High-Risk Human Papillomavirus DNA Detected in Primary Squamous Cell Carcinoma of Urinary Bladder

A retrospective search of the files of University of Miami/Jackson Memorial Hospital (Miami, Florida) was conducted for cases of primary SCC of urinary bladder. Cases were reviewed and the diagnosis confirmed by 2 independent pathologists, including 1 uropathologist (M.J.). The diagnosis of SCC was rendered when keratinization was noted within the infiltrating carcinoma and pure areas of urothelial carcinoma were not identified (Figure 1, A). Basaloid morphology, koilocytic changes, and in situ urothelial carcinoma/dysplasia were not identified in these cases. With these selection criteria, 38 cases of primary bladder SCC were identified. These cases were studied in our previous work, in which 14 (37%) were found to be diffusely positive for p16 (Dako M7247, 1:50; clone 484, Dako, Carpinteria, California) by immunohistochemical testing³  (Figure 1, B). Specimen types included transurethral resections and cystectomy specimens and were derived from both males (8 cases) and females (6 cases). Tumors were clinically deemed to be primary from bladder and did not involve other sites. Analysis of the presence of HR-HPV was performed in all cases by using both in situ hybridization as well as a third-generation signal amplification Invader assay.

Conventional in situ hybridization studies were performed on 4-μm-thick sections, which were cut from formalin-fixed, paraffin-embedded tissue blocks. The in situ hybridization procedure was carried out according to the manufacturer's instructions (Dako HPV types 16/18 biotinylated DNA probe Y1412 with the Dako GenPoint Catalyzed Signal Amplification Kit K0620 for detection). In short, tissue sections were deparaffinized by using xylene baths, dehydrated with 100% ethanol, followed by tissue digestion using 0.8% pepsin immersion. Subsequently, the tissue was treated with target retrieval solution for 20 minutes at 95°C, cooled, washed, and dried. Probe was then applied and the tissue was denatured at 92°C for 5 minutes. Hybridization was carried out at 37°C for 18 hours. Detection of hybridized probe was then performed according to the kit instructions.

Detection of HR-HPV DNA by signal amplification, using Invader technology, was also performed in all cases. Formalin-fixed, paraffin-embedded tissue samples were placed consecutively in CitriSolv (2 × 1 mL; Fisher Scientific, Pittsburgh, Pennsylvania) and then in 100% ethanol (2 × 1 mL) for deparaffinization. Tissue was dried briefly in a 56°C bead bath. Tissue digestion and DNA extraction were performed according to manufacturer's specifications by using the MagNA Pure LC and the MagNA Pure LC DNA Isolation Kit II (Roche, Indianapolis, Indiana). Screening for HR-HPV and subsequent identification of HPV types 16 and 18 were accomplished with Cervista HPV HR and Hologic Cervista HPV 16/18 reagents, respectively (Hologic, Bedford, Massachusetts). High-risk HPV–positive samples were amplified with type-specific primers.⁴  Products of amplification, using HPV type-specific primers, were DNA sequenced with Big Dye version 1.1 (Life Technologies, Carlsbad, California) on an Applied Biosystems 3100 instrument (Applied Biosystems, Foster City, California).

