Colorectal cancer (CRC) is 1 of the leading causes of cancer death worldwide, accounting for more than 150,000 new cases and 55,000 deaths annually in the US alone.1 Accurate prognostication for individual patients with CRC is therefore essential for planning and optimizing treatment strategies. Traditional clinical and histologic markers, including tumor differentiation, depth of invasion, and lymph node status, are commonly used today, but the accuracy of these parameters is still unsatisfactory in some clinical settings. For example, the variability of overall survival of patients with Stage II-III disease at 5 years may be as high as 40%.2 Thus, it is not rare that patients with similar tumor grade and disease stage may display significantly different clinical outcomes. This can lead to extreme difficulty in choosing the correct adjuvant treatment protocol, and may also lead to over- or undertreatment in many patients. Therefore, substantial benefit is likely to result from the characterization of additional prognostic factors more closely related to tumor cell biology (reviewed by Klump et al.3). Moreover, the identification of novel biological molecular markers related to cancer progression may potentially improve tailored application of adjuvant therapy for subgroups of high-risk patients and promote the discovery of novel and specifically targeted interventions.
Newer putative biological markers of interest are associated with cell growth, differentiation, and other factors involved in tumor formation and progression. Because uncontrolled cell proliferation is a main hallmark of cancer, alterations in the expression or activity of proteins that are intimately involved in cell cycle regulation are of particular interest. Under normal circumstances, progression of the cell cycle is largely dependent on the timely activation of cyclin-dependent kinases (Cdks).4–6 Cdks regulate checkpoints that integrate mitogenic and growth inhibitory signals, thereby coordinating cell cycle transitions. The regulation of these kinases is tightly controlled by specific Cdk inhibitors at different phases of the cell cycle. Two families of Cdk inhibitors regulate the different Cdks.7, 8 Members of the Cdk4 (INK4) family, including p15INK4A, p16INK4B, p18INK4C, and p19INK4D, form complexes with Cdk4 and Cdk6 that prevent the binding of cyclin proteins. The CIP/KIP (kinase inhibitor protein) family includes p21CIP1, p27Kip1, and p57Kip2, which inhibit the kinase activity of the G1 Cdk2/cyclin A and Cdk2/cyclin E complexes and thus prevent transition to the S phase. It is recognized by now that some of these inhibitors have important roles in tumor suppression.9 Recent evidence also suggests that among the different Cdk inhibitors, deregulation of p27 probably plays a major role in the progression and tumor behavior of many malignant diseases.9–16
p27 is a 198 amino acid protein whose gene maps to the short arm of chromosome 12, and was originally discovered in cells arrested by TGF-α.17–19 It is by now well established that p27 serves as an important regulator for both cellular proliferation and tissue differentiation. Thus, p27 levels are high in nonproliferative tissue compartments, such as skeletal muscle, cartilage, and mesenchymal fibroblasts, but are low in the progenitor compartments of many rapidly renewing epithelial tissues.20 The essential role of p27 in the regulation of normal cell growth and differentiation is also demonstrated by the finding of increased body size, organomegaly, and spontaneous pituitary tumors formation in mice lacking p27, and by altered differentiation programs in p27 knockout cells.21–24
In the cell cycle, p27 negatively regulates G1-S phase transition. p27 protein levels oscillate during the cell cycle, being highest at the G0/G1 phase and lowest at S phase. However, whereas protein levels change during the cell cycle, p27 mRNA levels remain remarkably constant throughout the cell cycle.7, 25 Thus, the regulation of p27 is mainly determined by its rate of degradation rather than by transcription or translation.25 It has been shown that the degradation of p27 is carried out by the ubiquitin system,26 in which covalent ligation to ubiquitin targets proteins for degradation by the proteasome.27 This liberates Cdk2/cyclin E and Cdk2/cyclin A from their negative control and promotes cell proliferation. The specificity of the ubiquitin system in targeting