Abstract
The interaction between the HPV (human papilloma virus) 16 E7 and other cell growth factors, such as p53 and NFκB in laryngeal cancer is not clearly understood. The aim of this study was to examine the expression of these three proteins in tumor and non-tumor laryngeal tissues from patients with laryngeal squamous cell carcinoma. These three proteins were dominantly expressed in the nucleus and their levels were higher in the tumor tissue than in the non-tumor tissue, although the comparison between the tumor and non-tumor tissues of p53 staining did not reach significance. The intensity of the nuclear stain of E7 and p53 was stronger than that of p65, a subunit of NFκB. Correlation analysis revealed that there was a positive relationship between the level of HPV16 E7 and the expression of p65. The correlation between E7 and p53 was also significant, although to a lesser degree. The finding of nuclear localization of p65 suggests that NFκB is constantly activated in the laryngeal cancer cells, whereas the sequestration of p53 in the nucleus may represent a mutated form of p53, which is probably inactivated by HPV16 oncoproteins. In conclusion, this study suggests that the nuclear localization of NFκB and p53 may play a role in the development of human laryngeal squamous cell carcinoma infected with HPV16.
H
The E7 of HPV16 is a 98-amino acid oncoprotein localized predominantly in the nucleus. In cooperation with the ras oncogene, it is able to induce full transformation of rodent and human primary cells, which are carcinogenetic when injected into mice (Phelps et al. 1988; Mansur and Androphy 1993). In addition, E7 in combination with E6 can efficiently immortalize human primary cells (Phelps et al. 1988; Mansur and Androphy 1993). A number of growth factors or survivor molecules have been linked to the development of human tumors. NFκB is one such molecule, with the ability to inhibit cell death and to maintain and promote the growth of cells (Huxford et al. 1999; Jung and Dritschilo 2001). The tumorigenic nature of E7 to extend the lifespan of human cells via interaction with various oncogenes suggests that there might be a connection between E7 and NFκB. In addition, the involvement of Bcl-2 family members in alteration of apoptosis in human laryngeal cancer may also provide a hint of an association of NFκB with the pathogenesis of laryngeal cancer (Whisler et al. 1998), because the activation of NFκB induces the expression of anti-apoptotic Bcl-2 (Tamatani et al. 1999,2000). There is also a close association between the activity of the NFκB and the level of p53. Overexpression of p53 inhibits NFκB activity, and a high level of NFκB suppresses p53 transactivation (Webster and Perkins 1999; Shao et al. 2000). Although increasing evidence has appeared to suggest that the interaction occurs among E7, p53, and NFκB, information about this aspect in human laryngeal cancer is still lacking. The aim of the present study is to examine the expression of these three proteins in tumor tissue obtained from patients with laryngeal cancer and in corresponding non-tumor tissue from the same subjects.
Materials and Methods
Laryngeal tissue samples were collected from the operating theater as soon as the laryngectomy specimen was removed from the operative field. They were then snap-frozen and stored in liquid nitrogen for future immunohistochemical (IHC) analysis. Eighteen patients with laryngeal squamous cell carcinoma were analyzed and all patients were from the Chinese population in Hong Kong. In addition to tumor tissues, their corresponding non-tumor laryngeal tissues were obtained as controls. All tumor and non-tumor tissue specimens were confirmed by pathological examination. We obtained informed consent from all patients for subsequent use of their resected tissues. Anti-HPV16 E7, p53 and p65 (NFκB) antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA) and all were polyclonal antibodies. According to the information provided by the manufacturer, HPV16 E7 was not reactive with E7 from other types of HPV and the p53 antibody would recognize both wild-type and mutant p53 proteins.
