% crystal violet staining.

Wound healing assay

FaDu and Detroit 562 (5 × 10⁵) were seeded in 6-well plates and incubated in MEM containing 10% FBS and 1% P/S overnight until a confluent monolayer was formed. Scratches were then made using a 200 μL pipette tip, and cells were washed with PBS to remove cellular debris. Mitomycin C (10 μM, 73274, NovoBiotechnology Co., Ltd., Beijing, China) was supplemented to 6-well plates to inhibit cell proliferation. At 0 and 48 h after scratching, images of the injured cell monolayer were taken using a microscope, and the area of wound healing was analyzed using ImageJ software.

Transwell assays

Corning 6.5 mm Transwell with 8.0 μm Pore Polyester Membrane Insert, Sterile (3464, Corning, Corning, NY, USA) was used. For the invasion assay, the HPSCC cells suspended in 200 μL of MEM containing 0.5% FBS were seeded in the upper chamber coated with Matrix-Gel (C0371, Beyotime, Shanghai, China) diluted at a ratio of 1:8 with serum-free MEM. For migration assays, HPSCC cells were seeded into the upper chamber of a Transwell without pre-coating. While 700 μL MEM containing 30% FBS was placed in the lower chamber. After 18 h, HPSCC cells were fixed with PBS containing 4% PFA for 20 min, followed by staining with 0.5% crystal violet stain and counting under a microscope.

Hematoxylin-eosin (HE) staining

The HE staining kit (E607318, Sangon, Shanghai, China) was used on lung tissue sections of nude mice to assess tumor metastasis. Specifically, after paraffin-embedded sections of lung tissue were deparaffinized, the nuclei were stained with hematoxylin stain for 5 min. This was followed by differentiation with 1% hydrochloric acid-ethanol for 20 s and treatment in PBS/PBST for 30 s. After washing with 95% ethanol for 10 s and counter-staining with eosin staining solution for 1 min, the sections were subjected to dehydration with ethanol and clearing with xylene. The number of lung metastatic nodules was examined by microscopy.

Immunohistochemical staining (IHC)

Paraffin-embedded sections of nude mouse lung tissue were deparaffinized, hydrated, and subjected to antigen retrieval in citrate buffer (pH = 6.0) for 20 min at 100°C. The primary antibodies used in IHC were anti-KLF7 (1:100, CSB-PA126129, Cusabio, Wuhan, Hubei, China), anti-THBS1 (1:100, A2125, BioAb, Inc., New Taipei City, Taiwan, China), anti-Slug (1:2000, 12129-1-AP, ProteinTech Group, Chicago, IL, USA), anti-Vimentin (1:5000, 10366-1-AP, ProteinTech Group), and anti-HDAC6 (1:100, 12834-1-AP, ProteinTech Group). After incubation with the goat anti-rabbit IgG secondary antibody, HRP (1:3000, 65–6120, Invitrogen) for 1 h, the reaction was visualized with the DAB staining solution (PR30010, ProteinTech Group). Sections were counterstained with hematoxylin, dehydrated, cleared, and sealed. The results of the staining were observed using a microscope.

IHC scoring results were calculated by multiplying the positive staining score by the staining intensity score. For positive staining, it was categorized as 0, negative; 1, <20%; 2, 20–50%; 3, 51–75%; and 4, >75% positive cells. For staining intensity, the grading system was divided into the following categories: 0, no staining; 1, pale yellow; 2, tan; 3, dark tan.⁴³

