Effects of BRD4 inhibitor JQ1 on the expression profile of super-enhancer related lncRNAs and mRNAs in cervical cancer HeLa cells

Introduction

The incidence of cervical cancer (CC) ranks second in women’s malignant tumors worldwide (Siegel, Miller & Jemal, 2018). However, according to Chinese tumor data, the incidence of CC ranks first among gynecological tumors in China, and its control and prevention remain challenging (Chen et al., 2016). In 2013, Whyte et al. (2013) first introduced the concept of super-enhancers (SEs), which has since become a prominent research hotspot in recent years. SEs are a large cluster of transcriptional activity enhancers that enrich high-density master transcription factors, cofactors and enhancer histone modification marks. Functionally, SEs can drive and control the expression of cell identity genes, offering a possible explanation for cell type-specific expression patterns and showing great potential for application in the research of developmental biology, cancer and other disease pathogenesis (Creyghton et al., 2010; Hnisz et al., 2013; Lovén et al., 2013; Whyte et al., 2013). Although the study of SEs in the context of tumors is still in its infancy (Sengupta & George, 2017), the mechanisms through which they regulate malignant behavior are drawing increasing interest from researchers, despite the ongoing challenges in SE identification (Huang et al., 2021).

Previous studies have shown that most drugs play an antitumor role by affecting the gene expression profile of malignant tumors, of which long non-coding RNAs (lncRNAs) were shown to play a pivotal role (Zhu et al., 2018). LncRNAs are non-coding RNAs with a length of more than 200 nucleotides and lack protein-coding function. They play an important role in various physiological and pathological processes of organisms, with their abnormal expression closely related to the occurrence and development of tumors. In 2017, Soibam (2017) identified a set of lncRNAs related to SEs for the first time and named them SE-related lncRNAs (SE-lncRNAs), thus introducing a new field of research on SEs and SE-lncRNAs. SE-lncRNAs are lncRNAs that form a triple helix of RNA: DNA: DNA with the SE region, which can recruit regulatory factors to the super-enhancer region, thus affecting the chromosome structure and acting as a space amplifier to promote tissue-specific gene expressions related to the SE (Anderson et al., 2016; Soibam, 2017). Given that most SE-lncRNAs are transcribed from the SE region, they often serve as the function executors of SEs, playing a critical role in transcriptional regulation (Mousavi et al., 2013; Xiang et al., 2014).

In recent years, there has been a growing research interest in investigating the impact of drugs on the expression profile of lncRNAs in malignant tumors. Bromodomain and extraterminal (BET) proteins are a new class of proteins that can interpret epigenetic codes and play an important role in regulating gene transcription through identifying and binding acetylated histones or non-histones proteins. Since acetylated chromatin can promote transcriptional activation and promote tumorigenesis, inhibiting the BET protein family may inhibit tumor progression and become a potential new drug target for tumor therapy (Barbieri, Cannizzaro & Dawson, 2013; Noguchi-Yachide, 2016). JQ1 is a BET bromodomain inhibitor, which mainly inhibit bromine domain protein 4 (BRD4) as well as BRD2/3 and BRDT (bromodomain testis-specific protein) (Filippakopoulos et al., 2010). It can specifically compete with BRD4 and bind to all bromine domains of the BET family but not to bromine domains outside the BET family, thus releasing BRD4 from chromatin and inhibiting the transcriptional regulation of BRD4 (Filippakopoulos et al., 2010). At present, it is believed that JQ1 can induce cell apoptosis, inhibit invasion ability and then inhibit the growth of tumor cells, but the specific mechanism is unknown (Liu et al., 2019). In vivo and in vitro studies have shown that JQ1 has inhibitory effects on various tumor cells, especially hematopoietic tumor cells (Baldan et al., 2019). However, the potential antitumor effects of JQ1 on cervical cancer cells and its impact on the expression profile of SE lncRNAs remain unknown. To explore the anti-tumor mechanism of JQ1 on cervical cancer, this study used high-throughput sequencing technology to investigate the transcriptome changes of cervical cancer cells treated with JQ1 to reveal the signaling pathway and biological function of JQ1 on cervical cancer cells and to provide transcriptome evidence for the application of JQ1 in the potential treatment of cervical cancer.

