Aberrant cell cycle regulation in cervical carcinoma.

Carcinoma of the uterine cervix is one of the most common malignancies among women worldwide. Human papillomaviruses (HPV) have been identified as the major etiological factor in cervical carcinogenesis. However, the time lag between HPV infection and the diagnosis of cancer indicates that multiple steps, as well as multiple factors, may be necessary for the development of cervical cancer. The development and progression of cervical carcinoma have been shown to be dependent on various genetic and epigenetic events, especially alterations in the cell cycle checkpoint machinery. In mammalian cells, control of the cell cycle is regulated by the activity of cyclin-dependent kinases (CDKs) and their essential activating coenzymes, the cyclins. Generally, CDKs, cyclins, and CDK inhibitors function within several pathways, including the p16(INK4A)-cyclin D1-CDK4/6-pRb-E2F, p21(WAF1)- p27(KIP1)-cyclinE-CDK2, and p14(ARF)-MDM2-p53 pathways. The results from several studies showed aberrant regulation of several cell cycle proteins, such as cyclin D, cyclin E, p16(INK4A), p21(WAF1), and p27(KIP1), as characteristic features of HPV- infected and HPV E6/E7 oncogene-expressing cervical carcinomas and their precursors. These data suggested further that interactions of viral proteins with host cellular proteins, particularly cell cycle proteins, are involved in the activation or repression of cell cycle progression in cervical carcinogenesis.

INTRODUCTION

Carcinoma of the uterine cervix is one of the most common malignancies among women worldwide. The incidence of cervical carcinoma is on the decline, but still 370,000 new cases and 190,000 resulting deaths occur each year.1 The development of cervical carcinoma is a multi-step carcinogenesis. However, the steps in the progression from low-grade cervical intraepithelial neoplasia to carcinoma remain unknown. Human papillomaviruses (HPVs) have been identified as the major etiological factor in cervical carcinogenesis.2 Epidemiological evidence indicates that the majority of cervical neoplasia is attributable to HPV infection. Also, although certain HPV genes are capable of immortalization and can cooperate in the process of transformation, not all non-invasive lesions progress to the full malignant phenotype. This suggests that other cofactors are required.

Recently, it has become increasingly clear that HPVs replicate by interfering with normal cell cycle control mechanisms. Malignant transformation is also intimately related to these processes, and it is likely that the oncogenic potential of papillomaviruses lies in their ability to alter cell cycle checkpoints, thereby leading to the accumulation and transmission of genetic abnormalities. As we know, the E6 and E7 oncoproteins of high-risk HPVs, particularly HPV 16 and 18, bind respectively to the p53 and Retinoblastoma (Rb) tumor suppressor proteins, which are involved in the regulation of growth control. Moreover, the E6 proteins of high-risk HPVs bind more effectively to the p53 protein, leading to their degradation via a ubiquitin-mediated pathway, than those of lower-risk HPVs.

The development and progression of cervical carcinoma has been shown to be dependent on various cellular genetic and epigenetic events, especially alterations of the cell cycle machinery at various checkpoints. The precise control of the cell cycle in mammalian cells is regulated by the activity of cyclin-dependent kinases (CDK1, CDK2, CDK4, CDK6) and their essential activating coenzymes, the cyclins (cyclins A, B, D, E). The kinase activities of these CDKs are regulated by the abundance of their partner cyclins, phosphorylation by various kinases, de-phosphorylation by cell cycle phosphatases, and interaction with CDK-inhibitory proteins (CDKIs).3-5

The CDK family is an important group of molecules that regulate cell proliferation. In addition, two classes of mammalian cyclin-dependent kinase inhibitors (CDKIs) have been described: the CIP/KIP family, comprised of p21, p27, and p57, and the INK4 family, comprised of p15, p16, p18, and p19.6 The INK4 molecules specifically inhibit cyclin D complexes by interaction with the CDK4 and CDK6 components. The KIP family is promiscuous, affecting cyclin E, cyclin A/CDK2, and cyclin B/CDK1 by binding both the cyclin and CDK subunit.3

In this article, we review the current knowledge of the role of cell-cycle regulatory proteins and their alterations in cervical cancer. CDKs, cyclins, and CDKIs generally function within several defined pathways, including the p16INK4A-cyclin D1-CDK4/6-pRb-E2F pathway, the p21WAF1-p27KIP1-cyclinE-CDK2 pathway, and the p14ARF-MDM2-p53 pathway.7 Each of these components plays either a positive or a negative role in cell-cycle control mechanisms in cervical carcinogenesis. Alterations in CDKs, CDKIs, and cyclins can lead to uncontrolled proliferation and might contribute to malignant transformation of the uterine cervix.

The p16INK4A-cyclin D1-CDK4/6-pRb-E2F pathway

p16INK4A

The p16INK4A gene maps to 9p21, contains three exons, and encodes a nuclear phosphoprotein with a molecular weight of 16 kDa. The p16 protein functions in the negative regulation of the cell cycle through the inhibition of cyclin-dependent kinases 4 and 6 and interactions with cyclin D1.8 In the absence of p16, CDKs bind to cyclin D1, and the Retinoblastoma protein (pRb) is phosphorylated. Phosphorylation of pRb leads to its deregulation at the G1/S checkpoint, and cell proliferation is switched on. In a variety of human malignant tumors and cell lines, the p16INK4A gene is inactivated by various genetic mechanisms, including point mutations, homozygous deletions, and hypermethylation of CpG islands in the p16INK4A promoter. Kim et al. found a high percentage of p16 exon 2 mutations in cervical cancer specimens.9 Dong et al. and Nuovo et al. also described hypermethylation of the p16 promoter, documenting inactivation of the gene as a frequent epigenetic event in cervical carcinoma.10,11 However, mutations in p16 in these lesions have not been confirmed.10 In some reports, mutation and deletion of p16 are suggested to be rare events: no gene alterations were detected in most studies of primary cervical carcinomas or cervical cancer cell lines, indicating that they are not required for the development of this tumor type. These data are concordant with those of previous reports.12-14

