Emerging Strategies to Inhibit the G1/S Transition for Cancer Therapy.

Introduction – G1/S control and the successes and shortcomings of current CDK4/6 inhibitors
The decision for eukaryotic cells to divide takes place at the G1/S transition of the cell cycle, and its deregulation is a hallmark of proliferating cancer cells (1,2). Entry into S phase is driven by cyclin dependent kinases (CDKs), a family of serine/threonine kinases that mediate many events throughout cell growth and division (3,4). CDKs usually function in cells in association with cyclins. Once active, the G1/S CDKs phosphorylate the retinoblastoma (RB) protein and related proteins p107 and p130, which releases their inhibition of E2F transcription factors and stimulates expression of E2F-dependent genes required for DNA replication and mitosis (1,2,5) (Fig. 1). In the canonical model of G1/S control, CDK4 and CDK6 together with Cyclin D first phosphorylate RB, which subsequently leads to Cyclin E expression and further RB hyperphosphorylation and inactivation by CDK2-Cyclin E complexes. G1/S CDKs are opposed by two families of protein inhibitors, which bind to CDK4/6 (INK proteins such as p16) or CDK2/4/6-cyclin complexes (CIP/KIP proteins such as p21 and p27). While it is becoming apparent that many cancer cells do not adhere to this canonical model (6), nearly all cancer cells have deficiencies in the CDK-RB-E2F control system, most commonly due to overexpression of CDKs and cyclin activators or deletions and mutations in CDK inhibitor proteins (3,7). These anomalies lead to aberrantly and prematurely high CDK activity. Consequently, chemical inhibition of CDKs has been a longstanding goal in cancer therapeutics for the past few decades.
CDK4/6 inhibitors are the first CDK inhibitors to have progressed to the clinic (8,9). These drugs, including palbociclib, ribociclib, and abemaciclib, have shown significant efficacy in hormone receptor-positive, HER2-negative breast cancer (10–16) (Fig. 2). In early clinical trials, CDK4/6 inhibitors were combined with fulvestrant, a selective estrogen receptor degrader, or letrozole, an aromatase inhibitor that blocks the production of estrogen in the body. Later clinical trials showed little differences in progression-free survival between the use of fulvestrant or letrozole with palbociclib (17).
This success of CDK4/6 inhibitors in breast cancer has led to many pre-clinical and clinical studies aiming at identifying other cancer types in which these inhibitors could be used. Based on years of research on the G1/S control pathway (9), one hypothesis was that RB wild-type tumors with high levels of Cyclin D may also respond favorably to CDK4/6 inhibitors. Unfortunately, in the MATCH trial, palbociclib did not improve outcomes for non-breast tumors with CCND1, CCND2, or CCND3 amplification (18). Still, some promising cancer types for the use of CDK4/6 inhibitors as single agent include liposarcoma (which often amplifies CDK4 (19)), mantle cell lymphoma (which harbors CCND1 translocation and is dependent on Cyclin D1 (20)), teratoma (21), and non–small cell lung cancer (particularly among tumors harboring a KRAS mutation (22,23)). Notably, the CDK4/6 inhibitor trilaciclib has been used in RB-mutant small-cell lung cancer (SCLC) to induce cell-cycle arrest in non-cancer cells in the tumor microenvironment. In this application, CDK4/6 inhibition arrests cycling hematopoietic cells to protect them from chemotherapy to provide myeloprotection (12,24,25).
Despite their initial promise and success in breast cancer, CDK4/6 inhibitors have not yet been approved for the treatment of any other cancer type. Even in breast cancer treatment, while the combination of CDK4/6 inhibitors with fulvestrant or letrozole significantly extends progression-free survival and overall survival, the additional benefit is typically limited to approximately one year, after which disease progression occurs (11–13). Hormone receptor-positive, HER2-negative breast cancer is therefore far from a solved problem, and the mechanisms of inherent and acquired resistance are being actively studied in both patients and preclinical models (26–28). A common thread among these resistance mechanisms is that cancer cells inactivate the G1/S control pathway through genetic or epigenetic changes that manipulate the concentrations of G1/S regulatory proteins. Examples include a loss of RB expression, an increase in Cyclin D or Cyclin E expression, and an increase in inhibitor-resistant CDK6 activity (26,27,29). In addition, while generally tolerated, the dosage of CDK4/6 inhibitors is still limited by some toxicities, most notably hematological and gastrointestinal (30). These toxicities likely result from target inhibition but also from off-target interactions. Considering the high homology in the active site among CDK paralogs, a challenge throughout the years has been to develop highly specific ATP-site inhibitors (Fig. 3). Therefore, a major goal of the field has been to find additional ways to target the G1/S transition that augment efficacy, limit resistance, and decrease side effects.
We discuss below a number of different strategies that have emerged in the past few years, including strategies to target other CDKs, to design alternative chemical approaches to inhibit CDKs, to combine CDK inhibition with inhibition of growth and growth signaling, and to target proteins that lie outside but impinge on the central G1/S control pathway. Our goal here is not to exhaustively present all the molecules that are being tested clinically and pre-clinically (which has been done in recent reviews (8,31–34)), but to discuss new concepts and their molecular basis using salient examples.