Materials and Methods

Construction of protein-protein interaction (PPI) network and screening of key DEGs mRNA

The STRING database (http://string-db.org/) was used for network construction, and the Cytoscape software was used for subsequent analysis. Gene sets containing more than one node were selected using the Cytoscape plug-in MCODE, and each module was ranked according to the results of MCODE analysis, with the top one being the key module. The key mods were selected and enriched for the GO function and KEGG pathways using the R package “clusterProfiler”.

The main flow chart of study design is shown in Fig. 1.

Results

JQ1 effectively inhibited the proliferation of HeLa cells in the CCK-8 assay

We first explored the optimal concentration of JQ1 inhibiting HeLa cells through the CCK8 experiment. The growth morphology of HeLa cells treated with different concentrations of JQ1 are shown in Fig. S1. JQ1 showed a significant proliferation inhibitory effect on HeLa cells in a dose- and time-dependent manner, as shown in Table 1. We also observed that the cells viability of JQ1 on HeLa cells decreased significantly with increasing concentration at the same treatment duration (P < 0.05), and the cells viability decreased with increasing treatment time under the same concentration (P < 0.05). Additionally, under the experimental conditions of three concentrations of 0.01, 0.1 and 1 μmol/L of JQ1 on HeLa cells at 24, 48, 72 and 120 h, we found that 1 μmol/L of JQ1 at 72 and 120 h had the same cell viability and the strongest cell proliferation inhibition. According to our CCK-8 assay and previous research reports (Ni et al., 2021), JQ1 treatment under a concentration of 1 μmol/L for 72 h was used in the experimental group. The growth morphology under DMSO and 1 μmol/L for 72 h is shown in Fig. 2.

Table 1:

The effects of JQ1 on the proliferation of cervical cancer HeLa cells at different concentrations and at different times.

Groups	24 h	48 h	72 h	120 h
DMSO	100.02 ± 2.58	99.97 ± 4.35	100.02 ± 1.75	100.00 ± 1.69
0.01 μmol/L	94.77 ± 1.88	87.65 ± 1.62	87.8 ± 2.3	79.80 ± 1.99
0.1 μmol/L	76.23 ± 2.13	79.02 ± 3.12	79.42 ± 4.27	39.85 ± 1.04
1 μmol/L	70.18 ± 1.69	70.18 ± 2.15	49.78 ± 1.47	39.73 ± 1.54
P	<0.001	<0.001	<0.001	<0.001

DOI: 10.7717/peerj.17035/table-1

15 differentially expressed SE-lncRNAs were screened

Gene annotation was performed on the experimentally obtained expression matrix, and a total of 3,111 lncRNAs were obtained via six samples with a sum of expression over one as the gene screening condition, of which 88 up-regulated genes and 105 down-regulated genes were detected. Subsequently, SE-lncRNAs were screened and matched according to Soibam’s report (Soibam, 2017), and a total of 162 SE-lncRNAs were identified, as shown in Table S2. Using FDR = 0.05 and LogFC = 1 as the difference condition, DEA with limma package revealed 15 SE-lncRNAs which were differentially expressed, of which eight SE-lncRNAs was down-regulated and 7 SE-lncRNAs was up-regulated, as shown in Fig. 3, Table 2.

Table 2:

Differentially expressed SE-lncRNAs detected by matched analysis.

Genes	LogFC	AveExpr	t	P-value	FDR	Diffsig
SNHG14	−2.091	7.901	−7.895	<0.001	0.006	Down
LINC01106	−1.386	9.314	−7.376	<0.001	0.008	Down
LINC02615	−2.079	6.818	−6.98	<0.001	0.01	Down
TMLHE-AS1	−1.892	6.651	−5.526	<0.001	0.019	Down
HLA-F-AS1	−1.553	7.5	−5.082	0.001	0.026	Down
PAX8-AS1	−2.433	6.148	−4.83	0.002	0.029	Down
ITGB2-AS1	−5.411	3.227	−4.913	0.001	0.028	Down
CARMN	−3.529	2.314	−4.97	0.001	0.028	Down
NDUFA6-DT	1.934	7.001	6.236	<0.001	0.013	Up
TM4SF19-AS1	1.317	7.468	5.928	<0.001	0.015	Up
TTLL1-AS1	4.472	2.777	7.335	<0.001	0.008	Up
DNAJC3-DT	1.086	8.633	5.309	<0.001	0.022	Up
BTG1-DT	1.276	7.144	5.148	0.001	0.025	Up
ADGRB3-DT	3.953	2.517	5.979	<0.001	0.015	Up
LINC02021	3.785	2.997	5.708	<0.001	0.017	Up

