The function of p27^KIP1 during tumor development

Abstract

Timely cell cycle regulation is conducted by sequential activation of a family of serine-threonine kinases called cycle dependent kinases (CDKs). Tight CDK regulation involves cyclin dependent kinase inhibitors (CKIs) which ensure the correct timing of CDK activation in different phases of the cell cycle. One CKI of importance is p27^KIP1. The regulation and cellular localization of p27^KIP1 can result in biologically contradicting roles when found in the nucleus or cytoplasm of both normal and tumor cells. The p27^KIP1 protein is mainly regulated by proteasomal degradation and its downregulation is often correlated with poor prognosis in several types of human cancers. The protein can also be functionally inactivated by cytoplasmic localization or by phosphorylation. The p27^KIP1 protein is an unconventional tumor suppressor because mutation of its gene is extremely rare in tumors, implying the normal function of the protein is deranged during tumor development. While the tumor suppressor function is mediated by p27^KIP1's inhibitory interactions with the cyclin/CDK complexes, its oncogenic function is cyclin/CDK independent, and in many cases correlates with cytoplasmic localization. Here we review the basic features and novel aspects of the p27^KIP1 protein, which displays genetically separable tumor suppressing and oncogenic functions.

Introduction

Extracellular environments initiate cell cycle division or arrest by activating or deactivating cycle dependent kinase (CDK) complexes. Since timely regulation of CDK complexes is critical for proper cell cycle regulation, multiple signals can integrate to control the activity of cyclin/CDK complexes. Activity of cyclin/CDK complexes is regulated by the accumulation of cyclins and by phosphorylation/dephosphorylation of specific complex components (Norbury and Nurse, 1992; Sherr et al., 1993; Malumbres and Barbacid, 2005; De Clercq and Inzé, 2006). Another important regulation of G1 cyclin/CDK complexes lies in their association with cyclin dependent kinase inhibitors (CKIs), such as p27^KIP1. Cyclin dependent kinase inhibitors are often associated with diseases when mutated or deregulated. The p27^KIP1 protein binds to various cyclin/CDK complexes throughout the cell cycle, and is one exemplary CDK inhibitor whose misregulation, and not genetic mutation, is found in diverse cancer types (Tsihilias et al., 1999; Besson et al., 2008; Chu et al., 2008).

