Quiescence, the ability to temporarily halt proliferation, is a conserved process that initially allowed survival of unicellular organisms during inhospitable times and later contributed to the rise of multicellular organisms, becoming key for cell differentiation, size control and tissue homeostasis. In this Review, we explore the concept of cancer as a disease that involves abnormal regulation of cellular quiescence at every step, from malignant transformation to metastatic outgrowth. Indeed, disrupted quiescence regulation can be linked to each of the so-called ‘hallmarks of cancer’. As we argue here, quiescence induction contributes to immune evasion and resistance against cell death. In contrast, loss of quiescence underlies sustained proliferative signalling, evasion of growth suppressors, pro-tumorigenic inflammation, angiogenesis and genomic instability. Finally, both acquisition and loss of quiescence are involved in replicative immortality, metastasis and deregulated cellular energetics. We believe that a viewpoint that considers quiescence abnormalities that occur during oncogenesis might change the way we ask fundamental questions and the experimental approaches we take, potentially contributing to novel discoveries that might help to alter the course of cancer therapy.
The myriad of possible mechanisms underlying cancer progression are more likely to be uncovered if they are approached from different perspectives. Here, we argue that cancer could be seen as a disease encompassing abnormal cellular quiescence rather than abnormal cell proliferation. Quiescence, the ability to temporarily halt proliferation, is an evolutionarily conserved process that initially allowed unicellular organisms to ‘wait’ for favourable environmental conditions, and, later on, was essential for acquisition of multicellularity and tissue specialization (O'Farrell, 2011). Even though quiescence might function as a tumour-suppressive mechanism under normal, healthy conditions, timely induction and loss of quiescence is also crucial for malignant transformation and tumour progression, as anomalies in quiescence regulation are involved in all the main traits of cancer, which have been systematically organized as the ‘hallmarks of cancer’ (Hanahan and Weinberg, 2000, 2011) (Fig. 1A).
In this Review, we present the current view of quiescence and its links to each hallmark of cancer. Although quiescence and proliferation could be seen as two sides of the same coin, a perspective more focused on the abnormalities associated with quiescence might change the way we ask fundamental questions in cancer research and our experimental approach, leading to advances in the field.
What is quiescence, and what are its impacts?
Quiescence – the ability to remain quiet – has a broad definition in cancer and cell biology, and can be divided into cellular quiescence and population quiescence (Taylor et al., 2013; Wells et al., 2013). Cellular quiescence refers to individual cells that are not actively proliferating but retain the ability to do so (Coller et al., 2006). Population (bulk) quiescence refers to a normal or tumour tissue that does not display a net increase in the number of cells because any cell proliferation is counteracted by cell death (Hinsull and Bellamy, 1981; Naumov et al., 2006). Here, we will focus on cellular quiescence (hereafter referred to as quiescence).
Quiescence is a tightly regulated process that goes beyond the absence of proliferation (Coller et al., 2006; Matson and Cook, 2017). Quiescent cells might not be dividing, but they are definitely not out of action: to remain quiescent, cells not only repress genes needed for DNA replication and cell division, but also sustain the transcription of anti-apoptotic, anti-senescence and, often, anti-differentiation genes, indicating that quiescent cells do acquire new properties (Coller et al., 2006; Matson and Cook, 2017; Sang et al., 2008).
Different organs deploy different strategies regarding quiescence. Cell renewal in some organs (including the skin, intestine and the haematopoietic system) relies on a specialized, mostly quiescent and undifferentiated cell type: the adult stem cells. Upon appropriate stimuli, stem cells proliferate asymmetrically, with one cell remaining as a quiescent stem cell and the other one proliferating and giving rise to the differentiated progeny that is characteristic of that organ (Rieger and Schroeder, 2012; Van Der Flier and Clevers, 2009). Other organs such as the liver present a different strategy: hepatocytes themselves re-engage the cell cycle, proliferate and replace cells lost in physiological process or even upon partial hepatectomy (Michalopoulos and Bhushan, 2021; Yang and Xu, 2011).
For how long must a cell be arrested to be considered quiescent? Let's not forget that cell cycle arrest per se does not characterize quiescence – the arrest must be reversible (Coller, 2011; Coller et al., 2006; Fiore et al., 2018; Sagot and Laporte, 2019; Sang et al., 2008). Once this premise has been fulfilled, the minimum arrest time to be considered quiescent depends on the usual cell cycle length for that particular cell, which in turn is linked to cell-intrinsic properties (such as organism and cell type) and cell-extrinsic factors (such as cell context) (Matson and Cook, 2017). For human cells, whose cell cycle is estimated to take 24 h (based on the average doubling time in culture; Cooper, 2000), the arrest would have to last at least for a few days.
How long are cells able to remain quiescent? Although experimental data indicate that quiescence might reach different depths (Kwon et al., 2017), and that the longer a quiescent cell remains arrested, the harder it might be for it to re-engage in proliferation (Fujimaki and Yao, 2020), quiescent cells can resume their cell cycle after years or even decades (Zhang et al., 2013). This is of particular importance for physiological processes, such as tissue homeostasis (So and Cheung, 2018), regeneration (Kalamakis et al., 2019; Rumman et al., 2015) and long-term immune memory (Chang and Radbruch, 2021), as well as for pathological conditions, including dormancy and reawakening of disseminated tumour cells (Sosa et al., 2014; Zhang et al., 2013).
