Autophagy is a self-degradation process that plays an important role in balancing energy sources and eliminating harmful metabolites such as misfolded proteins and reactive oxygen species (16, 17). Recent studies have shown that FOXK2 functions as a transcriptional inhibitor of autophagy. Interestingly, FOXK2 was used to balance the activity of another forkhead transcription factor, FOXO3, which induces a set of overlapping autophagy and atrophy targets in muscle. FOXK2 also limits the expression of autophagy genes by recruiting Sin3A-HDAC complexes. Notably, under nutrient-rich conditions, the mammalian target of rapamycin (mTOR) phosphorylates and activates FOXK2 to limit basal levels of autophagy (18). Chen demonstrated that the FOXK2 protein plays an important role in transcriptional regulation, linking DNA damage and autophagy. DNA damage induces CHK2-mediated FOXK2 phosphorylation at Ser61 and traps FOXK2 in the cytoplasm by binding to 14-3-3γ ( Figure 2 ). Trapping FOXK2 in the cytoplasm relieves the inhibition of autophagy-related genes (ATGs) and promotes autophagy (19) ( Figure 2 ). The autophagy-related proteins, ULK1, Vps34, and FOXO3, were markedly upregulated by FOXK2 knockdown and downregulated by FOXK2 overexpression. Although FOXK2 can inhibit autophagy, tumor cells can relieve the inhibitory effect of FOXK2 on autophagy through various mechanisms to protect their malignant behavior under conditions of nutritional deficiency, radiotherapy, chemotherapy, and stress.

A series of tumor-metabolism related genes are regulated by FOXK2 at the transcriptional level. Some studies have demonstrated that fasting/starvation-induced FOXK2 induces aerobic glycolysis by upregulating the required enzymatic machinery (including hexokinase-2, phosphofructokinase, pyruvate kinase, and lactate dehydrogenase), while simultaneously suppressing further oxidation of pyruvate in the mitochondria by increasing the activity of pyruvate dehydrogenase kinases 1 and 4 (20). FOXK2 was upregulated in the muscle and adipose tissue of wild-type C57BL/6J mice that were starved for 16 h, and in skeletal muscle after exercise. In response to FOXK2 overexpression, glucose uptake increased in myoblasts and myotubes, an effect that was reversed by FOXK2 knockdown. Additionally, FOXK2 overexpression induced glycolysis in adipocytes, suggesting that it is a key transcriptional regulator of glycolytic enzymes. Interestingly, transcriptome analysis of adipocytes either overexpressing or lacking FOXK2 revealed that differentially expressed glycolytic pathway genes were co-regulated by FOXK2, indicating additional redundancies in their functions (20). The in vivo relevance of these findings has been confirmed, as mice lacking FOXK2 expression exhibited reduced glucose uptake, supporting a role for FOXK2 in regulating aerobic glycolysis ( Figure 2 ).

Hypermethylation of promoter regions inhibits gene expression, and genome-wide hypomethylation leads to genome instability and the activation of oncogenes (21). In mouse embryonic stem cells, FOXK2 binds methylated DNA (22). Park et al. found that urinary nicotine equivalents are significantly associated with FOXK2-methylated CpG sites (23). Although the role of DNA methylation in transcriptional regulation is still unclear, a report by Baymaz et al. showed that FOXK2/polycomb repressive deubiquitinase (PR-DUB) and MBD6 share a subset of genomic target genes, establishing a link between DNA methylation and the recruitment of transcriptional complexes (24). In this study, MBD5 and MBD6 interacted with PR-DUB, and MBD6 was recruited to the site of DNA damage after microirradiation in a PR-DUB-independent manner. In conclusion, these studies demonstrated that FOXK2 was involved in DNA methylation.

Several studies have revealed that ncRNAs function as tumor regulators by targeting transcription factors, including FOXK2 ( Table 1 ). MicroRNAs (miRNAs) are a group of small, single-stranded ncRNAs (18–25 nucleotides in length) that regulate gene expression by binding to the 3’UTR of the corresponding target mRNA (31). By targeting FOXK2, miR-602 promotes esophageal squamous cell carcinoma proliferation and metastasis and regulates cell cycle progression. More importantly, in nude mice, systemic delivery of the planned miR-602 antagomir reduces tumor growth and increases FOXK2 protein expression (25). In hepatocellular carcinoma (HCC), FOXK2 is a direct target of miR-1271. Re-introduction of miR-1271 decreased FOXK2 mRNA and protein levels. FOXK2 expression was also reported to be inversely correlated with miR-1271-5p expression in HCC samples (26). Additionally, the expression of FOXK2 may be regulated by miR-1271 in non-small-cell lung cancers (28).