ATP-competitive inhibitors of other cell-cycle CDKs in pre-clinical models and in the clinic

CDK4-selective inhibitors
Growing evidence suggests that selective CDK4 inhibition may provide benefits over dual CDK4/6 inhibition, including especially the potential for less toxicity. While CDK4 and CDK6 have greater than 70% sequence homology and many functional similarities, they have differences in their expression pattern and interacting partners (35). In particular, functional genetic experiments have pointed to a more critical role for CDK6 in the hematopoietic lineage, which gives a rationale for why selective CDK4 inhibition may avoid inducing neutropenia, the main side effect of dual CDK4/6 inhibitors (36–39). In contrast, many solid tumors express higher levels of CDK4 than CDK6 and thus rely more on CDK4 activity (including breast cancer and lung adenocarcinoma) (40,41). Hence, selective CDK4 inhibitors may have therapeutic efficacy with more limited side effects in certain contexts.
The development of CDK4-specific inhibitors has taken advantage of the slight structural differences between CDK4 and CDK6, which result in distinct conformational dynamics (42). These ATP-competitive molecules, including atirmociclib (43), BGB-43395 (44), BEBT-209 (45), and CDDD2–94 (46) (Fig. 2), support the hypothesis that CDK4-specific inhibitors can have more limited side effects. Treatment with atirmociclib showed lower potency against hematopoietic stem/progenitor cells and resulted in less neutropenia compared to palbociclib in a dog model (43). Furthermore, treatment with atirmociclib showed potent tumor growth inhibition in 3D organoids and xenograft models of breast and prostate cancer (43). Clinical trials are currently underway to evaluate its efficacy in cancer patients (FourLight-3 trial, NCT06760637), and it is possible that higher doses of CDK4 inhibitors may be usable in patients (compared to CDK4/6 inhibitors), which might translate into more efficient tumor inhibition or delayed resistance. It is worth noting, however, that adverse events are still frequent in patients treated with atirmociclib (~80%), including neutropenia, leukopenia, anemia, and lymphopenia (47), which suggests that less selective CDK inhibitors (discussed below) may not be well tolerated by patients.

CDK2/4/6 inhibitors
While the motivating idea behind CDK4-selective inhibitors is to limit the side effects common to CDK4/6 inhibitors, the rationale behind pan CDK2/4/6 inhibitors is to overcome resistance mechanisms. A number of pre-clinical studies have identified increased CDK2 activity as a mechanism of resistance to CDK4/6 inhibitors (26,48–50), which could be delayed or stopped through the concomitant inhibition of CDK2 (51–53). A proof-of-concept study has shown that the CDK2/4/6 inhibitor ebvaciclib (PF-06873600, or F3600) (Fig. 2) has efficacy in multiple tumor models (52,54). Because immune cells such as T cells proliferate during immune responses against cancer, one limitation of potent cell-cycle inhibitors such as CDK2/4/6 inhibitors may be their anti-proliferative effects on immune cells. However, immune responses did not seem to be affected in syngeneic models (52). Additionally, MYC activity was found to predict ebvaciclib efficacy across multiple cell lineages (52). Ebvaciclib has been evaluated in a Phase 1/2a clinical trial in patients with breast and ovarian cancers, but it was reported that efficacy showed patient variability and that tolerability was less than expected from preclinical studies (55). Clinical trials with another CDK2/4/6 inhibitor, NUV-422, were also terminated (56) because patients developed uveitis. Despite these setbacks, other such inhibitors, including RGT-419B (31,57), are being developed.