DOI: 10.7717/peerj.17035/table-2

A total of 418 mRNAs related to SE-lncRNAs were screened using correlation analysis

A total of 13,439 mRNAs were annotated with the screening condition of summed expression over 1 from the six samples. 1,229 differentially expressed genes were identified with FDR = 0.05 and LogFC = 1, including 488 up-regulated genes and 741 down-regulated genes. Further, upon restricting the conditions to FDR = 0.05 and LogFC = 2, a total of 424 differentially expressed mRNAs were identified, among which 139 genes were up-regulated and 285 genes were down-regulated, as shown in Table S3 and Fig. 4. Subsequently, correlation analysis was performed on the matrix of differentially expressed SE-lncRNAs and the matrix of differentially expressed mRNAs, and a total of 418 mRNAs were screened using the criteria P ≤ 0.05, of which 395 mRNAs had positive correlation with 12 SE-lncRNAs and 408 mRNAs had negative correlation with 15 SE-lncRNAs, which are shown in Table S4. Because of space limitations, the heatmap of the correlation coefficients of the top 50 mRNAs with 15 SE-lncRNAs is shown in Fig. 5.

GO functional annotation and KEGG pathway enrichment analysis of differentially expressed mRNAs

Here, 418 mRNAs were analyzed for GO functional annotation and KEGG pathway enrichment using the R package “clusterProfiler”, and the results of the top 15 pathways for each enrichment analysis are shown in Fig. 6 and Table S5. The main five enrichment pathways for GO biological processes (BP) were cell-cell adhesion via plasma-membrane adhesion molecules, extracellular matrix organization, extracellular structure organization, external encapsulating structure organization, and homophilic cell adhesion via plasma membrane adhesion molecules. The main five enrichment pathways of GO molecular function (MF) were collagen-containing extracellular matrix, cell-cell junction, basement membrane, basolateral plasma membrane, and basal part of cells. In addition, the main five enrichment pathways of GO cellular component (CC) were cell adhesion mediator activity, cell-cell adhesion mediator activity, extracellular matrix structural constituent, G protein-coupled neurotransmitter receptor activity, and insulin receptor binding. Further, the main five enrichment pathways of KEGG were bladder cancer, bile secretion, cGMP-PKG signaling pathway, cAMP signaling pathway, and AGE-RAGE signaling pathway in diabetic complications.

PPI protein network construction and key module screening of differentially expressed mRNAs

The PPI network of DEGs was constructed using the STRING database, and the corresponding node data were imported into Cytoscape software for additional analysis, and the corresponding protein network diagram is shown in Fig. 7. The core modules in the PPI network were identified using the MCODE plug-in, and the first three modules are shown in Fig. 8. The core module 1 contained 10 genes, namely TNFSF10, OASL, GBP1, IRF7, TRIM22, IFI44, IFNB1, IFITM2, IFITM1, and IFI44L. All genes in core module 1 were imported to perform GO and KEGG analyses. The results showed that genes contained in core module 1 influenced BP processes such as response to viruses, defense response to viruses, defense response to the symbiont, regulation of viral life cycle, regulation of viral process, etc.; influenced CC processes such as lysosomal membrane, lytic vacuole membrane, vacuolar membrane, Cajal body, late endosome membrane, etc.; influenced MF processes such as cytokine activity, cytokine receptor binding, type I interferon receptor binding, spectrin binding, GTP binding, etc.; and influenced KEGG processes such as influenza A, the NOD-like receptor signaling pathway, lipid and atherosclerosis, the Cytosolic DNA-sensing pathway, the RIG-I-like receptor signaling pathway, etc.