All cases were morphologically similar and, immunohistochemically, stained diffusely and strongly positive for p16 (Figure 1, A and B).³  Detection of HR-HPV by the in situ hybridization method was negative in all cases (0 of 14) (Figure 1, C and D). The presence of HR-HPV DNA was detected, however, in 3 of 14 cases (21.4%) in which the Cervista HPV HR screening test was used (Table 1). Subsequently, 2 of these positive cases demonstrated an HPV 16 genotype with the Hologic Cervista HPV 16/18 reagents, while the genotype of the third positive case could not be determined with Hologic Cervista HPV 16/18 reagents (Table 2). The Cervista HPV HR screening test was repeated in the third case and the result was again found to be positive. To identify the HPV genotype of the third HR-HPV–positive sample and to confirm the Invader results in all positive samples, HPV DNA was amplified and sequenced by using type-specific primers^5,6  from all 3 positive samples. The HPV type-specific primers used were targeted to HPV types phylogenetically similar to HPV 16 (31, 33, 35, 52, and 58) since the Cervista HPV HR screen implicated 1 (or more) of these types.^6,7  The 2 samples demonstrated to be HPV 16 positive by Cervista HPV 16/18 reagents were found to be HPV 16 positive by sequencing. The type-specific region amplified by PCR and sequenced is shown in Figure 2. The DNA sequence of the 2 samples was identical and is aligned to the other phylogenetically similar HPV types in Figure 2, a. The third HPV HR–positive case was found to be HPV 35 by DNA sequencing, confirming the Cervista HPV HR screening test result (Figure 2, b) and the negative result from the Hologic Cervista HPV 16/18 reagents. Thus, the presence of HR-HPV types in primary SCC was demonstrated by 2 independent methods: Invader signal amplification and HPV type-specific PCR and DNA sequencing. The 3 HPV DNA–positive cases were all from female patients with ages of 52, 64, and 70 years.

Risk factors for the development of bladder cancer can be classified into 3 types: genetic and molecular abnormalities, chemical or environmental exposures, and chronic irritation. Genetic factors that have been shown to be involved in the genesis of bladder cancer include the upregulation of oncogenes, including those encoding p63, p21, and Ras proteins as well as epidermal growth factor receptors, and the down-regulation of tumor suppressor genes including p53 and retinoblastoma gene.^8,9  Other molecular events known to be involved in bladder tumorigenesis are aberrations in cell cycle regulatory proteins, such as Ki-67 and cyclin D1.¹⁰  Polymorphisms of the cyclin D1 gene, for example, have been shown to be associated with increased risk of development of bladder cancer.¹⁰  Other causal factors include chronic irritation, indwelling catheters, Schistosoma haematobium infestation, and pelvic radiation.

Although the carcinogenic role of HR-HPV infection has been clearly and repeatedly documented in epithelial carcinomas, including those of uterine cervix, anogenital tract, and carcinomas of the head and neck, its role in the development of bladder cancer has been controversial. Based on its well-documented carcinogenic potential in the development of carcinomas of other mucosa-lined surfaces, it is feasible that HR-HPV may play a role in the development of at least a subset of bladder cancers.

In gynecologic malignancies, the presence of HR-HPV is often associated with overexpression of the p16 protein. The molecular mechanism by which this occurs is as follows: different cell cycle phases are regulated by cyclin-dependent kinases (CDKs), which are themselves activated by the binding of cyclins. In the G₁ phase of the normal cell cycle, cyclins bind to CDK4 or 6, and the resulting cyclin D–CDK complexes act on the retinoblastoma protein (RB), which controls cell cycle progression through the G₁-S checkpoint.¹¹  When RB is in its hypophosphorylated form, it binds to the E2F transcription factor, an interaction that blocks the normal activation of genes required for entry into S phase and thus prevents cell cycle progression.¹²  Phosphorylation of RB protein blocks the binding to E2F and thus enables the cell to progress into late G₁ and S phase. p16^INK4a, a protein product of the cyclin-dependent kinase inhibitor 2 gene, binds to CDK4/6 complexes and, in doing so, inhibits phosphorylation of RB. Under these circumstances, RB remains hypophosphorylated and bound to E2F and, thus, cell cycle progression is halted. p53 protein, another major tumor suppressor protein, prevents cell growth and encourages apoptosis in the presence of DNA damage. p53 also upregulates the p21 protein, which subsequently blocks the formation of the cyclinD/CDK complex and, therefore, prevents the phosphorylation of RB. Because RB is in its hypophosphorylated form, it can bind E2F, and cell cycle progression is halted. In the presence of HR-HPV, virally encoded oncoproteins E6 and E7 can form specific complexes with tumor suppressor gene products. Specifically, the viral protein E7 binds to RB in its hypophosphorylated form and inactivates this protein. The HPV-encoded E6 protein can likewise associate with the p53 tumor suppressor protein, an interaction that results in degradation of this protein and, thus, loss of its tumor suppressor activity. Thus, in the presence of HR-HPV, the virally encoded oncoproteins E6 and E7 cause the degradation and functional inactivation of p53 and RB, respectively, and this inactivation is thought to be important in the carcinogenesis of HPV-related tumors.¹³ 

