Background
Head and neck cancers encompass various histologic types, with squamous cell carcinoma being the most common. Head and neck squamous cell carcinoma (HNSCC) is highly invasive and prone to metastasize cervical lymph nodes (CLN), leading to a higher risk of recurrence and metastasis, which is the primary cause of mortality [
1,
2]. Despite advancements in diagnostic and therapeutic methods, the 5-year survival rate for HNSCC patients has not significantly improved in the last 30 years and remains around 50% [
3,
4]. Therefore, it is crucial to comprehend the regulatory mechanisms governing invasive metastasis in HNSCC to develop innovative therapeutic strategies and improve the prognosis of patients.
Aberrant transcription of oncogenes driven by genetic and epigenetic alterations plays a critical role in the tumorigenicity and metastasis of cancers [
5,
6]. Recent evidence suggests that super-enhancers (SEs) are important non-coding regulatory elements that determine the identity of different cell types. SEs enhance the expression of genes that are crucial for maintaining cellular identity during both organism development and disease progression [
7‐
9]. SEs are DNA cis-regulatory elements that exhibit superior transcriptional activity, which are clusters composed of multiple neighboring enhancers. SEs are primarily composed of transcriptional activation-associated histone modifications (Histone 3 lysine 27 acetylation, H3K27Ac; Histone 3 lysine 4 monomethylation, H3K4me1), cofactors (bromodomain-containing protein 4, BRD4; mediator complex subunit 1, MED1), chromatin regulator (p300), RNA polymerase II (RNA Pol II), and cell type-specific transcription factors (TFs) by chromatin immunoprecipitation-sequencing (ChIP-seq) to define [
9,
10]. Several studies have utilized the Ranking of Super-Enhancers (ROSE) algorithm to identify SEs in different types of cancers. These studies have confirmed that SEs are responsible for regulating the overexpression of important oncogenes (such as CCAT1, TP63, MYC, SOX2, KLF5, and FOSL1) during the development and progression of cancers. This regulation helps maintain the characteristics of cancer cells and promotes their malignant progression [
11‐
17].
SEs are genomic regions that play crucial roles in controlling gene expression and cellular function. When TFs bind to SEs, BRD4 recognizes and attaches to acetylated histones, which in turn bind to the Mediator complex at the promoter region of the target gene. After transcription initiation, cyclin-dependent kinases 7 (CDK7)-containing TFIIH and CDK9-containing positive transcription elongation factor b (P-TEFb) are recruited by BRD4. These kinases phosphorylate the carboxy-terminal structural domain of RNA Pol II, facilitating the localization of RNA Pol II at the transcription start site and regulating gene transcription elongation [
18,
19]. By disrupting SEs or inhibiting the SE-associated transcriptional program, it is possible to selectively inhibit the transcriptional activity of cancer cells driven by SEs [
20‐
22]. Interestingly, Wu et al. discovered that there was less overlap of genes with altered expression by comparing the transcriptomic data of different types of cancer cells treated with JQ1, suggesting that, in different cellular environments inhibition of SE-associated transcription preferentially regulates vital cancer cell- or tissue-specific genes [
23]. Currently, therapeutic strategies that directly target SEs by disrupting the SEs or inhibiting the SE-associated transcriptional program have shown promising results in various cancers [
18,
20,
24].
In this study, we identified insulin-like growth factor 2 mRNA binding protein 2 (IGF2BP2) as a hub SE-associated gene in HNSCC. We observed a significant up-regulation of IGF2BP2 expression, which promotes tumorigenicity and metastasis in HNSCC. Mechanistically, we found that Krueppel-Like Factor 7 (KLF7) binds to the IGF2BP2-SE and drives its transcriptional activation and expression during the malignant progression of HNSCC. Finally, we demonstrated that targeting the SE-driven transcriptional program effectively reduced the tumorigenic potential and metastatic ability of HNSCC.