CDK2 inhibitors
While CDK2/4/6 inhibitors would certainly target resistance mechanisms to CDK4/6 inhibition, so far these inhibitors appear toxic, and it may be more useful to specifically target CDK2. While CDK2 is not essential for the proliferation of many cell types and CDK2 knockout mice are sterile but otherwise viable, the hyper-activation of CDK2 can be a key driver of proliferation in some cancer cells (6,58,59). These observations suggest that CDK2 inhibition may have limited side effects and possibly potent anti-tumor effects in specific contexts. For these reasons, considerable effort has been made to develop CDK2 inhibitors, and research has been further motivated in the last few years as the limitations of CDK4/6 inhibitors have been realized. Structural similarities with other CDKs, particularly CDK1, have made it difficult to design CDK2-selective inhibitors (Fig. 3) (60). Nevertheless, a number of compounds have been developed by pharmaceutical companies using extensive medicinal chemistry approaches (32,61). These compounds have now been used in a number of preclinical studies and are in clinical trials (Fig. 2). Accumulating evidence has suggested that tumors with high levels of Cyclin E1 or tumors that rely on MYC may be more uniquely sensitive to CDK2 inhibitors (31,62). Similarly, cancer cells with high levels of the CDK4/6-specific inhibitor p16 may have low CDK4/6 activity and may be more sensitive to CDK2 inhibition (63–65). In these tumors, CDK2 activity may be viewed as a driver of tumorigenesis and cancer cells may be “addicted” to CDK2 activity, which may translate into greater clinical benefits (as suggested in pre-clinical studies (64–67)). In other contexts, CDK2 inhibition may be more useful when CDK2 activity is upregulated as a mechanism of adaptive resistance to CDK4/6 inhibitors. In the latter cases, cancer cells with intrinsic plasticity in their cell-cycle regulatory networks and CDK2 inhibition may only have transient inhibitory effects (32,68).
While CDK2 and CDK4/6 inhibitors both inhibit the G1/S transition by preventing RB phosphorylation, there are unique differences and challenges in targeting CDK2. In addition to the selectivity challenges due to structural similarity with CDK1 (Fig. 3), CDK2 phosphorylates many substrates beyond RB and has cellular functions independent of the canonical G1/S pathway (32,69,70). CDK2 inhibitors induce a G2-M arrest at higher concentrations, and the CDK2 dependence of this arrest suggests it may relate to inhibiting CDK2 functions in S phase (64–66). However, this G2-M arrest may also be due to potential off-target inhibition of CDK1 (65,66). Nevertheless, preclinical studies of novel CDK2-selective inhibitors supports the idea that they may be effective as anti-cancer agents in some contexts, and results further indicate that such inhibitors may be used to delay or prevent resistance to CDK4/6 inhibitors (66,67) (Fig. 2). Based on these pre-clinical observations, clinical trials with CDK2-selective inhibitors have been recently launched (32) (Fig. 2).

CDK7 inhibitors
In addition to targeting the canonical cell-cycle kinases CDK2, CDK4, and CDK6 for cancer therapy, recent work has also targeted the related multifunctional kinase CDK7. CDK7 acts as a CDK-activating kinase (CAK), phosphorylating and activating other CDKs, including the cell-cycle CDKs. In addition to this role in the cell cycle, CDK7 also acts as a component of the general transcription factor TFIIH, involved in transcription initiation and DNA repair (71–74) (Fig. 2). As such, CDK7 plays a pivotal role not only in cell-cycle regulation but also in controlling gene expression, particularly genes involved in proliferation and survival. Inhibiting CDK7 thus disrupts both cell-cycle progression and transcriptional programs, and CDK7 is a compelling target for cancers that rely on high transcriptional activity (34,75). These observations have led to the development of CDK7 inhibitors including THZ1 (76), a covalent inhibitor that shows potent activity in various preclinical cancer models including SCLC (77) and triple-negative breast cancer (78). Several additional CDK7 inhibitors are in early-phase clinical trials (34,79) (Fig. 2). The analysis of mice without CDK7 due to gene knockout in adult tissues shows that they have few phenotypes in non-cycling tissues but premature aging phenotypes in proliferating tissues, possibly due to loss of stem/progenitor cells (80). This observation suggests that patients may experience similar issues if treated with potent CDK7 inhibitors for long periods of time, which may limit the use of such inhibitors.

