E2Fs are critical regulators of genes required for cell cycle progression and play an integral role in the control of cell proliferation. Previous studies have suggested that E2Fs are required to control cell proliferation differently in carcinogenic environments than in normal cells (29). As shown in recent studies, the proliferative potential of CSCs seems to be strongly correlated with cell cycle regulation by E2Fs. The transcriptional activators E2F1-3 regulate cell proliferation by activating genes essential for G1/S-phase progression in CSCs ( Figure 2A ). ENPP1 (ectonucleotide pyrophosphatase/phosphodiesterase 1) has been proved to induce the stemness features of cancer cells (70, 71). Bageritz et al (72) also discovered that E-NPP1 could facilitate glioblastoma CSC proliferation by controlling cell cycle progression. And E-NPP1-deficient glioblastoma CSCs led to a decrease in the transcriptional function of E2F1, and classical E2F1 target genes facilitating G1/S transition were downregulated, resulting in the accumulation of cells in G1 phase and inhibition of their proliferation. Similarly, E2F1 plays a pivotal role in regulating the proliferation status of chronic myeloid leukemia (CML) stem/progenitor cells (SPCs). In terms of cell cycle regulation, the E2F1 signaling pathway was found to be deregulated in CML SPCs, and inhibition of E2F1 led to cell cycle arrest and induced blockade of proliferation (21). Moreover, the dissociation of E2F1 from the RB protein was found to restore its transcriptional activity, which is required for cell cycle progression from G1 to S phase (73). In glioblastoma CSCs, the bromodomain inhibitor JQ1 was found to cause cell cycle arrest and suppress CSC proliferation by preventing the release of E2F1 from the RB/E2F1 complex (74). Additionally, glioblastoma CSCs display the dependence of HELLS (helicase, lymphoid-specific) to maintain proliferation. Both E2F3a and E2F3b were found to interact with HELLS, thereby increasing the expression of cell cycle progression-related genes to maintain glioblastoma CSC proliferation, while proliferation was impaired after E2F3 gene knockdown (75). Remarkably, although the function of E2F repressors is generally opposite that of E2F activators in the regulation of the cell cycle, the effects of E2F repressors are inconsistent with the predicted inhibitory effect on proliferation. In fact, current experimental evidence supports the role of E2F repressors in promoting cell cycle progression and proliferation of CSCs. For example, high levels of E2F7 detected in liver CSC populations were found to be vital for the maintenance of CSC proliferation via E2F7 activation of AKT1-cyclin D1 signaling and downstream cell cycle mediators to promote cell cycle progression and proliferation (26) ( Figure 2A ). The unexpected positive effect of repressive E2Fs on stem cell attributes of CSCs, such as proliferation by modulation of the cell cycle could potentially be a consequence of repression of negative regulators such as miRNAs (see below). However, E2F1 is also an effective activator of apoptosis (34), which may lead to the opposite effect of E2F1 on proliferation of CSCs. As demonstrated by El-Khattouti et al (76), the apoptosis of CD133 (+) melanoma subpopulation cells induced by the anticancer drug caffeic acid phenethyl ester (CAPE) requires the participation of E2F1. Functional analysis of E2F1 in the CD133(+) melanoma subpopulation demonstrated that overexpression of E2F1 induced both degradation of the antiapoptotic protein Mcl-1 and activation of the ASK1/JNK and p38 pathways, which are involved in regulating the localization of Bax to mitochondria to trigger apoptosis (76) ( Figure 2A ). Overall, E2Fs mainly regulate the proliferation of CSCs in a cell cycle-dependent manner, and E2Fs are seem to necessary factors in the control of CSC proliferation. But E2F1 can promote the apoptosis of CSCs, whether this is attributed to the CSC-type specific or other mechanisms is worth further study.

