The integration of HR-HPV increases the expression of cyclins A and E in cytologies with and without low-grade lesions.

BACKGROUND: Cyclin-A and cyclin-E are regulators of G1-S phase of normal cell cycle. Integration of human papilloma virus high-risk (HR-HPV) could alter this mechanism, and its overexpression has been associated with poor prognosis in cervical cancer. AIM: To determine the expression of cyclin-A and cyclin-E, types of HR-HPV and physical state of DNA in cytologies with the diagnosis of low-grade squamous intraepithelial lesion (LSIL).
MATERIALS AND METHODS: 115 cytological specimens in liquid base (liquid-PREP(™)) were analyzed. 25 specimens were with no signs of SIL (NSIL) and without HPV; 30 with NSIL with low-risk HPV (LR-HPV); 30 with NSIL with HR-HPV; and 30 with both LSIL and HR-HPV. The expression of cyclins was evaluated by immunocytochemistry; and the detection of viral DNA was done by polymerase chain reaction (PCR) and restriction fragment length polymorphism (RFLPs) for genotyping or sequencing of HPV. The physical state of HPV was evaluated by in situ hybridization with amplification with tyramide.
RESULTS: In the cytologies NSIL with LR-HPV, the expression of cyclin-A and cyclin-E was found respectively in 23.3% and 33.3% of the specimens. Among the specimens of NSIL with HR-HPV, 33.3% expressed cyclin-A and 40% cyclin-E, while 100% of the LSILs expressed the 2 cyclins. On the other hand, 100% of the samples NSIL with LR-HPV presented an episomal pattern. Of the specimens of NSIL with HR-HPV, 56.6% exhibited an episomal pattern, 23.3% integrated and 20%, mixed. Among the LSILs, 90% were mixed and 10% integrated.
CONCLUSIONS: The cyclins A and E are present in the LSILs that occur predominantly in mixed state in the presence of HR-HPV.

Introduction

Invasive carcinoma of the uterine cervix involves precursory stages known as squamous intraepithelial lesions (SIL).[1] Cytologically, SILs are divided into low-grade SIL (LSIL) and high-grade SIL (HSIL). LSIL represents an earlier diagnosis in cervical carcinogenesis. 80% of these lesions are associated with high-risk human papillomavirus (HR-HPV).[2] In the state of Guerrero, Mexico, 10 different types have been encountered: 16, 18, 31, 33, 35, 39, 45, 52, 58 and 59. HPV-16 is the most frequently found in cervical carcinoma (68.1%) and in HSIL (27.4%).[3]

SIL emerges after a long period of viral persistence, as a result of viral genome integration into the host cell's genome, provoking E2 function loss and overexpression of E6 and E7, prerequisites for development of HSIL and invasive carcinoma.[4] It has been proposed that in situ hybridization (ISH) may detect the presence and physical state of HR-HPV DNA. The diffuse signal of viral DNA indicates an episomal state while the punctate signal indicates integration into the cellular genome.[56] In LSIL with HPV-16, the episomal state has been encountered in 15.4% of the cases; the integrated state in 7.7%; and the mixed state in 76.9%.[7]

Cyclins participate in various phases of the cellular cycle. Cyclin-E is synthesized in the late G1 phase and is indispensable for moving into phase S. In normal cells, the cyclin-E expression diminishes rapidly as the cell enters into phase S. In premalignant and malignant lesions of the uterine cervix with expression of HR-HPV E7, the levels of cyclin E/cdk2 have been found to be increased.[8] It has been reported that expression of HPV-16 E7 could induce transcription of the promoter of cyclin-A through the binding site to E2F. This observation could suggest that activation is implicated in cyclin-A levels and this association could be necessary for cellular transformation.[9] It has been reported that cyclin-E and cyclin-A expression is an indicator for poor outcome in cervical carcinomas associated with HR-HPV.[10]

The aim of this study was to determine the immunoexpression of cyclins A and E, the physical state of HR-HPV DNA, in cytologies with and without LSIL, and to identify possible biomarkers of early cervical lesions.