Conflicting results about the expression of p16 protein have been obtained from studies of the role of p16 in the pathogenesis of cervical cancers.9,13,15,16 Overexpression of the p16 protein is a characteristic of displastic and neoplastic alterations of the cervical epithelium. The portion of p16-positive samples increases as the tumor progresses from the CIN I to the invasive carcinoma stage.16 Some reports have shown that p16 expression is detectable by immunohistochemistry in cervical neoplasia and that this expression may be a direct result of HPV infection with resultant inactivation of pRb, which is known to bind p16.15,16 Squamous cell carcinomas (SCCs), high-grade squamous intraepithelial lesions (HSILs), and adenocarcinomas (ACs) of the cervix have shown increased p16 immuno-staining.17 Klaes et al. examined the utility of p16 immunostaining to assess inter-observer agreement in the diagnosis of cervical intraepithelial neoplasia (CIN). In that study, 194 cervical cone biopsies were examined and stained with an antibody to p16. A significant increase in inter-observer agreement when examining p16 expression was observed. The p16 expression was restricted to cervical cancer, CIN 2-3, and those cases of CIN 1 associated with high-risk HPV types.

Kelley et al., however, reported that no alterations in p16 expression were detected in cervical cell lines, including 8 HPV-positive cell lines, 2 HPV-negative cell lines containing mutant Rb, and one tumorigenic cell line derived from normal cervical cells following transfection with HPV-16 and v-H-ras (CX16-2HR).14 These results confirm that mutational inactivation of p16INK4 is a rare event in tumor samples with compromised Rb activity. Volgareva et al. reported that all stages of CINs and carcinomas analyzed were heterogeneous with respect to p16 expression. Samples that showed expression of p16 in 25% of cells or more were identified in a side-by-side study with samples that stained poorly or lacked staining. These samples are evidence that p16-negative cervical neoplasms and carcinomas do exist.18

The molecular basis for overexpression of p16 in cervical squamous carcinoma remains unclear. Some explanations include a genetic mutation or a supragenetic event at the protein level, such as increased half-life of the p16 protein. The interaction of p16 with pRb is thought to be central to the role of p16 in controlling cell cycle progression. It has been suggested that HPV infection leads to HPV-E7 binding to pRb, which in turn results in increased p16 expression. Some studies have suggested that this process only occurs in cases of infection by high-risk types of HPV: p16 transcription may also be directly induced by the transcription factor E2F released from pRb after binding of the viral oncoprotein E7.19

p16 has been used as a biomarker for dysplasia in the diagnosis of cervical squamous lesions and has the potential to be used as an additional screening tool.8,20,21 The overexpression of p16 is closely associated with high-risk HPV infection and high-grade CIN.22 A recent report showed that in cervical biopsy specimens, the staining pattern of p16 and a high percentage of p16-positive cells are closely related to infection with high-risk HPV types 16 and 18, and with CIN 2/3. Several studies of p16 in liquid-based cytology specimens suggested that p16 could be used in cervical screening as a marker for persistent high-risk HPV infection, and HSIL could also be useful in resolving ambiguous cases involving a differential diagnosis of cervical neoplasia (Table 1).23,24

Table 1

Expression Status of p16 in Squamous Cervical Carcinoma

cGIN, cervical glandular intraepithelial neoplasia; CA, Cervical adenocarcinoma; SCC, squamous cell carcinoma; CIN, cervical intraepithelial neoplasia; LSIL, low-grade squamous intraepithelial; HSIL, high-grade squamous intraepithelial; IC, invasive carcinoma; CIS, carcinomas or carcinoma in situ; GCA, genital condyloma acuminate; DC, dysplasia of the cervix; LR-HPV, low-risk HPV; MSP, methylation-specific PCR; PCR-SSCP, polymerase chain reaction-single strand conformation polymorphism.

Cyclin D1

The cyclin D1 (PRAD-1, CCND-1) gene maps to 11q13 and shows the characteristics of a cellular oncogene. Expression of cyclin D1 moderately oscillates throughout the cell-cycle, reaching peak levels in G-phase.7 Cyclin D1 serves as a key sensor and integrator of extracellular signals in early to mid-G1 phase, mediating its function through binding the CDKs, histone acetylase, and histone deacetylases to modulate local chromatin structure around the genes that are involved in regulation of cell proliferation and differentiation.25 In addition to CDK-binding functions, a body of evidence now indicates that D-type cyclins have CDK-independent properties. These properties are important for cellular growth, metabolism, and cellular differentiation. Cyclin D1 forms physical associations with more than 30 transcription factors and transcriptional co-regulators.25 Several nuclear receptors, including the androgen receptor (AR), estrogen receptor (ER), thyroid hormone receptor, and peroxisome proliferator-activated receptor (PPAR), bind directly to cyclin D1 within cultured cells.26,27 Both basal and ligand-dependent transactivation of nuclear receptors is regulated by cyclin D1.