When virally encoded E7 protein binds native RB, the E2F complex is released and the transcription of RB-regulated genes is activated, leading to progression of the cell cycle through the G1/S phase. Thus, in the presence of the viral protein E7, the inactivation of RB, due to the phosphorylation induced by the cyclinD/CDK complex, is mimicked. When this happens, E2F accumulates, leading to a compensatory upregulation of p16. This increased expression of p16 can be detected at the protein level by using immunohistochemistry in HPV-related tumor samples and is in fact used as a surrogate marker for the presence of HPV-related tumors in the female genital tract.

In a recent study,³  we reported that more than one-third (37%) of primary SCCs of urinary bladder demonstrate diffuse p16 immunoreactivity independent of sex or tumor differentiation. From this study, we concluded that in tumors involving the pelvic cavity, the presence of p16 immunoreactivity alone cannot be used to distinguish between tumors of gynecologic or urothelial origin. Our observed p16 overexpression in many of the primary bladder SCCs further made us question whether this p16 overexpression is due to HPV-dependent mechanisms.

The potential role that HPV may play in the development of bladder carcinomas has been previously studied in a few small series, but results of these studies^14–21  have been conflicting and no large-scale series or series based in the United States has been conducted to our knowledge. Additionally, most previous studies include only a small subset, if any, of pure SCC of urinary bladder. In studies based outside of the United States, the incidence of identification of HPV DNA in primary urothelial tumor tissue has ranged from 34%¹⁵  (in a single study) to 0%^16–21  of cases, and conclusions have varied from suggesting an important causal role of this virus in the development of urothelial carcinoma to dismissing the idea because of the lack of identifiable HPV-positive cases. Most of these previous reports have not identified the presence of HPV in association with primary urothelial carcinoma^14–17,21  and most remaining reports identify HPV in only a very low percentage of cases.^18–20  In previous studies that did identify HPV DNA within tumor samples, most identified only rare or individual cases and therefore suggested a minor role of this virus in the oncogenesis of urothelial carcinoma. We hypothesized, from our previous observation in which a subset of primary bladder SCCs overexpressed p16 by immunohistochemistry, that HR-HPV DNA may be present in this cancer subtype. Therefore, we set out to determine whether HR-HPV DNA could be detected within cells of primary bladder SCC that were p16 positive (by immunohistochemistry) by using in situ hybridization and a third-generation signal amplification Invader assay, followed by PCR amplification and sequencing for confirmation. Using the Invader technology, followed by specific amplification and DNA sequencing, we found that 21% of our cases contained HR-HPV DNA within tumor tissue.

When using conventional in situ hybridization for the detection of HR-HPV DNA, none of our cases were positive. Conventional in situ hybridization for the detection of HR-HPV from clinical samples has been shown previously to have a poor sensitivity. Sensitivity rates of in situ hybridization compared to target amplification methods were reported to be 50% for the detection of HR-HPV types 16 and 18 and only 11% for the detection of HPV type 33.²²  Therefore, the lack of detection of HPV by in situ hybridization in our 3 positive cases can easily be explained by the known limitations of this assay.