Materials and Methods

Study subjects

115 female residents of the state of Guerrero, Mexico, were participants in this study approved by the Ethics Committee of the Autonomous University of Guerrero in the period 2010-2012. Each one of the participants signed informed consent and responded to a questionnaire with the purpose of obtaining sociodemographic, clinical and obstetrical information.

Specimen collection

All women included in this study provided exo-endocervical exfoliated cell samples collected by sampling the ectocervix with an Ayre spatula and endocervix with a cytobrush, making sure that cytologic material from the transformation zone was taken. Smears were used for cytomorphological examination using conventional Papanicolaou and cytological specimens in liquid base liquid-PREP™ (LPT) and read by an experienced cytopathologist and classified according to the Bethesda System.[11]

Subsequently, the specimens were classified into four groups for the cytological study: (a) no signs of SIL (NSIL) and HPV (i.e., NSIL without HPV) (25), (b) NSIL with LR-HPV (30), (c) NSIL with HR-HPV (30) and (d) with LSIL and HR-HPV (30).

HPV detection and typification

The DNA was extracted in accordance with the standard SDS-proteinase K-phenol-chloroform method.[12] DNA amplification was done in a 2400 GeneAmp polymerase chain reaction (PCR) system (Applied Biosystems, Foster City, CA, USA). The products of PCR were analyzed by electrophoresis and were displayed on 1.5% agarose gels dyed with ethidium bromide. 1 pg and 1 ng HPV-6 recombinant plasmid DNA were used as positive controls; sterile water was used as the negative control. The products of PCR were subjected to digestion with restriction enzymes BamHI, DdeI, HaeIII, HinfI, PstI, RsaI y Sau3AI (Invitrogen, Carlsbad, CA, USA). The viral type was determined by restriction fragment length polymorphism (RFLPs).[13] When specimens were being analyzed with the GP5+/6+ system, they were subjected to sequencing in an automated system (310 ABI-PRISM Genetic Analyzer, Applied Biosystems, Foster City, CA, USA). The obtained sequences were compared with HPV types' known sequences available on the NCBI website (http://www.ncbi.nim.nih.gov).[14]

Expression of cyclins A and E by immunocytochemistry

The expression of cyclins was determined by the streptavidin biotin peroxidase immunocytochemical method, utilizing the Cytoscan HRP/DAB Cell detection system (Cell Marque Corporation, Hot Springs, AR, USA). The monoclonal antibodies used were: cyclin-A (6E6; 1:100; Novocastra, Newcastle-Upon-Tyne, UK) and cyclin-E (13A3; 1:30; Novocastra, Newcastle-Upon-Tyne, UK). The cytology slides in liquid base were subjected to antigen retrieval (Immuno DNA Retriever with citrate, Bio SB Inc., Santa Barbara, CA, USA) for 6 minutes at 120°C. The primary antibody was added for 1 hour, and then the secondary antibody coupled with biotin was added. Streptavidin peroxidase was also added. For development, the chromogen DAB was used and Mayer's hematoxylin was used as a contrast dye. The cell line HeLa (HPV-18), which overexpresses cyclins, was used as positive controls. The same line, but omitting the primary antibody, was used as negative controls.

Expression of cyclins A and E

Expression was evaluated using the following criteria: Weak, if the colored reaction product did not obscure hematoxylin-counterstained chromatin; moderate, if the colored reaction product partially obscured counterstained chromatin; and strong, if the colored reaction product completely obscured counterstained chromatin.[15] In our experience, reproducibility of interpreting weak staining was poor, and we therefore scored cases with weak and no staining together as negative. Moderate and strong cyclins staining was scored as positive and was based on the presence of staining. We also evaluated the percentage of positive nuclei (0%, 1-10%, 11-50% and > 51%).[16]