Genetic aberrations in the regulatory circuits that govern transit through the G1 phase of the cell cycle occur frequently in human cancer, and overexpression of cyclin D1 is one of the most commonly observed alterations.25 One model suggests that the overexpression of cyclin D1 may serve as a driving force through its cell-cycle regulating function. Cyclin D1 is amplified and/or overexpressed in a substantial proportion of different human tumors.

The role of cyclin D1 in cervical carcinogenesis is not clearly understood, and controversial results have been described. Cho et al. found that cyclin D1 levels were significantly lower in HPV-positive HSIL, invasive SCC, or adenocarcinoma compared to HPV-negative cases and normal cervical epithelium, consistent with other authors' results.28-30 Contrary to this, the results of Nichols et al. described elevated cyclin D1 mRNA levels in invasive cervical cancer that were not associated with increased protein amounts. Additionally, underexpression of cyclin D1 was shown to be regulated at the transcriptional level in another study.29-32 Almost all research data suggests that cyclin D1 expression is not found to correlate significantly with clinicopathological factors (histological type, tumor grade and stage, patient's age).33,34 Conversely, despite an underexpression of cyclin D1 in cervical carcinomas compared to normal cervical epithelium, positive immunostaining, which was found in 28-32% of invasive carcinoma samples, was significantly associated with poor disease-free and overall survival.29,34 The latter report is the only one showing cyclin D1 gene amplification in cervical carcinomas (24%), but this was not correlated with overexpression.

Contradicting results have been reported about overexpression of cyclin D1 in immunostaining studies of cervical carcinoma. One study showed overexpression of cyclin D1 in 3% of patients, while underexpression was reported in another study.31,33 No significant increase in cyclin D1 protein levels, despite overexpression of cyclin D1 mRNA demonstrated by in situ hybridization, was reported in cervical carcinoma.28,29 These discrepancy might be attributed to the use of different antibodies, different scoring criteria for the detection assay, and the varying tumor tissue characteristics in different studies. In addition, some studies showed that the level of cyclin D1 was significantly lower in CIN and SCC compared with normal epithelium and that these levels correlated significantly with HPV positivity.28,33 Also, there is a low prevalence of G1 cyclins in cell lines with a mutated Rb gene, and DNA tumor virus infection can supplant tumor cell requirements for cyclin D1 protein.

Cyclin D1 levels are reported as significantly lower in HPV-positive LSIL, HSIL, invasive SCC, or AC compared to HPV-negative cases and normal cervical epithelium.28,30,35 Cyclin D1 and HPV E7 possess similar binding regions for pRb and pRb-related pocket proteins, and inactivation of pRb either by the cyclin/CDK complexes in G1 or by interaction with the high-risk HPV oncoprotein E7 may result in a decreased expression of cyclin D1 (Table 2).

Table 2

Expression Status of Cyclin D1 in Squamous Cervical Carcinoma

CIN, cervical intraepithelial neoplasia; SCC, squamous cell carcinoma; LSIL, low-grade squamous intraepithelial; HSIL, high-grade squamous intraepithelial; LR HPV, low risk HPV; HR HPV, high risk HPV; CI, cyclin index; RT-PCR, real-time polymerase chain reaction. a not published. *Authors of this review.

CDK4

The D-type cyclins (D1, D2, and D3) and their catalytic partners CDK4 and CDK6 act early in the G1 phase of the cell cycle.3 Mitogen-induced signal transduction pathways promote the activation of cyclin D/CDK complexes at many levels, including gene transcription, cyclin D translation and stability, assembly of D cyclins with their CDK partners, and import of the holoenzymes into the nucleus, where they ultimately phosphorylate their substrates. The cyclin D-dependent kinases (CDK4 and CDK6) can phosphorylate Rb family members (Rb, p107, and p130), thus helping to inactivate their transcriptional corepressor activities.

Aberrantly expressed CDK4 could play an important role in cervical tumorigenesis. It is postulated that CDK4 oscillates between the INK4 and KIP inhibitors, blocking their suppressor activity. In cervical cancer, the demonstrated lower levels of INK4 molecules and the high levels of CDK4 would favor binding of the more abundant KIP inhibitors to these kinases, undermining their inhibition of cyclin E. Thus, in this situation Cyclin D is expendable. The E7 would deregulate pRb initially, unleashing E2F-induced cyclin E expression; the overexpressed CDK4 would tether the KIP molecules, allowing cyclin E to become sufficiently active to phosphorylate and inactivate pRb and p27, perpetuating its own activity and that of E7.4,36,37

Yoshinouchi et al. found overexpression of CDK4 in 72.6% of cervical cancer specimens;38 this value was consistent with previous studies of cervical carcinoma.33,34 In another study, CDK4 gene amplification was described in 25% of cervical cancers, whereas no mutations in exon 2 of the CDK4 gene were found.34 To determine whether alterations of p16 might be involved in HPV-positive cervical cancers, Yoshinouchi looked for gene alterations and changes in the ability of the p16 protein to interact with CDK4 in 5 cervical cancer cell lines. No alteration of this gene was detected, and the p16 and CDK4 proteins were normally expressed. Additionally, the ability of p16 to interact with CDK4 was not abrogated in these cell lines. These cell lines were HPV-positive and carried wild-type p53 genes. These findings suggest that phosphorylation of pRb by CDK4 is not critical in the carcinogenesis or in the establishment of HPV-positive cervical cancer cell lines, since the HPV viral-transforming proteins E6 or E7 inactivate p53 and pRb tumor suppressor protein function, resulting in deregulated progression of the cell cycle.38 So far, research studies have not found a correlation between amplification and overexpression of CDK4 and patient age, histological tumor type, or tumor grade or stage.33,34

