The possible role of cell cycle regulators in multistep process of HPV-associated cervical carcinoma

Background Human papillomavirus (HPV) 16 and 18 are associated with cervical carcinogenesis through an interaction between HPV oncogenic proteins and cell cycle regulatory genes. However, the exact pathogenetic mechanisms are not determined yet. Methods We investigated 43 invasive squamous cell carcinoma (ISCC), 38 CIN III, 11 CINII and 18 CINI for cyclin D1, cyclin E, CDK4, p53, mdm-2, p21waf, p27, p16INK4A, Rb and Ki-67 aberrations using immunohistochemistry and molecular techniques. Twenty samples of normal cervical tissues (NCT) were taken as a control. Results There was a significant increase in the expression of Ki-67, cyclin E, CDK4, p16INK4A, Rb (p= 0.003, 0.001, 0.001, 0.01) and a significant decrease in p27KIP1 from NCT to ISCC (p = 0.003). Increased cyclin D1, p21waf, p53, mdm-2 expression, homozygous deletion (HZD) and promoter methylation (PM) of the Rb were detected in CINIII and ISCC only. On univariate analysis; tumor size, differentiation, lymph node status, FIGO stage, Ki- 67, cyclin D1, p53 and p27KIP1 are significantly associated with reduced overall survival (OS) while on multivariate analysis; only FIGO stage, Ki-67, cyclin D1, p53 and p27KIP1 were significant. Conclusion 1) Aberrations involving p27KIP1, cyclin E, CDK4, p16INK4A are considered early events in HPV 16 and 18-associated cervical carcinoma, whereas cyclin D1 and p53 pathway abnormalities are considered late events. 2) Immunohistochemical tests for p16INK4Aand cyclin E, could help in early diagnosis of cervical carcinoma. 3) Only FIGO stage p53, cyclin D1, p27KIP1 and Ki-67 are independent prognostic factors that might help in predicting outcome of cervical cancer patients.


Background
Carcinoma of the uterine cervix emerges from a defined series of prenoplastic lesions with increasing cellular dysplasia referred to as cervical intraepithelial neoplasia (CIN) grade I, II and III [1]. A large body of knowledge supports the view that high risk HPV types (HR-HPV) are strongly associated with invasive squamous cell carcinoma (ISCC) and its precursors as they have the ability to transform normal cervical cells into neoplastic ones [2]. HPV 16 and 18 are the most commonly detected HR-HPV types in these lesions, and therefore their detection was proposed as a useful surrogate marker to diagnose cervical dysplasia and carcinoma in situ (CIS) [2,3].
Progress in cervical cancer research provided evidence that integration of HR-HPV DNA into the host cell genome results in elevated expression levels of the E6 and E7 proteins with subsequent interaction between these oncogenic proteins and the cell cycle machinery [2,4]. Expression of E6 and E7 proteins of HR-HPV types induces immortalization of cells through their inhibitory effects on the tumor suppressor proteins pRb and p53; respectively, altering the cell cycle control and leading to chromosomal instability [2]. Undermining of pRb growth-inhibitory role with release of E2F transcription factors render the cells independent of mitogenic stimuli [5], whereas inactivation of the p53 precludes its role as one of the major factors controlling cell proliferation since it arrests the cell cycle in response to DNA damage or direct the damaged cell to an apoptotic pathway. This function is abolished via p53 degradation by the E6 protein of oncogenic HPV types or via cytoplasmic sequestration by complexing to mdm-2 protein [6].
However the E6-p53 and E7-Rb model is not sufficient to inevitably produce cervical carcinoma although it has resulted in the identity of the viral gene's actions on numerous cellular proteins and processes normally involved in cellular growth and proliferation. This is evidenced by the spontaneous clearance of HPV infection and the long delay between the onset of persistent infection and the emergence of malignancy. Recent studies show an association between HR-HPV types and cell cycle regulators [2,7]. The cell cycle is governed by a family of cyclins, cyclin dependent kinases (CDKs) and their inhibitors (CDKIs) through activating and inactivating phosphorylation events. Attention has been focused on altered expression of G 1 cyclins and Cdks because the major regulatory events leading to cell proliferation and differentiation occur within the G 1 phase of the cell cycle. The D-type cyclins reach maximum levels of expression and form functional kinase complexes with CDK4 or CDK6; during the mid-G 1 phase, whereas cyclin-E is expressed and associated with Cdk2 in an active complex near the G 1 -S boundary [5,8]. Active CDK/cyclin complex can be regulated by binding to CDKI (p16 INK4A , p21 waf1 and p27 KIP1 ) and inhibit cell cycle progression from G1 to S phase [7].
Deregulation of p16 INK4A , p21 waf1 and p27 KIP1 has recently been reported in various human tumors. In HPV-associated cervical carcinoma, the expression and function of these proteins is supposed to be impaired by the action of viral oncopoteins E6 and E7 [9]. However, it is unclear how and when factors that are innate to the HPV-infected cells including genetic aberrations launch the host cell into an irreversible progression to cancer [7]. Therefore, we attempted to assess the contribution of cyclin D1, cyclin E, CDK4, p53, mdm-2, p21 waf , Rb, p27 KIP1 , p16 INK4A and Ki-67 expression to the development of HPV16/18-associated cervical carcinoma and to determine at what stage of carcinogenesis these aberrations start to manifest. The prognostic value of these aberrations was also investigated in relation to the standard clinicopathological prognostic factors and overall survival.

