Cancer comprises more than 100 diseases that progress over time and are characterized by uncontrolled cell division. According to the World Health Organization, cancer accounted for nearly 10 million deaths in 2020. The epithelial–mesenchymal transition (EMT) is important for cancer development [1], and it involves the transformation of cuboidal, non-motile epithelial cells into a loosely organized, fibroblast-like mesenchymal phenotype with reduced intercellular adhesion, loss of apical-basal polarity, and increased motility and invasiveness.

There are three types of EMTs. Type 1 occurs during embryogenesis, while type 2 occurs during wound healing and fibrosis. Type 3 occurs in cancer and represents the first step toward cancer progression to the metastatic stage; here, cancer cells acquire the ability to erode the extracellular matrix, migrate, and eventually enter the bloodstream [2]. EMT activation is involved in the malignant progression of many cancers and can induce various cancer cell features, including the acquisition of a stem cell-like phenotype, enhancement of cancer cell metastasis, resistance to chemotherapy, and antigenic escape ^[3][4][3,4]. Despite the vast amount of data showing the importance of EMT in cancers, its character in vivo is debated since it is difficult to perform genetic fate-mapping or lineage-tracking of cancer in human cancer tissue. In addition, through the analysis of several studies on embryonic development, wound healing, and human tumors, David Tarin raised questions about the existence of EMT and its role in carcinogenesis in adult organisms. In contrast, in the perspective of Thompson et al., they assert that multiple studies have collectively provided evidence of coordinated molecular changes between the epithelial and mesenchymal states. They also state that a comprehensive molecular analysis of individual cells in actual tumors is needed to prove the existence and importance of EMT in carcinoma progression in vivo [5]. Currently, multiple studies have emerged to delve into this issue and it is proved in studies that targeting EMT-associated factors in cancer is a promising strategy [6].

Targeting TFs that engage in the EMT process as a cancer treatment strategy has been discussed in multiple studies ^[7][8][7,8]. However, there have been few reviews of TFs targeted therapies based on a systematic classification of TFs. Based on the research of Wingender et al. [9], because TFs are found to be able to recognize regulatory elements in promoters and enhancers on different DBDs by their own DNA–protein recognition code, TFs are classified based on the different DNA-binding domains (DBDs) including the basic domain, zinc-coordination DBD, helix-turn-helix domain, etc.; these domains are called a rank of Superclass in Wingender’s study. Wingender’s team classified the structure of DBD into several ranks (superclass, class, family, subfamily, genus, species) that serve as the main framework, and the folding of TF and the way it establishes the DNA interacting interface characterized as a class (for example, bZIP factors, Basic helix-loop-helix factors (bHLH) factors, etc.); TF in a class are subsumed to families or subfamilies (e.g., Jun, FOS, FOX, and c-Myb) ^[10][11][10,11].

2. The Role of EMT-TFs in Cancer

In tumor progression, classical EMT is characterized by decreased intercellular adhesion, loss of epithelial markers (such as E-cadherin and claudins), and acquisition of mesenchymal markers (such as vimentin and N-cadherin) [12]. The process is regulated by several key TFs, including Snail family proteins (including Snail1, Snail2), Zinc finger E-box binding (ZEB) homeobox family proteins, and Twist family proteins, which suppress the expression of genes linked to the epithelial state and simultaneously promote the expression of genes associated with the mesenchymal state. EMT has been proven to play a crucial role in various stages of embryonic development and result in the accumulation of extracellular matrix in fibrosis, as well as driving the progression of carcinomas towards a metastatic state [13]. Several factors cooperate to induce EMT which leads to inflammation and fibrosis in cancer. For example, TGF-β1, TNF-α, and hypoxia work together to initiate EMT by activating Snai1 through various mechanisms, with NF-κB activation playing a central role [13]. In addition, the partial activation of EMT by EMT-TFs are also reported to promote increased motility of cancer cells, whether through collective migration in cell clusters or as individual cells, and facilitate invasion and dissemination [14]. Interestingly, apart from their crucial role in the classical EMT program that regulates cancer invasion, EMT-TFs exhibit multiple characteristics in other aspects of cancer progression, including tumor initiation and chemoresistance. Firstly, EMT-TFs are associated with features that facilitate malignant progression, including evasion of senescence, DNA repair, and anti-apoptotic phenotypes [15]. Furthermore, EMT-TFs regulate the expression of pro-inflammatory and immunosuppressive cytokines in cancer cells, thereby modulating the tumor microenvironment [16]. Additionally, EMT-TFs seem to exhibit non-redundant functions that are often specific to different tissues and tumor types. For instance, the effects of Snai1 and ZEB1 on metastasis can vary depending on the type of cancer. Similarly, different EMT-TFs within the same family, such as ZEB1 and ZEB2, can have contrasting roles in tumor aggressiveness [17].