To release cells from G1 arrest and to promote entry into S-phase, pRb is phosphorylated, and thereby inactivated, by the cyclin D1/CDK4 complex. This inactivation may also be achieved by the interaction of pRb with the viral oncoprotein E7, leading to the hypothesis that up-regulation of positive regulators upstream of pRb might be dispensable in high-risk HPV-infected cervical carcinomas. Interestingly, very recent work has indicated that cyclin D/CDK4 complexes also phosphorylate Smad3, negatively regulating the functions of transcriptional complexes that mediate cell growth inhibition by proteins of the TGF family.39 Importantly, several lines of evidence indicate that cyclin D/CDK complexes play a second noncatalytic role in G1 progression by sequestering proteins of the Cip/Kip family, including p27KIP1 and p21CIP1, two potent inhibitors of CDK2.40,41

pRb/E2F

The retinoblastoma tumor suppressor gene (Rb) encodes the nuclear phosphoprotein pRb (p105), which has been found mutated or deleted in several types of human cancer. p105Rb and other Rb family members p107 and p130 regulate the activity of E2F transcription factors.42 Complexes consisting of E2F and hypophosphorylated p105Rb repress the transcription of genes that are required for cell cycle progression, and repression is relieved by CDK-mediated phosphorylation of p105Rb.43 pRb is subject to regulation by many factors, including E2F and cyclin D1. The hypophosphorylated pRb, complexed with a transcription factor, serves as a transcriptional activator of cyclin D1 by binding to its promoter. On the other hand, inactivation of pRb by phosphorylation via the cyclin D/CDK complex in late G1 would not only unleash E2F transcription factors, but would also decrease cyclin D1 expression.3,44 D-Type cyclins interact with pRb pockets through their NH2-terminal L-X-C-X-E motifs. These are similar to DNA virus oncoprotein region 2, such as region 2 from HPV E7.45 Thus, the oncoprotein may occupy the pRb pocket and displace E2F factors, preventing pRb/E2F from inducing cyclin D1 transcription and undermining its normal growth-suppressive function.

Some immunohistochemical studies of pRb in human cancer and in normal tissues, including the uterine cervix, have shown that pRb is expressed in mature and differentiated cells. Both binding and degradation of the Rb proteins by the HPV E7 protein are essential for sustained proliferation of HeLa cervical carcinoma cells, and E7 repression triggers senescence at least in part by activating the Rb pathway in both HeLa and HT-3 cells.42 Salcedo et al. observed that the majority of cells in the proliferating basal layer of the normal epithelium displayed low levels of pRb. pRb immunostaining in invasive cervical lesions is frequently lower than in SIL.43 This low expression of Rb may result from Rb gene mutations or downregulation mechanisms, but may also be related to pRb inactivation resulting from complex formation with high-risk HPV E7 oncoproteins.

Rb immunoreactivity was not affected by high-risk HPV infection in most studies of cervical squamous cell carcinoma.46 In contrast, Rb staining was even more frequently found in SIL and invasive cancer (93%) compared to normal epithelium or reactive atypia (78%).47 Mutations in the Rb gene seem to be rare events in cervical cancer.46,48 In most studies, Rb gene expression did not strictly correlate with the HPV status,49 although Scheffner et al. reported a higher Rb mutation rate in HPV-negative than in HPV-positive cervical cancer cell lines. These mutations resulted in aberrant proteins that were not phosphorylated and unable to complex with the adenovirus E1A oncoprotein.50 Cyclin D1 is the regulatory subunit of the holoenzymes that phosphorylate and, together with sequential phosphorylation by cyclin E/CDK2, inactivate the cell-cycle inhibiting function of the pRb protein. The hypophosphorylated form of pRb, complexed with E2F, serves as an activator of cyclin D1 transcription by binding to its promoter, thus driving cyclin D1 in early and mid G1-phase of the cell cycle. As D-type cyclins and HPV E7 possess similar binding regions for pRb and pRb-related pocket proteins, inactivation of pRb either by the cyclin/CDK complexes in G1 or by interaction with the high-risk HPV oncoprotein E7 may result in a decreased expression of cyclin D1 (Fig. 1).43,51

Fig. 1

Schematic presentation of the HPV viral oncoproteins E6/E7 and cell-cycle regulatory proteins in cervical carcinogenesis.

The p21WAF1/CIP1-p27KIP1-cyclin E-CDK2 pathway

p21WAF1/CIP1

p21WAF1/CIP1 is a cyclin-dependent kinase inhibitor that associates with a class of CDKs and inhibits their kinase activities, leading to cell cycle arrest and the dephosphorylation of pRb. A large body of evidence suggests that p21WAF1/CIP1 plays an important role in cell fate decisions during growth and differentiation. The p21WAF1/CIP1 protein is a p53-inducible protein that inactivates the cyclin/CDK complexes, blocking the cell cycle progression in the G1-S transition. p21WAF1/CIP1 is expressed in cells undergoing either G1 arrest or apoptosis by p53-dependent or -independent mechanisms.52,53 Mutations in the p21WAF1/CIP1 gene were very rarely detected in cervical carcinoma.

