Cells consult their extracellular environment, and the growth-regulating signals within it, during a discrete window of time in G₁ phase of the cell cycle. Once cells have moved through this window, they proceed through S, G₂, and M phases. However, if growth factors are removed before a specific time point, cells will fail to proceed toward S phase, and instead remain in G₁ or revert to G₀ phase (Pardee, 1974). This total dependence on extracellular signals, followed by entrance in late G₁ into a state of relative independence, indicates that an important decision must be made before the end of G₁. Precisely at this point, a cell must ‘make up its mind’ whether it will remain in G₁, retreat from the active cycle into G₀, or advance into late G₁ and S phase. This critical decision is made at the restriction point or R-point (Malumbres and Barbacid, 2001; Pardee, 1974; Weinberg, 2014).

Time-lapse microscopy of asynchronously cycling Swiss 3T3 cells showed that cycling cells are sensitive to serum withdrawal for only the first 3 to 4 h after mitosis, implying that the R-point is within this interval (Zetterberg and Larsson, 1985). Subsequent studies revealed that the R-point occurs 3 to 4 h after mitogenic stimulation in most mammalian cells (Zetterberg et al., 1995). The concept of the R-point heralded the subsequent discovery of the cell-cycle engine (Nurse, 2000; Pardee, 1974; Sherr and Roberts, 1999). In addition, the R-point is deregulated in most cancer cells (Pardee, 1974; Weinberg, 2014). Therefore, understanding how cells make a fate decision at the R-point should provide insight into how the cell cycle is controlled and how tumors develop (Blagosklonny, 2006).

The DNA-binding transcription factor RUNX3, which plays pivotal roles in lineage determination, defends against oncogenic K-RAS–induced tumorigenesis (Ito et al., 2015). Deletion of Runx3 in mouse lung results in development of lung adenomas and accelerates oncogenic K-Ras–induced progression into adenocarcinomas (Lee et al., 2013). In mouse embryonic fibroblasts, Runx3 deletion perturbs the R-point, leading to transformation (Chi et al., 2017). Recent work showed that RUNX3 functions as a pioneer factor that plays key roles in R-point–associated induction of immediate early genes, including p21^Waf/Cip (hereafter p21) and p19^ARF (p14^ARF in human, hereafter ARF). In this review, we summarize how RUNX3 contributes to the R-point decision in collaboration with histone modifiers, chromatin-remodeling complexes, the basal transcriptional machinery, and Polycomb group (PcG) proteins.

When extracellular mitogenic signaling is maintained up to the R-point, transcription of R-point–associated target genes is activated (Chi et al., 2017). For a silent gene to be induced, the chromatin structure of its chromosomal locus must be opened. Chromatin structures and functions are modulated by covalent modification of specific residues within the amino-terminal tails of histones; the unique combination of modifications has been described as the histone code (Mills, 2010). Trithorax group (TrxG) proteins establish histone modifications that activate transcription, whereas PcG proteins establish histone modifications that repress transcription. TrxG proteins consist of two categories: histone modifiers (Strahl and Allis, 2000) and nucleosome remodelers (Vignali et al., 2000). TrxG histone modifiers include the mixed-lineage leukemia (MLL) protein, which methylates histone H3 at lysine 4 (H3K4-me), a histone mark that favors transcriptional activation. Nucleosome remodelers of TrxG complex contain SWI/SNF complex, which facilitate the binding of transcription factors and basal transcription machinery (Imbalzano et al., 1994). PcG complexes include two categories: Polycomb repressor complexes 1 and 2 (PRC1 and PRC2). The PRC2 complex contains Enhancer of Zeste Homolog 2 (EZH2), which trimethylates histone H3 at lysine 27 (H3K27-me), a characteristic of inactive chromatin (Cao et al., 2002).

Gain of PcG and loss of TrxG is a common theme in human cancer, indicating that PcG and TrxG are involved in regulation of tumor suppressors: PcG suppresses and TrxG activates tumor suppressors. ARF, which induces cell-cycle arrest and apoptosis by facilitating p53 activity in response to aberrant oncogene activation (Efeyan and Serrano, 2007; Kamijo et al., 1997; Palmero et al., 1998), is regulated by PcG and TrxG. During normal proliferation, PcG complexes (PRC1 and PRC2) and histone deacetylases (HDACs) bind the ARF locus, thereby inhibiting senescence. In response to oncogenic RAS, TrxG-mediated chromatin dynamics override PcG-mediated repression, thereby inducing ARF/p53-mediated senescence (Mills, 2010).