Materials and methods
Acquisition and analysis of ChIP-seq data
All H3K27Ac ChIP-seq data were obtained from the GEO database. The normal oropharyngeal mucosa H3K27Ac ChIP-seq data were obtained from the GSE112021 dataset [
25]. The H3K27Ac ChIP-seq data of CAL27, Detroit562, and HN12 cells were obtained from the GSE128275 dataset [
26]. The H3K27Ac ChIP-seq data of SCC25 cells were obtained from the GSE103554 dataset [
27]. SRA files of the above ChIP-seq data were downloaded from the GEO database. After that, they were transformed into FASTQ format and subjected to quality control. The alignment of sequences with the human reference genome GRCh38 was performed using BWA. With the aid of MACS2, enriched regions within the ChIP-seq data were identified. SEs were further identified using ROSE with a stitching distance of 12.5 kb and visualized using the IGV Genome Browser. The SEs were assigned to genes with transcription start sites flanking the 50 kb window of the SEs. BEDGRAPH files from the GSE211473 dataset for UPCI-SCC-090, UM-SCC-104, and FaDu cells H3K27Ac ChIP-seq were converted to BIGWIG format and visualized on the IGV Genome Browser with the human reference genome hg19. HNSCC cell-specific SE-associated genes were further analyzed and enriched for disease gene networks using the 'DOSE' R package.
Acquisition of high-throughput chromosome conformation capture (Hi-C) data
Hi-C data of UPCI-SCC-090, UM-SCC-104, and FaDu cells were obtained from the GSE211296 dataset. Subsequently, these data was visualized using the WashU Epigenome Browser (
https://epigenomegateway.wustl.edu/browser/), with alignment to the human reference genome hg19.
Acquisition and analysis of RNA-seq data
We obtained RNA-seq data and clinical data for TCGA-HNSCC from the UCSC Xena website (
https://xenabrowser.net). Additionally, we downloaded gene expression matrices and clinical data from three other HNSCC datasets, namely GSE30784, GSE42743, and GSE41613, from the GEO database [
28,
29]. To ensure uniformity, we converted the probe ID to gene symbols and extracted the mRNA transcriptome expression matrices. Subsequently, we integrated these matrices with the corresponding clinical data by utilizing R software. Our analysis required the transformation of the expression data of TCGA-HNSCC from count values to log2 (TPM + 1) values. To visually represent the data, we utilized the R packages 'ggpubr' and 'ggplot2' to generate boxplots. Additionally, we employed the 'survival' package for COX regression survival analysis and plotted the resulting survival curves using the 'survminer' package.
Weighted gene coexpression network analysis (WGCNA)
To construct the weighted gene coexpression network, genes with a standard deviation (SD) > 1 were selected using the 'WGCNA' package [
30]. Next, all samples were clustered to identify any missing values or outliers, and outlier samples were removed. The optimal soft threshold was then determined to ensure that the gene network followed a scale-free distribution. By converting the expression matrix into an adjacency matrix and further into a topological overlap matrix, we facilitated the hierarchical clustering of genes based on dissimilarity, leading to the formation of a clustering tree. The DynamicTreeCut algorithm was applied to partition the clustering tree into different modules, with the merging of modules that had a dissimilarity of less than 0.2. Lastly, we computed the correlation coefficient and
P-value between each gene module and HNSCC. The gene module displaying the highest correlation coefficient and the smallest
P-value was identified as the most pertinent gene module.
Gene set enrichment analysis (GSEA)
Three HNSCC transcriptomes, TCGA-HNSCC, GSE30784, and GSE42743, were divided into high- and low-IGF2BP2 expression groups based on the median expression of IGF2BP2, respectively. We performed GSEA using the 'clusterProfiler' R package to compare the high- and low-IGF2BP2 expression groups. We assessed the enrichment of biological functions and pathways associated with IGF2BP2 expression using Hallmark and Kyoto Encyclopedia of Genomes (KEGG) gene sets.