Circular RNAs (circRNAs) are a special group of RNAs consisting of hundreds of nucleotides that regulate gene expression at the transcriptional or post-transcriptional level by binding to target miRNAs (32). In cervical cancer, circ-ITCH sponges miR-93-5p and upregulates FOXK2 by suppressing cell proliferation, migration, and invasion (29). Aberrant expression of circ-UBAP2 inhibits the proliferation, migration, and invasion of clear cell renal cell carcinoma (ccRCC) cells by sponging miR-148a-3p. The promotion of cell proliferation and invasion by the restoration of FOXK2 expression in circ-UBAP2 knockdown cells confirmed the circ-UBAP2-miR-148a-3p-FOXK2 axis in ccRCCs (30).

Long noncoding RNAs (lncRNAs), which are usually divided into five categories, are primarily transcribed by RNA polymerase II and lack an obvious open reading frame (33). Like miRNAs and circRNAs, lncRNAs can regulate the expression of FOXK2. For example, the lncRNA SNHG7 enhances the development of HCC by sponging miR-122-5p to regulate FOXK2 (27). This finding suggests that lncRNAs and FOXK2 are involved in the occurrence and development of cancer.

PTMs regulate protein structure and function by covalently adding functional groups, peptides, or other complex molecules to extend and diversify protein properties (34, 35). To date, more than 300 types of PTMs have been identified (36). Accumulating evidence indicates that the functions of FOXK2 are altered in cancers by PTMs, which affect the expression of FOXK2 target genes ( Table 2 ).

Phosphorylation is one of the most widely studied PTMs and orchestrates various cellular functions (42). CHK2 specifically phosphorylates FOXK2 at Ser61 in response to DNA damage and induces FOXK2 binding to the 14-3-3γ protein, which, in turn, traps FOXK2 in the cytoplasm (19). Simultaneously, FOXK2 phosphorylation leads to chemo-resistance in cancer cells by modulating autophagy (19). During the cell cycle, the phosphorylation level of FOXK2 increases periodically, and hyperphosphorylation occurs in mitotic cells. Cyclin/CDK complexes are important for generating the hyper-phosphorylated form of FOXK2 during G2-M. Ser368 and Ser423 were identified as the main binding sites for FOXK2 phosphorylation by CDK-CLNA complexes (37). Insulin signaling modifies phosphorylation at Ser415/Ser419 sites, resulting in the reciprocal translocation of FOXK2, which depends on the Akt, mTOR, and GSK3 pathways (38) ( Figure 3 ).

Ubiquitination and deubiquitination are dysregulated in cancer (43). FOXK2 binds to the SIN3A and PR-DUB complexes, which contain an important tumor suppressor protein, the deubiquitinase BAP1. FOXK2 recruits BAP1 to DNA, promotes local histone deubiquitination, and alters target gene activity (44). The binding of PR-DUB to target genes depends on FOXK2, and loss of the complex leads to increased H2AK119ub1 levels at target genes and their subsequent downregulation (39) ( Figure 3 ).

Similar to ubiquitylation, sumoylation is a highly conserved enzymatic cascade in which a SUMO protein is enzymatically conjugated to the ϵ-amino group of certain lysine residues (45). Unlike ubiquitin modification, SUMO modification is primarily responsible for modifying substrates rather than directly degrading them (46). Recently, sumoylation (Lys 527 and Lys633) was associated with the regulation of FOXK2 activity, and its role in mediating the cytotoxic response to paclitaxel through the tumor suppressor FOXO3 in breast cancer cells was recently reported (40) ( Figure 3 ). Unfortunately, there have been few studies on FOXK2 sumoylation. Hence, the precise sumoylation sites on FOXK2 and the biological mechanism of FOXK2 sumoylation require further study.

Acetylation is another reversible form of PTM that plays a key role in regulating gene expression and controlling cellular processes (47). It was not until recently that Wang et al. provided the first evidence that SIRT1 mediates the deacetylation of FOXK2. In this study, SIRT1 inhibition enhanced FOXK2-induced chemosensitivity to cisplatin via acetylation at K223 of FOXK2, and FOXK2 K223 deacetylation reduced the chemosensitivity of cancer cells to cisplatin (41) ( Figure 3 ). Further investigations are needed to shed light on the roles of FOXK2 acetylation at other sites and determine how they crosstalk with other PTMs.