Soon after mitogenic stimulation (within 1 h after stimulation), histone H4 and RUNX are acetylated by p300 acetyltransferase. BRD2 contains two bromodomains (BD1 and BD2), and each bromodomain interacts with a distinct protein: BD1 interacts with acetylated RUNX3, whereas BD2 interacts with acetylated histone H4. The p300–RUNX3–BRD2– histone complex is formed 1 to 2 h after serum stimulation, and binds to the promoters of the genes encoding p21 and ARF (Lee et al., 2019a). The complex dissociates 4 h later (Lee et al., 2019a). These observations suggest that a large complex containing RUNX3, BRD2, p300, and histone forms at the promoters of p21 and ARF at the R-point; within this complex, BRD2 grips both RUNX3 and histone through its two BDs. A schematic diagram of the complex formed at target loci soon after serum stimulation is shown in Figure 1.

BRD2 participates in multiprotein transcription complexes such as Mediator, recruits the SWI/SNF chromatin-remodeling complex (Denis et al., 2006), and allows RNA polymerase II to transcribe through the nucleosome (LeRoy et al., 2008). Interestingly, TAF1 and TBP form a complex with RUNX3 and BRD2 1 h after serum stimulation (Fig. 1). Similarly, MLL1/5 (activating histone modifiers) and the components of the SWI/SNF complex, Brg-1 and BAF-155, interact with RUNX3 and BRD2 simultaneously 1 to 2 h after serum stimulation, and subsequently dissociate (Fig. 1). These results demonstrate that the p300–RUNX3–BRD2–histone complex interacts with MLL1/5 as well as the SWI/SNF and TFIID complexes before the R-point.

ARF is a target of RUNX3 (Lee et al., 2013) that is critical for the life and death of cells; thus regulation of its expression could represent the R-point decision. Lee et al. (2019a) showed that MLL1/5, SWI/SNF, and TFIID are recruited to the ARF promoter locus, but not when the RUNX3 binding site is deleted. Consistent with this, MLL1/5, SWI/SNF, and TFIID are not recruited to the ARF promoter in H460 cells, which do not express RUNX3 but are recruited after ectopic expression of RUNX3. These results demonstrated that the MLL1/5, SWI/ SNF, and TFIID complexes are recruited to the ARF promoter, and that their recruitment is guided by RUNX3. The large RUNX3-containing complex formed before the R-point has been designated as the R-point–associated RUNX3-containing activator complex (Rpa-RX3-AC) (Fig. 1).

In HEK293 cells, H3K27-me3 (a repressive histone modification) is highly enriched in the ARF promoter region prior to serum stimulation. Notably, H3K27-me3 is replaced by H3K4-me3 (an activating histone modification) 1 to 4 h after serum stimulation, whereas H3K27-me3 is recruited to region 8 h later. Similarly, H4K12-ac (an activating histone modification) is detected in the ARF promoter region 1 to 4 h after stimulation and is subsequently erased. Consistent with this, induction of ARF leading to p53 stabilization occurs 1 to 2 h after stimulation (Lee et al., 2019a). By contrast, in HEK293 cells in which the RUNX binding site was deleted from the ARF promoter, the H3K27-me3 modification was maintained, and H4K12 was not acetylated for a long time after serum stimulation (Lee et al., 2019a). These results demonstrate that RUNX3 interacts with its binding site and opens and activates the chromatin structure of target loci 1 to 2 h after serum stimulation, and then subsequently closes and inactivates the loci 8 h later.

For induction of a silent gene, a special DNA-binding transcription factor must access the inactive chromatin and locally open it to make it competent for binding of other factors. These special transcription factors, which initiate regulatory events, are known as “pioneer transcription factors” (Zaret and Carroll, 2011). Because the R-point–associated activation of the ARF locus is dependent on RUNX3 protein and the presence of a RUNX binding site in the ARF promoter, RUNX3 is considered as a pioneer factor of the R-point (Fig. 1).