HNSCC specimens
Two sets of patients were used in this study: the HNSCC tissue microarray and the Sun Yat-sen University (SUSY) HNSCC patient cohort. The HNSCC tissue microarrays were acquired from two different sources, Shanghai Xinchao Biotechnology Co. (HOraC060PG01) and US Biomax (OR601c). The SUSY HNSCC patient group consisted of 109 HNSCC tissue specimens and 33 surgically removed adjuvant non-cancerous tissues (ANCTs) from the Department of Oral and Maxillofacial Surgery at the Hospital of Stomatology Sun Yat-sen University. Clinical and pathological information of the patients was gathered, and no preoperative treatment was given to any of the patients. Informed consent was obtained and the research was approved by the Medical Ethics Committee of the Hospital of Stomatology Sun Yat-sen University, following the guidelines set by the Declaration of Helsinki. The time from surgery until death from any cause or the last follow-up was considered as overall survival (OS), while the time from surgery until tumor recurrence (either local or distant) or the last follow-up was considered as disease-free survival (DFS).
Immunohistochemistry (IHC)
All paraffin-embedded tissues were cut into 4.0-µm sections and underwent dewaxing and dehydration. Sodium citrate buffer was used for antigenic repair, after incubating the sections with 3% H2O2. Following that, the sections were blocked using anti-goat serum (#AR0009, BosterBio) and incubated overnight at 4◦C with primary antibodies against anti-IGF2BP2 (1:250, #11,601–1-AP, Proteintech, China), anti-KLF7 (1:300, #PA5-81,206, Thermo Fisher), anti-Pan-Keratin (1:1000, 4545S, Cell Signaling), or anti-BRD4 (1:200, #ab128874, Abcam). Subsequently, the sections were washed, incubated with a secondary antibody, and stained using the DAB detection Kit (GK600510, Gene Tech) with diaminobenzidine (DAB). Finally, the sections were re-stained with hematoxylin (D006, Nanjing Jiancheng Biotech). The IHC scores were calculated on a continuous scale of 0–300. This was achieved by multiplying the proportion of positive cells (ranging from 0 to 100%) by the intensity of staining, which was classified as 0 (no staining), 1 (weak), 2 (moderate), or 3 (strong). HNSCC tissues with IHC scores higher than 150 were classified as the high-expression group, while those with scores less than or equal to 150 were classified as the low-expression group.
Cell culture and treatment
The present study utilized human HNSCC cell lines SCC25 and CAL27, along with human embryonic kidney-derived 293 T cells, which were procured from the American Type Culture Collection (ATCC). The SCC25 cells were cultivated in DMEM/F12, supplemented with 10% fetal bovine serum (FBS, #086–1500, WISENT) and 400 ng/mL hydrocortisone. On the other hand, CAL27 and 293 T cells were maintained in DMEM supplemented with 10% FBS, and all cells were incubated under constant conditions at 37 °C with 5% CO2. The small molecule inhibitors THZ1 (#V2557) and JQ1 (#V0411) were obtained from InvivoChem, whereas OTX-015 (#202,590–98-5) and CPI-637 (#1,884,712–47-3) were purchased from MedChemExpress. Upon reaching an approximate fusion rate of 60–70% overnight, the culture medium was replaced with varying concentrations of small molecule inhibitors for continuous cell culture and subsequent investigations.