Alternative chemical approaches to CDK inhibition
An additional strategy to achieving higher efficacy and less toxicity in G1/S therapies may be to use nontraditional chemical inhibitory mechanisms to target the cell-cycle CDKs. The most widespread approach toward developing chemical CDK inhibitors has been the optimization of molecules that competitively bind the ATP site of the active CDK-cyclin heterodimer (Fig 4A–B). The last few decades have seen a large number of such Type I kinase inhibitors synthesized from diverse scaffolds (9,61). Following the success of palbociclib and related CDK4/6 inhibitors, comparable molecules, many of which have been progressed from a similar fused heterocycle core, have been developed for CDK2 inhibition. The underlying challenge of Type I inhibitors is achieving sufficient kinase selectivity in binding the highly conserved ATP site (Fig. 3). Especially in the case of CDK2, the high homology with CDK1, which is thought to be a source of toxicity in preclinical studies and clinical trials, is problematic. To address the limitations of Type I kinase inhibitors, several alternative approaches have been explored in the last few years including allosteric inhibitors, Type II inhibitors that target the inactive CDK monomer, molecules that inhibit substrate docking, and proteolysis-targeting chimeras (PROTACs) (Fig. 4C–G). In addition to achieving higher selectivity, potential advantages of these alternative strategies include resistance to mutations and better pharmacodynamic properties that may result from manipulating the cellular levels of CDKs and their regulatory complexes.
The development of alternative CDK inhibitors has taken advantage of our extensive understanding of their structure and function. For example, an allosteric site was identified in CDK2 upon probing for binding of the hydrophobic ligand and fluorophore ANS (81). Subsequent studies optimized potency through modifications to the anthranilic acid scaffold (82). This allosteric site is in the CDK2 N-lobe adjacent to the ATP site and lies between the DFG motif and the C-helix (Fig. 4C). Remarkably, molecules that bind this site inhibit Cyclin A binding but do not impact binding at the ATP site. Structural analysis suggests that these small molecules in the allosteric site interact with and rotate the C-helix in a manner that is incompatible with cyclin association. It has also been shown that molecules that act as Type II kinase inhibitors and bind preferentially to monomeric, inactive CDK2 also negatively impact Cyclin A binding (83–85) (Fig. 4E). Although they bind in the ATP site, these molecules extend toward and reposition the C-helix (85). Importantly, data for both the allosteric and CDK monomer-binding inhibitors show better selectivity for CDK2 over CDK1 compared to traditional Type I inhibitors (60,82). Although efficacy has yet to be shown in cancer cells, these observations support further development of compounds that act as CDK-cyclin inhibitors. A more traditional allosteric inhibitor site was recently identified following a screen of tryptoline acrylamides that react stereoselectively with a cysteine in Cyclin E2 (86). The site sits at the CDK2-Cyclin E interface and is formed in part by the activation loop (Fig. 4D). These covalent modifiers inhibit CDK2 activity in vitro and in cells, likely by repositioning the activation loop such that substrate binding is impacted. Considering the modification site is at the CDK-cyclin interface, molecules that access this allosteric site could be developed to contact specific cyclins and target specific CDK-cyclin complexes.
Cyclins function both by driving the CDK into its active conformation and also by recruiting target substrates. For recruitment, cyclins contain docking sites that are thought to improve catalysis and tune substrate selectivity of different CDK-cyclin complexes (87). These docking sites can also be targeted to inhibit cell proliferation. The best characterized docking site lies near the so-called MRAIL helix and binds an RxL sequence motif in both substrates and CIP/KIP inhibitors (88–90) (Fig. 4F). Early proof of concept studies with cell penetrating peptides and RxL peptide mimetics found that targeting the MRAIL site inhibits proliferation selectively in cancer cells with a deregulated RB-pathway (91,92). The transcription factor E2F1 contains an RxL motif and is phosphorylated by CDK2-Cyclin A in S phase to downregulate its activity (93–95). One possibility is that RxL-derived cyclin inhibitors further deregulate E2F1 activity, ultimately leading to apoptosis following an inability to properly complete DNA replication and exit S phase. Additional peptide mimetics and cyclic macromolecules have since been developed to target the RxL site with improved potency and specificity (96,97). In the case of the macrocycles, targeting Cyclin A and Cyclin B leads to the hyper-activation of E2F and of Cyclin B, respectively. This drives cells prematurely into mitosis to activate the spindle assembly checkpoint and eventually trigger cell death. The potential to engineer cyclin specificity in these compounds by leveraging slight structural differences in the different cyclin molecules is exciting, because it would allow targeting diverse cell-cycle regulatory mechanisms. Cyclin D also contains a docking site that is distinct from the MRAIL site and that it uses to target RB and its related family members for phosphorylation (98). Inhibitors of this interaction may selectively inhibit phosphorylation of RB family proteins in a manner that could be beneficial over broad CDK4/6 inhibition using Type I kinase inhibitors.
Another way to inhibit an oncogenic protein is to selectively target it for degradation by leveraging the cell’s endogenous ubiquitin-proteasome system. Proteolysis Targeting Chimera proteins (PROTACs) have been developed that target CDK4/6 or CDK2 for degradation (Fig. 4G) (61,99–106). These bifunctional, chimeric molecules have a competitive ATP inhibitor linked to a small molecule or peptide mimetic that binds an E3 ubiquitin ligase. By inducing proximity of the kinase and E3 ligase, PROTACs stimulate ubiquitylation and subsequent destruction of the enzyme. The first PROTACs to degrade CDK4/6 were synthesized by covalently linking palbociclib and ribociclib to pomalidomide, which recruits the E3 ligase cereblon (CRBN) (99–101). Remarkably, several of these early PROTACs displayed selectivity for CDK6 degradation over CDK4 degradation in cells despite comparable affinity of the kinase inhibitor component for both paralogs. Some evidence suggests this selectivity arises because CDK6 can form a ternary complex in cells with CRBN and the degrader molecule (99). On the other hand, several PROTACs subsequently generated using ligands for other E3 ligases, including VHL and IAP, also showed similar preference to degrade CDK6 (102,103). It may be that these other ligases also preferentially interact with CDK6 in the PROTAC complex, or it may be there are yet to be determined mechanisms in cells that facilitate CDK6 ubiquitylation or proteolytic degradation. In terms of inhibiting cell proliferation, most data point to similar efficacy of CDK4/6 degraders compared to the noncovalent competitive inhibitors from which they are constructed, and the sensitivity of different cancer cell types are similar (99,104,107). For example, model cell lines that have high levels of Cyclin D and low levels of p16 are sensitive to both classes of inhibitors, while cell lines with high Cyclin E levels are insensitive to both (107).
Several CDK2-directed PROTACs have also been developed and also use ATP-competitive inhibitors to induce proximity of the kinase with CRBN (61,105,106). These PROTACs show promising anti-proliferative activity in CDK2-sensitive cancer models, and at least one compound, NKT3964, has entered clinical trials (NCT06586957). Notably, some of these PROTACs show increased selectivity for CDK2 over CDK1 compared to the parent CDK2 inhibitor. As in the case of the CDK4/6 PROTACs, selectivity may result from interactions between ligase and kinase. There have also been reports of molecular glue degraders of CDK2 and Cyclin E (108,109). Like PROTACs, molecular glues induce proximity between their target and an E3 ligase, but molecular glues necessarily promote protein-protein interactions, which can confer selectivity, and have better drug properties (110). Interestingly, these molecules, which also show high selectivity and efficacy in preclinical studies of CDK2-dependent cancer models, do not interact with the kinase active site (108,109). CDK7 degraders have also been developed (111,112). One compound, built from the selective ATP-site inhibitor YKL-5–167, shows improved efficacy in cell proliferation assays compared to the parent CDK7-targeting molecule (111).