p21WAF1/CIP1 expression usually correlates with favorable prognoses in ovarian, gastric, colorectal, and superficial bladder cancers and in esophageal squamous cell carcinoma. However, in cervical cancer, the conclusions about p21 expression and its prognostic importance vary considerably. Many authors found increased p21 expression in invasive carcinomas33 and an increase in the number of p21-expressing cases during the progression from normal epithelia through precancerous lesions to invasive cervical cancer.29,54 However, others detected an underexpression of p21 in micro-invasive and invasive cervical cancer compared to normal cervical epithelium.55,56 Van de Putte et al.57 did not find any expression of p21 in normal squamous cervical epithelium, in agreement with Giannoudis and Herrington58 and Skomedal et al.,33 but in contrast with other reports.29,55 Longer antibody incubation and the use of chronic cervicitis as normal controls may partially account for the differences. However, increased expression of p21 in SCC compared to normal epithelium was found in all cases in the abovementioned studies.29,33,55,59 An increase in p21 levels in invasive cancer seems at odds with its function of cell cycle control. However, the p21 level is probably increased in a futile attempt to overcome its impaired or bypassed function. In squamous cell carcinoma, its function could be impaired through the inactivation of p21 by the HPV-16 E7 oncoprotein. Other possibilities include inactivating mutations, mutation of its targets, overexpression of its targets sufficient to titrate p21, or overexpression of proteins in more downstream pathways in cell cycle control.29

On the other hand, Lu et al. reported that expression of p21WAF1/CIP1 was correlated with a favorable prognosis in adenocarcinoma of the uterine cervix.60 In most studies, however, there was no significant association of p21 expression and clinicopathological factors or prognosis.29,55,61,62 Although the literature on the prognostic significance of p21 expression in other human cancer types is extensive, no clear picture can be drawn.

p27Kip1

p27Kip1 is a negative regulator of the G1 phase of the cell cycle. The p27Kip1 gene is a tumor suppressor gene and is frequently lost in tumor cells. It has been implicated in the negative regulation of cell proliferation in response to extracellular signals and is induced upon serum deprivation. In normal epithelial cells, increased expression of p27Kip1 mediates the arrest of cells in the G1 phase of the cell cycle when induced by TGF-β, contact inhibition, or growth in suspension.63 p27 associates mainly with the cyclin E/CDK2 complex and, through this complex, inhibits pRb phosphorylation. Overexpression of p27 blocks the cell from entering S-phase.64 Although p27Kip1 is a putative tumor-suppressor gene, mutation or homozygous deletion of this gene is rarely found in human cancer.65 Evidence that p27 may be involved in human tumor progression comes largely from studies that have directly measured the expression of p27 protein in clinical tumor samples using immunohistochemical assays. A decrease or absence of p27 protein expression has been shown to be associated with more aggressive clinical behavior in a variety of human tumors, including breast, lung, colon, gastric, and ovarian carcinoma.65

Goff et al. found that p27 was strongly expressed in the normal cervix, in both the epithelium and the stroma.66 Other investigators have also found high levels of p27 expression in normal cervix.33,67 Shiozawa et al. revealed that the expression of p27 in the normal squamous epithelium of the uterine cervix was observed mainly in the cells of the intermediate and superficial layers, but the expression of p27 was weak in the parabasal cells, which replicate most actively.64 This is consistent with a previous study by Troncone et al.68 In contrary, some studies have indicated that p27 expression is significantly lower in patients with cervical carcinoma compared with normal epithelium and dysplasia.67,68 Huang et al. showed that expression of this protein was absent in 55.3% of the invasive carcinoma tumors.65 The trend of reduced p27 expression in microinvasive and invasive carcinomas further supports the notion that p27 plays a tumor suppressor function during neoplastic transformation in cervical epithelium.

Sgambato demonstrated that a decrease in p27 expression is associated with the development of cervical cancer and may play an important role in the early stage of cervical tumorigenesis.69 In fact, in that study, p27 expression was already reduced in preinvasive lesions of the cervix and became progressively more evident during the progression from low- to high-grade SIL and from SIL to carcinoma. This result of p27 expression in cervical carcinomas is in agreement with similar findings in several other types of human malignancies63 and with previous reports of cervical cancers.64,65,68

The lower levels of p27 in invasive cervical carcinomas compared with normal epithelium might be explained simply as a direct consequence of the increased cell proliferation, reflecting the fact that the time spent by the neoplastic population in the early part of the cell cycle is reduced.68 Shiozawa et al. demonstrated that p27 seemed to be involved in the regulation of growth and differentiation in the normal squamous epithelia of the uterine cervix.64 Thus, the apparently normal function of p27 may be preserved in preinvasive CIN lesions. However, invasive SCC samples showed a paradoxic overexpression of p27, which may represent aberrant regulation and/or function of the p27 protein.

Most studies indicate that low levels of p27 are not found to be predictive for reduced survival and do not correlate with age, menopausal status, tumor stage, grade of differentiation, histological type, or HPV infection.33,62,65,67 However, Huang et al. demonstrated that loss of p27 expression was significantly associated with lymph node metastasis.65 Contrary to that, in a study by Dellas et al., high levels of p27 were associated with poor survival in invasive cervical carcinomas of clinical stage IB.70

In summary, p27Kip1 expression is often shown to be detectable only in a small percentage of cervical cancers33,67 and to be lower in cervical carcinoma compared with normal epithelium and precancerous lesions.62,64,65,68 No relationship was observed between expression of p27Kip1 and proliferative activity in cervical cancer as well as in other cancers, suggesting that deregulated expression of p27Kip1 might contribute to tumor formation through mechanisms other than increased cell proliferation.64,68

Cyclin E

Cyclin E exists in two isoforms with high homology, designated cyclin E1 and E2. No major differences in expression or function between cyclin E1 and E2 have been found, and their expression has been assumed to be governed by the same molecular circuitry.71,72 Cyclin E, whose catalytic partner is CDK2, is another rate-limiting regulator of the G1 phase of the cell cycle,73 and increased expression of cyclin E has been found in several types of tumors. Appropriate regulation of this cyclin is essential for S-phase transition and numerous processes that determine the accuracy of chromosome replication. It can play a role similar to that of cyclin D1, driving the cell cycle by phosphorylation of Rb, p107, and p130 and the subsequent release of E2F and transcription of key proteins. Cyclin E appears in late G1 after passage through the restriction point. The level of cyclin E peaks in late G1 and disappears again in early S phase.

