Current Opinion in Oncology

INTRODUCTION

Loss of control of the mammalian cell cycle drives cellular transformation and promotes tumourigenesis. As a result, eukaryotic cells have developed multiple checkpoints that regulate cell cycle progression. Aberrant expression and activity of the proteins that mediate these cell cycle checkpoints leads to the development of cancer and greatly affects the efficacy of anticancer therapies. One such cell cycle regulatory protein, p21 [also known as wild-type p53-activated fragment 1 (WAF1), CDK-interacting protein 1 (CIP1), senescent cell-derived growth inhibitor 1 (SDI1), melanoma-derived antigen 6 (MDA6) and cyclin-dependent kinase inhibitor 1A (CDKN1A)], functions as a cyclin-dependent kinase inhibitor. In response to DNA damage or other cellular stressors, p21 expression is increased, resulting in the activation of cell cycle checkpoints until repair has taken place, so that cells can survive and maintain genetic fidelity. Since its discovery, an abundance of research has unravelled the complex mechanisms regulating p21 expression and function and provided insights into the role of p21 as a positive and negative regulator of tumour development and progression.

STRUCTURE

The wild-type p53-activated fragment (WAF1) gene is located on human chromosome 6p21 and encodes a 21-kDa protein (p21) that was originally identified through subtractive hybridization screening for transcriptional targets of the p53 tumour suppressor protein ^[1]. Simultaneously, p21WAF1 was independently discovered from a yeast two-hybrid screen as a 21-kDa protein that interacts with and inhibits cyclin-dependent kinase 2 (p21CIP1) ^[2], as SDI1 ^[3], and as MDA6 ^[4]. p21 is a member of the Cip/Kip family of cyclin kinase inhibitors (CKIs), which includes p27^Kip1 and p57^Kip1. These family members share significant sequence homology in their N-terminus, which has been demonstrated to harbour cyclin/CDK-binding domains that are necessary and sufficient to inhibit CDK activity and cause G1 arrest when overexpressed in cells. They also share a C-terminal nuclear localization signal (NLS), but have no other domains in common. Unlike other Cip/Kip family members, the unique carboxyl-terminal domain of p21 harbours a second cyclin binding site as well as a domain that interacts with the proliferating nuclear antigen (PCNA), a subunit of DNA polymerase δ.

The Cip/Kip family members are highly conserved throughout evolution. However, there are structural differences between the family members that support the idea that each of these inhibitors serves a distinct function in the cell. Interestingly, in Drosopohila, the dacapo (Dap) protein has been identified as a homologue of p27^Kip1 that functions as a key regulator of the exit from the cell cycle. However, although Dap contains low sequence homology to the eukaryotic Cip/Kip family members, there is striking conservation in the CDK-interacting domains and the PCNA motif, which is specific to p21, indicating that Dap may be more closely related to p21 in some of its regulatory functions. A recent study on the role of Dap as a regulator of premitotic S phase and genomic stability showed that dap −/− flies enter the meiotic cycle with high levels of cyclin E/CDK2 and accumulate DNA damage in ovarian cysts, independent of the double-strand breaks that initiate meiotic recombination ^[5]. These data suggest that CKIs, such as p21, may play essential and nonredundant roles in meiotic cell division.

REGULATION OF p21 EXPRESSION LEVELS

Since its discovery as a transcriptional target of p53, much progress has been made towards unravelling the mechanisms that govern p21 expression at both the transcriptional and protein levels.

Transcription

The transcriptional regulation of p21 has been exhaustively studied. p21 expression was originally found to be induced in the p53 tumour suppressor protein and was decisively shown to be the central mediator of p53-induced G1 arrest ^[1]. The WAF1 gene harbours several p53-response elements in both its 5’-terminus and in the body of the gene. Thus, in response to DNA damage, activation of p53 induces p21 protein expression and causes G1 arrest in a p21-dependent manner. In addition to its regulation by the classical p53 pathway, p21 expression is also modulated at the transcriptional level by an array of oncogenes and tumour suppressor proteins that induce p21 expression via the binding of different transcription factors to specific cis-acting elements located in the p21 promoter. For example, transforming growth factor β (TGF-β), nerve growth factor (NGF) and the tumour suppressor protein, breast cancer gene 1, induce p21 transcription through Sp1–3 sites in the p21 promoter ^[6–8]. Alternatively, the oncoprotein, c-Myc, has been shown to inhibit the transcription of p21 through its interaction and subsequent repression of multiple transcription factors ^[9]. Furthermore, c-Myc has been shown to interact with the carboxyl-terminus of p21, which disrupts its interaction with PCNA and decreases p21-mediated inhibition of DNA synthesis ^[10]. Due to the importance of c-Myc in cancer, this finding has wide ranging implications on the physiological outcome of cells in response to DNA damage. For instance, in tumour cells expressing high levels of c-Myc, the p53-dependent induction of p21 would be blunted, allowing for the proapoptotic arm of p53 to prevail. Thus, the regulation of p21 transcription is complex, and in addition to p53 and serum, a variety of other transcription factors, such as activator protein 2, E2 transcription factor (E2F), signal transducer and activator of transcription (STATs), and CCAAT/enhancer binding protein alpha, can induce p21 transcription in response to different signals ^[11].