Clinical samples
The study included 110 fresh-frozen tissue samples that were selected from a total of 200 based on the positivity for HPV16 and/or18 and on the availability of clinical and follow-up data. The studied cases included 43 invasive squamous cell carcinoma (ISCC), 38 cervical intra-epithelial neoplasia (CIN) III, 11 CINII and 18 CINI cases were collected from patients who were diagnosed and treated at the hospitals of Kasr El-Aini School of Medicine and the National Cancer Institute (NCI), Cairo University during March 1999-July 2003 after a recent evidence of abnormal cervical cytology and/or conization biopsy (obtained by large loop excision of the transformation zone). Cases of ISCC were examined by two independent pathologists, classified and graded according to the World Health Organization (WHO) criteria and staged according to criteria of the International Federation of Gynecolology and Obstetrics (FIGO) [10]. The clinicopathological features of the studied cases are illustrated in table 1. Twenty normal cervical tissues (NCT) obtained from patients undergoing hysterectomy for medical conditions not related to the cervix were included in the study as a control. All control samples showed normal cervical cytology and histology and all were negative for HPV by PCR. In ISCC and CIN samples, only cases with ≥ 75% tumor cells were included in the study. A written consent was obtained from all patients prior to enrollment in the study, and the ethical committee of the NCI approved the protocol which was in accordance with the ethical guidelines of the 1975 Declaration of Helsinki.

Cell lines
The HPV16 (SiHa), HPV18 (HeLa) positive cervical carcinoma and the mdm2 positive (SA-1) cell lines (from Institute of Human Genetics, Academic Medical Centre, Amsterdam, The Netherlands) were maintained in Dulbecco's Modified Eagle Medium (DMEM; Sigma, USA) supplemented with 10% fetal bovine serum (Sigma, USA) and used as controls.

DNA extraction
High molecular weight DNA was extracted from fresh tumor and normal samples as well as cell lines, according to standard protocols [11].

Immunohistochemistry
Tumors and normal tissue samples were put in 10% neutral buffered formalin and processed routinely for preparation of hematoxylin and eosin-stained slides and for immunohistochemical studies. Five micron thick sections were cut onto positive-charged slides and used for immunohistochemical detection of cyclin D1, cyclin E, CDK4, p53, mdm-2, p21 waf ,p27 KIP1 , p16 INK4A , Rb, and Ki-67. The standard streptavidin-biotin-peroxidase detection technique was performed using the antibodies illustrated in table 2. Briefly, after deparaffinization in xylene and rehydration through graded alcohols, sections were microwaved in 0.01 M citrate buffer (pH 6.0) for 15 min. Endogenous peroxidase activity was blocked by immersion in 0.03% H 2 O 2 in methyl alcohol for 30 min and 10% normal rabbit (for mouse primary antibodies) or goat serum (for rabbit primary antibodies) was applied to avoid non-specific reaction. Sections were then incubated with the primary antibodies or with non-immunized mouse or rabbit serum for the negative control at 4°C. After washing with PBS, biotinylated anti-muse or rabbit IgG was applied for 30 min at room temperature. The peroxidase-conjugated-streptavidin solution was applied for 30 min and visualized using 0.05% 3'-3' diaminobenziine (DAB). Counterstaining was performed with Meyer's hematoxylin [13].
Only a distinct brown nuclear staining was scored positive. In NCT, cyclin D1; cyclin E and CDK were focaly expressed in the parabasal cells only, there was no immunostaining for p53 and mdm-2, p21 waf was focally detected, p27 KIP1 was normally detected in > 50% of the cells and Rb was detected in all NCT samples. This was the normal expression pattern for the studied protein [8]. An aberrant expression of cyclin D1, cyclin E, CDK4, p21 waf and p16 INK4A was defined as staining in excess of normal tissues (> 20% for cyclin D1; cyclin E, CDK4 and > 10% for p21 waf and p16 INK4A ) ( Figure 1). p53 and mdm-2 overexpression was considered with > 10% positive cells. Loss of p27 KIP1 was defined as staining in < 50% of the cells and altered expression of Rb was defined as absence of nuclear staining in all sections examined [8].