Even though gene amplification and post-transcriptional modification are the common causes of aberrant cyclin E expression in different malignancies, high-risk-HPV oncoprotein-associated mechanisms might contribute to the aberrant cyclin E expression in cervical lesions. Untethering of E2F factors by the action of the HPV E7 protein could result in the overexpression of cyclin E. This connection, however, is not enough to explain the highly expressed cyclin E in cervical epithelium. The HPV oncoprotein E7 was demonstrated in complexes with cyclin E and cyclin A, thus eliminating phase-dependent variation in activity. Also, the HPV E7 was found to interact with p21 and thereby block p21-mediated inhibition of cyclin E-associated kinase activity.5,75

In normal cervical squamous epithelium, cyclin E expression was not immunohistochemically detectable or was very weak.28,64,67,74,76 However, cyclin E expression was increased in both low- and high-risk HPV-infected squamous and glandular lesions.28,70 Cyclin E expression correlated strongly with morphologic features of the HPV-infected lesions, and has been observed in intermediate, and partly in superficial, cells from LSIL and HSIL, implying sequential dysregulation.77 A positive correlation between cyclin E expression and cell proliferation rate in precancerous lesions was demonstrated by some studies.70,78 In contrast, Kanai et al. showed lower proliferation rates in cyclin E-positive than in cyclin E-negative squamous cervical carcinomas, and Quade et al. found cyclin E staining more frequently in more differentiated areas of invasive cancer.74,79 HPV-infected cancer was found to express cyclin E significantly more frequently than HPV-negative cases, but this expression was not related to prognostically important factors including age, tumor size, and tumor type.67 This is in accordance with the only other study on the prognostic significance of cyclin E in squamous carcinoma.

Cyclin E overexpression or over-activity may be the direct result of the presence of the HPV E6 and E7 proteins in invasive disease. The presence of HPV E7 leads to inactivation of pRb and increased transcription of cyclin E. The HPV E7 oncoprotein inactivates p27, leading to decreased inhibition of cyclin E.75 Southern et al. reported an up-regulation of cyclin E and cyclin A expression in cases of both low-risk and high-risk HPV infection.76

CDK2

CDK2 probably functions as an effector through phosphorylation of key substrates such as the transcription factor p53, the cell cycle regulating phosphatases cdc25A/cdc25C, MDM2, BRCA1, and the transcription factor E2F1.80 CDK2 can have an apoptosis-sensitizing effect and can arrest the cell cycle at different stages.

E-type cyclins (E1 and E2) govern the activity of the single catalytic subunit of CDK2. Unlike various combinations of D-type cyclins that are expressed in different cell types, cyclin E-CDK2 complexes are periodic and maximal at the G1- to S-phase transition.81 Cyclin E-CDK2 also preferentially phosphorylates pRb at different sites from the cyclin D-dependent kinases, and these modifications may differentially impact on the interactions of pRb with E2Fs, histone deacetylases, and other chromatin-remodeling proteins.82 One of the most significant consequences of pRb inactivation is activation of cyclin E/CDK2 subunits, often as a result of increased cyclin E expression. Cyclin E/CDK2 complexes can themselves participate in maintained inactivation of pRb in tumor cells.

The HPV oncoprotein E7 may abrogate p21-mediated sequestration of cyclin E and render its catalytic partner CDK2 resistant to p27.83 Expression of CDK2 was observed in parabasal cells of normal squamous epithelium, which was inversely correlated to p27 expression. Shiozawa et al. revealed that no p27 expression was observed in atypical epithelial cells of SIL with elevated expression of Ki-67, cyclin E, and CDK2.64 The expression of cyclin-dependent kinases (CDK2 and cdc2) was also increased in a considerable number of invasive cervical carcinomas.64,79 In those cases, p27 expression was also high and retained the ability to bind to CDK2, but the p27/cyclin E/CDK2 complex still possessed phosphorylating activity.64 CDK2 also regulates the cell cycle through its interactions with cyclin A2; thus its functions extend beyond G1 regulation to govern events in S and G2/M as well. Whereas maximal periodic cyclin E-CDK2 activity is detected at G1/S, low levels of cyclin A-CDK2 activity are first detected in late G1 phase.