These are not trivial questions to answer, as there are no universal markers for quiescence in complex organisms (Coller, 2011; O'Farrell, 2011). The best way to identify quiescent cells relies on the absence of proliferation traits, such as cell-endogenous activating molecules of the cell cycle, DNA synthesis and mitotic features (O'Farrell, 2011). The lack of specific markers for quiescence – such as those available for the identification of proliferating cells – is at least in part due to the vast morphological and functional heterogeneity found in quiescent cells (Coller, 2011; Fiore et al., 2018) but could also be due to the fact that there has been less research aimed at the elucidation of mechanisms governing quiescence, as compared to the number of studies focused on proliferation.
Quiescence and the hallmarks of cancer
Cancer is a general name for a group of diverse diseases, each presenting several unique features (or molecular circuits leading to the same feature) that make every tumour unique. However, there are also common features that bring all cancers together, and these traits have been complied as the hallmarks of cancer, which are essentially a series of premises for transformation of normal cells into a metastatic cancer (Hanahan and Weinberg, 2000, 2011). Despite being organized as eight principles and two enabling features, the more we understand cancer progression, the more we acknowledge that many of these hallmarks – if not all – are highly interrelated and typically do not occur separately; hence, they cannot be disentangled from one another. This is also clear when we consider how these traits can affect – or be affected by – quiescence. Below, we present sections mirroring the hallmarks of cancer, put forth originally by Hanahan and Weinberg (2000, 2011), but viewed instead from the standpoint of quiescence and its contributions to each of these hallmarks.
Sustaining proliferative signals – inability to remain quiescent
The commitment to proliferate or to become quiescent is mutually exclusive (Matson and Cook, 2017), and failure to sustain quiescence has been recognized as a hallmark of cancer itself (O'Farrell, 2011; Roche et al., 2017). Most cells in our body are not dividing, and the extracellular signals that trigger physiological proliferation are tightly regulated (Barr et al., 2016; Bissell and Hines, 2011; Fiore et al., 2018). Mitogenic signals are biochemically integrated within the cell, where the balance between pro- and anti-proliferative signals results in a binary decision (to cycle or to remain quiescent), which is controlled by a bistable switch involving retinoblastoma (Rb) and transcription factors of the E2F family (Box 1) (Blagosklonny and Pardee, 2002; Matson and Cook, 2017; Yao, 2014).
The restriction point of the cell cycle was first identified almost 50 years ago (Pardee, 1974). Cells rely on growth factors to proliferate, but once a certain point – the restriction point – is passed, the cell is committed to proliferation, and mitogenic signalling is no longer required to complete cell division (Blagosklonny and Pardee, 2002; Pardee, 1974). Molecularly, the restriction point is characterized by hyperphosphorylation of the tumour suppressor Rb. Hypophosphorylated Rb binds to E2F transcription factors, inhibiting their transcriptional activity. Upon Rb phosphorylation, the E2F proteins are released and drive the expression of positive cell cycle regulators, including cyclins and E2F proteins themselves (Blagosklonny and Pardee, 2002). Proliferation and quiescence stimuli, which favour or prevent Rb hyperphosphorylation, respectively, are integrated by pathways upstream the Rb–E2F switch and determine cell fate with regard to whether a cell divides or remains quiescent (Yao, 2014).
Cancer cells usually have some of the pro-proliferation pathways switched on and some of the anti-proliferation pathways switched off (Hanahan and Weinberg, 2011). For example, the receptor tyrosine kinase (RTK)–RAS–mitogen-activated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K) pathways are amongst the most constitutively activated oncogenic pathways across many different tumour types (Sanchez-Vega et al., 2018). Conversely, these pathways are known to be widely inactivated in quiescence (Beliveau et al., 2010; Coller et al., 2006; Fiore et al., 2017; Gookin et al., 2017). Members of these pathways – such as cyclins, cyclin-dependent kinases (CDKs), PI3Ks, AKT kinases and MAPKs – are considered to be proto-oncogenes and are known to promote proliferation, and their upregulation or increased activity in cancer cells drive quiescence escape (Barkan et al., 2010; Fiore et al., 2017; Leung and Brugge, 2012; Sanchez-Vega et al., 2018). Similarly, antagonists of these pathways – tumour suppressors such as p53 (also known as TP53), CDK inhibitors (CKIs), RB1 and PTEN – promote cell cycle arrest in normal tissues (Barr et al., 2016; Coller et al., 2006; Kwon et al., 2017; Worby and Dixon, 2014), but they are often lost or inactivated in cancer cells (Jia and Zhao, 2019; Millis et al., 2016; Sanchez-Vega et al., 2018), contributing to quiescence escape. Furthermore, tumours produce their own growth factors, and cancer cells are hypersensitive to them (Hanahan and Weinberg, 2011; Jia and Zhao, 2019; Sanchez-Vega et al., 2018). Thus, the spatiotemporal regulation of proliferation found in homeostatic tissues is lost because cancer cells are no longer able to remain quiescent.
Evading growth suppressors – bypassing quiescence-inducing signals
Complementarily to the features described above, cancer cells not only become self-sufficient in mitogenic signals, they also deploy strategies to shut down growth suppressors, which is a tactic to hinder quiescence. A plethora of cell-intrinsic and cell-extrinsic signals not only supress proliferation but also actively induce quiescence (Fiore et al., 2018). These signalling pathways involved in proliferation suppression and/or quiescence induction go awry in cancer (Bissell and Hines, 2011; Fiore et al., 2018), resulting in cancer cells becoming much less responsive to quiescence-inducing signals (Fiore et al., 2017).