RNA stability

It has long been known that inflammation and cellular stress can upregulate p21, and one mechanism responsible for this response was recently elucidated ^[12▪▪]. The ZO-1–associated nucleic acid binding protein (ZONAB)/DbpA is a Y-box transcription factor that is regulated by components of intercellular junctions that are affected by cytokines and tissue damage. In response to several types of cellular stress, ZONAB binds to specific sites in the 3′-UTR of p21 mRNA, resulting in mRNA stabilization and enhanced translation. Furthermore, Rho-stimulation induced binding of ZONAB to p21 mRNA influences Ras-induced p21 expression. These findings demonstrate RNA stabilization as a unique mechanism that links the cellular stress response to p21 expression and cell survival.

Proteasomal degradation

Under normal growth conditions, p21 is an unstable protein with a relatively short half-life. At the time of synthesis, heat shock protein 90 (HSP90) is recruited to p21 via the WAF1/CIP1 stabilizing protein 39 (WISp39) adapter protein and protects p21 from proteasomal degradation ^[13]. Confirming the importance of HSP90 for p21 stability, treatment with geldanamycin, a HSP90 inhibitor, decreased the half-life and steady-state levels of p21. Also, p21 was not upregulated in response to DNA damage in cells lacking WISp39, suggesting that the stabilization of p21 is necessary for its upregulation, regardless of transcriptional activation.

To date, a majority of the research on p21 turnover suggests that it is primarily degraded through the ubiquitin-proteasome pathway. Several distinct E3 ubiquitin ligase complexes, including SCF^SKP2, cullin4A-RING E3 ubiquitin ligase (CRL4)^CDT2, murine double minute 2 (MDM2) and MDMX, CRL2^LRR1 and anaphase-promoting complex/cyclosome (APC/C)^CDC20, have been shown to mediate the turnover of p21. Interestingly, each E3 ligase complex seems to preferentially bind p21 at different stages of the cell cycle or in different locations within the cell to regulate specific pools of p21. For example, CRL4^CDT2 promotes the ubiquitination and degradation of p21 only when bound to PCNA in S phase ^[14], whereas APC/C^CDC20 ubiquitinates p21 in prometaphase to allow for full activation of cyclin/CDK1 during mitosis ^[15]. Due to the cytoplasmic localization of the E3 ligase subunit, leucine-rich repeat protein 1 (LRR1), CRL2^LRR1 specifically targets the cytoplasmic pool of p21 for proteasomal degradation ^[16▪▪]. Interestingly, the canonical ring-finger E3 ligases that target p53 for degradation, MDM2 and MDMX, also promote the turnover of nuclear p21 by brining p21 directly to the proteasome, independent of ubiquitin, via the binding of MDM2/p21/14–3–3τ directly to the C8 subunit of the 20S proteasome ^[17].

Post-translational modifications

There is mounting evidence of the importance of post-translational modifications in controlling p21 expression levels, cellular localization and activity. The identification of post-translational modifications that alter the localization of p21 may be of particular importance considering that nuclear accumulation of p21 is associated with growth inhibition, whereas its oncogenic activities are primarily associated with cytoplasmic localization. Most notably, the prosurvival kinase, Akt, phosphorylates p21 at Thr145, which disrupts the binding of p21 to PCNA, decreases its inhibitory effect on cyclin CDK complexes and results in the cytoplasmic accumulation of p21 ^[18,19]. As a result, in tumours that have constitutive activation of Akt due to amplification of receptor tyrosine kinases, among other mechanisms, the phosphorylated, cytoplasmic form of p21 can no longer initiate its growth-inhibitory functions.