Consistent with crucial roles for cyclin E/CDK2 downstream of pRb, many tumor cells are extremely sensitive to inactivation of cyclin E/CDK2 whether or not they express pRb. This conclusion has been drawn from a multitude of studies demonstrating antiproliferative effects of overexpression of p27Kip1, a protein inhibitor of CDK2, or of dominant-negative CDK2 subunits. Further, injection of antibodies against the CDK2 activators cyclin E and cyclin A blocks proliferation, as does treatment of many different cells with CDK2 inhibitors.84,85

The p14ARF-MDM2-p53 pathway

p14ARF

The CDKN2A gene on human chromosome 9p21 encodes two distinct proteins, p14ARF and p16INK4A, which arise from the same gene by alternative mRNA splicing.86 p14ARF has a mass of 15 kDa and is translated from mRNAs bearing a unique first exon, called 1-beta, located 15 kb upstream of the exon 1-alpha of p16INK4A. Both transcripts share common exons 2 and 3. p14ARF functions as a cell cycle regulator, stopping cell growth at the G1-S border and also at G2-M.87 p14ARF interaction inhibits the MDM2 oncoprotein, thereby blocking formation of the MDM2-p53 complex and preventing MDM2-induced p53 degradation.88

Kanao et al. reported that the overexpression of p14ARF and p16INK4A is strongly associated with HPV-positive cervical cancers and that reduced expression of p14ARF and p16INK4A is correlated with HPV-negative cervical cancers.86 These findings may indicate that impaired p14ARF and p16INK4A mRNA expression contribute to tumor development in HPV-negative cervical cancers by failure to support p53 and Rb instead of their inactivation by HPV E6 and E7.

In human fibroblasts, it has been shown that p14ARF-induced cell cycle arrest is p53-dependent and is abrogated by the co-expression of the HPV E6 protein.89,90 Therefore, upregulation of p14ARF is a consequence of inactivation of p53 rather than a sign of cell cycle inhibition. Brooks et al. demonstrated that the expression levels of p14ARF were clearly elevated in 84% of cervical cancers and 79% of CIN III relative to matched normal tissue by RT-PCR.91 Interestingly, immunocytochemical analysis revealed that p14ARF was abundantly expressed in almost all cases of both cervical SCC and CIN III and was overexpressed with p73 in a great majority of cases. The frequent simultaneous overexpression of p14ARF with p73 is consistent with E2F1-driven expression of both genes. p73 has structural and functional homologies to p53, including sequence-specific DNA binding and transactivation.92

MDM-2

The cellular protein MDM-2 is encoded by the mdm-2 oncogene and has been shown to bind to the p53 protein, inactivating its ability to function as a transcription factor.93 MDM-2 is a zinc finger protein composed of 491 amino acids. The MDM-2 protein interaction with p53 to form the p53/MDM-2 complex covers the p53 N-terminal acidic or activation domain required for activation of transcription.94 This is the same region of the p53 DNA binding domain bound by the HPV E6 protein, thus blocking transcriptional activation. Whether MDM2 acts solely through the inhibition of p53 or has additional activities remains to be determined.

The mdm-2 gene can be de-regulated in soft-tissue sarcomas, bladder carcinomas, renal-cell carcinomas, and breast carcinomas.95-97 Dellas et al. analyzed the altered patterns of MDM-2 and p53 expression in cervical neoplasms and detected MDM-2 as well as p53 in all histological grades of cervical neoplasia.95 The fraction of positive cells for MDM-2 ranged from 0 to 31% in CIN cases and from 0 to 81% in invasive carcinomas. The percentage of p53-stained neoplastic cells ranged from 0 to 40% in CIN cases and from 0 to 69% in invasive carcinomas. Dellas et al. demonstrated that aberrant MDM-2 and p53 expression is frequently found in cervical neoplasia. Skomedal et al. found that MDM2 may protect against HPV-induced p53 protein degradation.95

In contrast to the above-mentioned results, Lie et al. reported that there is no correlation between HPV status and expression of the cell cycle regulators p53, MDM-2, and p21.54 Inactivation of p21 and p53 protein may be important, and MDM-2 abnormalities seem to play a minor role in the development of high-grade CIN. Ikenberg et al. found that p53 mutation and amplification of the mdm-2 oncogene are rare, even in HPV-negative primary cervical carcinomas.98 In contrary, Tsuda et al. found strong expression of MDM-2 was higher in invasive cancers (32.1%) than in CINs (7.1%).12 Troncone et al. reported that immunostaining for MDM-2 and p21WAF1 was not abrogated in invasive cervical cancer by high-risk HPV genomic sequences, and there were no significant differences in the expression of p53, MDM-2, and p21WAF1 between the HPV DNA-positive and -negative groups, as similar levels of expression were observed.99

p53

The TP53 gene, present on chromosome 17p, acts as a tumor suppressor gene, controlling entry into the S-phase of the cell cycle, and p53 plays many important roles in cell proliferation. Mutation of this gene inactivates its suppressor activity and is related to tumor progression.100 The p53 tumor suppressor gene encodes a transcriptional factor central in the regulation of cell growth, DNA repair, and apoptosis induction. Its activity requires the induction of several target genes, including MDM-2 and p21WAF1.99

The TP53 gene is lost or mutated in over 50% of human cancers. However, in contrast to many other human tumor forms, TP53 mutations are only rarely detected in cervical cancer.101,102 A great amount of research data suggests that the inactivation of p53 is believed to play a major role in the carcinogenesis of the uterine cervix. Two different mechanisms may explain the loss of p53 function in cervical cancer-a somatic gene mutation which leads to an inactive form and/or the enhanced protein degradation promoted by the E6 oncoprotein of HPV type 16 and type 18. Horner et al. found that sustained inactivation of the p53 pathway by the E6 protein is required for maintenance of the proliferative phenotype of HeLa cervical carcinoma cells.103 Troncone's results are consistent with several other reports, indicating that p53 immunostaining occurs in the vast majority of cervical cancers. This report also confirms that higher p53 protein levels are not dependent on the absence of high-risk viral infection as shown by the similar immunoreactivity scores of the HPV-positive and HPV-negative groups.99,104,105 Tsuda et al. reported that the frequency of p53 overexpression was 28.6% of 42 CINs and 28.3% of 53 invasive carcinomas.12