Cell transfection
To knock down IGF2BP2, we used short hairpin RNA (shRNA) targeting the IGF2BP2 gene, which was cloned into the pLKO.1 plasmid (Addgene). For overexpressing IGF2BP2 and KLF7, we cloned the full-length open reading frames (ORFs) of the human-derived IGF2BP2 and KLF7 into the pCDH-CMV-MCS-EF1-copGFP-T2A-Neo plasmid (System Biosciences, SBI) and pCDH-CMV- MCS-EF1-copGFP-T2A-Puro plasmid (System Biosciences, SBI), respectively. The KLF7 overexpression plasmid was tagged with the HA protein. To deplete IGF2BP2-SE, we designed CRISPR/Cas9 constructs with small guide RNA (sgRNA) targeting the IGF2BP2-SE region (E1, E2, and E3), which were cloned into the pU6-gRNA-Cas9-puro plasmid (Addgene). Lentiviruses were produced by transfecting the target plasmid with the packaging plasmids psPAX2 and pMD2.G (Addgene) into 293 T cells. SCC25 cells or CAL27 cells were incubated with the viral particles for 24 hours. After 48 hours of viral infection, cells were screened using puromycin (SCC25 cell: 1 mg/mL, CAL27 cell: 0.5 mg/mL) or G418 (SCC25 cell: 0.4 mg/mL, CAL27 cell: 0.2 mg/mL) medium for 1 week, and then the concentration was halved to continue screening the cells for 1 week. For siRNA transfection, HNSCC cells were grown overnight until reaching a fusion rate of approximately 50%. The appropriate negative control (si-NC) or small interfering RNA (siRNA) was added using the Pepmute Transfection Reagent (#SL100566, Signagen) according to the manufacturer's instructions. The sequences of the shRNA, sgRNA, and siRNA oligonucleotides are listed in Additional file
1: Table S1.
Western blot
Cells were lysed with a mixture of RIPA buffer (#CW2333S, CWbio) and a protein inhibitor cocktail on ice for 15 minutes. After centrifugation at 12,500 rpm for 20 minutes at 4 °C, the supernatants were collected and assayed for protein concentration using the BCA Protein Assay Kit (#CW0014S, CWbio). Equal quantities of proteins were then separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE; #PG112, Epizyme Biotech) and transferred to polyvinylidene fluoride (PVDF) membranes (#ISEQ00010, Millipore). The membranes were subsequently blocked with 5% bovine serum albumin at room temperature for 1 hour. Next, the primary antibodies were incubated at 4 °C overnight. The primary antibodies used in this study were anti-IGF2BP2 (1:1000, #14672S, Cell Signaling Technology), anti-β-ACTIN (1:1000, #19069S, Cell Signaling), anti-BRD4 (1:1000, #ab243862, Abcam), anti-MED1 (1:1000, #ab60950, Abcam), and anti-KLF7 (1:500, #SC398576, Santa Cruz Biotechnology). Following a 1-hour incubation period with the secondary antibodies, chemiluminescent images of the membranes were recorded accordingly.
Quantitative real-time PCR (qRT-PCR)
According to the manufacturer's instructions, total RNA was isolated from the cells using RNAzol® RT (#RN190, Molecular Research Center). The Hifair III 1st Strand cDNA Synthesis kit (#11141ES60, Yeasen) was used to reverse-transcribe 1 μg of total RNA into complementary DNA. The qPCR reaction was performed using SYBR Green Master Mix (#11201ES08, Yeasen) on a LightCycler 480 II (Roche). Gene expression was calculated by 2
−ΔΔCt normalized to β-ACTIN. The primer sequences used for qRT-PCR are listed in Additional file
1: Table S2.
Cell proliferation and colony formation assay
SCC25 and CAL27 cells were inoculated into 96-well plates overnight at a density of 2000 cells per well. Cell proliferation was measured at specific time points using the Cell Counting Kit-8 (CCK-8, #2003ES80, Yeasen). Following the provided, 100 µLof serum-free medium with 10% CCK-8 reagent was added to each well and the plates were incubated at 37 °C for 1 hour. The absorbance values were measured at 450 nm using a microplate reader (Biotek), and growth curves were plotted based on absorbance and time. For the cell colony formation assay, 500 SCC25 cells or 10,000 CAL27 cells were inoculated in 6-well plates and cultured for 7–10 days. Afterward, the cells were fixed using 4% paraformaldehyde and stained with 0.4% crystal violet. Direct counting was performed to determine the number of colonies in the SCC25 cells. The colony number for CAL27 cells involved the examination of three random fields under a microscope, magnified at 5 × .