Discussion

Transcriptome sequencing is an important bioinformatics method widely used in cancer screening (Koeppel et al., 2018). In this research, transcriptome sequencing was used to investigate the effect of JQ1 on the expression profile of cervical cancer cell lines. First, we detected the inhibitory effects of JQ1 on HeLa cells via CCK8 assay. In line with our expected scenario and the finding in some previous studies, JQ1 demonstrated a significant inhibitory effect on the proliferation of HeLa cells, which was both time- and concentration-dependent (Wang et al., 2016; Zanellato, Colangelo & Osella, 2018). In our exploratory experiments, we found that the inhibitory effects of JQ1 treatment at a concentration of 1 μmol/L for 72 h were close to (50.22 ± 1.47)%, which was thereby set as the experimental dose. To the best of our knowledge, only a few studies had previously explored the anticancer mechanism of JQ1 in cervical cancer (Rataj et al., 2019; Wang et al., 2021). More importantly, the clinical application of JQ1 was limited due to its short plasma half-life, despite those in vitro studies of various solid tumors reported JQ1 as having a strong anticancer effect (Bagratuni et al., 2020; Carra et al., 2020; Li et al., 2018). JQ1 is a small molecule inhibitor of BRD4 that highly selectively binds to the BRD4 structural domain to exert competitive inhibition. It is currently being studied extensively in various solid tumors (Shi et al., 2018b). Although the application of JQ1 as a single drug for tumor treatment is still being investigated in experimental research, it showed significant inhibitory effects on various malignant tumor cells with different genetic backgrounds (Zhang et al., 2021b, 2021c). Previous studies also showed that JQ1 has a radiosensitizing effect on radiotherapy for cervical cancer (Ni et al., 2021), and few studies have found that JQ1 can inhibit malignant behaviors such as invasion and metastasis of cervical cancer (Wang et al., 2021). In particular, there are also few studies on the mechanism of BRD4 involvement in the carcinogenesis and development of cervical cancer (Zhao et al., 2021). To address these deficiencies, we designed the present experiment to explore the potential anticancer mechanism of JQ1 against cervical cancer at the gene expression level.