The Cervista HPV HR test (Hologic) has been approved by the US Food and Drug Administration to be used for human cervical samples to assess the presence or absence of high-risk HPV types. The Invader technology used in this study (Cervista HPV HR and Cervista HPV 16/18) is a signal amplification assay based on 2 simultaneous isothermal reactions.²³  In the first reaction, a specific probe and an Invader oligonucleotide anneal to the DNA target sequence to generate a 1-bp overlapping structure with a 5′ flap, if the desired sequence is present. Then, Cleavase enzyme (Hologic, Bedford, Massachusetts) can specifically cleave the overlapping primary probe, which releases the 5′ flap plus 1 nucleotide. These specifically cleaved 5′ flaps can then combine with a second probe, namely, a fluorescent resonance energy transfer (FRET) probe, in a secondary overlapping cleavage reaction generating a fluorescent signal. This assay, in which a signal is amplified but nucleic acid is not amplified, has been shown to have a high specificity and lower false-positive rate relative to other molecular HPV detection methods. In fact, comparisons of this third-generation Invader assay with the HPV Hybrid Capture 2 Assay (Digene Corporation, Gaithersburg, Maryland) for the detection of HR-HPV have been done in liquid-based cytology specimens. Ginocchio et al²³  reported that the sensitivity, specificity, positive predictive value, and negative predictive value for the Invader assay were 95.9%, 97.6%, 97.5%, and 96.1%, respectively. Overall, the sensitivity and negative predictive value were similar to those of the HPV Hybrid Capture Assay, but the Invader assay demonstrated superior specificity and positive predictive value, owing to fewer false-positive results. Additionally, this comparison study pointed out that the rate of false-negative results with the Invader assay may be reduced because of the built-in internal control for each sample. With this internal control, false-negative results due to insufficient sample quantity are identified. The analytic performance of the Cervista HPV HR test has also been firmly established for cervical cancer screening. In these reports, Invader results correlated well with those of PCR and sequencing.^24–26  More importantly, however, the Cervista HPV HR test has been shown to be clinically both sufficiently sensitive and specific.²⁷ 

The Hologic Cervista signal amplification technology has been approved for detection of DNA from HR-HPV types in human cervical samples. In this study, we have applied the methodology for HPV detection in paraffin-embedded tissue to demonstrate that 21.4% of primary bladder SCCs contained HR-HPV DNA. The genotype of 2 of these cases was found to be that of HPV 16 with the Hologic Cervista HPV 16/18 reagents and was confirmed by HPV type-specific PCR and DNA sequencing. Additionally, the third HR-HPV DNA–positive case was found to harbor HPV 35, again by HPV type-specific PCR and DNA sequencing. These findings indicate that in fact there is a subset of primary bladder SCCs that contain HR-HPV DNA. Based on the well-documented carcinogenic role of this virus in other mucosal-lined sites, it is possible that HR-HPV may play a role in the development of at least a subset of primary bladder SCCs. Because the in situ hybridization studies failed to identify the presence of HPV in any of the 3 cases that were positive, using the signal amplification method, and because of the previously documented low sensitivity of this assay, we would not recommend using in situ hybridization as the sole method for detection of HPV in this setting.

All cases used in this study had been previously shown to be positive for p16 by immunohistochemistry.³  The association between the presence of HPV oncoprotein E7 and the subsequent overexpression of the p16 protein by infected cells has led to the widespread use of immunohistochemical positivity for p16 as an indicator of HPV infection, particularly types 16 and 18. However, p16 overexpression has also been identified in other neoplasms that have not been shown to be HPV related, including dermatofibrosarcoma protuberans, gastric adenocarcinomas, serous endometrial carcinomas, and Hodgkin and non-Hodgkin lymphomas. p16 is also expressed in many normal human tissues including thymus, proliferative endometrium, breast ducts, gastric antral cells, esophageal squamous epithelium, salivary glands, and some neuroendocrine cells.²⁸  Therefore, p16 overexpression may be seen outside of the presence of HPV. The fact that all cases used in this study were known to overexpress p16 by immunohistochemistry, and yet only 21.4% could be demonstrated to contain HR-HPV DNA, suggests that there are other, non-HPV–related mechanisms that contribute to p16 overexpression in bladder SCC. This also suggests that the use of p16 immunohistochemical staining cannot be used as a surrogate marker for the presence of HPV in this setting. Whether positivity for p16 can be used as a screening method for the presence of HPV was not addressed in this study, as all cases were positive for p16 initially.