Tissues stored in liquid nitrogen were thawed at room temperature (RT). After being thawed, they were fixed in a 10% neutral buffered formalin solution, underwent tissue processing, and were embedded in paraffin. Formalin fixation before embedding was less than 30 hr throughout, which is important for preserving the antigenic determinants analyzed in this study. Tissues were sectioned at 4-μm thickness. Immunostaining was performed on paraffin sections according to the standard procedure of an ABC kit from Vector Laboratories (Burlingame, CA). In brief, the staining methods were as follows. Tissue sections were deparaffinized and rehydrated through three changes of xylene and graded alcohol. After tissue sections were boiled in citrate-based antigen unmasking solution for 1 min and cooled in Milli-Q water, the endogenous peroxidase activity in the tissue sections was quenched with a 3% hydrogen peroxide solution for 5 min. A 1.5% normal blocking serum, supplemented with avidin solution (Avidin/Biotin blocking kit; Vector Laboratories) was used to block tissue sections for 30 min. Preparations were then incubated with a primary antibody overnight at 4C. The primary antibodies were prepared in 1.5% normal blocking serum supplemented with biotin solution from the Avidin/Biotin blocking kit and were used at a working dilution of 1:200. After tissue sections were washed with PBS, a biotinyl-labeled secondary antibody, horse anti-goat IgG, was applied for 30 min. After tissue sections were washed with PBS, ABC reagent (Avidin/Biotin kit; Vector Laboratories) conjugated with horseradish peroxidase was applied for 30 min. The staining of HPV16 E7, p53, and p65 antigen was visualized by either DAB or NovaRED substrate (Vector Laboratories). The reaction was stopped by rinsing the sections in tapwater. All incubations were done in a humidified environment at RT except where indicated in the text. Finally, sections were counter-stained with Vector Gill's hematoxylin. After dehydration through graded alcohol and clearing with xylene, sections were mounted with DPX permanent mountant. Negative controls were prepared by using non-immune serum instead of the primary antibody. The specificity of the antibodies used was verified by corresponding blocking peptides, which were from the same source (Santa Cruz Biotechnology) as the antibodies. In some experiments, the relevant blocking peptides that specifically block the primary antibody were used to show the specificity of the antibody. The intensity of the antigen staining was scored according to the standard described in Table 1.
Scoring standard for antigen staining
To detect apoptotic cells, in situ labeling of the 3′-end of the DNA fragments generated by apoptosis-associated endo-nucleases was performed using the DeadEnd apoptosis detection kit (TUNEL assay) from Promega (Madison, WI). Briefly, the sections, after being dewaxed in xylene and rehydrated in ethanol, were incubated with 20 μg/ml proteinase K at RT for 15 min. The slides were then incubated with a terminal transferase enzyme and biotinylated nucleotide mix at 37C for 60 min to allow the end-labeling reaction to occur. Endogenous peroxidase activity was blocked by treating the slides with 0.3% hydrogen peroxide in PBS, pH 7.2. Horseradish peroxidase-labeled streptavidin solution was then applied on the slides and incubated for 30 min at RT. After incubation the color was developed with the peroxidase substrate, hydrogen peroxide, and the stable chromogen diaminobenzidine. The slides were then mounted and examined with a light microscope. For control purposes, tissues treated with DNase I were used as a positive control. DNase I is an agent that can cause fragmentation of the chromosomal DNA, an early sign of apoptotic cells. Cells were defined as apoptotic if the entire nuclear area of the cell labeled positively. Apoptotic bodies were defined as small positively labeled globular bodies in the cytoplasm of the cells, which could be detected either singly or in groups. To estimate the apoptotic index that represented the percentage of apoptotic cells in a given area, apoptotic cells and bodies were counted in 10 high-power fields and this figure was divided by the number of cells in the same high-power fields.
All values were expressed as mean ± SEM. The relationships among variables were analyzed with the Wilcoxon test using the software StatView (Abacus Concept; Berkeley, CA). A p value of less than 0.05 was taken as statistically significant. Spearman correlation and linear regression were applied to determine the relationships among different markers.
Results
Expression of E7, p65, and p53
The E7 protein was detected in 59% of tumor tissues and was almost exclusively located in the nucleus (Figures 1A and 2), with only two tissues being positive in both the nucleus and the cytoplasm. The level of nuclear E7 was significantly higher in the tumor tissue than in the non-tumor tissue (Figures 1A, 1B, and 3). The p65 protein was mainly detected in the nucleus (82% of the cases) in both tumor and non-tumor tissues (Figures 1C and 2). As for the E7 protein, the expression of nuclear p65 protein was much higher in the tumor tissue than in the non-tumor tissue (Figures 1C, 1D, and 3). Eighty-two percent of cases were positive for p53 protein, which was found in the nucleus in most cases (Figures 1E and 2). Its level in the nucleus was increased in the tissue samples from the tumor area compared with that from the non-tumor region (Figures 1E, 1F, and 3). However, the difference did not reach significance, and this was probably due to the large variation of its distribution among the tissue samples tested. When we compared the staining intensity of these three proteins in the nucleus, we found that the expression of both E7 and p53 was at a similar level and that both of them were much higher than p65 (Figures 1A–1C and 3). Our co-IHC staining of the E7 protein and the p53 proteins demonstrated that both could be co-localized in the nucleus, and in some cases they appeared to overlap in a given place in the nucleus, suggesting that they might bind each other (Figure 1G). A similar result was obtained for E7 and p65 (Figure 1H).