Self-renewal is the process by which cells divide while maintaining an undifferentiated state. This process requires maintenance of stem cell transcription factor expression, activation of self-renewal signaling pathways and strict control of cell metabolism (77–79). Accumulating evidence suggests the substantial engagement of distinct E2F members in each of the abovementioned self‐renewal programs and the contributions of E2Fs to stemness acquisition and maintenance of the self-renewal ability of CSCs. Several studies have confirmed that E2F1 can directly transcriptionally activate the stem cell transcription factors NANOG, KLF4, SOX2 and SCF to act as a regulator of CSC self‐renewal ( Figure 2B ). For example, using liver tumor-initiating stem-like cells (TICs), Chen et al (80) demonstrated that E2F1 transcriptionally activated NANOG expression, which in turn inhibited oxidative phosphorylation (OXPHOS) and activated fatty acid oxidation (FAO) through metabolic reprogramming, maintaining the TIC self-renewal ability. In ovarian CSCs, E2F1 acts as an upstream transcription factor to activate KLF4 expression and participates in EIF5A2-mediated modulation of CSC self-renewal properties (81). In addition, YAP1 is known to be a major mediator of Hippo signaling pathway and plays an important role in CSC self-renewal (82). Schaal et al (83) proved that E2F1 transcriptionally activates the Sox2 promoter via YAP1 mediated in response to the signaling events caused by the binding of nicotine to α7 nAChR, which enhances the self-renewal of lung CSCs. In the same context, E2F1-3a was also found to induce SCF expression at the transcriptional level under nicotine exposure. Research on the underlying mechanism has mainly involved E2F1. When the binding of nicotine to nAChRs initiates a signaling cascade, it results in the induction of E2F1-mediated SCF promoter activation and stimulates c-Kit expressed on stem-like cells, facilitating the self-renewal of lung cancer side population cells (84). Additionally, the NOTCH and WNT/β-catenin signaling pathways are essential for the self-renewal of CSCs (57). Limited studies have demonstrated regulation of the NOTCH and β‐catenin pathways by E2Fs to promote the acquisition of self-renewal capability. In prostate cancer cells, this involvement of MUC1-C (mucin 1 C-terminal transmembrane subunit) in lineage plasticity is associated with induction the CSCs state and self-renewal capacity (85, 86). A recent study confirmed that MUC1-C could directly bind to E2F1 activated the BAF remodeling complex, induced NOTCH1 signaling pathway and promoted NANOG expression, which in turn led to the promotion of CSCs stemness (86). An additional study reported that TREX1 (three prime repair exonuclease 1) negatively regulates the self-renewal of osteosarcoma CSCs. With respect to underlying mechanisms, TREX1 suppression allowed the activation of β-catenin signaling in the dependence of E2F4, thus possibly enhancing the self-renewal ability of osteosarcoma CSCs by upregulating OCT4 (87) ( Figure 2B ). Intriguingly, a recent finding suggests crosstalk among E2F2, the NAD⁺ biosynthesis pathway and CSC self‐renewal. Gujar et al (25) found that NAD⁺ metabolism governs glioblastoma CSC self-renewal via the NAMPT-E2F2-ID axis. NAMPT is the rate-limiting enzyme in the NAD salvage pathway, and E2F2 acts downstream of NAMPT and controls ID1 gene transcription to drive glioblastoma CSC self-renewal. E2F2 is required for the link NAD⁺ metabolism and the self-renewal transcriptional program in glioblastoma CSCs ( Figure 2B ).

Recent studies have highlighted that E2Fs promote CSC self‐renewal by enhancing the activity of stem cell transcription factors, regulating self‐renewal signaling pathways, and modulating cell metabolism. However, to unequivocally define E2Fs as regulators of CSC self‐renewal, further studies are needed to clarify that E2Fs are required for the regulation of CSC self‐renewal.

Metastasis is the process by which malignant tumor cells travel from the primary site through lymphatic vessels, blood vessels, or body cavities and spread to other sites (88). In metastasis, CSCs show more invasive and metastatic phenotypes than non-stem tumor cells and play the lead role (89). Because CSCs have the ability to initiate tumors and a high metastatic capacity, they are considered to be the cells responsible for repopulating metastatic tumor (89, 90). In recent years, emerging research evidence has documented that some circulating tumor cells (CTCs) are CSC-like cells with strong metastatic potential that can drive macroscopic tumor growth in distant tissues (91–95). Interestingly, in a mouse model of metastatic breast cancer, the lung colonization capacity of CTCs and the numbers of metastatic lesions in the lung were strikingly reduced in mice on both the E2F1 and E2F2 knockout backgrounds (24, 96) ( Figure 2C ). Moreover, E2F1 loss also caused loss of VEGFA expression and tumor angiogenesis defects (24). CSCs are mostly located around blood vessels and rely on the microvasculature for metastasis (97, 98). These findings raise the possibility that cell invasiveness and tumor angiogenesis regulated by the activity of E2F1/2 might be effectively involved in the metastasis of CSC-like cells, thus facilitating tumor metastasis. Moreover, a recent study by Teng et al. (99) found that upregulation of E2F7 resulted in increased activation of the VEGFR-2 signaling pathway and promoted metastasis, invasion, and angiogenesis in hepatocellular carcinoma (HCC). Concomitantly, high expression of E2F7 in HCC stem cells (26) provides the potential for E2F7 to enhance CSC metastasis.