Current evidence revealed that lncRNAs have important regulatory roles in the cancer biology of CC (Aalijahan & Ghorbian, 2019; He et al., 2020; Shi, Zhang & Liu, 2018a; Zhang et al., 2021a). Thus, this theoretical basis inspired our speculation that JQ1 might regulate the expression of target genes through lncRNAs to control the malignant behavior of Hela cells. Although some lncRNAs are well characterized functionally, the vast majority of such molecules remain functionally uncharacterized. According to different functional modules, lncRNAs can be finely classified into immune-related lncRNAs (Ye, Chen & Lu, 2021; Zheng et al., 2020), ferroptosis-related lncRNAs (Jiang et al., 2022; Li et al., 2022), metabolism-related lncRNAs (Lang, Huang & Cui, 2022; Lu et al., 2022), etc. Further, we speculated that super-enhancer-associated lncRNAs might be involved in this regulatory mechanism and were significantly regulated by enhancers, especially super-enhancers (SEs) (Bian et al., 2021; Peng et al., 2022; Wang et al., 2020; Yan et al., 2021). Previous studies suggested that most of SE-lncRNAs significantly regulated the expression of cancer-related marker genes (Bian et al., 2021). In this study, we focused on the effects of JQ1 on the expression profile of SE-lncRNAs. The very important premise hypothesis of our study was that super-enhancers could mediate the network regulation of various mRNAs in HeLa cells through SE-lncRNAs, while JQ1 altered the behavior of HeLa cells by interfering with the expression of SE-lncRNAs, thereby altering the expression of important target genes, such as mRNAs. To verify this hypothesis, JQ1 treatment under a concentration of 1 μmol/L for 72 h was delivered to HeLa cells, and whole transcriptome sequencing was applied to detect changes in gene expression profiles. Our differential expression analysis of genes had identified 193 differentially expressed lncRNAs, of which 162 genes were identified as SE-lncRNAs and 15 SE-lncRNAs were differentially expressed. In line with our finding, a plethora of evidence supports that JQ1 can alter the expression profile of lncRNAs in cancer cells and exert a critical role in the proliferation inhibition of cancers (Baek et al., 2022; Choi et al., 2022; Liu et al., 2021). The general consensus is that lncRNAs exert physiological regulatory effects by regulating mRNAs. To systematically identify functional lncRNAs, correlation analysis was proposed for the identification of lncRNA-mediated mRNAs by combining global and local regulatory direction consistency of expression. According to this computational approach, a total of 418 differentially expressed mRNAs were screened in this present study. BRD4 is widely known for its role in super-enhancer organization and transcriptional activation of several major oncogenic genes, including c-MYC and BCL2. Therefore, BRD4 inhibitors are being regarded as promising therapeutic agents for cancer treatment (Liu et al., 2022). Mechanically speaking, BRD4 binds to acetylated histones at enhancers and promoters via its bromodomains (BDs) to regulate transcriptional elongation (Rahnamoun et al., 2018). In some human cancers, BRD4 has been found to be recruited into enhancers co-occupied by mutated p53 and to support the synthesis of enhancer-directed transcripts (eRNAs) in response to chronic immune signaling (Rahnamoun et al., 2018). BRD4 selectively associated with enhancer-directed transcripts that were produced from BRD4-bound enhancers. The interaction between BRD4 and eRNAs increased BRD4 binding to acetylated histones in vitro and augmented BRD4 enhancer recruitment and transcriptional cofactor activities, thereby regulating the expression of downstream target genes such as lncRNAs and mRNAs (Duan, Yu & Chen, 2023). JQ1 inhibits the abnormal expression of cancer related lncRNAs and mRNAs by blocking the interaction between BRD4 and eRNAs, thereby exerting anti-tumor effects (Duan, Yu & Chen, 2023).

Next, GO functional annotation and KEGG pathway enrichment analysis were conducted to explore the biological functions of these differentially expressed mRNAs. At the GO biological process level, most differential genes affected the functions of cell-cell adhesion via plasma-membrane adhesion molecules, extracellular matrix organization, extracellular structure organization, external encapsulating structure organization and homophilic cell adhesion via plasma membrane adhesion molecules, and more. Cell-cell adhesion molecules are adhesion molecules in the immunoglobulin superfamily widely distributed between cells in various tissues. It is now known that tumor invasion and metastasis are associated with changes in the expression of their adhesion molecules (Lewczuk, Pryczynicz & Guzińska-Ustymowicz, 2019). On the one hand, the decrease in the expression of certain adhesion molecules of tumor cells can weaken the intercellular adhesion, and the tumor cells detach from the surrounding cells, which is at the forefront of tumor invasion and metastasis. On the other hand, the expression of certain adhesion molecules of tumor cells enables tumor cells that have entered the bloodstream to adhere to the endothelial cells of blood vessels, resulting in hematogenous metastasis (de Méndez & Bosch, 2011). Common cell adhesion molecules, such as MMP1 and ADAMTS8, are reported to be associated with invasive metastasis of cervical cancer and to reduce patient prognosis (Kurnia et al., 2022). Dysregulation of MMP1 transcription promotes tumor metastasis because of its role in extracellular matrix degradation in tumor invasion (Kurnia et al., 2022). Our GO results showed that MMP1 is an important target gene and plays an important role in GO BP. ADAMTS8 usually functions as a tumor suppressor, as it has been found to be frequently downregulated in colorectal cancer (Li et al., 2020). ADAMTS8 was found to inhibit tumor progression in lung cancer, and low ADAMTS8 expression was reported to be an important factor associated with poor patient prognosis (Zhang et al., 2022). In our study, we observed a significant difference in the expression level of ADAMTS8 after JQ1 treatment, and we speculated that JQ1 might affect the prognosis of cervical cancer by altering the expression status of ADAMTS8. KEGG pathway enrichment analysis showed that most differential genes affected the pathway such as bladder cancer, bile secretion, cGMP-PKG signaling pathway, cAMP signaling pathway and AGE-RAGE signaling pathway in patients with diabetic complications, which were also significantly associated with cancer progression. GO and KEGG analyses showed that differentially expressed mRNAs were widely involved in the biological process of cervical cancer, while PPI analysis further confirmed the functional modules of some important genes. The first core module for differentially expressed mRNAs involved 10 genes that play an important role in signaling pathways such as the NOD-like receptor signaling pathway.