The p27^KIP1 protein was first identified as an inhibitor of cyclin E/CDK2 complexes during TGFβ-induced G₁ arrest (Sheaff et al., 1997). The p27^KIP1 protein inhibits cyclin/CDK activity by binding cyclin/CDK complexes through its N-terminal, blocking ATP binding, and physically occluding the catalytic cleft of the CDK (Russo et al., 1996; Hong et al., 2009; Zhang et al., 2009). A major regulatory mechanism of controlling the p27^KIP1 inhibitory function is to regulate p27^KIP1 protein levels through transcriptional, translational, and post-translational mechanisms (Hengst and Reed, 1996; Carrano et al., 1999; Boehm et al., 2002; Bagui et al., 2009; Shin et al., 2009; Trabosh et al., 2009). Recent study revealed that a novel regulatory mechanism during postphosphorylation and polyubiquitination of proteins could be another regulatory mechanism for p27^KIP1. Pin1, peptidyl-prolyl isomerase, recognizes and stabilizes p27^KIP1 when phosphorylated on Thr187 by inducing its conformational change (Zhou et al., 2009). Misregulation that results in increased degradation of p27^KIP1 can be related to cancer development (Hershko, 2008; Mishra et al., 2009). In addition to TGFβ, other environmental factors such as serum starvation or contact inhibition can increase p27^KIP1 protein whose role becomes essential during cell cycle arrest and differentiation. For instance, serum deprivation increases the amount of p27 phosphorylation during muscle differentiation or quiescence approacing G₀ state of colon carcinoma cells and mitogen stimulation then causes cells to enter G₁ with the translocation of p27 to the cytoplasm (Rodier et al., 2001; Boehm et al., 2002; Deng et al., 2003; 2004; Jin et al., 2009). Phosphorylation, as such, is the mechanism primarily used for regulating p27^KIP1 activity. The p27^KIP1 protein possesses multiple tyrosine, serine, or threonine phosphorylation sites (Figure 1). The inhibitory actions of p27^KIP1 on cyclin/CDK complexes are weakened by phosphorylations directed by some signal transduction pathways (Larrea et al., 2008; Morishita et al., 2008; Tossidou et al., 2008; Jin et al., 2009). It appears that different signalling pathways direct the fate of the protein through differential phosphorylation patterns as Thr187 phosphorylation leads p27^KIP1 to a SCFSkp2 ubiquitin ligase complex and promotes the polyubiquitination and degradation while T198 phosphorylation of p27^KIP1 by ribosomal S6 kinase 1 (RSK1) promotes cell motility (Grimmler et al., 2007; Larrea et al., 2009). This suggests that signal transduction pathway utilizes regulatory mechanisms of p27^KIP1 by phosphorylation to control the activity of cyclin/CDK complexes. Moreover, recent studies suggest that post-translational modification of p27^KIP1 by phosphorylation can play diverse roles for p27^KIP1 in regulations on its half life and subcellular localization. While it is evident that nuclear p27^KIP1 through inactivating cyclinE/CDK2 complex acts as a tumor suppressor, tumorigenic properties of p27^KIP1 have also been proposed in recent years especially when located in cytoplasm. (Blagosklonny, 2002; Sicinski et al., 2007; Hidaka et al., 2009). Pro-tumorigenic activities include enhancing cell mobility and assisting assembly of cyclin D/CDK4 complexes, thereby promoting metastasis and cell cycle progression (Besson et al., 2004; Jeon et al., 2008; Kim et al., 2008; Vervoorts and Lüscher 2008). Since cytoplasmic localization can be associated with cell migration promoting activity of p27^KIP1, removal of p27^KIP1 from nucleus to degradation is less oncogenic than translocalization of nuclear p27^KIP1 to the cytoplasm. The puzzling dual function of p27^KIP1 is intriguing and worth additional attention because it answers some mysteries surrounding p27^KIP1, such as existence of the p27^KIP1 protein in malignant cancers. While potential pro-oncogenic activity of p27^KIP1 awaits further investigation, several research groups have independently generated null mutants in order to better understand biological function of p27^KIP1 and to provide better evidence of tumor suppressor activity of the protein.

Figure 1.

Schematic diagram of p27^KIP1 phosphorylation and binding sites. Phosphorylation sites (upper) and binding sites (lower) of full length p27^KIP1 protein are illustrated with the specific site numbers for kinases, or with the binding regions for cyclin/CDK proteins. Nuclear localization sequences (NLS) are also depicted. Note that the threonine subject to AKT phosphorylation, T157, is located within the NLS.

The p27^KIP1 null (-/-) mouse shows an overall increase in cell proliferation, resulting in approximately 30% increase in body size, multiple organ hyperplasia, and disorganization of sensory epithelia in the retina and inner ear. In addition, the null mice also display female sterility due to defective ovarian and uterine function (Fero et al., 1996; Chien et al., 2007; Kim et al., 2007). The tumor suppressive activity of p27^KIP1 using null mice was first demonstrated by development of adenomas in the intermediate lobe of the pituitary gland, and by their inclination to develop tumors more easily when challenged with chemical carcinogens or irradiation (Nakayama et al., 1996). It is likely that combined loss of p27^KIP1 with other tumor suppressor genes further enhances tumorigenesis (Di Cristofano et al., 2001; Glover et al., 2009). A 50% reduction in p27^KIP1 protein level predisposes p27^KIP1 heterozygous (+/-) mice to tumors in multiple organs, when combined with administration of carcinogens, or when genetically combined with additional oncogenes or tumor suppressors. In humans, p27^KIP1 deficiency has also been found to be associated with sporadic tumorigenesis. The first human cases reported to have abnormally low amounts of nuclear p27^KIP1 were associated with increased tumor aggressiveness and a relatively poor clinical outcome for breast and colon cancer (Loda et al., 1997).