Cell growth, senescence, and G1/S inhibition
Mechanistically, CDK4/6 inhibitors halt breast cancer cell proliferation by blocking the kinase activity of the CDK4/6-Cyclin D complex, which is essential for phosphorylating and inactivating the retinoblastoma (RB) protein to drive G1/S progression (28). When CDK4/6 is inhibited, RB remains active, suppressing E2F-dependent transcription and causing a G1 cell cycle arrest. These drugs do not directly kill cancer cells but are cytostatic and stop them from dividing (28). Despite this survival, tumor growth slows or regresses in many patients. One explanation is that prolonged CDK4/6 inhibition leads to cellular senescence, a state of arrest accompanied by changes in gene expression and secretory activity (113,114). Senescent tumor cells may not only cease dividing but are also likely to attract immune clearance, contributing to the therapeutic response (115).
Yet why would CDK4/6 inhibitors induce senescence? These drugs block G1/S progression but do not directly cause DNA damage, which is the classical trigger for senescence in response to many cancer therapies, such as DNA-damaging agents (114). One possibility is that CDK4/6 inhibitors promote senescence through continued cell growth during G1 arrest (116–120). In cell culture, this has been shown to be the case, as CDK4/6 inhibition induces a cytostatic arrest, but cells continue to grow. As a result, they become excessively large, and their proteomes begin to resemble those of senescent cells (119). Importantly, recent studies have demonstrated that excessive cell size alone, independent of DNA damage, is sufficient to cause durable cell-cycle arrest (121–124). This cell-cycle arrest is unlikely to be due to increases in RB activity because RB concentration continually decreases during palbociclib-induced G1 arrest due to being targeted for degradation by the E3 ligase UBR5 (125–127). Instead, large cell size can induce a proteome imbalance, replication stress, or mitotic failure. For example, after four days of CDK4/6 inhibition followed by drug withdrawal, the largest cells remain durably arrested, while smaller cells are able to re-enter the cell cycle (119,120). This observation shows a clear size-dependent threshold for irreversible arrest. Thus, the initially reversible G1 arrest can convert into a senescent-like state due to excessive growth. In this way, rapidly growing cancer cells may be more susceptible to CDK4/6 inhibitor–induced senescence than other, more slowly growing cells in the tissue. While the link between growth and senescence could explain the efficacy of cytostatic CDK4/6 inhibitors, it is important to note that whether these inhibitors induce senescence clinically is currently unclear. The examination of samples from patients being treated with palbociclib showed increased cell-cycle arrest and reduced Ki67 straining, which is a marker of senescence (128). However, in some cells this apparent senescence may be reversible as Ki67 levels can recover following cessation of CDK4/6 inhibitor treatment (129).
Although combining cancer therapies that inhibit both cell proliferation and cell growth might seem intuitively beneficial, this strategy could be counterproductive when using CDK4/6 inhibitors. This is because of the cytostatic nature of CDK4/6 inhibitors, which arrest cells in G1 without directly inducing cell death. An increasing amount of evidence suggests that the long-term efficacy of CDK4/6 inhibitors depends on the induction of cellular senescence. In this context, combining CDK4/6 inhibitors with drugs that suppress anabolic growth pathways, such as mTOR inhibitors (e.g., everolimus, rapamycin) or PI3K inhibitors (e.g., alpelisib), could blunt this response by preventing the cell enlargement and the associated proteomic imbalances that likely trigger senescence. Indeed, senescence requires active biosynthesis and mTOR signaling. In cultured cells, inhibiting mTOR in combination with inhibiting CDK4/6 prevents the induction of the senescent state and instead cells are in a quiescent state which can re-enter the cell division cycle when the CDK4/6 inhibitor is removed (119,122–124). Similar results were found for CDK7 inhibitors, which also arrest cultured cells without stopping cell growth (124). Thus, while growth-targeting agents may be effective in other cancer settings, they could antagonize the senescence-based mechanism underlying CDK4/6 inhibitor therapy and diminish its therapeutic impact. It is also possible that the interaction between growth regulation and senescence is more complex in patients. Consistent with this view, a Phase Ib clinical trial examined the effect of combining Gedatolisib, a pan-PI3K/mTOR inhibitor, with the CDK4/6 inhibitor palbociclib and letrozole in HR+/HER2− advanced breast cancer patients (130). This trial demonstrated efficacy at least comparable to standard therapy with no additional safety concerns, thereby supporting advancement of the combination therapy to a Phase III study. Yet, another possibility is that the effect of CDK4/6 inhibitors might be enhanced by combination with senolytics that specifically induce death in senescent cells (131). Further clinical studies are required to determine how cell growth relates to CDK4/6 inhibition in patients.