For instance, laminins present in the basement membrane are key inducers of epithelial cell quiescence and differentiation (Fiore et al., 2018; Weaver et al., 1997). During malignant transformation, the extracellular matrix (ECM) undergoes extensive remodelling, resulting in a decrease in laminin and an increase in collagen and/or fibronectin content. This shifts the previously quiescence-inducing microenvironment towards a proliferation-permissive scenario (Bonnans et al., 2014; Crist and Ghajar, 2021; Fiore et al., 2018). In parallel, owing to other abnormal molecular features (for example, loss of tumour suppressor genes), tumour cells not only become insensitive to laminin-induced quiescence, but highly aggressive cancer cells also have their molecular circuitries so corrupted that laminin binding can actually contribute to cell proliferation, rather than to cell cycle arrest (Friedland et al., 2007; Ramovs et al., 2017; Zahir et al., 2003). For example, upon binding to their ECM substrate, integrins, which are the cellular receptors for laminins, are known to activate focal adhesion kinases (FAKs), SRC, ILK and other signalling pathways linked to cell survival, growth and proliferation (Abdel-Ghany et al., 2002; Gehler et al., 2013). Whereas normal cells can block these downstream proliferative signals upon binding to laminins by, amongst other strategies, downregulating PI3K–AKT (Fiore et al., 2017) or MEK–ERK (Beliveau et al., 2010) signalling, cancer cells often present mutations leading to constitutive activation of both pathways (Millis et al., 2016; Sanchez-Vega et al., 2018) and, thus, are usually no longer able to inhibit proliferation. An additional example of a mechanism responsible for triggering quiescence involves TGF-β, which has tumour-suppressive or tumour-promoting effects, depending on cell context (see Box 2 for more details).
The role of TGF-β family members in cell cycle inhibition and epithelial to mesenchymal transition (EMT) during morphogenesis is well established (Massagué, 2008; Morikawa et al., 2016; Nelson et al., 2006). Normal quiescent cells are often responsive to this signalling pathway, but members of this pathway (including TGF-β receptors) are frequently mutated across multiple cancer types (Coller et al., 2006; Jia and Zhao, 2019; Sanchez-Vega et al., 2018). Such mutations could make cancer cells less sensitive to TGF-β-mediated, quiescence-inducing signalling (Morikawa et al., 2016; Tang et al., 2003). However, in more advanced stages of tumour formation, the tumour-suppressive role of TGF-β signalling is not only bypassed by cancer cells, but can also lead to the opposite effect; indeed, increased TGF-β signalling has been linked to drug resistance, cell migration, invasion and metastasis (Brown et al., 2017; Kobayashi et al., 2011; Morikawa et al., 2016; Tang et al., 2003). The molecular mechanisms underlying this apparent disparity are highly cell- and context-dependent, as they are affected by other signalling pathways and microenvironmental features (for an excellent review, please refer to Massagué, 2008). Inactivation of core components of TGF-β signalling (such as the TGF-β receptors) impairs quiescence, whereas alterations affecting only the cytostatic branch of this pathway hamper quiescence induction and maintenance, but nevertheless still allow EMT, which is key for metastasis (Massagué, 2008). TGF-β-induced quiescence itself could contribute to disease progression even at late stages. An interesting mechanism has been reported in squamous cell carcinoma, where a quiescent population survives treatment with 5-fluorouracil (a cytotoxic chemotherapeutic drug that disrupts RNA synthesis), driving disease relapse. Differential gene expression analysis of quiescent versus proliferating cells has shown that TGF-β is responsible for quiescence induction in these cells (Brown et al., 2017).
The combination of the loss of extrinsic quiescence-inducing signals and the abnormal molecular signalling displayed by transformed cells leads to the evasion from growth suppressors and, consequently, the escape from quiescence observed in cancer.
Resisting cell death – quiescence as a mechanism of survival
Programmed cell death is one of the last resources in tumour suppression – when everything else fails, cell death comes into play. Non-cycling cells are more resistant to cell death (Borst, 2012; Naumov et al., 2002; Sosa et al., 2014), which might be due to mechanisms associated with the quiescent state (Phan and Croucher, 2020). These include switches in cell metabolism (Coller, 2019; Takubo et al., 2013), autophagy modulation (Vera-Ramirez et al., 2018; Wang et al., 2018), increased anti-apoptotic and survival signals (Friedland et al., 2007; Ranganathan et al., 2006; Zahir et al., 2003), decreased levels of pro-apoptotic factors (Coller, 2011; Hanahan and Weinberg, 2011; Ranganathan et al., 2006) and reduced immunogenicity (Malladi et al., 2016), as well as quiescence-specific niche contexts (Carlson et al., 2019; Sistigu et al., 2020).
Furthermore, standard chemotherapy often targets proliferative cells; hence, the ability to enter a quiescent state safeguards cancer cells, which later can resume the cell cycle. In this sense, an important mechanism known to induce cell cycle arrest that has been linked to drug resistance in cancer cells relies on TGF-β signalling (Box 2).