EMT is closely related to the migration and invasion of CSCs. Of note, EMT combines two of the most important attributes to facilitate metastasis: invasiveness and stemness (89, 100). To the best of our knowledge, E2F1-8 extensively activate the EMT program in a variety of cancers, leading to the acquisition of a very invasive and metastatic phenotype (101–106), which may further result in the acquisition of a metastatic CSC phenotype ( Figure 2C ).

To date, the exact mechanism of several E2Fs in the regulation of CSC metastasis is still unclear, although their involvement seems essential. Further research is necessary to understand how E2F-dependent regulation of EMT and angiogenesis are interconnected in CSC metastasis.

CSCs are the main culprit of resistance to chemotherapy and radiotherapy. The mechanisms underlying drug resistance involves in efficiently inducing drug efflux through ATP-binding cassette (ABC) transporters (including MDR1, MRP1, and ABCG2), activating DNA repair programs, enhancing autophagy, and/or inhibiting apoptotic signals to avoid death (67, 107). In a variety of human cancers, the resistance mechanism of ABCG2 protein overexpression in CSCs has been used as an effective functional mechanism to identify and isolate CSCs (19). E2F1 was further reported to drive chemotherapeutic drug resistance in lung and breast cancer cells via activation of ABCG2 expression (108). This observation demonstrates new scenarios that E2F1 may participate in ABCG2-mediated regulation of CSC resistance. Additionally, an increasing amount of evidence shows that tumor cells under hypoxic conditions are induced to transform into CSCs phenotype and mediate tumor radiotherapy or chemotherapy resistance (68). Meanwhile, many cancer literatures have investigated the close link between E2Fs and hypoxia. For example, E2F1 can transactivate the RAD51 promoter under hypoxia to facilitate homologous recombination repair and drug resistance of prostate cancer cells (109). E2F4 was reported to repress BRCA1 expression with hypoxia-induced (110). Concomitantly, DNA repair gene BRCA1 is a critical tumor suppressor that helps CSCs acquire drug resistance under hypoxic conditions (111). And E2F1 and E2F3 can mediate autophagy activation of cancer cells with hypoxia (112, 113). Moreover, autophagy caused by hypoxia commonly results in enhancing CSCs chemoresistance (114, 115). Thus, this set of observations indicate the possibilities that E2Fs could induce the DNA repair and autophagy enhancement through hypoxia regulation, thereby increasing CSC drug resistance. Other studies have also suggested that E2F2 mediates the radioresistance of glioblastoma CSCs (25). The NAMPT-E2F2-ID1 pathway was found to be upregulated in glioblastoma CSCs after radiation and represented a protective response to radiation ( Figure 2D ).

In summary, these studies, together with the limited research focus on E2Fs and CSC resistance, accentuate the necessity to redirect our attention to the study of E2Fs, because these proteins perform unique transcriptional functions to mediate resistance mechanisms in CSCs.