proteins for degradation is mainly defined by its ubiquitin ligase complexes.27 The machinery involved in targeting p27 for degradation is an SCF (Skp1-Cullin-F-box protein)-type ubiquitin ligase complex that contains Skp2 as the specific substrate-recognizing subunit.28–30 SCF complexes belong to a large family of ubiquitin ligases that contain several constant subunits (called Cullin-1, Skp1, and ROC1) and a variable subunit known as an F-box protein.31 Each F-box protein binds a specific subset of protein substrates and thus promotes their ligation to ubiquitin and subsequent degradation by the proteasome.5 Skp2 is an F-box protein that was originally discovered, along with Skp1, as a protein associated with the S-phase kinase Cdk2/cyclin A32 and hence its name (S-phase kinase-associated protein 2). It was also reported that expression of Skp2 is required for entry of cells into the S-phase and that Skp2 expression was greatly increased in transformed cells in culture.32 In the normal cell cycle, levels of Skp2 are low in G0/G1 and increase in S-phase. The role of Skp2 as the main rate-limiting regulator for the degradation of p27 has been clearly shown in both intact cells and in cell-free systems.28–30
More recently, the essential role of Cks1 (cyclin kinase subunit 1) in facilitating the ubiquitin-mediated proteolysis of p27 via interaction with the Skp2 ubiquitin ligase complex was discovered.33, 34 Cks1 is a member of the highly conserved family of Cks/Suc1 proteins, which interact with Cdks.35 Whereas it was known for more than a decade that Cks proteins are essential for CDK function and cell division in yeast, their precise functions remained poorly understood. Cks1 was found to be essential for efficient Skp2-dependent destruction of p27.33, 34 The critical role of Cks1 in targeting p27 for degradation by Skp2 was indicated in these studies by demonstrating the lack of p27 ubiquitination and breakdown in the absence of Cks1 in vitro and the slow cell proliferation and accumulation of p27 in Cks1 nullizyous mice in vivo.
Low or absent p27 expression has been frequently observed in many human cancers, including breast, prostate, gastric, lung, and CRC.9–16 Moreover, down-regulation of p27 was strongly associated with aggressive tumor biology and poor prognosis.9–16 Deregulation of p27 levels in cancer is complex, but recent studies have shown that the accelerated rate of ubiquitin-mediated degradation is the major pathway leading to these observations (Fig. 1). In this article, we review the current knowledge on the impact of p27 expression on tumor progression and clinical outcome, the mechanisms leading its deregulation, and the potential applications of these findings in CRC patients.
Prognostic Implications of p27 Levels in CRC
In the past few years a number of studies have examined the expression of p27 and its prognostic significance in CRC (Table 1). The first study, published by Loda et al.,13 examined the expression of p27 in tumor samples obtained from 149 patients who underwent surgery for CRC. The expression of p27, as determined by immunohistochemistry, was compared with commonly used clinicopathologic parameters and overall survival. p27 expression was significantly associated with survival by actuarial analysis, with a median survival of 241 months in patients whose tumors displayed p27 in more than 50% of cells, 149 months in tumors expressing p27 in less than 50% of cells, and 69 months in patients that lacked p27 expression. Moreover, when p27 expression was compared with survival in patients with Stage II disease, patients with more than 50% positive tumor cells for p27 had a median survival of 219 months compared with survival rates of 140 and 69 months in patients with tumors that expressed less than 50% positive cells or did not express any p27, respectively. In these patients, lack of p27 expression was associated with a relative risk of dying from CRC of 32.1 (P = .004). Multivariate analysis demonstrated that both disease stage and p27 were independent prognostic predictors of overall survival, whereas age, tumor differentiation, and gender were not. When this study was published, 2 other reports that examined the prognostic role of p27 in breast cancer were published in the same journal.10, 11 These studies also showed that p27 expression was a strong and independent prognostic marker for disease-free and overall survival in breast cancer.