Furthermore, quiescent cells do display some particular features – in addition to the lack of proliferation – that contribute to drug resistance. Using a model of hepatocarcinoma dormancy in vitro, it has been observed that quiescence induces a p38-dependent signalling pathway, leading to increased activation of BiP (also known as HSPA5) and PERK (also known as EIF2AK3), which are proteins involved in endoplasmic reticulum stress response that prevent the activation of Bax, a pro-apoptotic factor that allows for cytochrome c release from mitochondria. This dormancy-related signalling pathway configures a mechanism of resistance to cell death induced by the chemotherapy drugs etoposide and doxorubicin (Ranganathan et al., 2006). Complementarily, Rac–NF-κB signalling has been linked to resistance to cell death in 3D models of quiescent mammary acini subjected to treatment with TRAIL (also known as TNFSF10), a cytokine that triggers apoptosis, or Taxol, a chemotherapeutic that inhibits mitosis, and to anchorage-independent survival of breast cancer cells (Friedland et al., 2007; Zahir et al., 2003). Another key adaptation that contributes to survival is a metabolic shift towards autophagy, which occurs in both normal and cancer cells when they are quiescent (see Box 3).
Autophagy is a metabolic adaptation used as a survival strategy, allowing cells to cope with metabolic stress by degrading intracellular components to recycle precursors of macromolecules, which can then be used as building blocks for other cellular components and for energy production (Chang, 2020). Molecular factors controlling autophagy, such as ATG7, p62 (also known as SQSTM1), BECN1, LC3A (also known as MAP1LC3A) and LC3B (also known as MAP1LC3B), are upregulated in normal and cancerous quiescent cells, and this shift towards increased autophagy contributes to cell survival (Coller et al., 2006; García-Prat et al., 2016; Ho et al., 2017; Mcleod et al., 2012; Schneider and Cuervo, 2014; Sosa et al., 2014; Vera-Ramirez et al., 2018). For example, loss of autophagy in haematopoietic stem cells leads to accumulation of mitochondria and a metabolic shift towards OXPHOS, which in turn promotes cell differentiation and impairs stem cell self-renewal (Ho et al., 2017). Thus, autophagy-mediated mitochondria clearance maintains stem cell quiescence and stemness by avoiding a metabolic shift required for expansion and differentiation (Ho et al., 2017). Similarly, in 3D and in vivo models of breast cancer, dormant cells, unlike their proliferative counterparts, display high levels of autophagy (Vera-Ramirez et al., 2018). Pharmacological or genetic inhibition of autophagy in these models leads to cell death only in dormant cells, due to the accumulation of damaged mitochondria and oxidative stress, highlighting the role of autophagy in maintaining cellular homeostasis during quiescence (Vera-Ramirez et al., 2018). Conversely, using breast cancer models of lung metastasis, a recent report has shown that autophagy impairment switches cancer cells towards a more proliferative phenotype, increasing the size of the metastatic lesions but not necessarily the number of foci (Marsh et al., 2020). Here, autophagy induction could actually reduce metastatic burden (Marsh et al., 2020).
Although these data seem controversial at first, inhibition of autophagy might cause cancer cells to leave their quiescent state instead of evoking cell death, which would ultimately increase tumour burden. However, increased autophagy could contribute to dormancy of disseminated cells, which might be detected initially as an apparent reduction in metastatic burden. Performing experiments designed to transiently modulate autophagy activity and to accommodate the subsequent recovery time, which would allow the awakening of dormant cells, could help to confirm this idea.
Quiescence as a mechanism of survival is of particular importance for residual cancer cells in the primary tumour or for disseminated tumour cells (DTCs), where their quiescent status ultimately contributes to chemoresistance, long-term survival of malignant cells and metastasis outbreak (Naumov et al., 2003; Sosa et al., 2014; Wang et al., 2018; Wells et al., 2013). Quiescence is, in fact, recognized as the major mechanism underlying drug resistance, and neoadjuvant therapies selectively targeting dormant tumour cells might efficiently eradicate residual cancer cells (Borst, 2012; Damen et al., 2021; Naumov et al., 2003; Phan and Croucher, 2020). Of note, breast cancer cells lying dormant in the perivascular niche die upon treatment with integrin-blocking antibodies, indicating the relevance of integrin–ECM interactions in sustaining viability in quiescent cells (Carlson et al., 2019).
Enabling replicative immortality – quiescence avoids senescence, and immortalization is linked to loss of quiescence
The transcriptome of quiescent cells includes the expression of anti-senescence genes (Coller et al., 2006; Matson and Cook, 2017; Sang et al., 2008), which along with other anti-senescence strategies, such as autophagy modulation (García-Prat et al., 2016), could contribute to cell immortalization, as senescence is considered a degenerative step towards cell death (Terzi et al., 2016).