MicroRNAs (miRNAs) are small noncoding RNAs that function as posttranscriptional regulators. Previous works have described relationships between the expression of miRNAs and E2Fs. E2F activity can be regulated by miRNAs; in contrast, miRNAs themselves are targets of E2F proteins (116). Subsequent studies found that this feedback relationship plays a pivotal role in CSCs ( Table 1 ). Regarding E2F1, oncogenic miR-20b-5p was found to increase the proliferation of breast CSCs by upregulating E2F1 protein expression (117). Concomitantly, E2F1 was found to act as a downstream target of miR-185-3p and promote CSC stemness properties through the LINC00511/miR-185-3p/E2F1/Nanog axis in breast cancer (118). Additionally, regulation of the E2F1 pathway by miRNAs is a novel strategy for combating CSCs. For example, hsa-mir183/EGR1–mediated upregulation of E2F1 is required for CML SPC survival and proliferation. Downstream of this event, inhibition of E2F1 was found to reduce the proliferation of CML SPCs and induce p53-mediated apoptosis, but healthy SPCs were not affected by inhibition of the E2F1 pathway (21). The researches above indicated that E2F1 seems to be a positive regulator of CSC-traits. However, E2F1 can also negatively regulate the characteristics of CSC. In gastric CSCs, upregulated miRNA-20a was found to contribute to the self-renewal and proliferation of CSCs by targeting the inhibition of E2F1 protein expression and subsequently activating Wnt/β-catenin signaling (119, 120). Taken together, E2F1-mediated regulation promotes the properties of breast cancer and CML CSCs, but has an inhibitory effect on the characteristics of gastric CSCs. And targeting blockage of E2F1 anticipated is expected to be a potential therapeutic applicable to breast cancer and CML. However, whether the promotion and inhibition of E2F1-mediated CSCs regulatory activities are tumor-type specific or miRNAs regulation, which requires further evidences to predict outcome.

A variety of miRNAs regulate the biological characteristics of CSCs by directly targeting E2F2 ( Table 1 ). For instance, miR‐125b was found to suppress the proliferation of CD133-positive glioblastoma CSCs by direct downregulation of E2F2 (121). Elevated Let-7b also repressed E2F2 expression in glioblastoma CSCs, resulting in reductions in tumorsphere growth and CSC populations (122). In triple-negative breast cancer (TNBC), miR-4319 was found to negatively regulate the self-renewal and metastasis of CSCs through targeted inhibition of E2F2 (123). Concomitantly, overexpression of miR-638 also inhibited the self-renewal, proliferation, and invasion abilities of breast CSCs by suppressing E2F2 (124). Moreover, E2F2 expression was shown to inversely correlate with miR-99 expression and directly with elevated level of vimentin in lung cancer biopsies. By inhibiting E2F2, miR-99a repressed the EMT process, accompanied by suppression of stemness features, consequently decreasing the CSC population (125). Collectively, targeting E2F2 is expected to become a therapeutic target for a variety of CSCs.

Fewer studies have been devoted to studying the role of miRNAs related to the remaining E2Fs (E2F3-E2F8) in CSCs ( Table 1 ). For example, miR-449b was found to inhibit the proliferation of SW1116 colon CSCs through downregulation of E2F3 expression (126). miR128-1 was found to suppress the proliferation and tumorigenicity of glioblastoma CSCs by targeting E2F3 (127). Furthermore, E2F7 was found to be the direct target of miRNA-302a/d. miRNA-302a/d negatively regulates the self-renewal ability and cell cycle entry pathway of liver CSCs through direct repression of the target gene E2F7 and its downstream AKT/β-catenin/CCND1 signaling pathway (26).

On the other hand, E2Fs play a role in CSCs by regulating the transcription and maturation of various miRNAs ( Table 1 ). Recent research findings have suggested that E2F6 functions as a competing endogenous RNA (ceRNA) for miR‐193a. Upregulation of E2F6 mRNA expression, in turn, was found to upregulate the stemness marker c-Kit through E2F6-mediated silencing of miR‐193a and to promote ovarian cancer stemness and tumorigenesis (22, 128). In colon CSCs, E2F7 was found to transcriptionally inhibit miR-199b expression to promote USP47 expression, thereby increasing colon CSC stemness and accelerating the occurrence of colon cancer (23). In addition, E2F7 silencing was shown to decrease the production of ALDH1⁺ cells and repress antagonistic effects of ALDH1⁺ cells on 5-fluorouracil (5-FU) treatment.

The above findings show that E2Fs and miRNAs regulate one another and play important roles in the proliferation, self-renewal, metastasis, and drug resistance of CSCs. However, there are still some E2Fs (E2F4, E2F5 and E2F8) that are relatively infrequently studied in CSCs, and the related miRNAs have not been studied, which may be the direction of future research.