Cell migration and invasion assays
Serum-free medium was used to prepare SCC25 cell suspension at a density of 7.5 × 105/mL and CAL27 cell suspension at a density of 6.0 × 105/mL for conducting in vitro cell migration and invasion. After thorough mixing, 200µL of cell suspension was added to the upper chamber of the Transwell (for cell migration assay, 8 µm pore size, Corning) and Transwell coated with 0.33 mg/mL Matrigel (for cell invasion assay, 10 mg/mL, #354,234, Corning), while 800 µL of complete culture medium was added to the lower chamber. After incubation for 48 hours, 4% paraformaldehyde fixed the cells that passed through the cell chamber filter attached to the surface and stained with 0.4% crystal violet. To ensure comprehensive analysis, five independent fields of view were randomly selected for photographing and recording. The total number of cells in each field of view was then counted.
Approval for animal experiments in this study was granted by the Institution Animal Care and Use Committee at Sun Yat-sen University. Female BALB/c-Nude mice, aged 4–6 weeks, were acquired from the Animal Experiment Center of the East Campus of Sun Yat-sen University and housed under pathogen-free conditions. To assess the influence of IGF2BP2 on tumor growth and metastasis in living organisms, the BALB/c-Nude mice were divided into three randomized groups: the control group, the shIGF2BP2 group, and the IGF2BP2 overexpression group, with 5 mice in each group. An orthotopic xenograft model was established by injecting 50μL suspension containing 1.0 × 106 CAL27 cells into the tongue of anesthetized nude mice using a 27G insulin needle. Following cell inoculation, the nude mice were positioned in lateral recumbency with their tongues pulled out to prevent asphyxiation, and were exposed to a warming lamp to maintain body temperature. Mice with a weight loss exceeding 15% of their baseline were euthanized. After two weeks, the mice were anesthetized and euthanized by cervical dislocation. To investigate the effects of JQ1 treatment on tumor growth and metastasis in living organisms, nude mice were randomly assigned to two groups, each containing 6 mice, after intraperitoneal injection of 1.0 × 106 CAL27 cells into their tongues to establish an orthotopic xenograft model for one week. The treatment group received a daily intraperitoneal injection of JQ1 at a dosage of 50 mg/kg, while the control group received only the vehicle for 14 days. After the experiment, the mice were euthanized by cervical dislocation following anesthesia. Tumor size was measured with vernier calipers, and tumor volume was determined using the formula: tumor volume = length × width × width/2. The tongues and CLNs of the nude mice were then collected, fixed, dehydrated, embedded, sectioned, and subjected to hematoxylin–eosin (H&E) staining and IHC staining.
ChIP-qPCR
ChIP analysis was performed using the SimpleChIP Plus Enzymatic Chromatin IP Kit (Magnetic Beads) (#38191S, Cell Signaling) following the manufacturer's protocol. An approximate amount of 4 × 10
6 cells was included in each ChIP reaction mixture. The cells underwent crosslinking with 1% formaldehyde at room temperature for 10 minutes, followed by glycine neutralization for 5 min. Subsequent steps involved resuspending the cells, digesting the chromatin with Micrococcal Nuclease, and sonication to obtain suitable DNA fragments. Dilution of the chromatin complexes occurred in ChIP dilution buffer, and immunoprecipitation was performed using these antibodies: anti-BRD4 (#ab243862, Abcam), anti-MED1 (#ab60950, Abcam), anti-H3K27Ac (#ab4729, Abcam), anti-normal rabbit IgG (#2729S, Cell Signaling), and anti-HA Tag (#66,006–2-lg, Proteintech). Elution of the chromatin from the antibody/protein G magnetic beads utilized ChIP elution buffer and was later transferred to a centrifugation column for DNA purification. Measurement of the immunoprecipitated DNA samples was done through qPCR. The resulting data were presented as a percentage of input DNA. For ChIP-qPCR primer sequences, please refer to Additional file
1: Table S2.