In situ hybridization

Detection of the viral genome was done with a system of tyramide signal amplification (GenPoint Dako Cytomation, Carpinteria, CA, USA). The monolayer smears were submitted to digestion for 1 minute with proteinase K (1:1000). A drop of test reagent (biotinylated viral DNA) with probes for 13 HR-HPV genotypes (16, 18, 31, 33, 39, 45, 51, 52, 56, 58, 59 and 68) and with individual probes directed at HPV-6 and -11 was applied afterwards. The slides were denatured for 10 minutes and subjected to hybridization for 20 hours (Hybridizer Dako, Carpinteria, CA, USA). They were then placed in an astringent solution; primary streptavidin peroxidase was added, afterwards biotinyl–tyramide and then secondary streptavidin. DAB was added and finally Mayer's hematoxylin (Merck). Positive reaction was visualized with a brown color inside of the nucleus and according to the type of signal was classified as diffuse (episomal state), punctate (integrated state) or mixed (diffuse and punctate).[5] SiHa cell lines (HPV-16) were used as positive controls that showed an integrated state; the same cell lines without the probe were used as negative controls.

Statistical analysis

Fisher's exact test was used for comparison of frequencies. To determine the relationship of cyclin expression and cervical lesion or physical state of DNA, multinomial logistic regression models were evaluated. The statistical analysis was done with the software STATA, version 11.1. A value of P < 0.05 was considered significant.

Results

The age of the women was between 20 and 67 years old, with a median age of 44. The majority of women with LSIL and HR-HPV were between 51 and 60 years old (80%) and have had three or more sexual partners (P < 0.001) (data not shown).

Frequency of HPV genotypes and viral DNA physical States

A total of 9 HR-HPV genotypes were identified: 16, 18, 31, 35, 39, 45, 52, 58 and 59. HPV-16 was most frequent in NSIL (60%) and LSIL (53%) cytologies (P < 0.001). 90% of the LR-HPV specimens exhibited HPV-6 and only 10% corresponded to type 11 (data not shown). In 100% of the specimen with NSIL with LR-HPV, viral DNA was found in episomal form in a scant number of cells, while for the specimens NSIL with HR-HPV, the episomal state was observed in 56.6% of the cases [Figure 1a], mixed in 20% and integrated in 23.3%. For the specimens with LSIL and HR-HPV, 90% presented a mixed pattern and 10% an integrated pattern [Figure 1b].

Figure 1

Physical state HPV-16 DNA and expression of cyclin A and E; (a) Specimen NSIL, showing normal cells that presented as an episomal pattern (black arrow); (b) Specimen LSIL showing cells with karyomegaly presented as an integrated pattern (black arrow); (c) Positive for cyclin A expression, cells intermediate with karyomegaly (black arrows); (d) Expression of cyclin E positive, cells intermediate with binucleation and karyomegaly (black arrows). (Technique: In situ hybridization with tyramide amplification, ×400 and ×1000; Streptavidin biotin peroxidase, ×400)

Expression of cyclins A and E and their relationship with cytological diagnosis

Expression of cyclins A and E increased gradually within the study groups. In 100% of the LSIL specimens, these cyclins' expression was found (P < 0.001) [Table 1]. Also in the LSILs, the cyclin's expression was moderate to strong in some intermediate cells with koilocytosis, karyomegaly and discrete perinuclear halos [Figure 1c and d], compared with the cytologies NSIL with HR-HPV where the expression of these proteins was found weak to moderate for cyclin-A (33.3%) and cyclin-E (36.7%) in some normal intermediate cells (data not shown).

Table 1

Expression of cyclin A and E and their relationship with the cytological diagnosis

Association of expression of cyclins A and E with cytological diagnosis

Expression of cyclins A and E is strongly associated with LSIL (OR: 393 and 188.3 respectively, P < 0.001) in comparison with the women with NSIL and without HPV [Table 2] in the same comparison. In the specimens with NSIL with HR-HPV, this association was lesser for the cyclins A and E, with ORs of 5.9 and 5.8, respectively, in comparison with women with NSIL and without HPV [Table 2].