Lee et al. (2019a) showed that BRD2 interacts with RNF2 and BMI1 (components of PRC1) 4 to 8 h after mitogenic stimulation. At that time, RUNX3 associates with Cyclin D1 and HDAC4, but dissociates from BRD2. Also, EED and EZH2 (components of PRC2) interact with RUNX3 8 h after stimulation. Consistent with this, ChIP analysis revealed that Cyclin D1, HDAC4, and EZH2 are recruited to target chromatin loci after the R-point, at a time when H4K12-ac is absent and H3K27-me3 is enriched at the target chromatin loci. After 4 h, RNF2 briefly interacts with RUNX3 and BRD2, and binds to target chromatin loci, suggesting that PRC1 and Rpa-RX3-AC form a transient complex immediately before the Rpa-RX3-AC is destroyed (about 3 h after mitogenic stimulation). The transient construction has been designated as the R-point– associated RUNX3-containing transient complex (Rpa-RX3-TR) (Fig. 2).

At 4 to 8 h after mitogenic stimulation, RUNX3 and BRD2 are present in separate complexes: RUNX3 is in complex with Cyclin D1, HDAC4, and PRC2, which remain bound to target chromatin loci, whereas BRD2 forms the BRD2–PRC1 complex, which was released from the loci. Because the RUNX3– Cyclin D1–HDAC4–PRC2 complex closes and inactivates target chromatin loci, it has been designated as the R-point– associated RUNX3-containing repressor complex (Rpa-RX3RE) (Fig. 2).

The R-point transition is governed by R-point–associated proteins (R-proteins), which include cyclins, cyclin-dependent kinase (CDKs), p21, p27, E2F, and pRB, with pRB serving as the primary molecular regulator (Weinberg, 2014). The complexes of CDKs and their cyclin partners are responsible for sending out signals from the cell-cycle clock to a number of responder molecules that carry out the actual work of moving the cell through its growth and division cycle. During much of the G₁ phase of the cell cycle, CDK4 and CDK6 (CDK4/6) are guided by and depend upon their association with D-type cyclins (D1, D2, and D3). After the R-point in late G₁, the E-type cyclins (E1 and E2) associate with CDK2 to enable the phosphorylation of appropriate substrates required for entry into S phase, making the R-point decision irreversible.

In the case of Cyclin D1, which is the best studied of the three D-type Cyclins, growth factor-activated signal cascades result in rapid accumulation of Cyclin D1. Conversely, removal of growth factors from the growth medium results in an equally rapid collapse of the Cyclin D1 level. Therefore, D-type Cyclins must serve as regulators of the R-point, conveying signals from the extracellular environment to the cell-cycle clock operating in the nucleus. In early and mid-G₁ phase, Cyclin D–CDK4/6 phosphorylates pRB prior to passing through the R-point gate, and Cyclin E–CDK2 further phosphorylates pRB and drives the cell cycle into S phase. Therefore, the R-point decision must be made after formation of the Cyclin D– CDK4/6 complex and before formation of the cyclin E–CDK2 complex. In other words, during the R-point decision, which requires substantial time for collection of extracellular signals and intracellular information, the Cyclin D–CDK4/6 complex must be formed, whereas the Cyclin E–CDK2 complex must not be formed. In that case, how is the timing of formation of these two complexes regulated in an R-point manner? Important in this respect, Michieli et al. (1994) revealed that p21 is induced in a p53-independent manner in mouse embryonic fibroblasts, reaching a peak 2 h after mitogenic stimulation, and is then gradually downregulated. Because p21 was originally identified as a CDK inhibitor and a downstream target of p53, the biological meaning of the mitogen-stimulated p53-independent transient induction of p21 was not understood at the time. Subsequent studies revealed that p21 activates Cyclin D–CDK4/6 by stimulating the association of its component proteins, whereas it inhibits Cyclin E–CDK2, Cyclin A–CDC2, and Cyclin B–CDC2 (Cheng et al., 1999; LaBaer et al., 1997). However, it still remained unclear why p21 exerted opposite effects on different types of CDK complexes. This became clearer when Chi et al. (2017) and Lee et al. (2019a) reported that the R-point–associated RUNX3-containing activator complex (Rpa-RX3-AC) is formed 1 h after mitogenic stimulation and activates p21 transcription (Chi et al., 2017; Lee et al., 2019a). The induced p21 facilitates association of Cyclin D–CDK4/6 and inhibits association of Cyclin E–CDK2 (Cheng et al., 1999; LaBaer et al., 1997). Therefore, p21 promotes entry into the R-point at early/mid-G₁ phase by activating Cyclin D–CDK4/6, but prevents further progression through the R-point by inhibiting Cyclin E–CDK2 (Sherr and Roberts, 2004; Weinberg, 2014). When cells decide to undergo cell-cycle progression, Rpa-RX3-AC dissociates, and p21 is downregulated, allowing the Cyclin E– CDK2 complex to form. Therefore, p21 may play a key role in maintaining R-point stage until the R-point decision is made.