Dual-luciferase reporter assay
To validate the presence of KLF7 binding sites on the IGF2BP2-SE, we performed the cloning of a 322 bp segment that encompasses the wild-type (WT) KLF7 motif-binding sequence (obtained from chr3: 185,824,281–185,824,602) into the BglII site of the pmirGLO plasmid. This cloning process aimed to generate the luciferase reporter gene, designated as PmirGLO-WT. To create a mutated version of the KLF7 binding sites, the original sequence (GGGGCGGGG) was altered to AATAATTAT using the seamless cloning site-directed mutagenesis technique. The resultant plasmid was termed PmirGLO-mut. Following the plasmid construction, SCC25 and CAL27 cells were cultured in 12-well plates until they reached a fusion rate of approximately 60%. The cells were then transfected with pmirGLO, pmirGLO-WT, and pmirGLO-mut using the LipofectamineTM 3000 Transfection Kit (#L3000015, Thermo Fisher). After 48 hours of transfection, we measured the cellular luciferase activity using the Dual-Luciferase Reporter Gene Assay Kit (#11402ES60, Yeasen), following the manufacturer's instructions. To assess the translational efficiency of the reporter gene, the firefly luciferase activity was normalized with Renilla luciferase.
Statistical analysis
Statistical analysis was carried out using GraphPad Prism 9.0 software (GraphPad Software, San Diego, CA, USA), and the data were presented as mean ± standard deviation (SD). To determine the normal distribution of the data, the Shapiro–Wilk test was performed. For comparing two groups, the student's t-test was employed, whereas for comparing three or more groups, ANOVA was used. The Log-Rank test and Kaplan–Meier method were utilized for survival analysis of clinical specimens. Additionally, Fisher's exact test was conducted to assess the differences in CLN metastasis in animal experiments. To examine the relationships between genes, Spearman's correlation was applied. A significance level of P < 0.05 was considered for all statistical analyses: *P < 0.05, **P < 0.01, ***P < 0.001.
Discussion
A thorough understanding of the intricate molecular mechanisms responsible for the invasive metastasis of HNSCC, a highly aggressive cancer prone to spreading, remains an unresolved challenge. Fundamental research has shown that oncogenes in several human cancers acquire cis-regulatory SEs throughout their malignant progression, thereby influencing their transcriptional expression [
36‐
38]. Thus, it becomes paramount to investigate how SEs regulate crucial oncogenes, as it may yield new opportunities for therapeutic interventions and enhance the prognosis of HNSCC patients. Our investigation unraveled the association between HNSCC cell-specific SEs and multiple oncogenes, suggesting their acquisition during the development of HNSCC tumors. Furthermore, we unequivocally demonstrated that IGF2BP2, an essential protein-encoding gene associated with SEs, enhances the proliferation, invasion, and metastasis of HNSCC cells in both in vitro and in vivo experiments. The enrichment of BRD4, MED1, and H3K27Ac modification on the IGF2BP2-SE was found to synergistically activate its transcriptional program. Moreover, our findings indicate that the transcription factor KLF7 directly binds to the SE as well as the promoter regions of IGF2BP2, thereby enhancing its transcriptional activity and driving the expression of IGF2BP2 (Fig.
7K).
SEs are known to be specific to certain cells and tissues, and they play a crucial role in the overexpression of key oncogenes in cancer cells. This, in turn, contributes to the initiation and maintenance of tumor properties [
39,
40]. For instance, in esophageal adenocarcinoma, SEs that are specific to this type of cancer regulate the overexpression of pro-carcinogenic transcription factors like ELF3, KLF5, GATA6, and EHF. These transcription factors activate the STAT3 and PI3K/AKT signaling pathways, thereby promoting cell proliferation [
41]. Another study conducted on normal liver cells and hepatocellular carcinoma tissues revealed that cancer-specific SEs drive the overexpression of the oncogene SPHK1, leading to the proliferation and metastasis of hepatocellular carcinoma cells [
36]. These findings emphasize the role of SEs in promoting the expression of oncogenes, activating pro-oncogenic signaling pathways, maintaining the identity of cancer cells, and facilitating the progression of malignancy. In a previous study, we also demonstrated the involvement of SE-associated FOSL1 in the tumorigenicity and metastasis of HNSCC by impacting stemness and EMT [
17]. In our current research, we have identified IGF2BP2 as a gene associated with SEs specific to HNSCC cells, based on the analysis of H3K27Ac ChIP-seq and transcriptome data.