Another key event in cell immortalization involves the telomeres – the keepers of genome integrity. Telomeres are shortened at every round of cell division, and when they arrive at a critical size, DNA-damage checkpoints are activated, culminating in cell senescence or apoptosis (Maciejowski and de Lange, 2017; Shay and Wright, 2010). However, in pre-cancerous cells, in which DNA damage-induced cell cycle arrest is compromised, telomere shortening might trigger telomere crisis, characterized by extensive DNA fragmentation, random repair and hypermutation (Maciejowski and de Lange, 2017; Shay and Wright, 2010). Most cells undergoing telomere crisis succumb to cell death, but those that are able to escape from it typically do so by reactivating telomerase, which alleviates the DNA-damage checkpoint and cell death signalling triggered by short telomeres (Dagg et al., 2017; Maciejowski and de Lange, 2017; Shay and Wright, 2010). Therefore, restoration of telomerase activity in cells undergoing telomere crisis rescues cells from cell death and gives rise to an immortalized cell with a highly rearranged genome that is ready to re-enter the cell cycle (Dagg et al., 2017; Maciejowski and de Lange, 2017; Shay and Wright, 2010). Cancer incidence considerably increases as we age (National Cancer Institute – Age and Cancer Risk, https://www.cancer.gov/about-cancer/causes-prevention/risk/age; Nordling, 1953), not only due to the replicative mutations acquired over time (Tomasetti et al., 2017), but also because malignant cells might arise from preceding quiescent cells that overcame telomere crisis (Maciejowski and de Lange, 2017; Shay and Wright, 2010). Additionally, an aged microenvironment containing cells with dysfunctional telomeres can induce loss of quiescence and tumorigenesis (Lex et al., 2020; Song et al., 2012). Quiescence not only protects telomeres from attrition by restraining proliferation, but also prevents their recombination or fusion by anchoring telomere clusters to the nuclear envelope, as has been reported in yeast (Coulon and Vaurs, 2020; Maestroni et al., 2020). This is likely to occur in mammals as well, as specific nuclear compartments that contain clustered telomeres anchored to the nuclear envelope have been observed in adult neural stem cell populations, but not in their differentiated or proliferating counterparts (Cebrián-Silla et al., 2017), ensuring telomere maintenance and homeostasis in quiescent cells. In sum, upon successive failures of tumour-suppressive mechanisms, including maintenance of adequate quiescence itself, the outcome of telomere crisis is the loss of quiescence in a highly mutated malignant cell.
Inducing angiogenesis – endothelial cells escape quiescence, and the vascular environment regulates the tumour cell cycle
Tumours promote angiogenesis by producing and secreting factors that either trigger the sprouting of new vessels from nearby endothelia or recruit bone marrow-derived endothelial precursors (Folkman, 2007; Yano et al., 2017). In both cases, endothelial cells (or their precursors) must exit their quiescent state and expand to provide the building blocks to assemble new blood vessels.
Blood vessels provide the nutrients needed to sustain tumour growth and offer a route for metastatic dissemination, but the vascular niche also has an active role in suppressing or stimulating tumour growth (Butler et al., 2010). Upregulation of angiogenic factors and inflammatory cytokines released either by tumour or inflammatory cells activates endothelial cells and triggers the sprouting of vessels, thereby configuring a microenvironment that awakens quiescent cells in established tumours (Butler et al., 2010; Yano et al., 2017), as well as dormant DTCs (Ghajar et al., 2013). The niche found in sprouting vessels is rich in growth factors, such as BMP2, BMP4, FGF2, PGF, PDGFβ, POSTN, TGF-β1, inflammatory cytokines and chemokines, including IL-8, IL-6, G-CSF (also known as CSF3) and GM-CSF (also known as CSF2), as well as adhesion molecules, such as ICAM1 and VCAM1 (Butler et al., 2010; Ghajar et al., 2013). Conversely, the niche found in mature vessels, which is rich in anti-angiogenic thrombospondins, supports cancer cell quiescence (Butler et al., 2010; Ghajar et al., 2013; Yano et al., 2017).
Taken together, the studies described above demonstrate that cancer cells themselves or tumour-promoting inflammatory cells trigger loss of quiescence of progenitor or endothelial cells, promoting angiogenesis, and that the sprouting of new blood vessels provides an anti-quiescence, pro-proliferation microenvironment for cancer cells. Combined, both processes linked to loss of quiescence foster tumour growth, cancer progression and metastasis.
Activating invasion and metastasis – disseminated cancer cells escaping quiescence
Metastasis is the most striking (and deadly) event during cancer progression (Loberg et al., 2007). Cancer is often considered a disease of abnormal proliferation, but especially for metastasis, quiescence plays a crucial role. After seeding into secondary organs, cancer cells undergo dormancy (quiescence), which might last from only a few days to up to a few decades, before metastatic outgrowth (Sosa et al., 2014). Indeed, most DTCs actually remain dormant for a lifetime (Bissell and Hines, 2011; MacKie et al., 2003; Sistigu et al., 2020; Sosa et al., 2014). The precise mechanisms driving quiescence induction and reawakening at secondary sites are still poorly understood, but they likely involve cell-intrinsic (including epigenetic regulation and shifts in metabolism) and cell-extrinsic (alterations in the microenvironment, such as ECM remodelling and immunosurveillance and/or immunoevasion) events (Anderson et al., 2019; Phan and Croucher, 2020; Sosa et al., 2014). Metastatic seeding can take place early in cancer progression, occurring before diagnosis, or even preceding the assembly of a clinically detectable primary tumour (Hosseini et al., 2016; Riethmüller and Klein, 2001). Thus, there is a need for effective therapeutic strategies targeting these prevalent and widespread metastatic seeds (Anderson et al., 2019). Dormancy is a crucial component of the natural disease course of malignant tumours (Giancotti, 2013; Sistigu et al., 2020). It allows cancer cells to adapt to the new (and potentially inhospitable) microenvironment (Phan and Croucher, 2020; Valcourt et al., 2012), contributes to chemoresistance (Borst, 2012) and ultimately is a major driver of disease relapse (Phan and Croucher, 2020; Riethmüller and Klein, 2001). Although the molecular mechanisms regulating metastatic dormancy and reawakening are mostly unknown (Anderson et al., 2019; Eckhardt et al., 2012), some of the underlying pathways are starting to be elucidated. For example, astrocyte-derived laminin-211 inhibits YAP1 signalling in dormant breast cancer cells in the brain (Dai et al., 2022), and TGF-β signalling induces and sustains dormancy in several tumour types (Box 2). Undoubtedly, well-timed quiescence (dormancy) induction and cell cycle reactivation are required for cancer progression and metastasis.