3. The Epigenetic Regulation Pathways of TFs Involved in EMT

There are various transcriptional pathways that regulate the EMT, which ultimately leads to the downregulation of E-cadherin and dissolution of cell–cell adhesion. Many genetic or non-genetic regulation pathways were involved in the over-expression of these TFs, which in turn increases the expression level, turnover time, and activities of TFs. Among these pathways, there are certain key pathways that play crucial roles in this process, such as epigenetic regulation, while facilitating chromatin remodeling and transcription initiation through histone H3 acetylation.

3.1. SNAIL-Associated Regulation Pathway

The upregulation of Snai1 is regulated by multiple signaling pathways such as TGFβ, Wnt, and ISX ^[18][19][100,173]. Snail is reported to repress gene expression by binding to the E-cadherin promoter through its carboxy-terminal zinc-finger domains. This binding reduces cell–cell adhesion in cancer cells, facilitating their detachment from the primary tumor and promoting subsequent metastasis. In detail, Snai1 recruits the Polycomb repressive complex 2 (PRC2), which consists of methyltransferases enhancer of zeste homologue 2 (EZH2), G9a, and suppressor of variegation 3–9 homologue 1 (SUV39H1), as well as the co-repressor SIN3A, histone deacetylases 1, 2, and/or 3, and the Lys-specific demethylase 1 (LSD1), upon binding to the E-box sequence in the promoter region. All of these components work together to regulate histone modifications, specifically methylation and acetylation, at specific sites on histone H3, including lysine 4 (H3K4), lysine 9 (H3K9), and lysine 27 (H3K27). The methylation of H3K9 and H3K27 is associated with repressive chromatin, whereas the methylation of H3K4 and acetylation of H3K9 mark active chromatin. This creates a poised state for the promoter, enabling timely activation while maintaining repression in the absence of differentiation signals. The bivalent control of the E-cadherin promoter may contribute to the reversible nature of EMT. Apart from repressing epithelial genes, Snai1 also triggers the activation of genes associated with the mesenchymal phenotype. This mechanism may involve the presence of bivalent domains, which exhibit repressive H3K9 trimethylation and activating H3K18 acetylation. These bivalent domains facilitate the expression of the mesodermal gene goosecoid in response to TGFβ-related Nodal55 ^[4][20][4,83].

3.2. Twist-Associated Regulation Pathway

TWIST expression can be activated by several pathways such as Wnt and hypoxia-inducible factor 1α (HIF1α). Under hypoxic conditions, HIF1α can directly bind to TWIST through hypoxia-responsive elements in the TWIST proximal promoter, leading to the upregulation of TWIST expression, and promotes the EMT and the dissemination of tumor cells; additionally, in drosophila melanogaster epithelia, Twist expression is induced by mechanical stress in a β-catenin70-dependent manner. In cancer cells, Twist1 suppresses E-cadherin and stimulates N-cadherin expression in a SNAIL-independent manner. Twist1 accomplishes this by recruiting the methyltransferase SET8 (also known as SETD8 in humans), which mediates H4K20 monomethylation. This histone modification is associated with repression at E-cadherin promoters and activation at N-cadherin promoters, contributing to the induction of the EMT process ^[21][174].

3.3. ZEB-Associated Regulation Pathway

ZEBs are also a key regulator in promoting EMT as they repress epithelial cell markers and activate the expression of mesenchymal biomarkers. There are several pathways regulating the ZEBS expression including estrogen signaling cascades, TGFβ, and Wnt/β-catenin signaling pathways. In addition, Twist1 and Snail1 are noted for their cooperative role in regulating the expression levels of ZEB1. The ZEB-mediated transcriptional pathway involves the recruitment of the C-terminal-binding protein (CTBP) co-repressor, polycomb proteins, CoREST, and the Switch/sucrose non-fermentable (SWI/SNF) chromatin remodeling protein BRG1, which enables ZEB1 to bind to regulatory gene sequences at E-boxes and represses the expression of E-cadherin. Further, ZEB1 expression results in the suppression of various genes associated with the generation and maintenance of epithelial cell polarity. Notable examples of these genes include CDH1, Lgl2, PATJ, and Crumbs3. The expression of ZEB1/2 in epithelial cells induces EMT and promotes a mesenchymal phenotype, thereby facilitating tumor invasion and metastatic dissemination into a cancer stem cell state ^[4][20][22][4,83,175].