IGF2BP2 is situated on chromosome 3q27.2. It is a newly discovered N6-methyladenosine (m
6A)-reading protein that is upregulated in various cancer types. Its association with cancer progression and its negative impact on cancer prognosis are well-established. The m
6A-dependent manner in which IGF2BP2 operates enhances mRNA stabilization and translation. Furthermore, it plays a crucial role in regulating cell proliferation, metabolism, EMT, migration, and invasion [
42‐
47]. Particularly in the context of HNSCC, IGF2BP2 acts as a facilitator for cell migration, invasion, and EMT by increasing the stability of Slug mRNA in an m
6A-dependent manner [
48]. Our investigation, combining data from public databases and analysis of clinical samples, has demonstrated a significant increase in IGF2BP2 expression in HNSCC. This increase is closely associated with metastasis and poor prognosis in patients. GSEA analysis suggests that IGF2BP2 may also have a regulatory role in EMT and cell cycle signaling. These processes are crucial for maintaining the proliferative and invasive properties of HNSCC cells. In vitro experiments have supported the role of IGF2BP2 in promoting the proliferation, migration, and invasion of HNSCC cells. Moreover, an orthotopic xenograft model has further validated its contribution to HNSCC tumorigenesis and CLN metastasis. In conclusion, these findings shed light on the invasion and metastasis of HNSCC by identifying IGF2BP2 as an SE-associated gene.
Epigenetic characteristics linked to the assembly and functioning of SE involve increased levels of H3K27Ac modification, recruitment of BRD4, Mediator complex, RNA Pol II, CDK7-containing TFIIH, and CDK9-containing P-TEFb, as well as binding of CBP/p300 acetyltransferase [
7]. BRD4, upon binding to the acetylated chromatin, summons the Mediator complex, RNA Pol II, TFIIH, and P-TEFb to support the initiation and elongation of transcription, ultimately leading to the activation of target genes [
49,
50]. CDK7, a member of the CDK family, regulates RNA Pol II phosphorylation and governs transcription initiation, pausing, and elongation. It serves as a crucial element of the transcription complex, with a preference for binding to SEs, driving the expression of SE-associated genes [
51]. Histone acetyltransferases CBP/p300 operate as transcriptional coactivators, inducing an increase in H3K27Ac levels in promoters, enhancers, and SEs of target genes. This instigates the assemblage of diverse transcription components, thereby initiating gene transcription [
52]. In comparison to enhancers, SEs have heightened vulnerability to transcription-associated inhibitors. The interaction between the SE region and its corresponding transcription components can be specifically impeded through CRISPR/Cas9 interference or the usage of small molecule inhibitors that target SEs. This selective inhibition leads to the suppression of transcription and expression of SE-associated genes [
36,
53]. In our investigation, we divided the highly enriched H3K27Ac SE region of IGF2BP2 into three segments (E1, E2, and E3). By implementing CRISPR/Cas9 technology, we targeted and edited E1, E2, and E3 to disrupt the interactions between the SE and the promoter. Our findings demonstrate that inhibiting these three independent segments results in the suppression of transcriptional expression of IGF2BP2. BRD4 and MED1 co-localize on acetylated chromatin, particularly H3K27, which consequently affects the transcriptional activity controlled by SEs [
9]. In hepatocellular carcinoma, silencing BRD4 or MED1 repressed the transcription of SE-associated genes like SPHK1, E2F2, CCND1, MYCN, and MYC [
36]. Similarly, silencing MED1 hindered the transcription of TP63, MET, BIRC, and MMP3, which are SE-associated genes in HNSCC cells. Additionally, silencing BRD4 or MED1 significantly decreased the enrichment of BRD4 or MED1 in TP63-SE and MET-SE [
54]. Interestingly, the mRNA and protein levels of IGF2BP2 were also suppressed upon silencing BRD4 or MED1. ChIP-qPCR assays confirmed that silencing BRD4 or MED1 led to a decrease in the enrichment of BRD4 or MED1 on IGF2BP2-SE in HNSCC cells, respectively, resulting in the inhibition of IGF2BP2 transcriptional expression. Furthermore, treatment of HNSCC cells with small molecule inhibitors THZ1, JQ1, OTX-015, and CPI-637 specifically reduced the mRNA and protein levels of IGF2BP2 by inhibiting the SE-associated transcriptional program. Subsequent ChIP-qPCR analysis also verified that JQ1 or CPI-637 treatment significantly reduced the enrichment of BRD4 or H3K27Ac at IGF2BP2-SE in HNSCC cells, respectively. These findings support the notion that IGF2BP2-SE exerts a positive regulatory role in maintaining the functional characteristics of HNSCC by governing the transcriptional expression of IGF2BP2.