Potential targets outside the core CDK-RB G1/S control pathway

DYRK1A
The Dual-specificity Tyrosine phosphorylation-Regulated Kinase 1A (DYRK1A) impinges on the RB G1/S control pathway and antagonizes cell-cycle entry and progression into S phase through at least two distinct mechanisms. DYRK1A phosphorylates Cyclin D family members in their C-terminus to promote their degradation (132–134). DYRK1A also drives assembly of the DREAM complex, which is a critical repressor of cell-cycle gene expression during quiescence and early G1 (135,136). DYRK1A phosphorylates the DREAM subunit LIN52 to increase affinity of LIN52 for the RB homolog p130. DYRK1A has long been considered a potential drug target, because its upregulation has been linked to down syndrome and associated defects in brain development (137). However, the potential for DYRK1A inhibition in cancer therapy has been less explored (138).
While DYRK1A is not yet an established target for cancer therapy, there are cellular contexts in which DYRK1A inhibition could be beneficial. For example, in cell culture, DYRK1A inhibition increases Cyclin D levels and increases expression of cell-cycle genes through loss of DREAM function (133,136). DYRK1A inhibition has been shown to inhibit proliferation of precancerous lymphocytes, because quiescence, facilitated by Cyclin D3 phosphorylation and degradation, is a necessary step in their maturation (134). Another rationale for exploring DYRK1A inhibition is that DREAM-dependent quiescence, which is promoted by DYRK1A activity, has been observed as a mechanism of cancer cell dormancy that confers resistance to chemotherapies (139–141). Considering its role in promoting cancer cell-cycle exit, the development of agonists of DYRK1A activity or inhibitors of its negative regulators may be a viable strategy toward cell-cycle inhibition. One possibility is to inhibit FAM53C, which has recently been reported to be an inhibitor of DYRK1A that promotes cell-cycle entry (142). Taken together, these recent studies argue that further investigation of DYRK1A regulation and function in specific cancer cell cycles is needed to determine if DYRK1A can be an effective cancer therapeutic target.

Targets that are synthetic lethal with RB loss
Preclinical and clinical data and genetic screens have identified RB as a critical mediator of CDK inhibition (28), which is in line with RB family proteins being the key CDK substrate target that effects transition to S phase. Considering that some cancers lack functional RB and others lose RB as a mechanism of CDK inhibitor resistance, there is a critical need for therapeutics that have efficacy and even specificity for RB null tumor cells (143). One obvious target would be E2F family transcription factors, which are activated by the loss of RB. Consistent with this view, loss or deletion of activator E2Fs can compensate for RB loss in several contexts (144,145, 146). There have been some reports of chemical E2F inhibitors (147), yet the potency and specificity have not been well optimized, and there have been no clinical trials to date. Several studies have aimed to identify targets other than E2F whose inhibition would be synthetic lethal with RB deletion. These studies take advantage of the deregulated cell cycle, inappropriate gene expression, and chromosome instability that occurs upon RB loss. Both inhibitor and genetic screens have identified Aurora kinases as synthetic lethal targets in RB mutant small cell lung cancer and triple negative breast cancer (TNBC) lines (148–150). This work suggests that RB loss renders cells dependent on factors that are regulated by Aurora A and B and contribute to the spindle assemble checkpoint and chromosome segregation. Inhibitors of another mitotic kinase, PLK1, and the DNA-damage checkpoint kinase CHK1 were also found to selectively kill RB negative TNBC cells over their isogenic but RB positive counterparts. Other synthetic lethal targets identified include the CDC25 phosphatase (151) and PARP enzymes (152), which is likely due to the accelerated cell cycle and accumulation of DNA damage that occurs with RB loss. Whether or not these findings can be translated to the clinic has yet to be seen.