Table 2

Expression of cyclin A and E with the cytological diagnosis

Analysis of HR-HPV DNA state and its relationship with expression of cyclins A and E for diagnosis

Specimens from those women who exhibited mixed (90%) and integrated (10%) patterns and LSIL were related to greater frequency with the expression of cyclins A and E in comparison with the cytologies NSIL with HR-HPV (P < 0.001) [Table 3].

Table 3

Status physical of HR-HPV DNA and its relationship with the expression of cyclin A and E for cytological diagnosis

Analysis of HR-HPV DNA state and its relationship with expression of cyclins A and E and HPV type

The mixed and integrated state and the presence of HPV-16 are related more frequently to the expression of cyclins A (72%) and E (76%) [Table 4] and HPV-6 and -11 with episomal pattern were related to expression of these proteins. Nevertheless, the expression of these proteins was found weak in some normal intermediate cells (data not shown).

Table 4

Analysis of the status physical of HPV DNA and its relationship with the expression of cyclin A and E and HPV genotypes

Discussion

The prevalence and incidence of cervical cancer has decreased worldwide. This decrease is due to the introduction and improved penetration of cervical cancer screening programs,[17] the quality and reliability of which are further increased by HPV testing. However, several studies have shown that cytology has a limited sensitivity for detecting SIL because of subjectivity and interobserver and intraobserver variabilities;[18] in our study the cytological diagnosis was verified by PCR, RFLPs or sequencing.

Our data showed viral HR-HPV integration (10%) in LSIL. Furthermore, these lesions exhibited HPV-16 with greater frequency. In a cervical scraping study of 40 LSIL cases with HPV-16, the integrated pattern was found in 5% of them suggesting that viral integration is an early event in the progression of the disease; furthermore studies have demonstrated that cell populations with integrated HPV-16 posses a selective growth advantage compared to cells that maintain HPV-16 viral genomes as episomes.[19] We found mixed status viral DNA with greater frequency in LSIL (90%). This observation could indicate a low number of integrated and episomal copies that sustain the latent pattern of the virus and during the incubation period; the virus is first assumed to be episomal, after which a selection of integrated cell clones arises from a mixture of cells with both episomal and integrated HPV. These findings are similar to a study in which we found with greater frequency (76.9%) mixed forms in LSILs associated with HPV-16.[7]

It has been found that integration depends on several factors including the presence of HPV-AR (16 and 18) and also the presence of fragile sites genomic regions that are prone to chromosomal breaks that facilitate integration of foreign DNA;[20] it is suggested to further explore the relationship of the physical state of HR-HPV DNA with quantitative methods such as real-time PCR relative to the expression of cyclins A and E; this could be a complementary tool for cytological evaluation, as it could reduce the number of unnecessary colposcopy-guided biopsies in women with LSIL cytologic diagnosis and follow-up studies could be performed to identify patients who are at risk of developing HSIL and cervical cancer.[21] Moreover, in cytologies NSIL with HR-HPV, 23.3% showed the integrated pattern. This could suggest that these women have probability of progressing to LSILs. A rapid progression from NCIN to CINIII reported showed that the integrated and episomal forms of HPV-16 were found along with a heavy viral load.[19] These results approach reports which found the integrated pattern associated with HPV-16 in 11% and 20% of the cytologies NSIL,[722] unlike that reported in cases with LSIL and HPV-16 where the researchers did not observe viral integration.[23]

In our study, women with NSIL with LR-HPV presented only the viral episomal state. This agrees with another study in which only the episomal pattern in cases NSIL with HPV-6 was observed, as well as a low viral load, suggesting that the low number of copies of this virus could maintain E2 intact, by which the cellular cycle is not altered and morphological alterations are not observed.[24]

It is known that the HPV contributes to neoplastic progression predominantly through the action of two viral oncoproteins, E6 and E7, which interact with various host regulatory proteins such as cyclins A and E to influence the function or expression levels of host gene products, eventually leading to the disruption of the cell cycle.[25] In this regard, a number of evidence exists establishing that the cyclins are associated with the expression of oncoproteins and eventually also involved in methylation. Cyclin A promoter hypermethylation was found associated with the development of cervical cancer, but not in the LSILs,[26] though this relationship was not found in respect to cyclin E. It will be important to further explore the expression of cyclins and relate these methylation.