Experimental studies of JQ1 in other cancers have also demonstrated its anticancer effects. In a study of the Cal27 line of oral squamous carcinoma, JQ1 was found to have an important role in inhibiting tumor proliferation, invasion and metastasis, suggesting that JQ1 might be a new drug for the treatment of oral squamous carcinoma (Wang et al., 2016). Another study showed that excessive activation of BRD4 could promote epithelial-mesenchymal transition (EMT) in prostate cancer cells, thus promoting prostate cancer cell metastasis, which JQ1 could be used to effectively block this process, thereby exerting antitumor effects (Blee et al., 2016). In breast cancer, JQ1 was shown to inhibit tumor proliferation mainly by suppressing c-Myc expression, but the specific mechanism of its role in malignant biological behaviors such as invasion and metastasis has not been clarified and deserves in-depth study (Tian et al., 2019). The mechanisms by which BET inhibitors exert antitumor effects are becoming clearer as research progresses. BET inhibitors are now believed to exert their antitumor effects by specifically regulating the expression of many genes (Tian et al., 2019). c-Myc and Bcl-2 are key proteins regulating cell proliferation and survival, and abnormal alterations of associated genes are the most important biological features of cancer (Luoto et al., 2010). Studies in cervical cancer have shown that BRD4 inhibitors can enhance the response of tumor cells to chemotherapy by inhibiting the expression of HPV16 E6, and are thus expected to be potential new targets for cervical cancer therapy (Rataj et al., 2019). In addition, miRNAs may also be important molecules involved in the regulatory network of BRD4 biology. A study found that miR-152-5p was down-regulated and BRD4 was up-regulated in cervical cancer tissues. Additionally, luciferase localization analysis showed that miR-152-5p could directly bind to BRD4 in HeLa and CaSki cells (Zhao et al., 2021). Thus, the proliferation, invasion, and EMT transformation of HeLa and CaSki cells could be effectively inhibited by regulating the miR-152-5p/BRD4 axis (Zhao et al., 2021). In any case, our experiments demonstrate that JQ1could inhibit cervical cancer HeLa cells, leading to an altered expression profile of lncRNAs and mRNAs, one of the core mechanisms of JQ1 antitumor.

Some future directions of JQ1 need to be discussed. First, JQ1 has not been applied in clinic, and the main reason is that the half-life of JQ1 is very short. How to prolong the action time of JQ1 in vivo is an urgent problem to be solved. Secondly, the inhibitory effects of JQ1 are cytostatic, and the expected anticancer effect of JQ1 alone is limited. Therefore, how to select a reasonable combinate drug is worth further investigation. Third, JQ1 mainly targets BRD4, but the targets protein related to SE include BRD2, BRD3, or BRDT etc. In the future, the search for dual-target inhibitors based on the BRD4-related pathways is expected to reduce off-target and improve the efficacy of tumor therapy. Third, an important limitation of this study needs to be raised, although our study is an exploratory study. The results concerning the effect of JQ1 is based solely in in vitro experiments with just one cell line, HeLa, so further studies need to be carried out in different cell lines to validate our results. Finally, correlation analysis was applied for finding the co-expression relationship between SE-lncRNAs and mRNAs in our studies, to preliminarily construct a potential co-expression network between SE-lncRNAs and mRNAs, with the aim of laying the foundation for subsequent analysis. However, the regulatory relationship between lncRNAs and mRNAs still needs further validation. In the future, we plan to further validate the biological effects of key SE-lncRNAs through interference or overexpression tests of lncRNAs.

Conclusion

In conclusion, JQ1 can significantly inhibit the proliferation of HeLa cells and affect the expression profile of SE-lncRNAs and mRNAs.

Supplemental Information