Investigation of the Possible Cytopathological Effect of Human Papillomavirus Infection on p-16INK4a Overexpressed Urothelial Carcinomas of the Bladder in the Urine

Seiko Ogura1,2, Toshitetsu Hayashi3,4*, Keiko Yano1, Masami Sakurai5, Takaki Sakurai6 and Reiji Haba3 1Department of Diagnostic Pathology, Saiseikai Noe Hospital, Osaka, Japan 2Graduate School of Medicine, Kagawa University, Kagawa, Japan 3Department of Diagnostic Pathology, Faculty of Medicine, Kagawa University, Kagawa, Japan 4Department of Diagnostic Pathology, Takamatsu Red Cross Hospital, Japan 5Osaka City University, Japan 6Department of Diagnostic Pathology, Kyoto University Hospital, Japan


Introduction
Human papillomavirus (HPV) is well-known as the major etiological agent for uterine cervical cancer. Its possible relationship with cancers of the oropharynx, anus and vulva has also been investigated [1][2][3][4][5][6][7]. Over the past few decades, a considerable number of studies have been conducted to determine the role of HPV in urinary bladder tumors [8][9][10][11][12][13][14][15][16][17]. However, the most of their studies have focused on the detection of HPV infection, and little attention has been paid to the cytologic morphology correlated with HPV infection in urinary bladder tumor. On the other hand, detection of HPV-DNA and overexpression of p16 INK4a has been reported at a high rate in cervical squamous cell carcinoma, but there are only occasional reports of HPV-DNA and p16 INK4a expression in urothelial carcinoma of the bladder. The present study has been carried out to search for cytopathologic differences between HPV positive and negative cases with urothelial carcinoma and the possible role of p16 expression in the carcinogenesis of urothelial carcinoma or association with the HPV infection.

Cases
We examined 63 patients who underwent transurethral resection for papillary urothelial carcinomas, based on the 2004 WHO classification from May 2010 to September 2012 in the Department of Urology, Saiseikai Noe Hospital, Osaka, Japan. The present study included 91 tissue samples and 76 urine specimens (20 voided urines and 56 bladder washings), and was approved by the institutional "Ethical Review Board" of Saiseikai Noe Hospital. 2, 21%-70%; score 3, >71% (Figures 1-3). The samples displaying score 3 immunoreactivity were considered to have overexpression of p16 INK4a following the Nakazawa et al. criteria [18].

HPV DNA in situ hybridization
ISH was carried out on sections showing p16 INK4a overexpression on formalin-fixed, paraffin embedded tissue sections cut at a thickness of 4 µm. Following deparaffinization, the tissue was digested in protease solution for 7.5 minutes at 37°C. Slides were dehydrated in ascending order of ethanol concentration (70%, 95%, and 100%) for 1 minute each and air-dried. The probe used was Wide Spectrum HPV Biotinylated DNA probe, which detects HPVs 6,11,16,18,31,33,35,45,51, and 52 (Y1404, Dako North America Inc., Carpinteria, CA). It was applied to slides and coverslips were immediately placed over the probe solution and sealed with rubber cement. The slides were placed in a dry oven and denatured at 95°C for 5 minutes. Hybridization was carried out at 37°C overnight, and afterwards, the rubber cement and coverslip were removed and the slides immersed in the warmed stringent wash solution to incubate at 48°C for 30 minutes. Primary streptavidin-AP reagent was applied to the slides at room temperature for 20 minutes. BCIP/NBT substrate solution applied to the slides at room temperature for 2 hours. Tissues have been counterstained by Nuclear Fast Red solution. Positive ISH signal patterns were identified and classified by Cooper et al. [18,19] as follows: (1) punctate, when distinct dot-like intra-nuclear signals were stained (indicative of integrated HPV); (2) diffuse, when nuclei were completely stained (indicative of episomal HPV).

Comparison of cytopathological finding
Urine cytologic specimens consisted of 20 voided urines and 56 bladder washings. They were prepared by conventional methods, such as Cytospin (Thermo Shandon, Pittsburgh, PA) or smear preparations following centrifugation, and fixed in 95% ethanol or airdried for Papanicolaou staining or May-Grünwald-Giemsa staining. We compared the cytomorphological difference in urinary cytology between HPV-positive and HPV-negative cases. The cytopathological features of tumor cells were reviewed, including their background, arrangement, and cellular pleomorphism (variation of cell diameter occurring more than twice), and nuclear and cytoplasmic features. Furthermore, the presence of cytomorphological parameters associated with HPV infection considered previously by Bollmann et al. [20], such like koilocytosis, dyskeratocytes, abortive koilocytes, mild dyskeratosis, parakeratosis, mild nuclear hyperchromasia, mild nuclear variations, binucleation or multinucleation, measles cells, keratohyalin and keratohyalin-like granules, macrocytes, and cytoplasmic folding were reviewed.

Statistical analysis
Data were entered into JMP8.0 (SAS Institute, Inc. North Carolina, USA) software and analyzed. Significant differences between groups were found using the Chi-square test and Fisher's exact test. P-values<0.05 were considered significant.