Recent findings from knock-out models also show data supporting the pro-oncogenic property of p27^KIP1 protein. Two groups reported that p27^KIP1 heterozygous (+/-) mice were more susceptible to mammary and prostate tumors than p27^KIP1 null (-/-) mice (Muraoka et al., 2002; Gao et al., 2004). Analysis of knock-in mice with CDK mutant p27^KIP1, p27^KIP1 (CK-), that lacks CDK inhibitory function revealed that p27^KIP1 (CK-/ CK-) mice not only displayed the tumor development phenotype of p27^KIP1 null (-/-) mice, but also developed a whole range of hyperplasia and neoplasia, suggesting that the p27^KIP1 (CK-) protein function as an oncogenic protein (Besson et al., 2007). In human cancer cells, homozygous inactivation of the p27^KIP1 gene in sporadic tumors is extremely rare, but the correlation of cytoplasmic localization of p27^KIP1 protein with high tumor grade and poor prognosis was discovered (Slingerland and Pagano, 2000). The p27^KIP1 (CK-) protein was also found in cytoplasm, which correlated with the hypothesis of the p27^KIP inhibitory action of RhoA pathway.

Benefited from many clinical reports, we now understand that the p27^KIP1 protein can be a prognostic indicator for breast, colon, prostate, lung, esophageal, and gastric cancers. Tumorigenic activity due to abated p27^KIP1 is dose-dependent, meaning lower doses of the protein are associated with increased malignancy. Also, it is now understood that deregulation of p27^KIP1 protein, not loss of its gene, is the cause of reduced p27^KIP1 protein levels. Orderly regulation of p27^KIP1 includes proper activation of the oncogenic signal proteins, deregulation of which in turn is commonly found in human cancers. Therefore, better understanding of the regulatory mechanisms of p27^KIP1 function may provide good clinical value with prognostic and therapeutic applications. This fascinating protein has drawn the attention of clinicians as well as scientists, and as a result, understanding the exact cellular functions and mechanisms is currently under active investigation. This report will focus on the regulatory mechanisms and dual roles of p27^KIP1 in cancer biology.

Interaction of p27^KIP1 with cyclin/CDK complexes and its inhibitory function

Different from INK4 family proteins' binding to a CDK alone, CDK4 or CDK6, CIP/KIP proteins like p27^KIP1 interact with the cyclin E/CDK2 complex. The first α-helical loop of a CIP/KIP protein interacts with the cyclin, and the second helix binds to the catalytic cleft of the CDK subunit, thereby blocking ATP loading (Sherr and Roberts, 1999; Yil et al., 2007; Jung et al., 2008). The inhibition of ATP loading blocks activation of the cyclin E/CDK2 complex, and prevents progression through the cell cycle. The p27^KIP1 protein binds not only to the cyclin E/CDK2 complex, but also to the cyclin D/CDK4,6 complexes. However, the interaction with the cyclin D/CDK complexes is more complicated. It has become consensus that p27^KIP1 is a required assembly factor for the complex, but whether the binding is inhibitory is still questionable. Moreover, how it might switch between the two modes of inhibitory and non-inhibitory needs to be answered.

Arguing for a two state mechanism, Dr. Blain's group has assiduously sought-after, and shown that p27^KIP1 can be both a CDK4 bound inhibitor, and a bound non-inhibitor, depending on the growth state of the cell (Ray et al., 2009). They also have discovered that p27^KIP1 associates with cyclin D/CDK4 constitutively, and that a specific tyrosine phosphorylation converts p27^KIP1 from a bound inhibitor to a bound non-inhibitor under different growth conditions. To further support this, they showed that in vitro tyrosine phosphorylation, by the tyrosine kinase Abl, converts the bound inhibitor to a bound non-inhibitor. Larrea et al. reported a similar finding, demonstrating that phosphorylations at Thr157 and Thr198 are required for binding to cyclin D1 and CDK4, but are not sufficient to activate the cyclin D/CDK4 complex (Larrea et al., 2008). In addition, Larrea et al. showed that tyrosine phosphorylation by SRC activates the p27^KIP1 bound cyclin D1/CDK4 complex, but tyrosine phosphorylated p27^KIP1 does not affect assembly of the complex.