On the other hand, some studies indicate that immunohistochemical p53 overexpression is not associated with survival in cervical carcinoma.12,106 Other studies also show that there is no significant relationship between p53 expression and prognosis in squamous cell carcinomas of the uterine cervix. This is explained partly by the presence of HPV. The oncogenic potential of HPV appears to be mediated by the E6 and E7 proteins, which are known to bind and inactivate the p53 and pRb proteins, respectively. Furthermore, the HPV E7 oncoprotein has been shown to abrogate p21 and p27 function.103,106,107 In addition to inactivation of the p53 and pRb proteins, other cell cycle regulatory proteins may be involved in the carcinogenesis of cervical cancer.

Other factors

Cyclin A

Cyclin A appears in the nucleus precisely at the G1/S transition, accumulates throughout the rest of interphase, and disappears at the beginning of mitosis. Both cyclin E and cyclin A control the progression through the cell cycle primarily by activating CDK2. Cyclin A is important for the G1-S transition, for progression through the S phase and for the G2-M transition. Cyclin A binds to CDK2 in S phase and to CDK1 (cdc2) during the G2-M transition. This complex induces phosphorylation of pRb and the nuclear membrane protein laminin, augmenting nuclear membrane disruption. Therefore, cyclin A is essential in the G1-S and G2-M phase transitions.108 Forced expression of cyclin A provides anchorage independence to normal cells, and overexpression of cyclin A is a prognostic factor in different tumor types. The transcription of cyclin A is induced by E2F when it is untethered from pRb by the action of HPV E7 oncoprotein.75 Cyclin A is overexpressed in cervical intraepithelial neoplasia and SCC compared to normal epithelium.76,79 Van de Putte et al. found overexpression of cyclin A in SCC, consistent with previously described results.57 Zehbe et al. demonstrated an overriding of cyclin-dependent kinase inhibitors by high- and low-risk HPV: cyclin A and p21WAF1/CIP1 were coexpressed in the same cells, mainly in cases of low grade cervical intra-epithelial neoplasia (LCIN).109 This provides strong evidence that the p21WAF1/CIP1 G1 block can be overcome in HPV-infected cells in vivo. This conclusion is further strengthened by the result that cyclin A was never found during the G1 phase of cervical carcinoma-derived cell lines.

Cyclin B

During the normal cell cycle, cyclin B forms complexes solely with CDK1 (cdc2 or p34cdc2), forming the mitosis-promoting factor, which regulates the transition from G2 to M phase.3,110 The mitosis-promoting factor is later translocated from the cytoplasm into the nucleus in late S phase. Recent evidence has shown that p53 regulates a G2 checkpoint through cyclin B: p53 prevents the G2/M transition by decreasing cyclin B protein levels and attenuating the activity of its promoter.107

Cho et al. found that cyclin B/p34cdc2 is significantly increased in HPV-16 or -18 positive cervical lesions, including CIN and SCC, in comparison with HPV-16 or -18 negative lesions.107 Southern et al. demonstrated that expression of the cyclin B protein was up-regulated and persisted into the upper epithelial layers, in parallel with cyclin A expression, in HSIL infected with human papillomaviruses 16, 31, 33, 51, 58, 66, and 67.77 Kanai et al. also found elevated cyclin B levels in SIL and invasive carcinomas and concomitantly elevated levels of cdc2, indicating that cyclin B plays an important role in cervical carcinogenesis.79 Contrary to the above results, Hashiguchi, et al. examined HPV status and cyclin B1 in CIN and invasive cancer by immunohistochemistry. This study showed that expression of cyclin B1 was 0% in CIN and 12.2% in invasive cancers, and the relationship between HPV type and cyclin B1 expression was not significant.111 Further experimental data are needed to confirm the role of cyclin B in malignant transformation of the cervix.

CONCLUSION

Cervical cancer, which plagues women all over the world, is still a major focus of oncological research. Infection with HPV has been implicated as an important etiological factor in the development of uterine cervical cancer. Moreover, even in cases where HPV infection persists, cervical cancer most often does not appear until many years or more than a decade after infection. This long delay has supported the view that HPV infection may initiate, but not necessarily cause, progression to cervical carcinoma, and that oncogenic influences by other factors such as environmental factors, specific aspects of lifestyle, or genetic alterations are necessary to fully establish an HPV-induced malignancy. Because cancer is a multi-etiological disease, the time lag between HPV infection and the diagnosis of cancer also indicates that multiple steps as well as multiple factors may be necessary for the development of cervical cancer.

High-risk HPV types, specifically types 16 and 18, have been demonstrated to participate in cervical carcinogenesis through the expression of proteins, such as E6 and E7, interaction with different steroid hormones, and interaction with innate genetic susceptibilities to develop cervical cancer. The results from many studies showed aberrant alterations in the regulation of several cell cycle proteins, including cyclin D and E and p16, p21, and p27, to be characteristic features of HPV-infected and HPV E6/E7 oncogene-expressing cervical carcinomas and their precursors. These data suggest further that interactions of viral proteins with host cellular proteins, particularly cell cycle proteins, are involved in activation or repression of cell cycle progression in cervical carcinogenesis.