Table 1. p27 as a Prognostic Factor in Colorectal Cancer
Subsequent studies on different groups of patients with CRC have confirmed many of these preliminary observations.14, 20, 36–41 In general, most studies agree that high levels of p27 expression are associated with a well-differentiated tumor histology and good overall survival, whereas low levels are associated with aggressive tumor behavior and poor clinical outcome (Table 1). Palmqvist et al.36 evaluated the expression of p27 in 89 CRC and related p27 levels to clinicopathologic characteristics and overall survival. Tumors expressing less than 50% staining conferred impaired prognosis compared to high p27 expressors (P = .0069). p27 expression correlated significantly with depth of tumor invasion, but not age, gender, or tumor type. Interestingly, in this study it was found low p27 levels more commonly in right colon cancers than in other locations of the colon or rectum (P = .026). This finding was also observed by Zhang et al.,37 but not by others. Galizia et al.38 examined the predictive effect of p27 expression on clinical outcome in 104 colorectal patients undergoing either curative or palliative surgery. Levels of p27 were significantly higher in the 86 patients undergoing curative surgery compared to the 18 patients that underwent noncurative surgery. In those undergoing curative surgery, low levels of p27 (<20%) were significantly associated with poorer disease free survival by both univariate (hazard ratio [HR] = 26.555; P = .001) and multivariate analysis (HR = 11.167; P = .001). p27 was also found to be a reliable predictor of metastatic potential in early CRC.40, 41 It is estimated that 10% to 15% of all patients with early CRC will eventually develop metastases.40, 43 Thomas et al.42 examined the expression of p27 levels from the colon and the site of metastases in 23 patients with synchronous Stage IV disease and in 13 patients who initially had undergone colon resections for Stage I-III disease and subsequently underwent resection for liver or pulmonary metastases when those occurred more than 6 months after the initial operation (metachronous disease). In the synchronous group, 57% of the primary tumor and metastases pairs did not express p27 and the remainders were low expressors. In the metachronous group of patients, 54 of the primary tumors were low expressors and the rest expressed high p27 levels. In this group of patients, there was a significant reduction in p27 levels in the metastases compared to the primary tumor (P = .0023). Hirano et al.41 examined the correlation between altered expression of p27 and risk of metastasis in 66 patients with early CRC. The frequency of low p27 levels were significantly more common (P = .0016) in the primary tumors that eventually developed metastases (81% of cases) compared with tumors that did not. Interestingly, some studies have shown that in cancers where both invasive and in situ tumors coexist, loss of p27 is observed in both components.44 Taken together, these studies suggest that events leading to p27 down-regulation may precede tumor progression in transition from in situ disease to invasive disease and from localized disease to metastatic dissemination.
Loss of p27 Results from Accelerated Ubiquitin-Mediated Degradation Mediated by Its Specific Ubiquitin Ligase Subunits Skp2 and Cks1
Earlier studies examined p27 gene expression in human cancers, but rarely found this gene to be inactivated or mutated.19, 45, 46 In the study mentioned above by Loda et al.,13 they also examined the relation between p27 mRNA and protein expression in CRC and found a significant discrepancy in 59% of the cases. Because cell-cycle regulation of p27 cellular abundance occurs mainly at the posttranscriptional level, they analyzed the p27 degradation rate and found enhanced proteolytic activity resulting from increased proteasome-mediated degradation. These findings were also confirmed in other cancers such as lung cancer, lymphomas, and malignant gliomas.12, 47, 48
The in vitro discovery of the involvement of Skp2 as the specific ubiquitin ligase that targets p27 for degradation raised the question of whether the low levels of p27 in aggressive human cancers may be caused by increased expression of Skp2. To answer that question, we examined the relation between the expression of Skp2 and p27, using both immunoblot analysis and immunohistochemistry on tumor specimens obtained from patients with CRC.49 A strong inverse relation between Skp2 levels and p27 levels was observed in 85% of cases (P < .0001). Thus, decreased levels of p27 were strongly associated with increased levels of Skp2, whereas high levels of p27 coincided with low levels of Skp2. Moreover, a high ratio of Skp2/p27 was found in all poorly differentiated tumors, whereas this ratio was significantly lower in well-differentiated cancers. The role of Skp2 as the main regulator of p27 abundance in malignant tumors was also verified later in other cancersincluding breast, prostate, lung, and oral squamous carcinomas.50–53 The essential role of Cks1 as a cofactor for efficient Skp2 proteasome-dependent destruction of p27 in vitro prompted us to investigate its expression and relation to Skp2 and p27 in CRC.54 A strong correlation was found