CDK4/6 combination therapies
The concept behind the development of CDK4/6 inhibitors, based on decades of genetic and biochemical evidence (7), is that inhibition of these kinases and the subsequent activation of RB will lead to cell cycle arrest in cancer cells. While abemaciclib monotherapy is FDA-approved, CDK4/6 inhibitors are nearly always used in combination with endocrine therapy in breast cancer patients. This combination therapy may be effective due to the fact that estrogen signaling promotes the expression of Cyclin D1 (153). Thus, reducing estrogen signaling may contribute to lower CDK4/6 activity by reducing the levels of a key CDK4/6 partner in cancer cells. Thus, it is logical to think about developing CDK4/6 inhibitors in the context of combination therapies to target a wider variety of cancers.

CDK4/6 and HER2 dual inhibition
Another key oncogenic pathway in breast cancer (and other cancers) is activated by HER2 (Human Epidermal Growth Factor Receptor 2) (154). Activation of HER2 signaling promotes both cell growth and cell division, including via upregulation of Cyclin D (155,156). Pre-clinical studies have shown that CDK4/6 inhibition may cooperate with HER2 inhibition to suppress tumor growth (157). In addition, activation of CDK4/6 may mediate resistance to HER2 blockade (158). Based on this type of observations, combination therapies have been tested in patients. In the MonarcHER trial, for example, abemaciclib was combined with trastuzumab (a monoclonal antibody that serves as a HER2 inhibitor) and fulvestrant in patients with hormone receptor-positive and HER2-positive tumors (HR+/HER2+). This trial showed a tolerable safety profile and improved progression-free survival compared to chemotherapy plus trastuzumab (159), providing support to the combination therapy. Other trials have generated similar results (e.g., (160–162)). The early successes of these combination therapies bodes well for expanding the use of CDK4/6 inhibitors in breast and other cancers.

Inhibition of CDK4/6 and signaling pathways downstream of HER2 and RAS
The MEK/ERK pathway (Mitogen-activated protein kinase/Extracellular signal-regulated kinase) and the PI3K/AKT/mTOR pathway, two major signaling pathways that are downstream of HER2 and RAS, have been targeted in combination with CDK4/6 inhibitors due to their roles in promoting proliferation and tumor growth. Targeting MEK/ERK signaling alongside CDK4/6 inhibition disrupts complementary mechanisms that drive G1 progression. The MEK/ERK signaling pathway controls cell cycle progression in G1, including by upregulating Cyclin D1 expression (163,164). In addition, CDK4 activity is often critical in mutant KRAS-driven lung cancer (40). Thus, the combined inhibition of CDK4/6 and MEK/ERK signaling has long been thought to be an effective strategy against cancer, as this approach would target two parts of the CDK4/6-Cyclin D complex. This dual inhibition can trigger a deep cell-cycle arrest and senescence, which has been seen in multiple pre-clinical contexts (e.g. (23,165–169)). Managing toxicities and finding the right combination may be key issues in clinical trials, as there have not yet been any successful trials using combinations of CDK4/6 inhibitors and RAS/MEK/ERK inhibitors (e.g., (170)).
Manipulating the PI3K/AKT/mTOR pathway in CDK4/6-arrested cells presents a complex therapeutic dilemma. The PI3K/AKT/mTOR pathway controls cell growth. Senescence is thought to be induced in cells that are arrested in G1, but they keep growing in volume due to activity of the PI3K/AKT/mTOR pathway. In the context of G1-arrested CDK4/6-inhibited cells, it may not be beneficial to inhibit a pathway that would promote senescence (see (119,120) and discussed here (118) – see also above). However, the interplay between the PI3K/AKT/mTOR pathway and the CDK4/6 pathway is complex, and pre-clinical studies indicate that the combination therapy may be effective in multiple settings, including by altering the metabolism of cancer cells and inducing cell death (165,171–177). Clearly, cell growth is a key feature of cancer cells so that its manipulation could be useful for therapy. Clinical trials may identify the contexts in which inhibiting these two pathways can be beneficial to patients (e.g., (178)).