Determination of p53 gene status
Exons 5-9 of the p53 gene were amplified separately using p53 Gene Analysis Kit (Visible Genetics Inc., Toronto, Canada) in all studied cases. Briefly, 3 ul of genomic DNA were added to a 25-µl PCR mixture according to manufacturer's instructions and then cycled through 35 cycles consisting of 94°C for 30 s, 60°C for 30 s, and 72°C for 60 s in a thermo-cycler (Perkin Elmer, Norwalk, CT, USA). PCR products were then visualized by electrophoresis in an ethidium bromide-stained 1% agarose gel. Single strand conformation polymorphism (SSCP) was performed as previously described [14]. All PCR samples with aberrant conformers on SSCP were sequenced using nested CY5/CY5.5-labeled primers provided in the p53 Gene Analysis Kit (Visible Genetics Inc., Toronto, Canada) according to supplier's instructions and analyzed using the MicroGene Blaster Automated DNA electrophoresis unit (Visible Genetics Inc., Toronto, Canada)

Determination of cyclin D1, cyclin E and CDK4 gene status
was performed using the differential PCR assay (DPA) as previously described with GAPDH as an internal control [15].
Detection of DNA methylation in CpG islands of the Rb gene was done using a methylation specific PCR. Homozygous deletion (HZD) was assessed using a semiquantitative differential PCR as described by Nakamura et al. [16].

Determination of p16 gene mutation
The presence of p16 gene mutations was assessed using the SSCP/sequencing according to Tripathi et al. [17].
mdm-2 gene amplification was assessed using the primer sequence and PCR conditions of Oda et al [18] with phenyl alanine hydroxylase (PAH) as an internal control. The level of mdm-2 gene amplification was determined by comparing the intensities of the mdm-2 and PAH-PCR products for each of the samples with SA-1 cells as a positive control (seven folds). Samples showing more than 2 fold amplification were considered positive.

Expression of CDKIs
Nuclear immunoreactivity for p21 waf was sporadically detected in NCT, CINI and CINII. However, diffuse stain-ing for p21 waf was detected in CINIII and ISCC. There was a significant difference in the PI of p21 waf between CINII and CINIII as well as between CINIII and ISCC (p = 0.01) ( Figure 2c & Table 3).  (Table 3). All positive CIN cases were scored (++) except for a single case that was scored (+) whereas 11 ISCC cases were scored (+++), 5 were (++) and 3 were (+). There was a significant difference in the PI between CINIII and ISCC (p = 0.01) (Figure 2a).

Clinical correlations
There was a significant correlation between altered expression of Rb and p16 INK4A (p < 0.001), p53 and mdm2 overexpression (p = 0.003), cyclin D1 and CDK4 (p = 0.001) as well as between increased cyclin E and reduced p27 Kip1 expression in all studied groups (p = 0.011).
In the group of ISCC, there was significant relationship between p53,mdm2 and tumor size (p = 0.003), and between type (p = 0.01), as well as between p53 overexpression and lymphovascular invasion, and FIGO stage (p = 0.003). Similarly, there was a significant association between reduced p27 Kip1 , tumor size, tumor type (p = 0.047), positive lymph node, stromal invasion (p = 0.001&p = 0.024) and FIGO stage (p = 0.002); between p21 waf overexpression and tumor size (p = 0.001) and vaginal involvement (p = 0.042). Increased cyclin D1 was significantly associated with tumor size (p = 0.003) and FIGO stage (p = 0.008) whereas, increased cyclin E expression was significantly associated with tumor size (p = 0.001). A high Ki-67 index was significantly associated with tumor size, lymph node involvement and FIGO stage ( Table 1).