The p27^KIP1 protein can also bind to the nuclear pore-associated protein (mNPAP60), and interact with the nuclear export protein chromosome region maintenance 1 protein (CRM1) (Shin et al., 2005). Since CRM1 mediates nuclear export, CRM1 interaction with p27^KIP1 causes trans-localization into the cytoplasm and out of nucleus. Furthermore, interaction of p27^KIP1 with CRM1 also causes displacement of cyclin D1 from CRM1, leading to increased cyclin D1 levels in the nucleus, and progression through the cell cycle.

p27^KIP1 is phosphorylated at multiple sites by various signaling pathways

An ability of p27^KIP1, together with other cell cycle molecules, is to respond to diverse extracellular demands. In doing so it helps cells adjust to the new environment through proper cell cycle regulation, which is pivotal to maintaining normal cellular homeostasis. The ability to properly respond to different signaling pathways comes from accordingly regulating the p27^KIP1 protein by intricate, but unmistaken phosphorylations. Misregulation or functional inactivation of p27^KIP1 protein is caused when oncogenic kinases such as PKB and SRC are over-activated, leading to the development of malignancies. If these oncogenic signaling pathways can be inhibited, it is possible to restore the tumor suppressive functions of misregulated p27^KIP1 protein. Hence, understanding mechanisms of p27^KIP1 phosphorylation will provide additional perspectives for finding potential targets for cancer prevention and therapy.

The p27^KIP1 protein is short-lived, and activity of this protein largely depends on protein levels that are regulated mainly through proteasome-dependent degradation and/or transcriptional control. Recent studies show that potentially oncogenic serine/threonine kinases, PIM kinases, promote cell cycle progression by phosphorylating Thr157 and Thr198 of p27^KIP1 leading to its nuclear export and proteasome-dependent degradation. However, while the mechanisms involved are under active investigation, p27^KIP1 possesses seemingly contradictory actions in facilitating cell motility by interacting with proteins involved in functions aside from cell cycle regulations while located in the cytoplasm. The role of p27^KIP1 in facilitating cell motility through inhibition of RhoA activation in the cytosol, implies that its cytosolic localization does more than free cyclin/CDK complexes from inhibition. It also does more than subject this inhibitor to proteasomal degradation (Figure 2). The significance of multiple biological functions for p27^KIP1 protein is generally acknowledged. The onset of environmental cues delivers a new role to this cell cycle inhibitor as a result of post-translational modifications like phosphorylation or ubiquitination. Extensive studies have informed us that p27^KIP1 protein remodels itself predominantly through phosphorylation and consequently alters its interacting protein partners, changes its cellular function, localization, as well as protein expression levels.

Figure 2.

The p27^KIP1 protein is regulated by phosphorylation on multiple sites. Hypo-phosphorylated or unphosphorylated p27^KIP1 acts as a cyclin/CDK inhibitor. Multiple phosphorylated forms of the protein, through activation of various mitogenic signals, are present in cells. Phospho-p27^KIP1 diverges into discrete fates according to the location of phosphate groups. Respective functions of multifarious combinations of individual phospho sites are beyond comprehension at present. However, tyrosine phosphorylation in general renders phopho-p27^KIP1 non-inhibitory and serine/threonine phosphorylation results in cytosolic localization. Some phospho forms of the cytosolic p27^KIP1 are subject to proteasomal degradation, while others inhibit RhoA activation and facilitate cell migration.

Important roles for p27^KIP1 in guarding cells against breast cancer have advanced our understanding of the relationship between p27^KIP1 phosphorylation and its role in CDK inhibition. Phosphorylation of serine or threonine on p27^KIP1 by ERK1 targets the protein for ubiquitination, and phosphorylation on more than one threonine site by different kinases is involved in cytoplasmic localization. Cyclin E/CDK2 phosphorylates p27^KIP1 on Thr187 and leads to ubiquitin-dependent degradation. The p27^KIP1 protein is phosphorylated by AKT at Thr157 and Thr198 to become a better assembler of cyclin D/CDK4 complexes, and the binding of p27^KIP1 to cyclin D/CDK4 facilitates activation of cyclin E/CDK2 through sequestration of the inhibitory protein. Therefore, the differential binding of p27^KIP1 to distinct CDKs during G₁ cell cycle can also be attributed to the phosphorylation status of p27^KIP1. As p27^KIP1 phosphorylations are cell cycle dependent, the cyclin E/CDK2 inhibitory activity of p27^KIP1 is maximal during G₀, and falls as cells move through G₁ into S phase. At the same time, the cyclin D/CDK4 bound p27^KIP1 is maximal in early G₁. All of these findings suggest that anti-mitogenic signaling, which can alter phosphorylation status of p27^KIP1, can switch p27^KIP1 binding from CDK4/6 complexes to cyclin E/CDK2 complexes, promoting restoration of cell cycle control.