Patient characteristics
Clinicopathological characteristics of our samples are shown in

Comparison of cytologic findings of HPV positive UC and HPV negative UC
Cytopathologic findings could be evaluated in 29 samples (8 voided urine, 21 bladder washings) consisting of 11 HPV-positive cases and 18 HPV-negative cases. The cytomorphological findings are summarized in Table 3. No significant differences were found in the cytology of tumor cells when HPV-positive case and HPV-negative cases were compared ( Figure 5 and Table 3 with corresponding p-values). The cytomorphological parameters associated with HPV infection were observed in 2 (18.2%) cases of HPV-positive cases, and in three (16.7%) HPV-negative cases of benign squamous cells. The cytopathologic parameters observed were bi-or multinucleation, abortive koilocytes, and keratohyalin-like granules ( Figure 6). No significant differences were present in cytomorphological findings among HPV-positive and HPV-negative cases.

Discussion
In this study, detection of HPV-DNA by ISH was performed only on cases which had p16 INK4a protein as a surrogate marker of HPV infection in cervical tissue with overexpression. p16 INK4a binds to cyclindependent kinase4 and inhibits its activity [21,22]. In the cervix, HPV infection inactivates RB protein by HPV E7 protein, and then the free transcription factor E2F increases. The cell cycle proceeds from G1 phase to S phase and cell proliferation is enhanced. At this time, since the E2F protein induces the expression of p16 INK4a , it is overexpressed. Detection of HPV-DNA and overexpression of p16 INK4a has been reported at a high rate in cervical squamous cell carcinoma [23][24][25], but reports of HPV-DNA and p16 INK4a expression of the bladder have been few. Steinestel et al. [26] found the expression of p16 INK4a in 25 specimens of 27 (92.6%), but HPV-DNA was not detected. Moreover, Piaton et al. [27,28] showed that in three patients with p16 INK4a immunoreactive tumor cells and high risk-HPV in the urine, HPV genotyping and in situ hybridization for high risk-HPV were negative in tissue sections. However, Shigehara et al. [16] examined 106 cases of urothelial carcinoma, 4 cases of squamous cell carcinoma, 6 cases of adenocarcinoma, and one case of another bladder cancer and found HPV-DNA in 18 out of 117samples (15%). Among them, they reported score3 (>50% of the cells were positive) staining of p16 INK4a in10 samples [16]. In this study, we admitted overexpression in 29 of 91 papillary urothelial carcinoma samples. We conducted ISH, and HPV-DNA was detected in 4 of Score 0, no staining; Score1, <20%; Score2, 21%-70%; Score3, >71%  Differences in cellular findings caused by the presence or absence of HPV infection was not observed in papillary urothelial carcinoma in this study. Having an HPV infection suggests the presence of atypical squamous cells, koilocytes, dyskeratotic cells and multinucleated cells in cervical smears [20]. These cells appear mainly at the stages of CIN1 and CIN2 but are rare in the stage of invasive cancer, except when associated with CIN1 and CIN2 lesions. HPV infection is not only involved in squamous cell carcinoma, but also adenocarcinoma in the uterine cervix [30,31]. However, the cellular changes of HPV infection in glandular cells are not well established. In urothelial cells also, cellular changes due to HPV infection is not well understood. HPV-DNA detected by ISH method in this study showed the all integrate pattern; no episomal pattern was observed. Because of these results, finding 14 LGUC cases (28.6%) and 7 of 15 HGUC cases (46.7%). Previous reports showed either a high rate in Grade 3 [10,17] or a higher rate in Grade 1 [16]. In this study, HGUC showed a higher detection rate of HPV-DNA. This controversy is probably related to our selected cases having IHC overexpression of p16 INK4a . Criteria of overexpression of p16 INK4a in tissue specimens varied in the literature [3,16,25,27,29]. In our study, overexpressed positive findings meant that there was diffused distribution in more than 70% of the nucleus and cytoplasm of tumor cells. In Shigehara's report [16], among 18 HPV-DNA positive cases there were 6 cases of p16 INK4a from 20% to 50%. They were 3 cases of urothelial carcinoma, Grade 1>2, and 3 cases of urothelial carcinoma, Grade 1. It is therefore expected when the threshold of overexpression of p16 INK4a is lowered, HPV-DNA detection by LGUC will rise.    the cytologic differences between HPV-DNA positive and negative UC cases on urine was difficult. Also, in the urine cytology specimens, because HPV-DNA positive cases can focus on non-tumor cells unlike HPV-DNA negative cases, significant differences in the incidence of abortive koilocytes or bi-nucleation were observed. Atypical squamous cells that appeared in urine cytology specimens of HPV-DNA positive cases in this study were of the "Non-classic type" rather than "Classic type" which includes koilocytes and dyskeratotic cells. There are many established reports of the appearance of atypical squamous cells of nonclassic type in uterine cervical smears, and there are types of observed parameters or combination of parameters indicative of HPV infection [20,31]. However, atypical squamous cells that appeared in HPV-DNA positive cases were observed in both cases of the voided urine and it was not possible to limit them from the urinary bladder epithelium.
Cancer generated from uterine cervix, oral cavity, tonsil, pharynx, anus, vulva, vagina and penis are carcinogenic by HPV infection as observed by IARC. Although attempts have been made to study urinary bladder cancer associated with HPV infection, carcinogenic HPV in bladder cancer is still controversial. Our results support the theory that HPV infection in urinary tract does not add clinically relevant importance to the carcinogenesis and cytomorphology of urothelial carcinomas. Further investigations are needed to clarify the role and correlation between bladder cancer and HPV infection.