Discussion
The present study is the first to investigate the role of a large panel of cell cycle regulatory genes in Egyptian patients with cervical carcinoma both at the gene and protein levels. The studied population was homogeneous since all patients with CIN and ISCC were positive for HPV16 and/or 18. Tissue samples representing the different stages of transformation of normal cervical epithelium into ISCC have been analyzed for the possible role of the studied markers in etiopathogenesis of cervical carcinoma and their prognostic value. It has been shown that the E6 and E7 proteins of HR-HPV types disrupt cell cycle checkpoints, particularly affecting CDKI linked to the G1-and G2-checkpoints with consequent accumulation of genetic aberrations [19]. To our knowledge, the role of cyclins, CDKs and CDKI in ISCC and its precursor lesions is not well defined yet. Moreover no single study has investigated the contribution of these proteins together and in association with other cell cycle related genes such as p53, Rb and mdm-2. In the present study aberrant expression of CDKs and their inhibitors were detected at several points during the transformation of HPV-infected cervical epithelium.
Cyclin D1 amplification and protein overexpression were found in 46.5% and 41.9% of our ISCC cases as well as in18.4% and 13.15% of CINIII cases only. Our results are in agreement with Nicholas et al. [19] and Cheung et al. [15]. In contrast, Bae et al. [20] reported reduced cyclin D1 mRNA and protein expression in cases of CIN and ISCC compared to NCT. It is difficult to explain the controversy between our results and those of Bae et al. [20]. However, their study is one of the very few studies showing reduced cyclin D1 in cervical carcinoma. Moreover, they reported increased cyclin D1 expression with increasing severity of the lesion from high grade squamous intraepithelial neoplasia (HSIL) to ISCC which makes a higher expression in normal epithelium compared to the neoplastic one unlikely.
Our study shows a stepwise increase in the expression level of CDK4 from normal to tumor tissues indicating an important role for CDK4 at an early stage of transformation of HPV-infected cervical epithelium. Our results confirm the few available reports in this context [15,8].
Although it is well known that cyclinD1 is necessary for the activation of CDK, the concordance reported in the present study between CDK4 and cyclin D1 expression was 49% only. We also detected increased CDK4 expression in cases of CINI and CINII although these cases did not show simultaneous increase in cyclin D1. This could be explained by the involvement of other D-type cyclins such as cyclin D2 and 3 in the activation of CDK4 [21]. Alternatively, the D type cyclins may not be required at all for G1 progression in HPV-transformed cervical epithelium since binding of HPV E7 protein to Rb leads to release of E2F transcription factor omitting the role of cyclin D1 in cell cycle progression [19].
Our results denote an important role for cyclin E in the early stages of HPV-associated cervical carcinogenesis since protein overexpression and gene amplification were detected during progression from NCT into ISCC. We also noticed a significant association between increased cyclin E expression and reduced p27 kip1 . Our data confirm the results of Dellas et al [21] who demonstrated a feed back inhibitory loop between cyclin E and p27 kip1 . This suggests the presence of a synergistic effect between both genes which will eventually lead to enhanced progression through the cell cycle as a consequence of the proliferative effect induced by increased cyclin E expression and the loss of the inhibitory function of p27 kip1 .
Inactivation of CDKIs (p27 kip1 , p21 waf , p16 INK4A ) via reduced expression was reported in various human tumors [22]. In HPV-associated cervical carcinoma, the situation is less clear since some studies showed that the tumor suppressor activity of these proteins is overcome through the action of the viral oncogenes E6/E7 without any change in their expression level [22], others showed that this applies to p21 waf and p27 kip1 only whereas p16 INK4A is usually down regulated [22][23][24]. Moreover, the HR-HPV types were shown to impair the function but not  the expression of the p21 waf and p16 INK4A by rendering them insensitive to cyclin-CDK complex formation whereas p27 KIP1 is usually down-regulated [24].
In the present work, a stepwise decrease in p27 kip1 expression and a stepwise increase in p16 INK4A were found in cervical epithelium as it progressed from normal to a neoplastic one. Our results regarding p27 kip1 expression are in agreement with previously published data [8,[24][25][26]. In contrast, Shiozawa et al., [27] reported in their study a strong p27 kip1 expression in normal cervical epithelia which was markedly reduced to a negligible level in CIN samples. However, in ISCC cases they reported an increased expression of p27 kip1 and demonstrated that the p27 protein was bound to cdk2 and cyclin E. Consequently, the authors concluded that, p27 kip1 expression may be involved in the growth regulation of NCT however aberrant function of the p27 kip1 may occur in ISCC of the cervix. Although we can not find a proper explanation for this discrepancy in the results, we assume that the mechanism proposed by Shiozawa et al. [27] might represent an alternative pathway for p27 kip1 inactivation in ISCC of the uterine cervix through binding and sequestration by cdk2 and cyclin E which render it inactive just as the case with p53 and mdm2.