A recent report from Brian Hemming's group revealed a novel link between p21 and the tumour suppressive mamallian sterlie 20-like kinase (MST) signalling pathway through activation of the downstream nuclear Dbf2-related (NDR) kinases ^[20▪▪]. Specifically, NDR kinases control the protein stability of p21 by direct phosphorylation at Ser146, which destabilized p21. Therefore, inhibition or genetic depletion of MST3 or NDR stabilized p21 and led to the accumulation of cells in G1. These findings are the first to describe the existence of a novel MST3-NDR-p21 axis as an important regulator of G₁/S progression in mammalian cells and provide new insight into the potential regulation of p21 in organ development and tumourigenesis.

REGULATORY FUNCTIONS

Due to its central role downstream of numerous signal transduction pathways, p21 plays an important role in determining whether cells adapt or undergo apoptosis by altering gene expression via direct and indirect mechanisms in response to cellular insults.

Cell cycle inhibitor

p21 exerts its inhibitory control over the cell cycle primarily through direct binding to cyclins and CDKs. p21 binds the cyclin subunit through a conserved motif in the N-terminus and a weaker, redundant motif near the C-terminus. Furthermore, p21 has been shown to interact with the CDK1 and CDK2 through an additional CDK-binding site in the N-terminus. Interaction with both the CDK subunit and the cyclin subunit is required for optimal inhibition of CDK/cyclin kinase activity ^[21]. Binding to p21 disrupts the interaction between CDKs and their substrates, which generally bind to the same motif as p21. In the context of the cell cycle control, the primary targets of p21 are CDK1 and CDK2, whose activation is required for the phosphorylation of Rb, which leads to the activation of E2F-dependent gene expression and proteins directly involved in DNA synthesis.

Transcriptional regulation

As expected, p21 expression dramatically alters the expression of genes involved in cell cycle progression and senescence. A direct result of its function as a CDK inhibitor is the subsequent blockade of phospho-Rb and E2F transcription factors. p21 can also influence the transcription of genes through direct binding and inhibition of PCNA, which inhibits DNA polymerase activity, transcription and excision repair functions ^[22]. Recent evidence also indicates that p21 can regulate transcription through mechanisms that are independent of its classical interacting partners, CDKs and PCNA. Overexpression of a mutant p21 construct lacking the ability to interact with CDKs and PCNA was reported to directly interact with and markedly decrease E2F transcriptional activity ^[23]. However, the physiological relevance of this interaction and the direct influence of p21 on E2F-mediated transcription remain unknown.

p21 can also reduce STAT3-mediated transcription ^[24]. Coimmunoprecipitation experiments revealed that p21 bound directly to STAT3 and the cAMP response element-binding (CREB)-binding coactivator protein following stimulation with leukaemia inhibitory factor (LIF). Interestingly, overexpression of CREB-binding protein was sufficient to rescue the decrease in STAT3-mediated following p21 expression. Although these reports indicate a role for p21 as a transcriptional regulator, independent of its CDK and PCNA-binding functions, the possibility that these transcriptional effects are independent of other p21-regulated proteins has not been adequately ruled out.

PHYSIOLOGICAL FUNCTIONS OF p21

p21 is a critical regulator of many cellular processes involved in both normal and tumour development, such as differentiation, apoptosis, stem cell renewal and metastasis. As a result, p21 can have opposing effects on the balance between cell survival, proliferation and death.

Proliferation

In general, p21 is thought to inhibit cellular proliferation, as evidenced by the fact that overexpression of p21 in normal and tumour cell lines causes G1 arrest. However, under certain conditions, p21 has been shown to promote proliferation. In response to mitogen stimulation, p21 is induced to levels that are not sufficient to cause cell cycle arrest, but instead, promoted cyclin-CDK activity ^[25]. Thus, the role of p21 as a cyclin/CDK complex assembly activator or inhibitor depends on its expression level; at low concentrations, p21 promotes assembly, whereas at high concentrations, it is an inhibitor. In addition to expression levels, it has been reported that p21 is hyperphosphorylated on Thr57 by CDK2, specifically during G2/M phase ^[26]. Transient phosphorylation at this residue enhances its interaction with cyclin B, promoting the formation and activation of the cyclin B/CDK1 complex. Thus, p21 can play dual roles as either an activator or inhibitor of cell cycle progression.