Avoiding immune destruction – quiescence-induced immune evasion
The notion that the immune system can fight tumours and that immune cells infiltrate malignant tissues goes back to the 19th century (Shalapour and Karin, 2015). Later on, it became clear that a balance between anti- and pro-tumorigenic immune-related events dictates whether or not a cancer emerges and progresses (Shalapour and Karin, 2015). If tumours arise, pro-tumorigenic immunomodulation dominates (Grivennikov et al., 2010; Shalapour and Karin, 2015). Apart from pro-tumorigenic inflammation (discussed in a dedicated section below), strategies to escape immune surveillance are critical for cancer development. These might include recruitment of immunosuppressive cells to the tumour microenvironment, natural killer (NK) cell quiescence and cancer cell-autonomous mechanisms (Correia et al., 2021; Grivennikov et al., 2010; Phan and Croucher, 2020; Shalapour and Karin, 2015; Sistigu et al., 2020). The ability to become quiescent is advantageous here: a wealth of evidence indicates that immune evasion is highly dependent on quiescence, and that the immune-selective pressure could favour pre-malignant clones that originate from long-lasting cells with a more quiescent profile (Agudo et al., 2018; Boyd and Rodrigues, 2018; Grivennikov et al., 2010; Malladi et al., 2016).
Quiescent cells – regardless of their malignant status – downregulate major histocompatibility complex class I (MHC-I), the cell-surface receptor responsible for presenting self-antigens (peptides derived from intracellular components) to the immune system; hence, quiescent cells are antigen-presentation deficient, and therefore, they are less likely to be attacked by the innate immune system (Agudo et al., 2018; Boyd and Rodrigues, 2018; Malladi et al., 2016). In a mouse model of experimental lung metastasis, dormant breast cancer or lung cancer cells have been observed to enter a state of self-imposed immune-evasive quiescence. Upon expression of DKK1, a WNT inhibitor, these cancer cells become dormant, and this is accompanied by downregulation of NK-cell ligands on the cell surface, resulting in escape from NK surveillance (Malladi et al., 2016). Therefore, descendants from these cells, which might be less immunogenic, would be favoured (i.e. not targeted and not eliminated) during immunoediting (Boyd and Rodrigues, 2018; Malladi et al., 2016). Recently, a novel mechanism involving the induction of NK cell quiescence and metastatic outgrowth has also been described. Using spontaneous models of breast cancer metastasis to liver, it has been shown that activated hepatic stellate cells secrete CXCL12, which triggers NK cell quiescence and leads to increased metastatic burden (Correia et al., 2021). Taken together, this body of evidence highlights the interplay between quiescence, immune evasion and cancer progression.
Deregulating cellular energetics – metabolic reprogramming and the proliferation–quiescence switch
Proliferation imposes high energetic demands on a cell, and not surprisingly, there is overlap between signalling pathways involved in metabolism and the cell cycle (Kalucka et al., 2015; Koppenol et al., 2011; Locasale and Cantley, 2011). Activation of well-known oncogenes (including those encoding SRC, RAS proteins, MYC, PI3Ks, AKT kinases and hypoxia-inducible factors) or loss-of-function of tumour suppressors (such as p53, VHL and PTEN) not only heavily influence cell cycle progression but also modulate cell metabolism (DeBerardinis et al., 2008; Kalucka et al., 2015; Koppenol et al., 2011).
The most remarkable – and certainly the best studied – metabolic alteration in cancer is the Warburg effect (or aerobic glycolysis), whereby even in the presence of oxygen, tumours metabolize approximately ten times more glucose to lactate (i.e. fermentative glycolysis) than normal healthy tissues, which rely mostly on oxidative phosphorylation (OXPHOS) to obtain energy (Kalucka et al., 2015; Koppenol et al., 2011). Normal stem cells, but perhaps no other quiescent cells, also rely on glycolysis as their major energetic source, but interestingly, they do need to undergo a metabolic rewiring towards OXPHOS to differentiate (Takubo et al., 2013; Yu et al., 2013). Although anaerobic glycolysis in stem cells might be a result of the hypoxic microenvironment they experience, it also indicates that at least some types of quiescent cells exhibit a glucose metabolism, although not coupled with biomass production, closer to what is found in proliferating cells. In both cases, this metabolic adaptation is thought to protect the cells against damage by reactive oxygen species (ROS) that can be generated during mitochondrial respiration (Coller, 2019; Valcourt et al., 2012).