In a monitoring study on biopsies of premalignant and malignant uterine cervix lesions with HR-HPV, in which expression of cyclins A and E was evaluated, researchers observed expression of these proteins as the lesion grade increased, suggesting that upon these proteins being overexpressed in early lesions, they may be indicators of bad prognosis.[10] We encountered a significant relationship between LSILs and the expression of the cyclins A and E [Table 2].

In the women we studied with both LSIL and HR-HPV, 100% were positive for cyclins A and E. Besides, they had expression moderate to strong of some nuclei positives that related to the alterations characteristic of infection by HPV (koilocytosis, karyomegaly and discrete perinuclear halos), while the other study groups did not exhibit this characteristic. This results concur with the findings in LSILs in which 96.7% of the expression of cyclin-E was associated with HR-HPV.[8] This could be due to the fact that coupling of E7 with E2F results in prolongation of phase S, a condition which prevents inhibition of the cyclin-E/Cdk complex[9] and if you also add the presence of integrated DNA this could be associated with the fact to have found intermediate cells (undifferentiated) that they should not be in proliferation, which were the only ones that showed expression in positive nuclei with moderate to strong expression. With regard to the expression of cyclin-A this could be due to activation of cyclin-A by HR-HPV oncoproteins E6 and E7 seem altered by degradation of p53 by E6, a condition that prevents transcription of protein inhibitors of the Cdk-cyclin complex, such as p27Kip1, being initiated, and consequently the increase in expression of cyclin A as well. On the other hand, E7 in conjunction with Cdk2/cyclin-A is important in cellular transformation.[9] The specific role of cyclins A and E as a biomarker of LSILs may be a bit controversial. Complementary studies are required to evaluate quantitatively the presence of these cyclins in LSIL.

We also observed that cytologies NSIL with HR-HPV were positive for cyclins A and E in a minor percentage in comparison with cases with LSIL and HR-HPV, a result that has not been previously reported. These findings show that despite the fact that these cells do not present any morphological alterations, the elevated expression of these proteins could allow on a molecular level prior detection of deregulation of the cell cycle induced by HR-HPV. The immunoexpression could be due to the cyclin E/Cdk2 and cyclin A/Cdk2 complex being augmented in phase S. It is probable that cells found positive were encountered in this cell cycle phase stimulated by the presence of the HR-HPV oncoprotein E7. To date, there are no reports which have evaluated these proteins in women NSIL with LR-HPV. It has been shown that LR-HPV oncoproteins possess low affinity for cellular proteins pRb and p53, LR-HPV specimens having low immunoexpression of cyclins and therefore a low number of copies in episomal form.[27]

In our study, 16% of normal cytology without HPV showed weak expression of cyclins A and E in parabasal and basal cells. Similar results found that 8% of normal epithelial cells or metaplastic nuclear staining showed mild expression of cyclin-;[8] sporadic expression of cyclin-E in parabasal cells of normal squamous epithelium has also been observed,[28] while another study reported sporadic expression of cyclin-E and cyclin-A in <1% in the parabasal cell nucleus.[29]

Our study has limitations as to the quantitative evaluation of the expression of cyclins and in determining the number of copies integrated in the LSILs episomal; however, no studies have evaluated together where expression of cyclins A and E, HR-HPV types and physical state of the DNA in early stages of cervical carcinogenesis, so this study could be important for the explanation of the pathogenesis, recognizing that they have limitations as quantification, so it is suggested to consider this relationship.

In conclusion, the cyclins A and E were present in all cases of LSIL, this expression occurred predominately as mixed state in the presence of DNA of HR-HPV; these results may help understand the pathogenesis of HPV infection, so further large-scale studies are needed to examine the clinical usefulness of cyclins A and E immunoexpression and HR-HPV ISH signal patterns, alone or in combination, as markers for identifying LSIL.