In our study, p16 INK4A overexpression was detected during the early stages of cervical carcinogenesis although no mutation was reported in any of the studied cases. We therefore assume that inactivation of p16 INK4A gene in HPV-associated ISCC is possibly achieved via mechanism(s) other than gene mutations. Among which is the binding to-and sequestration by other cellular and/or viral proteins. Our results are consistent with Volgareva et al. [22] and Tringler et al. [28]  An interesting finding in this study is the significant increase of p21 waf which was recognized at CINIII and ISCC cases only. Although it is expected that increased expression of p21 waf should suppress the progression of cells in the cell cycle and consequently suppress cell growth, all the cases that revealed p21 waf overexpression had a high Ki-67 positivity index. Several explanations could be mentioned in this regard including: 1) the interaction between HPV E7 oncoprotein and p21 waf which abrogates its inhibitory effect on cyclin/CDK4 activities, 2) the occurrence of checkpoint adaptation after constant stimulation, 3) altered or inhibited binding to the CDK4/ cyclin D1 complex, or 4) mutations in downstream targets of p21 waf [3,8,10,28].
Our study is the first to assess RB gene HZD and PM in addition to altered protein expression in cervical carcinoma and its premalignant lesions. We detected altered Rb expression in the early stages of cervical carcinogenesis whereas PM and HZD were reported in CINIII and ISCC only. Therefore, we assume that altered protein expression in the early stages of neoplastic transformation could be attributed to mechanism(s) other than gene deletion or silencing by PM. Degradation by HPV-E7, sequestration by other cellular proteins or inactivation by yet unidentified mechanism(s) could be mentioned in this context. Our results regarding altered Rb expression are consistent with the few available reports in this context [3,8,28].
Since all ISCC cases are HPV-16 or 18 positive, it is not expected to find accumulation of the p53 nor the p53induced proteins in these cases. However, elevated levels of p53 and p21 waf proteins were seen in a relatively high fraction of cervical cancer together with mdm-2 in CINIII and ISCC cases. Our data are comparable to Tsuda et al. [29] and Huang et al. [30] who found p53 overexpression in 11.4% and 40.5% of HSIL compared to 56.6% and 30.7% of ISCC cases; respectively. Previous studies on ISCC showed that p53 overexpression is not usually associated with gene mutation but with sequestration by cellular or viral proteins of which mdm-2 is a highly possible candidate. This is in agreement with our results where p53 mutations were detected in 4 cases only whereas mdm-2 over-expression was present in 51.2% of ISCC, 13.7% CIN.
In gynecologic oncology, valid prognostic factors are necessary to define biologically similar subgroups for analysis of therapeutic efficiency. In the present study cyclin D1, p27, p53 and a high Ki-67 showed a prognostic relevance as increased expression of cyclin D1, p53 and Ki-67 as well as loss of p27 were highly correlated with the standard clinicopathologic prognostic factors for ISCC and they were significantly associated with reduced OS. Our data regarding the prognostic value of cyclin D1 and p53 are in agreement with Bae et al. [20] and Huang et al. [30], respectively who demonstrated that cyclin D1 and p53 overexpression are associated with decreased disease free survival (DFS) and OS rates. Similarly, our data regarding p27 are consistent with previous reports [16,22] which considered p27 KIP1 an independent prognostic factor that significantly correlates with poor survival in ISCC.
We conclude that, infection of cervical mucosa by HPV16 and 18 leads to deregulation of the cell cycle via altering the expression level of certain genes. Whereas p27 KIP1 , cyclin E, CDK4, p16 INK4A and Rb aberrations are early events; cyclin D1, mdm-2, p53 and p21 waf aberrations occur late in cervical carcinogenesis (Figure 8). Together, these alterations lead to acceleration of the cell cycle with increased proliferation rate, as indicated by a high Ki-67 PI, and acquisition of more genetic damage. Our results also provide evidence that the application of immunohistochemical tests for p16 INK4A , cyclin E, p27 KIP1 could help in early diagnosis of cervical carcinoma since alterations affecting the expression level of these proteins occur at an early stage of cervical carcinogenesis. Therefore they could be used as surrogate markers for early detection of ISCC and for monitoring patients with cervical dysplasia. On the other hand, p53, cyclin D1, p27 KIP1 and Ki-67 are independent prognostic factors in cervical cancer that could be used for predicting the clinical outcome of those patients.

Conclusion
It seems that aberrations involving p27 KIP1 , cyclin E, CDK4, p16 INK4A are early events in HPV 16-and 18-associated cervical carcinoma, whereas cyclin D1 and p53 pathway abnormalities are late events. The use of immunohistochemical tests for p16 INK4A and cyclin E, may be of help in early diagnosis of cervical carcinoma. To predict the disease outcome of cervical cancer patients, only FIGO stage p53, cyclin D1, p27 KIP1 and Ki-67 are independent prognostic factors.
The percentage of protein expression of different studied markers in the studied groups Figure 8 The percentage of protein expression of different studied markers in the studied groups.