3.4. Intestine-Specific Homeobox (ISX) and P300/CBP-Associated Factor (PCAF)

Intestine-specific homeobox (ISX) is a homeobox-containing protein that belongs to the paired subfamily and is homologous to Pax3, Pax7, and Prrx1 phylogenetically ^[23][176]. ISX was induced by the pro-inflammatory cytokine interleukin-6 and was highly expressed as a proto-oncoprotein in hepatoma cell and HCC samples ^[24][99]. Further, ISX transcriptionally regulated the downstream cell cycle proteins cyclin D1, E2F1, and indoleamine 2, 3-dioxygenases ^[25][26][102,103]. This phenomenon then dysregulated tyrosine catabolism and reduced the levels of immune checkpoint regulators (PD-L1 and B7-2) and epithelial–mesenchymal transition (EMT) regulators (Twist1 and Snail1), thereby affecting the survival time of patients with HCC ^[25][102]. Pathologic studies revealed that ISX exhibited a tumor-specific expression pattern, and it is significantly correlated with patient survival and tumor size, number, and stage ^[24][99]. Histone modification by acetylation is critical in the regulation of oncogenic gene expression and subsequent cancer progression ^[27][177]. Recently, Wang et al. discovered P300/CBP-associated factor (PCAF) acetylation of intestine-specific homeobox (ISX) regulates epithelial–mesenchymal transition (EMT) marker expression and promotes cancer metastasis ^[18][28][100,178]. PCAF acetylation of ISX at lysine residue 69 recruits acetylated bromodomain-containing protein 4 (BRD4) at lysine residue 332, and the resulting complex translocated into the nucleus and binds to EMT promoters, where acetylation of histone 3 at lysine residues 9, 14, and 18 initiates chromatin remodeling and subsequent gene expression in tumor cells ^[18][100]. Activated ISX then enhanced EMT marker expression—including TWIST1, Snail 1, and VEGF—and consequent cancer metastasis, but suppressed E-cadherin expression ^[18][29][100,101]. Evidence suggests that the PCAF–ISX–BRD4 regulation axis may hold the promise as a new therapeutic target for the discovery of new small molecular inhibitors, leading ultimately to more efficacious cancer therapy.

4. Therapeutic Implications of Targeting EMT-TFs

There are multiple therapeutic strategies for targeting EMT, including the inhibition of upstream signaling pathways such as TGFβ, NF-κB, Wnt, EGFR, and Notch. Additionally, targeting molecular drivers of EMT, such as the key TFs Snai1, ZEB, and Twist, and focusing on mesenchymal cells, integrins, and the extracellular matrix are other approaches. Moreover, there are several therapeutic agents that target TFs for cancer treatment, including small molecule inhibitors, micro RNA, and gene editing techniques ^[30][179]. Small molecule inhibitors undergoing clinical trials includes Curcumin (phase III, targeting NF-kB in brain tumor) ^[31][180], Metformin ^[32][181] (phase III, targeting ZEB1, Slug, Twist in breast cancer), Omo-103 ^[33][34][182,183] (phase I), and Disulfiram ^[35][184] (phase II, targeting ERK/NF-kB/Snail pathway in germ cell tumors). Furthermore, BRD4, a member of the bromodomain-contained protein family, is known to be involved in tumorigenesis via its binding to acetylated histones in several types of cancers. Blockade of the BRD4 interaction with HATs by small molecule inhibitors has been shown to effectively block cell proliferation in cancers, some of which have, in fact, been evaluated in human clinical trials. Small molecule inhibitors targeting the bromo- and extra-terminal (BET) domain protein, known as BETi, offer another novel strategy to inhibit the BRD4-MYC axis and subsequently suppress the downstream trans criptional pathway ^[36][185]. MicroRNAs have currently been recognized as novel target for EMT-TF; for example, miR-200 ^[37][38][186,187] (Phase II SWOG S0925, targeting ZEBs in prostate cancer), miR-186 ^[39][188] (pre-clinical, targeting Twist in cholangiocarcinoma cells), and miR-342 ^[40][189] (pre-clinical, targeting FOX in nasopharyngeal carcinoma). CRISPR/Cas gene editing ^[41][190] also hold the potential for regulating the EMT-TFs (Phase I, NCT02793856) ^[42][191].

5. Challenge of Targeting EMT-TFs in Cancer

Though targeting EMT-TFs would be attractive and several agents are undergoing clinical trials, currently there are no FDA (U.S. Food and Drug Administration)-proved therapies for targeting EMT-TFs since there are multiple technical problems remain challenging. First, the expression levels of various EMT-TFs are intricately interconnected through multiple feedback mechanisms. Second, certain EMT-TFs exhibit complementary and redundant functions. The specific role of each EMT-TF greatly relies on the cellular context and microenvironment. Moreover, there is a concern that targeting EMT-TFs with small molecule inhibitors may encounter side effects due to the essential role in normal cell survival and proliferation; further, there remains difficulty in selectively targeting TFs without affecting other TFs since that most of the TFs may interact with each other in multiple pathways. Therefore, the better understanding of the precise regulatory networks and functions of the EMT-TFs in different EMT contexts is needed ^[43][44][192,193].