Western blot assay

HPSCC cells or nude mouse lung metastatic tumor tissues were treated with SDS lysis buffer (BB-3205, Bestbio, Shanghai, China) containing a mixture of phosphatase inhibitors and protease inhibitors to obtain total protein. After quantification of protein concentration using Pierce BCA Protein Assay Kit (23227, Thermo Fisher Scientific), total protein was loaded into an SDS polyacrylamide gel and transferred to a PVDF membrane. After incubation with 5% BSA at room temperature for 2 h, the membranes were maintained with the primary antibody as indicated at 4°C overnight and then incubated with the secondary antibody at room temperature for 1 h. Protein bands were visualized with a High Sensitivity ECL Kit (HY-K2005, MedChemExpress). The primary antibodies used were as follows: anti-THBS1 (1:250, A2125, BioAb), anti-E-cadherin (1:1000, 3195, Cell Signaling Technologies), anti-N-cadherin (1:100, PA5-19486m Invitrogen), anti-Phospho-p38 MAPK (Thr180/Tyr182) (1:1000, 4511, Cell Signaling Technologies), anti-total p38 (1:1000, 9212, Cell Signaling Technologies), anti-p-ATF2 (Thr71) (1:1000, 44-295G, Invitrogen), anti-p-MK2 (Thr334) (1:1000, 3007, Cell Signaling Technologies), anti-KLF7 (1:1000, CSB-PA126129, Cusabio), anti-Slug (1:1000, 12129-1-AP, ProteinTech Group), anti-Vimentin (1:20000, 10366-1-AP, ProteinTech Group), anti-HDAC6 (1:1000, 12834-1-AP, ProteinTech Group), and anti-GAPDH (1:20000, AC001, BioAb).

ChIP

ChIP assays were performed using the SimpleChIP Enzymatic ChIP Kit (9003, Cell Signaling Technologies). Chromatin from treated HPSCC cells was fixed, lysed, and sonicated. The sonicated chromatin (2%) was taken as Input, and the remaining chromatin was immunoprecipitated using antibodies to KLF7 (1:20, MBS9129225, MyBioSource, Inc., San Diego, CA, USA), HDAC6 (1:50, 12834-1-AP, ProteinTech Group), H3K9ac (1:25, ab4441, Abcam, Cambridge, MA, USA), H3K14ac (1:100, MA5-32814, Invitrogen), H4K8ac (1:20, 701796, Invitrogen), and normal Rabbit IgG (1:50). Enrichment of the THBS1 or KLF7 promoter fragment in the precipitation complex was analyzed by qPCR and imaged using 2% agarose gel. The primer sequences of qPCR were as follows: forward 5′-TTCTAGCTGGAAAGTTGCGC-3′ and reverse 5′-CTGGAGAGCGACAGGAGC-3′ for the THBS1 promoter (chr15: 39,580,792-39,581,141) and forward 5′-AACCCTGCACAATTCACGTT-3′ and reverse 5′-GGGAAAGACAAGGAGGGTGA-3′ for the KLF7 promoter (chr2: 207,167,132–207,167,464).

Dual-luciferase reporter assays

Using the KLF7 position weight matrix from the JASPAR database (https://jaspar.elixir.no/) to scan the THBS1 promoter, the log likelihood scoring algorithm predicted the highest-scoring potential KLF7 binding site: GGGGCGGGG (-strand). Its corresponding sense strand coordinates in the hg38 reference genome are chr15: 39,580,905-39,580,913 (+strand). This site was subsequently mutated to CTTTATTTT (THBS1 promoter^Mut). FaDu and Detroit 562 cells were seeded in 24-well plates at a density of 1 × 10⁵ cells/well. Using Lipofectamine 3000 transfection reagent (L3000001, Invitrogen), 500 ng of pGL3 Basic Vector (212936, Addgene, Watertown, MA, USA) containing the THBS1 promoter or the THBS1 promoter^Mut were co-transfected into each well with 50 ng of pRL-TK vector (VT1568, Youbio, Changsha, Hunan, China). Forty-eight hours after transfection, cells were lysed, and luciferase activity was determined using the Renilla-Firefly luciferase dual assay kit (16186, Thermo Fisher Scientific). The firefly luciferase activity was normalized to Renilla luciferase for each well.

Quantification and statistical analysis

The sample size for each experiment was determined to be sufficiently large to guarantee the validity of the statistical results. Statistical analyses were performed using GraphPad Prism 10.3.0 (GraphPad, San Diego, CA, USA). All experiments were repeated at least three times, independently, and the results were expressed as the mean ± standard deviation (SD). The differences among the two groups and multiple groups were assessed using t-tests and the one-way/two-way ANOVA with a post hoc Tukey test, respectively. p < 0.05 was considered statistically significant. For the categorical variables shown in Table 1, Fisher’s exact test was used to assess statistical significance, with significance denoted by ∗ (∗p < 0.05).

Published: January 27, 2026