Immunohistochemical staining of HPV16 E7, p65, and p53 in tumor and non-tumor tissues. Expression of HPV16 E7 (
HPV16 E7, p65, and p53 protein levels in nucleus and the cytoplasm. Expression of HPV16 E7, p65, and p53 proteins in the nucleus and the cytoplasm was detected by IHC analysis and the results were semiquantified according to the standards listed in Table 1. The comparison between the nucleus and the cytoplasm was made by paired Wilcoxon test, ∗∗ p<0.01.
Nuclear HPV16 E7, p65, and p53 protein levels in tumor and non-tumor tissue. Expression of nuclear HPV16 E7, p65, and p53 proteins in tumor and non-tumor tissues was detected by IHC analysis and the results were semiquantified according to the standards listed in Table 1. The comparison between the tumor and non-tumor tissues was made by paired Wilcoxon test, ∗ p<0.05.
The Relationships Among Expression of E7, p65, and p53
To evaluate whether there was any link between the viral protein E7 and cell growth factors (p65 and p53), we performed Spearman correlation analysis, a nonparametric method, to quantitate the degree of linear association between two variables. We found a positive correlation between E7 and p65 in the tumor tissue (Figure 4A). To a lesser degree, there was also a correlation between E7 and p53 (Figure 4B). However, the data did not reveal a significant correlation between p53 and p65 (Figure 4C).
Correlation between the nuclear HPV16 E7, p65, and p53 proteins. The expression of nuclear HPV16 E7, p65 and p53 proteins was detected by IHC analysis and the results were semiquantified according to the standards listed in Table 1. Spearman correlation and linear regression were used to determine the relationship among the different proteins tested. HPV16 E7 correlated positively with both p65 (
Apoptotic Index
Apoptotic cells, as detected by the DeadEnd apoptosis detection kit from Promega, were very rare in both tumor and non-tumor tissues (data not shown). There was also no significant connection between the apoptotic index and the viral protein E7 or cell growth factors (p65 and p53) detected in the present study. A significant number of cells became apoptotic in the positive control tissues that were treated with DNase I, which is known to cause fragmentation of the chromosomal DNA.
Discussion
Increasing evidence has suggested that NFκB plays a role in the development of human cancers (Gilmore et al. 1996; Chen et al. 2002). The role and expression pattern of NFκB in human laryngeal cancer is not clear, although there are limited publications addressing the association between HPV and NFκB in human cancer. Duffey et al. (1999,2000) showed that THF-α induced activation and nuclear localization of p65, a subunit of NFκB, in squamous cell carcinoma cell lines of human pharyngeal origin and that such an activation of NFκB prevented cell death. Vancurova et al. (2002) found an increase in p50/p50 NFκB activities in studies of HPV6- and HPV11-related laryngeal papilloma tissue. To our knowledge, there is no report on the relationship and interaction between HPV16 E7 and NFκB in human laryngeal cancer tissues. In our present study, we demonstrated that the expression of p65 was significantly increased in the tumor tissue and that the expression was mainly confined to the nucleus, in which the E7 protein was preferentially housed. Overlapping of both proteins in the same nuclear location was found on many occasions, which probably suggests a potential interaction between them. In the inactivated condition, NFκB is located in the cytosol as a three-subunit complex consisting of two prototypical subunits of p50 and p65 and an inhibitory subunit called IkB (Gilmore et al. 1996). When suitably stimulated, the inhibitory subunit IkB dissociates from the other two prototypical subunits and thus enables the p50–p65 dimer to translocate to the nucleus, where it binds to its regulatory elements and activates relevant genes. According to this concept, our finding of the dominant nuclear location of p65 is an indicator that NFκB has been activated. It has been shown that the E7 protein accumulates mainly in the nucleus (Greenfield et al. 1991; Fujikawa et al. 1994), which is in agreement with our present study. The E7 protein plays a critical role in extension of human primary cell life and it is even able to immortalize cells without cooperation of HPV16 E6, another potential oncoprotein of HPV (Phelps et al. 1988; Mansur and Androphy 1993; Suzuki et al. 1996). Because both E7 and NFκB promote the survival of cells and stimulate cell growth, the overlapping expression of NFκB and HPV16 E7 in the nucleus found in this study may indicate occurrence of cross-talk between them. E7 does not directly target various tumor-promoting genes but it does alter proliferation and cell growth via degrading tumor suppressor Rb, abrogating surviving protein p21, and enhancing anti-apoptotic Bcl-2 (Boyer et al. 1996; Funk et al. 1997; Pei et al. 1998). Interestingly, NFκB is known to be a cellular signal in the Rb, p21, and Bcl-2 pathways. NFκB is required to induce Rb hyperphosphorylation and thus promote progression of G1 to 5. (Guttridge et al. 1999). Cells with constitutive levels of NFκB in their nuclei can upregulate p21, probably via increasing the transcription (Bren et al. 2001; Pennington et al. 2001). A decrease in NFκB level is suggested to deter the binding of NFκB subunits to the regulatory sites in the Bcl-2 promoter region and to subsequently result in downregulation of Bcl-2 (Feuillard et al. 2000; Kurland et al. 2001). In addition, the interaction between E7 and NFκB may create a more favorable environment for the persistent HPV infection and subsequent oncogenesis (Huang and McCance 2002). All these findings support the existence of a strong link between E7 and NFκB, which is in favor of our result that there is a positive correlation between the level of E7 and the expression of NFκB in the nucleus.