During the progression of breast cancer, the interaction of FOXK2 with ERα leads to decreased stability and ubiquitin-mediated degradation of ERα, which depends on BRCA1/BARD1 (10). Corroborating these data, a separate study reported that FOXK2 mediates the cytotoxic effects of chemotherapeutic agents (paclitaxel and epirubicin) on breast cancer cells (48). Some studies have shown that the expression of FOXK2 is gradually lost during breast cancer development. To further explore FOXK2 functions in breast cancer tumorigenesis, Shan et al. reported that FOXK2 interacts with the transcription co-repressor complexes NCoR/SMRT, SIN3A, NuRD, and REST/CoREST to repress a cohort of genes, including HIF1β and EZH2, and to regulate several signaling pathways, including the hypoxic response (49). The authors found that FOXK2 inhibited the proliferation and invasion of breast cancer cells and suppressed the growth and metastasis of breast cancer. Notably, FOXK2 is transactivated by ERα and transrepressed through reciprocal successive feedback by HIF1β/EZH2. Additionally, combining transcription inhibitory complexes, such as SIN3A, NCoR/SMRT, NuRD, and FOXK2, inhibits the transcription of a series of oncogenes, including HIF1, EZH2, and VEGFR. This affects a series of signaling pathways, including the cell cycle, DNA damage response, hypoxia response, p53 signaling, and the epithelial–mesenchymal transition (EMT), ultimately inhibiting the occurrence and development of breast cancer.

Similar to its role in breast cancer, FOXK2 has been implicated in suppressing tumorigenesis in non-small-cell lung cancer (NSCLC). FOXK2 regulates NSCLC cell growth by suppressing the expression of cyclin D1 and CDK (28). Additionally, the capacity for invasion and EMT progression of NSCLC cells were impaired following FOXK2 knockdown. Wang et al. analyzed the expression of FOXK2 in 151 patients with glioma and confirmed that the expression of FOXK2 was negatively correlated with tumor grade and prognosis (50). At the cellular level, FOXK2 inhibits the proliferation, invasion, migration, and upregulation of E-cadherin and α-catenin, and downregulates the expression of N-cadherin and vimentin, suggesting that FOXK2 modulates the expression of epithelial and mesenchymal markers closely linked to the EMT process (50). Consistent with these findings, Liu et al. reported that FOXK2 downregulation in gastric cancer cell lines inhibits cell proliferation and colony formation, in addition to the suppression of cellular migration and invasion (51). Zhang indicated that FOXK2 inhibits the malignant phenotype of ccRCC and acts as a tumor suppressor, possibly through the inhibition of EGFR (52). Interestingly, methylated FOXK2 has been observed in circulating leukocytes of smokers in an epigenome-wide association study of nicotine equivalents (23), suggesting that epigenetic modifications serve as additional mechanisms for silencing FOXK2 gene expression, which may have implications for the pathogenesis of lung cancer.

Despite its role as a tumor suppressor, some reports have attributed an oncogenic role to FOXK2. In HCC, overexpression of FOXK2 results in enhanced cell growth and migration. Further investigations revealed that FOXK2 facilitates the development of HCC via the activation of the PI3K/AKT pathway; overexpression of FOXK2 resulted in a significant increase in the expression of phosphorylated Akt, survivin, c-Myc, p27, and CyclinD1 proteins, which was reversed following FOXK2 knockout. Treatment with LY294002, an inhibitor of the PI3K/AKT pathway, markedly attenuated the cell growth induced by FOXK2 in HCC cell lines (26). Another study reported that FOXK2 downregulation inhibited cell proliferation, colony formation, migration, and invasion in HCC cells and suppressed EMT partly through the inhibition of the Akt signaling pathway (53). The gene targets involved in mediating the oncogenic effects of FOXK2 in HCC remain to be identified. In addition to HCC, FOXK2 acts as an oncogene in colorectal cancer (CRC). FOXK2 promotes CRC metastasis by activating EGFR and ZEB1 (54). FOXK2 protein levels are elevated in CRCs and correlated with DVL nuclear localization, and they positively regulate Wnt/β-catenin signaling by translocating DVL into the nucleus (11). This is consistent with a report on the oncogenic role of FOXK1 in CRC (55). Corroborating these data, Qian et al. have reported that FOXK2 promotes the proliferation of CRC cell lines and that the FOXK2 promoter is transcriptionally regulated by SOX9 oncogenic proteins (56).

In papillary thyroid cancer, stable knockdown of FOXK2 markedly inhibited cell proliferation, significantly increased the ratio of LC3-II/LC3-I, and reduced p62 expression, whereas overexpression of FOXK2 resulted in the opposite phenotype (57). FOXK2 positively regulates VEGF and VEGFR-signaling networks. Functional experiments have demonstrated that FOXK2 promotes angiogenesis by inducing VEGF-A transcription. We also found that FOXK2-induced VEGF-A could bind to VEGF-R1 to compensate for VEGF-R2 blockage, which promoted angiogenesis by activating ERK, PI3K/AKT, and P38/MAPK signaling in human umbilical vein endothelial cells (58).