Apoptosis

Similar to its effect on proliferation, p21 also plays a dual role in the apoptotic process. A majority of the available data indicate that p21 protects cells from undergoing p53-mediated apoptosis by inducing cell cycle arrest. In response to DNA damaging agents, such as camptothecin, wild-type colon cancer cells undergo cell cycle arrest and senescence, whereas isogenic cell lines lacking either p53 or p21 undergo apoptosis ^[27]. Similarly, overexpression of p21 has a protective effect in response to gamma-irradiation-induced apoptosis in colorectal cancer cells ^[28]. p21 can also shift the balance between survival and apoptosis through the regulation of key apoptotic proteins. For example, p21 associates with and inhibits the activation of the proapoptotic mitogen-activated protein kinases, p38 and c-Jun N-terminal kinase (JNK) ^[29]. Binding of p21 to JNK is shown to block its phosphorylation and activation by the upstream mitogen-activated protein kinase kinase 4 kinase. p21 is also a substrate for caspase-3, a critical mediator of the intrinsic cell death pathway ^[30]. Upon activation of caspase-mediated apoptosis, caspase-3 cleaves the C-terminal NLS from the nuclear pool of p21, resulting in a truncated, inactive form of p21 that accumulates in the cytoplasm and can no longer provide protection from cell death.

Differentiation

p21 expression is induced through both p53-dependent and independent mechanisms during differentiation and controls the exit from the cell cycle, which is required for terminal differentiation. Overexpression of p21 is able to induce differentiation in a wide range of both cancerous and normal cells. It has been well established in many normal and tumour models that vitamin D and its analogues promote cellular differentiation via increasing p21 expression levels ^[31]. In fact, several nuclear receptors, including retinoid receptors and vitamin D receptors, initiate p21 transcription in a p53-independent manner by binding directly to response elements in the p21 promoter. However, normal differentiation has been observed in p21 knockout mice, suggesting that p21 is not required for terminal differentiation and that other genes are able to compensate for the loss of p21 to ensure proper differentiation ^[32].

Stem cell self-renewal

Another critical aspect of p21 biology is its role in maintaining stem cell quiescence. The relative quiescence of haematopoietic stem cells (HSCs) is of great importance for maintaining the stem cell compartment. In p21 knockout mice, HSC proliferation and total numbers are dramatically increased ^[33], indicating that p21 controls the entry of stem cells into the cell cycle, and in its absence, the stem cell population is rapidly exhausted. Another recent report indicates that p21-induced cell cycle arrest is required for leukaemia stem cell self-renewal ^[34▪▪]. In a similar manner to that described for HSCs, the activation of p21 prevents the accumulation of excessive DNA damage and functional exhaustion of rapidly dividing leukaemia cells. Therefore, depending on the cellular context, the effect of p21 inhibition on stem cell populations could be viewed as a positive (depleting leukaemic stem cells) or as a negative (depleting haematopoietic stem cells) strategy. p21 also regulates tumour stem cells in vivo. Liu et al.^[35▪] recently reported that p21 controls the expression of genes involved in epithelial mesenchymal transition (EMT) and the putative cancer stem cell population in Ras or c-Myc driven mammary tumourigenesis. Genetic silencing of p21 enhanced features of EMT, while overexpression of p21 reduced EMT and decreased mammosphere formation in vitro. These findings indicate that the loss of p21 may promote breast cancer EMT and stem cell properties in vivo.

Senescence

Recently, several reports have suggested that p21 and other CKIs mediate tumour growth through the induction of autophagy and cellular senescence. The Lisanti laboratory has proposed a two-compartment model of cancer metabolism, whereby the induction of senescence and autophagy genes have differential effects in stromal cells and tumour cells. In particular, CKIs were reported to play an important role in the compartment-specific cell cycle arrest and senescence in breast cancer tumour growth ^[36▪]. Overexpression of p21 in senescent fibroblast cell lines promoted autophagy and mitochondrial dysfunction, which dramatically increased tumour volume in a xenograft model of breast cancer. Conversely, overexpression of p21 in breast cancer cell lines promoted the expression of markers of cell cycle arrest and autophagy, but in xenograft models, these senescent tumour cells showed reduced tumour growth. Thus, the effects of p21 on tumour growth may be compartment-specific, as the activation of p21 in normal fibroblasts promotes tumour growth, although p21 expression in tumour cells has the opposite effect. Further studies are necessary to determine the mechanisms by which p21 can promote autophagy and metabolic changes in normal and tumour cells.