The frequency of HPV16 found in our samples was 59%. In Europe and North America, the rate of HPV infection in laryngeal or pharyngeal squamous cell carcinoma varies between 22% and 83% (Smith et al. 2000), although most reports show the rate to be less than 50%. Studies on the Chinese population indicate that the frequency of infection is usually greater than 50%. Ma et al. (1998) found the rate of infection in Northeast China to be 58.8%. Although the frequency of HPV infection has not been reported in patients with laryngeal or pharyngeal squamous cell carcinoma in the Hong Kong Chinese population, Dickens et al. (1991) show that 59% of laryngeal papillomas, a premalignant lesion, are positive for HPV. Therefore, the discrepancies in the infection percentage may be due to the population tested.
Extensive studies have been conducted on the interaction between HPV16 E6 and p53 and the effect of HPV16 E7 on Rb. However, publications of the relationship between E7 and p53 are limited. The p53 protein was detected in 82% of cases, and in the majority it was located in the nucleus, where the E7 protein was also found. The p53 nuclear import or retention is essential for its normal function in growth inhibition or induction of apoptosis (Knippschild et al. 1996; Stewart and Pietenpol 2001). A defect in the regulation of p53 nuclear import or export may result in tumorigenesis. Dysfunctional transportation of p53 between the nucleus and the cytoplasm is known to occur in a subset of human tumors and, in such a situation, p53 is sequestered either in the cytoplasm or in the nucleus (Bosari et al. 1995; Lee et al. 1999). In normal cells, the wild-type p53 protein is kept at a low concentration by rapid degradation. Therefore, the wild-type p53 protein is usually undetected or detected at a very low level. However, when there are mutations in the p53 gene, they result in a dysfunctional protein product with a prolonged half-life that enables them to accumulate in the cell. Our result also suggests that the behavior of p53 protein in HPV-containing laryngeal carcinoma may differ from that in HPV-positive cervical cancer, because the rate of p53 detection in the latter is often much lower. The reason for this may be due to different levels of p53 mutation in these two types of cancers. Accumulation of p53 protein either in the cytoplasm or in the nucleus is considered a pathological index (Stewart and Pietenpol 2001). Therefore, it is believed that the p53 protein detected in the nucleus of laryngeal cancer infected with HPV16 E7 is a mutated form, losing its function as a tumor suppressor. This assumption is supported by an in vitro experiment showing that spontaneous mutation of the p53 gene occurred in HPV16 E7-immortalized human laryngeal epithelial cells and, furthermore, that this mutated p53 was located in the nucleus (Fujikawa et al. 1994). A study by Massimi and Banks (1997) also demonstrated that HPV16 E7 was able to bind to p53 and inhibit its transcriptional activity. However, there is a conflicting result showing that the E7 stabilizes p53 (Seavey et al. 1999). The percentage of the nuclear staining reaction for p53 in the present study (82%) is similar to the frequency of 77% reported by Jackel et al. (2000) in 88 patients with laryngeal squamous cell carcinoma.
Both p53 and NFκB are well known for their roles in the control of cell proliferation and growth. Colocalization of p53 and HVP16 E7 in the nucleus may reflect an interaction between them, which may lead to the mutation of p53 and the sequestration of p53 in an area closed to the E7 protein. The nuclear localization of NFκB (p65) is a clear sign indicating that NFκB is constantly activated and such an activated state of NFκB may serve as a signal to relay the effect of the E7 oncogenic protein to other downstream molecules such as Rb, p21, and Bcl-2, whose levels are known to be abnormal in human laryngeal cancer (Whisler et al. 1998; Hirvikoski et al. 1999; Mizokami et al. 1999; Jackel et al. 2000).