Lee et al. (2019a) reported a correlation between R-point progression and the RAS-RAF-MEK pathway, namely, that the transition from Rpa-RX3-AC to Rpa-RX3-TR occurs only after MEK activity is reduced. Ectopic expression of oncogenic K-RAS (Myc-K-RAS^G12V) facilitates the formation of Rpa-RX3-AC but inhibits the transition from Rpa-RX3-AC to Rpa-RX3-TR. Consistent with this, H4K12-ac and H3K4-me3 histone marks are maintained for a long time, and ARF and p21 (targets of Rpa-RX3-AC) are not downregulated but are instead maintained for long periods by expression of Myc-KRAS^G12V (Lee et al., 2019a). These results demonstrated that p21 and ARF loci are opened by Rpa-RX3-AC and closed at the R-point in normal cells, but are not closed in cells expressing oncogenic K-RAS due to failure of the Rpa-RX3-AC → Rpa-RX3-TR transition. This R-point decision-making program is tightly regulated by the concerted action of multiple pathways that mediate mitogenic signaling (Lee et al., 2019a). If the balance of mitogenic signals is deregulated, the cell cannot pass through the R-point. Therefore, the R-point constitutes an oncogene surveillance mechanism that can discriminate between normal mitogenic and constitutively activated oncogenic signals.

In mammals, there are three RUNX family genes (RUNX1, RUNX2, and RUNX3), which share a highly conserved Runt-domain. All three RUNX family genes show distinct tissue-specific expression patterns and are intimately involved in carcinogenesis (Ito et al., 2008; 2015). In addition, not only BRD2, but also other BRD family members (BRD3, BRD4, and BRDT) contain two distinct, highly conserved bromodomains that recognize acetylated lysine residues (Filippakopoulos et al., 2012). BRD family members are also mutated in various tumors (Belkina and Denis, 2012).

Recently, Lee et al. (2019b) showed that all RUNX family members form complexes not only with BRD2, but also with BRD3 and BRD4 2 h after serum stimulation, and that these complexes are subsequently destroyed. Importantly, when the constitutively active form of K-RAS (K-RAS^G12V) is expressed, all complexes were maintained for a long time, not only in the absence of serum stimulation, but also after serum stimulation. These results indicate that the RUNX and BRD families are involved in the lineage-specific R-point decision-making machinery, and thus contribute to oncogene surveillance (Fig. 2).

Considering the fundamental role of the R-point decision-making machinery during normal cell differentiation and proliferation, the decision-making machinery must be able to recognize the status and environment of the cell, enabling the cell to make an appropriate decision. Recent studies revealed that concerted action of RUNX family members, chromatin remodelers, histone modifiers, the basal transcriptional complex, Polycomb complexes, and cell-cycle regulators contribute to the R-point decision, in which RUNX proteins function as pioneer factors. These recently identified R-point decision-making processes provide important clues about fundamental questions regarding how chromatin is selectively opened and closed in a signal-dependent manner, as well as how cells recognize and defend against oncogenic RAS signals. Therefore, identification of RUNX family members as pioneer factors of the R-point and subsequent molecular events represents a significant advance in cancer biology, and could potentially provide a new strategy for the development of anti-cancer drugs.

S-C Bae is supported by a Creative Research Grant (2014R1A3A2030690) through the National Research Foundation (NRF) of Korea. J-W Lee is supported by Basic Science Research Program grant 2018R1C1B6001532.

The authors have no potential conflicts of interest to disclose.