Dormant cancer cells also display a glycolytic phenotype and, just like some normal quiescent cells, rely on glucose, glutamine and increased autophagy for survival (Kalucka et al., 2015; Koppenol et al., 2011; Sosa et al., 2014; Vera-Ramirez et al., 2018). Also similarly to normal stem cells, dormant cancer cells appear to require a certain degree of OXPHOS to wake up and proliferate. In animal models, cancer cells with depleted levels of mitochondrial DNA and impaired respiration exhibit a quiescent phenotype, but dormancy outbreak at primary or secondary sites is triggered by the acquisition of host mitochondrial DNA and restoration of respiratory capacity (Sansone et al., 2017; Tan et al., 2015).
Cancer cells with a quiescent-like phenotype are more glycolytic and present lower levels of ROS in experimental models for breast and intestinal tumours (Dey-Guha et al., 2011; Sebastian et al., 2022). Although these G0-like cells are relatively rare in breast cancer patients that have not undergone treatment, their number increases substantially after neoadjuvant chemotherapy (Dey-Guha et al., 2011), reinforcing the notion that residual quiescent cells – with particular metabolic features – can be responsible for disease relapse (Nicolini et al., 2022; Valcourt et al., 2012).
The similarities between metabolic adaptations sustaining quiescence in both normal stem cells and dormant cancer cells also include increased autophagy, which is required for cell survival and maintenance of a quiescent phenotype (Box 3). Furthermore, recent work has shown that the oxidative microenvironment present in skeletal muscle sustains DTCs in a quiescent state (Crist et al., 2022). The authors of this study also demonstrated that in organotypic cell culture and in mouse models of disseminated breast cancer, metabolic alterations, either in tumour cells themselves or in the surrounding non-cancerous cells, that reduce the levels of ROS or allow cancer cells to better cope with the oxidative microenvironment enable cancer cells to bypass dormancy and proliferate (Crist et al., 2022).
Taken together, metabolic adaptations commonly found in cancer are tightly linked to cell cycle progression. Furthermore, metabolic features of dormant cancer cells contribute to sustaining their viability and to chemoresistance, allowing for timely cell cycle re-entrance and disease relapse.
Genomic instability and mutation – loss of quiescence-induced genome protection
It is easy to appreciate how genome instability might lead to loss of quiescence – for example, through activation of oncogenes and deactivation of tumour suppressors – but reciprocally, loss of quiescence also contributes considerably to genomic instability. Extensive research has shown that quiescence protects cells against genomic instability through mechanisms that are, thus far, poorly elucidated, amongst them changes in cell metabolism (Takubo et al., 2013; Yu et al., 2013), chromatin organization and telomere homeostasis (Coulon and Vaurs, 2020; Shay and Wright, 2010; Zahedi et al., 2020) – events that have already been discussed above. Furthermore, the majority of mutations found in cancer arise from unrepaired and therefore propagated mistakes in DNA replication, which are errors that quiescent cells are obviously spared from (Lynch, 2010). Notably, the first cell cycle after quiescence appears to be particularly unlicensed (Box 4), leading to increased replication stress and DNA damage (Matson et al., 2019), reinforcing the idea that disruption of quiescence further increases genomic instability, which could impact DTCs that rest dormant for long periods of time upon their re-entry to the cell cycle.
During the G1 phase of the cell cycle, cells prepare for DNA replication by loading minichromosome maintenance complexes (MCMs) at the replication origins; this loading peaks at the G1-S transition (Mei and Cook, 2021). MCMs are essential components of the helicase complex that is responsible for DNA unwinding, and once DNA replication has started, to avoid re-replication stress and genomic instability, multiple mechanisms block additional MCM loading, thereby restricting licensing to G1 (Liu et al., 2007; Mei and Cook, 2021). Not all replication origins are fired at the beginning of S phase; many of them serve as a backup and are only activated when fork stalling occurs to rescue replication and avoid DNA breaks, and thus, genome instability (Liu et al., 2007; Matson et al., 2019; Mei and Cook, 2021). Therefore, underlicensing (low levels of MCM loading) might result in DNA damage owing to the persistence of stalled forks, which leads to double-strand breaks.
Tumour-promoting inflammation – eliciting escape of tumour cells from quiescence
Chronic inflammation increases the risk of cancer, and inflammatory processes affect all stages of tumorigenesis (Alečković et al., 2019; Grivennikov et al., 2010; Lex et al., 2020; McAllister and Weinberg, 2014; Phan and Croucher, 2020). Pro-tumorigenic and pro-metastatic processes triggered by inflammation involve loss of quiescence by both immune precursors and malignant cells (Alečković et al., 2019; Baram et al., 2021; Grivennikov et al., 2010; Kitamura et al., 2015; McAllister and Weinberg, 2014; Shalapour and Karin, 2015). To meet the demand for immune cells required to mount an inflammatory response, immune precursors, especially from the myeloid lineage, must leave quiescence, expand, differentiate and reach the targeted tissues (McAllister and Weinberg, 2014). Reciprocally, inflammatory cells, cytokines and chemokines can trigger cell cycle re-entry of primary tumour cells and dormant cancer cells at secondary sites (Alečković et al., 2019; Grivennikov et al., 2010; Kitamura et al., 2015). Although the exact mechanisms underlying how inflammation triggers cell cycle re-entrance are yet to be uncovered, it has been demonstrated that inflammation leads to extensive extracellular matrix remodelling (Bonnans et al., 2014) and produces growth factors and cytokines that endow tumour cells with stem cell features (Grivennikov et al., 2010). Microenvironmental cues are key to induce and sustain quiescence and hamper proliferation; therefore, inflammation-induced tissue remodelling might lead to both loss of quiescence-inducing signalling and gain of proliferation-inducing signals (Bissell and Hines, 2011; Egeblad et al., 2010; Fiore et al., 2018). This could be further amplified by growth factors and pro-tumorigenic cytokines that are produced by both immune and stromal cells in response to inflammation (Grivennikov et al., 2010; Shalapour and Karin, 2015). These pro-tumorigenic events are particularly well characterized for tumour-associated macrophages, which are at the core of inflammation-driven cancer progression (Baram et al., 2021; Gao et al., 2019; Linde et al., 2016; Phan and Croucher, 2020; Talmadge et al., 2007). Proinflammatory cytokines can activate a stem-like programme in pre-cancerous or cancerous cells (Grivennikov et al., 2010; Shalapour and Karin, 2015); however, quiescence, rather than solely stemness, has been proposed as the metastasis-initiating phenotype (Wang et al., 2015), suggesting that cancer progression hijacks normal cellular processes involved in quiescence regulation in stem cells. Tumour-promoting inflammation is also relevant for preparation of the pre-metastatic niche, dormancy escape and metastatic outgrowth (Baram et al., 2021; Gao et al., 2019; Grivennikov et al., 2010; Kitamura et al., 2015; Linde et al., 2016; Phan and Croucher, 2020; Shalapour and Karin, 2015).