between Cks1 levels and Skp2 levels (P<.0001) and inversely with p27 levels and loss of tumor differentiation (P<.0001), suggesting an important role for Cks1 overexpression in decreasing p27 levels in these cancers. Representative immunoblot and immunohistochemical studies for Skp2, p27, and Cks1 expression are shown in Figure 2. The role for Cks1 in the deregulation of p27 was also found in other cancers including breast, oral squamous, and gastric carcinomas.55–57 Next, we examined the prognostic implications of Skp2 and Cks1 by immunohistochemical analysis in 80 primary CRC.58 Skp2 or Cks1 were each found to be a novel independent marker for overall survival (Skp2: univariate analysis, P<.001; multivariate analysis, P = .003; odds ratio [OR] 3.48; confidence interval [CI] 1.54–7.84; Cks1: univariate analysis, P = .001; multivariate analysis, P= .019; OR 2.25; CI 1.00–5.05). Subgroup analysis showed that these proteins were also highly prognostic for survival in the 60 patients with Stage II-III disease (P<.001). Moreover, the addition of either staining for Skp2 or Cks1 provided additional prognostic information to that provided by p27 alone. Thus, patients with low p27 levels and high Skp2 levels had low overall survival rates, with a mean survival rate of 32 months, whereas patients with low p27 levels and low Skp2 levels had a significantly better prognosis (P<.001), with a mean survival rate of 54 months. For patients with high p27, a clear and statistically significant difference was found for patients with low Skp2 levels (mean survival of 68 months) compared to 46 months for patients with high Skp2 levels. Similarly, patients with low p27 levels and high Cks1 levels had extremely poor survival (27 months) compared to patients with low p27 levels and low Cks1 levels (63 months).
Whereas the use of clinical and histologic parameters for the determination of prognosis and treatment strategies for patients with CRC is still of great value, they may be distressingly inaccurate in many clinical situations, especially in patients with Stage II-III disease. This may be attributed, at least in part, to differences in the biological behavior of tumors that are determined by altered molecular regulatory mechanisms. Thus, the characterization of molecular changes in CRC in recent years has been the focus of great interest for both researchers and clinicians, because it may lead to the identification of new prognostic markers more closely resembling the biological nature of the cancer. This, in turn, may better delineate the patients with intermediate disease who may benefit from additional adjuvant therapy from those patients in whom additional treatment may not be warranted. Among the various alterations in gene and protein expression in CRC, deregulation of p27 has a profound adverse effect on tumor biology and clinical outcome in many cancers and, in particular, in CRC. The majority of studies have verified that the expression of p27 and its ubiquitin ligase subunits in CRC are accurate novel predictors of prognosis, independent of other commonly used clinicopathologic parameters. Nevertheless, despite the increasing evidence that shows that these markers may more accurately stratify and subgroup patients at risk, at present there are very few studies that have clinically implicated their application in the clinical decision process for appropriateness of adding adjuvant therapy. This is also true for other well-established molecular markers such as p53, cyclin D, and many others. Moreover, some studies have shown that a combination of some of these markers such as p27, p53, and VEGF or p27 and Skp2 may further improve the accuracy of prognostication.38, 41 Thus, it seems that future large-scale studies should address these issues and prospectively examine the utility of these markers in clinical decision making. Because laboratory evaluation of p27 and Skp2 is inexpensive and reliably performed by immunohistochemistry, we believe that routine examination of these proteins should be performed, at least for patients with intermediate disease.
Finally, there is growing interest now in the development of novel therapeutic interventions that may target important specific key elements that promote tumor progression. Because gene therapy at present is not applicable, the ubiquitin-mediated proteolysis pathway has become a focus of anticancer treatment.59 Recent studies have shown that bortezomib, a general inhibitor of the ubiquitin-proteasome system, both in in vitro and in vivo models and in Phase I-III trials, displayed clear evidence of antitumor activity against many cancers, including CRC, with manageable side effects.60–62 Thus, the discovery of drugs that may interfere with specific elements of the ubiquitin-proteasome pathway may potentially further increase the specificity of therapy against cancer cells while minimizing the effects on normal cells. Given the important and highly selective roles of Skp2 and Cks1 in targeting the destruction of p27, and the clear correlation between increased expression of these proteins and aggressive tumor behavior and poor prognosis, it seems that these regulatory proteins may well be considered in the future as novel targets for therapeutic intervention in CRC.