TFs are proteins that regulate gene transcription by forming transcriptional complexes with RNA Pol II and binding to specific DNA sequences of target genes. Depending on specific spatial and temporal conditions, TFs can either activate or repress gene transcription [
55]. TFs bind to oncogenic SEs associated with specific signaling pathways and enhance the transcriptional activity of oncogenes, thus promoting tumor development and progression [
56]. In our correlation analysis, using the JASPAR database prediction, as well as TCGA and GEO data, we found that KLF7 exhibited the strongest correlation with the expression of IGF2BP2. Furthermore, we have demonstrated that KLF7 promotes the transcriptional expression of IGF2BP2 by binding to its SE and promoter regions. KLF7 belongs to the Krüppel family of transcriptional regulators and is located on chromosome 2q33.3. It functions as a transcriptional activator [
57]. Notably, it is highly expressed in high-grade plasma ovarian cancer, pancreatic ductal adenocarcinoma, hepatocellular carcinoma, and breast cancer. Additionally, its expression is associated with clinical stage, pathological grade, and metastasis in these cancers [
58‐
61]. Similarly, our analysis of comprehensive public databases and clinical samples revealed that KLF7 is highly expressed in HNSCC and strongly correlates with malignant progression and poor prognosis in patients. Several cohorts also identified a significant correlation between KLF7 and IGF2BP2 mRNA and protein expression in HNSCC, highlighting a robust and compelling relationship. Finally, our findings suggest that high expression levels of KLF7 and IGF2BP2 are associated with shorter OS and DFS, underscoring the critical role of the KLF7/IGF2BP2 axis in promoting malignant progression in HNSCC.
Studies have provided evidence that blocking SE-associated transcriptional programs can disrupt oncogene transcription and hinder tumor growth, presenting a unique approach to treating cancer [
62‐
65]. Notably, the SE-associated transcriptional regulatory process involves various components, such as BRD4, CDK7, CDK9, and CBP/p300, which are potential targets for small molecule inhibitors. Numerous small molecule inhibitors targeting the SE-associated transcriptional program have been assessed in preclinical models and clinical trials, demonstrating promising activity against different types of advanced cancers [
20,
66,
67]. In line with these discoveries, our investigation shows that JQ1 treatment effectively suppresses the proliferation and invasive metastatic ability of HNSCC cells in both in vitro and in vivo experiments. While BRD4, CDK7, CDK9, and CBP/p300 are commonly found binding proteins on transcriptional regulatory elements, they are highly enriched in oncogenic SEs. Therefore, their inhibitory effects primarily impact the transcription of SE-associated oncogenes, which ultimately leads to the attenuation of the malignant characteristics in cancer cells. In the future, further exploration into the specific elements and underlying mechanisms of SE-associated oncogenes will aid in the identification of novel therapeutic targets and the development of more precise and potent cancer therapies.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.