Inhibition of CDK4/6 and T-cell based immunotherapies
In addition to their effect on cancer cells, CDK4/6 inhibition may also affect immune responses (see for example (179) and discussed in (180)). These early observations have been confirmed in multiple tumor contexts (107,156,181–183), and responses can be seen in both T cells and other immune cells (184). As discussed above with trilaciclib in the context of SCLC, CDK4/6 inhibitors can have direct effects on the proliferation of immune cells. Other mechanisms may include the senescence-associated secretory phenotype (SASP), defined as the set of factors secreted by senescent cells, as these factors can directly signal to immune cells in the tumor microenvironment (184,185). However, inflammatory responses such as SASP may sometimes also inhibit immune responses against tumors (167), and the reprogramming of the tumor microenvironment by CDK4/6 inhibitors may induce immune checkpoints such as PD-L1 limiting T-cell responses (186). Liver toxicity may also be of concern in some combination therapies (ribociclib and anti-PD-1 (187)). Combining CDK4/6 inhibitors with immune checkpoint blockade is an active area of clinical research (187,188).

Concluding thoughts and questions in the field
In recent years, targeting the G1/S transition has emerged as a promising strategy for cancer therapy and has become the standard of care for HR-positive, HER2-negative metastatic breast cancer. The G1/S checkpoint plays a critical role in regulating cell-cycle progression, and its dysregulation is a hallmark of cancer. While CDK4/6 kinase inhibitors targeting the ATP pocket is the current method for inhibiting this transition, their limitations, resistance mechanisms, and off-target effects have driven the search for alternative strategies. A growing body of basic research has identified new druggable targets that can also impact the G1/S transition in ways that may overcome the limitations of CDK4/6 inhibition and compliment current approaches. The most closely related approach is to develop more specific ATP pocket inhibitors that target exclusively CDK4 and reduce unwanted side effects arising from CDK6 inhibition. However, we are most excited about the exploration of new mechanisms to expand the set of targets to inhibit the G1/S transition such as the cyclin-dependent kinase CDK2, its partner Cyclin E, the E2F transcription factor, and the Cyclin D partner of CDK4/6. New technologies such as PROTACs and molecular glues may make some of these proteins, which do not contain traditional druggable enzymatic domains, accessible. Alternatively, nontraditional targets may also be accessed by manipulating the activity of their upstream regulators. For example, Cyclin D degradation could in principle be stimulated by upregulating kinases that phosphorylate its degron (e.g. DYRK1A or GSK3β), by inhibiting its deubiquitylases (e.g. USP2, USP13, or USP22), or by recruiting or activating its E3 ligases (e.g. AMBRA1 or FBXO4) (180,189,190). Emerging therapies targeting this wider range of G1/S regulators can offer a spectrum of therapeutic options that could be combined with existing treatments to limit resistance and increased therapeutic efficacy for treatment of a wider variety of cancers.
Combination therapy offers many unexplored possibilities, and our summary here of work combining CDK4/6 inhibitors with other therapies is not exhaustive. A number of additional studies investigate combining CDK4/6 inhibitors with PARP inhibitors (e.g. (183)) or radiotherapy (discussed in (156)). If the main effect of CDK4/6 inhibitors is a G1 arrest in cancer cells, then strategies targeting other stages of the cell cycle (e.g., induction of DNA damage in S or G2/M) may not be very effective. However, CDK4/6 has other functions in cells than phosphorylation of RB, and it may be sometimes difficult to predict how some combination therapies may affect cancer cells. It will be critical to investigate how all the most recent CDK2/4/6/7 inhibitors act in combination with the approaches discussed here for CDK4/6 inhibitors, as response may be different for unexpected biological reasons.
Despite the excitement surrounding these emerging therapies, challenges remain in translating them into clinical practice. Many of these new inhibitors are still in the early stages of development, and their safety, efficacy, and optimal usage alone or in combination with other therapies pose significant hurdles and need to be thoroughly evaluated. For example, while it may be intuitive to inhibit both cell growth and the G1/S transition together, this could limit the efficacy of cytostatic G1/S inhibitors that likely rely on continued cell growth to trigger senescence and thereby limit cancer proliferation. Additionally, some of these strategies may still have limitations related to toxicity or resistance. An important open question, which many ongoing trials may potentially answer soon, is whether inhibition of both CDK4/6 and CDK2 has sufficient toxicity that will preclude widespread use. If this is the case, further exploration of other inhibition combinations, for example a CDK2/4 dual inhibitor, may be needed. It is likely that a more refined understanding of tumor-specific vulnerabilities and the development of personalized treatment regimens may be crucial to determine which patients are likely to experience the successful application of these new G1/S-targeting therapies. Importantly, all these new therapies developed from the increased basic understanding of G1/S control mechanisms that has emerged over the past decades. As our understanding of the cell cycle and its regulation deepens, these alternative approaches, and others yet to be discovered, can emerge to complement or replace existing therapies.