DEREGULATION IN TUMOURIGENESIS

Despite its critical role as a tumour suppressor, the WAF1 gene is not a target for mutational inactivation in human cancer, likely because complete loss of p21 expression promotes cell death in response to stress. However, a p21 knockout mouse was successfully bred and developed normally, but p21−/− embryonic fibroblasts were deficient in their ability to promote G1 arrest in response to DNA damage ^[37]. p21 has also been reported to influence the spread of metastasis. Following exposure to γ-irradiation, p21 heterozygous and null mice developed more tumours than wild-type mice, and complete loss of p21 increased the number of metastatic tumours ^[38]. These findings suggest that p21 elicits tumour suppressive and antimetastatic functions.

Expression changes in p21 have been widely studied in a broad range of cancers and are differentially associated with clinical outcome in different tumour types, as might be expected due to its oncogenic and tumour suppressive effects. A thorough assessment of p21 expression changes in different cancer types has been previously reviewed ^[39]. In some instances, the loss of p21 correlates with poor patient prognosis, likely due to the loss of its growth-inhibitory functions and ability to suppress oncogenes. However, there are also instances in which heightened p21 expression correlates with disease progression. In particular, there is growing evidence that p21 functions as an oncogene when localized to the cytoplasm. Accumulation of cytoplasmic p21 has been reported for many cancers and is associated with tumour aggressiveness, metastasis and poor prognosis. One critical mechanism for the association between cytoplasmic p21 and metastasis is its ability to bind and inhibit Rho-activated kinase 1 (ROCK1), leading to the activation of the Rho–ROCK–LIM domain kinase 1–cofilin pathway. The resulting activation of the actin depolymerizing protein, cofilin, results in the disassembly of actin stress fibres, the formation of cell extensions and increased cell motility ^[40]. Thus, cytoplasmic p21 can promote metastasis through the regulation of the actin cytoskeleton. Despite the vast gains in our understanding of the molecular regulation and cellular function of p21, its diverse roles in the regulation of cell survival and growth make its precise role in tumourigenesis difficult to define.

CONCLUSION

Several preclinical and clinical anticancer agents, including vitamin D, butyrate, lovastatin and histone deacetylase (HDAC) inhibitors, such as trichostatin A (TSA), have been shown to elicit their antiproliferative and proapoptotic effects, at least in part, by enhancing p21 expression levels. Although our understanding of how p21 functions has greatly improved, it has also elucidated the dual role of p21 as both a classical tumour suppressor and an oncogene. Thus, it would be of great interest to use the molecular knowledge that has been gained to design therapeutics that preferentially block the ability of p21 to promote senescence, stem cell renewal and cyclin/CDK activation, while leaving its tumour suppressive functions intact. Due to its dual role in tumourigenesis, depending on the cellular context, it may be beneficial to either activate or inhibit p21 expression (Fig. 1). For example, p21-induced senescence is known to promote resistance to chemotherapy and radiation. Thus, in this context, a p21 inhibitor would sensitize tumour cells to standard therapies. On the contrary, p21 has valuable antiproliferative and proapoptotic effects that suggest that p21 activation could be useful anticancer therapeutics. In this case, it may be useful to identify therapeutic combinations between p21-inducing compounds and small molecules that inhibit the effectors downstream of p21 that drive proliferation. For example, CDK inhibitors have gained great interest and shown promising results in clinical trials. The use of CDK1 and CDK4/6-specific inhibitors in combination with a p21-inducing agent may promote p21-mediated apoptosis while negating its positive effect on cell cycle progression. Therefore, it will be of the utmost importance to use caution when assessing the efficacy of p21-targeted therapeutics. More work is required to determine whether targeting p21 as an anticancer strategy is a promising endeavour, but with a greater molecular understanding of how p21 functions in specific cellular contexts, its potential as an anticancer therapeutic target is limitless.

Acknowledgements

None.

Conflicts of interest

There are no conflicts of interest.

REFERENCES AND RECOMMENDED READING

Papers of particular interest, published within the annual period of review, have been highlighted as:

• ▪ of special interest
• ▪▪ of outstanding interest

Additional references related to this topic can also be found in the Current World Literature section in this issue (p. 101).

REFERENCES