p27^KIP1 exerts anti- and pro-tumorigenic activities

Decreases in p27^KIP1 expression levels have been implicated in the genesis and progression of many human malignancies. Just as a mouse knock-out system was used to determine anti-tumorigenic activities of p27^KIP1, a mouse knock-in model was used to insert one allele of p27^KIP1 that cannot bind cyclins and CDKs. This knock-in model has been generated to address the question of the tumor-prone phenotype with CDK inhibitory function. The knock-in mouse model shows an increase in spontaneous tumorigenesis in many tissues when compared with either wild-type or p27^KIP1 null animals (Kim et al., 2005; Susaki et al., 2009). However, many studies indicate that the function of p27^KIP1 cannot be solely attributed to its CDK inhibitory action. The critical function of p27^KIP1 in tumorigenesis therefore may extend beyond its ability to regulate CDKs. Indeed, the fact that p27^KIP1 levels in tumors do not always correlate with proliferative index and that the subcellular localization of p27^KIP1 is a negative prognostic factor in some tumors, indicates that p27^KIP1 may not function as a CDK inhibitor in these cases. Another study shows that the haploinsufficient p27^KIP1 heterozygous mouse is more susceptible to tumor formation than the p27^KIP1 null animals in mammary and prostate tumor models, suggesting an active contribution of the remaining p27^KIP1 allele to tumor development (Gao et al., 2004). The group interpreted the enhanced susceptibility to be a result of partial down-regulation of the cyclin/CDK inhibitory activity but maintenance of other pro-tumorigenic roles. Mice carrying the p27^KIP1 S10A allele, which renders p27^KIP1 to nuclear localization, show a partial resistance to urethane-induced tumorigenesis, despite reduced overall abundance of the p27^KIP1 protein (Besson et al., 2007). This study further supports the correlation between cytosolic localization and tumorigenic activity of p27^KIP1.

The current model generally accepted is that p27^KIP1 suppresses tumorigenesis by inhibiting cyclin/CDK activity in the nucleus, but exerts other functions in the cytoplasm that are potentially oncogenic. While both roles may be important for homeostasis, inactivation of the nuclear function and/or exaggeration of the cytoplasmic functions may promote tumor progression. The cytosolic functions include the regulation of the actin cytoskeleton and cell migration through modulation of RhoA activity. Cytosolic activities are generated by the C-terminal portion of the p27^KIP1 protein (Besson et al., 2004; Batsi et al., 2009). During mouse embryonic development, this function is critical for the migration of cortical neurons (Nguyen et al., 2006). Interestingly, cyclin/CDK binding is through the N-terminal portion of p27^KIP1, and this same region modulates differentiation of neuronal progenitors in vivo via stabilization of neurogenin-2, independent of cyclin/CDK interaction (Nguyen et al., 2006).

Conclusion

The p27^KIP1 cyclin dependent kinase inhibitor displays apparently contradicting roles by acting as a classic tumor suppressor in one instance, and a pro-tumorigenic oncogene in the other. As its primary function as a CKI is to bind and inhibit cyclin/CDK complexes, the p27^KIP1 protein functions throughout all cell cycle phases by interacting with different cyclin/CDK complexes. However, this cell cycle inhibitor has emerged to play important roles in other cellular functions, such as cell migration. Being a substrate of several important mitogenic signaling kinases, the p27^KIP1 protein is post-translationally modified by phosphorylation. Mitogen-induced phosphorylation causes alterations in expression levels, intracellular localization, and induction of p27^KIP1's diverse functions. The proper regulation of p27^KIP1 expression levels, localization, and differential functions, must be important in maintaining homeostasis and preventing cells from forming tumors. Understanding the diverse roles of p27^KIP1 protein during normal cell cycle progression and tumor development provides new insights into the field of tumor prognosis and therapeutics.

Abbreviations

CDKs

cycle dependent kinases

CKIs

cyclin dependent kinase inhibitors

CRM1

chromosome region maintenance 1 protein

Pin1

peptidyl-prolyl cis-trans isomerase NIMA-interacting 1

RSK1

ribosomal S6 kinase 1

SCF

skp1-cullin1-F-box protein

Skp2

S-phase kinase-associated protein 2