Interplay between a quiescence signature and the hallmarks of cancer
To further illustrate the associations between quiescence and cancer, we retrieved a quiescence gene expression signature (Coller et al., 2006) and looked for possible links between differentially expressed genes (DEGs) contained in this signature and the hallmarks of cancer (Tables S1 and S2). Indeed, many of the DEGs within the quiescence signature appear to take part in a range of biological processes linked to all the hallmarks of cancer, and interactions between the hallmarks themselves, through quiescence-related genes, are also present (Fig. 2). This further demonstrates that quiescence or quiescence-related events might underlie cellular and molecular traits required for cancer progression.
Taking quiescence into account for the comprehension of cancer biology and for the development of therapeutic interventions
What difference would it make if we consider cancer as a disease of imbalanced quiescence rather than imbalanced proliferation? The idea that proliferation is the default state of cells (Parr, 2012) points to the existence of specific mechanisms governing the entrance to and exit from quiescence. But what makes normal and cancer cells quiescent? Is it acquired intrinsic features, external signals or a combination of both? These are very important questions that are still poorly addressed, mainly because research efforts have mostly focused on the mechanisms involved in proliferation and overlooked those that underlie quiescence.
However, if indeed a relatively small population of quiescence-competent cells within a tumour were the drivers of tumorigenesis, immune escape, drug resistance, disease relapse and metastasis, then understanding the molecular mechanisms that orchestrate quiescence is key for targeting these cells. The feasibility of targeting quiescent (dormant) cancer cells has been suggested and experimentally demonstrated (McCarty, 2018; Vera-Ramirez et al., 2018). Yet, the fact that the mechanisms controlling quiescence might be shared by stem cells and cancer cells is a double-edged sword. Although the knowledge obtained from studying normal quiescent stem cells might apply to dormant cancer cells, by targeting one, you might hit both. For example, autophagy appears key to viability during quiescence in cancer cells (Sosa et al., 2014; Vera-Ramirez et al., 2018) but is also required to sustain quiescence and avoid senescence in a variety of normal cells, including adult stem cell populations (García-Prat et al., 2016; Ho et al., 2017; Mcleod et al., 2012; Schneider and Cuervo, 2014).
Therefore, it is imperative to identify vulnerabilities that are specific for quiescent cancer cells. To that end, experiments comparing major types of adult quiescent stem cells (such as haematopoietic stem cells, satellite cells and hair follicle stem cells) with quiescent cancer cells, in order to identify their similarities and differences, might be highly fruitful. Such discoveries cannot be achieved if the studies instead compare cancer cells with non-malignant cells when they are proliferating.
We have discussed here how abnormalities in quiescence regulation permeate all steps and traits needed for cancer development, using the hallmarks of cancer as guiding principles (Fig. 1). The ability to exit and resume the cell cycle opportunely is key to the rise of a deadly cancer. Loss of quiescence can be linked to sustained proliferative signalling, evasion of growth suppressors, pro-tumorigenic inflammation, angiogenesis and genomic instability, whereas quiescence induction can be related to immune evasion and resistance against cell death. In contrast, replicative immortality, metastasis and deregulated cellular energetics can be associated with both the acquisition and loss of quiescence. Notably, there are substantial overlaps in the signalling pathways governing the different hallmarks of cancer, as they share major molecular and cellular players, including those related to abnormal quiescence regulation (Fig. 2).
As cancer researchers, it is time for us to consider that the secret to better understanding and treating cancer might rely on mastering the specific and context-dependent mechanisms behind quiescence regulation, rather than solely those that underlie proliferation.
The authors are very thankful to Guido Lenz (Federal University of Rio Grande do Sul, Porto Alegre, Brazil) and Cyrus Ghajar (Fred Hutchinson Cancer Research Center, Seattle, USA) for their careful reading of the manuscript and insightful inputs.
The A.B.-C. lab is funded by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP; grant number #2019/26767-2). R.T. is supported by a postdoctoral fellowship from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES).
The authors declare no competing or financial interests.