• human papillomavirus;
  • integration;
  • tonsil cancer;
  • p16;
  • epidermal growth factor receptor


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. References

The prevalence of human papillomavirus (HPV) infection is high in the oropharyngeal mucosal regions, of which the tonsil is the most commonly affected. There may be a link between HPV and the pathogenesis of tonsillar cancer (TC), because of common anatomical characteristics between cervical and tonsillar cancer. We aimed to clarify whether HPV directly affects the oncogenesis and biologic behavior of TC by making a comparison between infection prevalence, physical status and viral loading numbers, and clinicopathologic prognostic factors. To compare HPV-related molecules between TC and tonsillitis (CFT), p16, survivin, HIF-1α, skp-1, cyclin A, cyclin B1, c-myc and EGFR were investigated. We observed a significant difference in HPV prevalence between 52 TCs and 69 CFTs (73.1% vs. 11.6%), and most of the HPVs were type 16 (87.2%) and nonepisomal (94.1%). Most TCs associated with HPV arose from the tonsillar crypts, and tended to be inverted and poorly differentiated. Compared with HPV-negative TC, HPV-positive TC showed a strong association with p16 overexpression (p < 0.0001), and an inverse association with EGFR amplification (p = 0.0478). HPV-16 integration status was strongly associated with c-myc amplification (p = 0.034) and HIF-1α overexpression (p = 0.022). HPV-16 integration could be directly related to tonsillar carcinogenesis initially in tonsillar crypts, followed by cell cycle aberration such as p16 overexpression related to the G1-S phase. © 2006 Wiley-Liss, Inc.

Human papillomavirus (HPV) may be postulated to play a role in the pathogenesis of palatine tonsillar cancer (TC), not only due to its morphological similarities to cervical cancer but also because the mucosal squamous epithelium, similar to that of the uterine cervix, is easily exposed to viral infection. HPV infection is prevalent in the oropharyngeal mucosal regions, and the tonsil is the most commonly affected anatomical region in the oropharynx.1, 2 Furthermore, TC frequently has a verrucous or papillomatous appearance, with tumor cells showing features of HPV infection, namely koilocytotic atypia. However, it remains uncertain whether HPV is directly associated with the malignant transformation of oropharngeal tumors.

The molecular biology of viral oncogenesis has been well established. The viral protein E6 promotes the degradation of p53 and E7 inactivates pRb, usually followed by viral integration into the host genome.3, 4 For viral integration, disruption occurs most frequently in an E2 open reading frame (ORF), the E1 ORF being disrupted in a minor proportion of patients. Both are important in viral replication and transcription. More specifically, the breakage of E2 allows for the dysregulation of the E6/E7 oncoprotein, which eventually leads to malignant transformation.5 In the non-integrated episomal state, the transcriptional level of E6/E7 is regulated by a promoter in the long control region in high-risk HPV-infected cells. It is also influenced by viral and cellular transcription factors, including the binding sites for the transcription factor Yin Yang 1.6 An investigation of the physical status of HPV would be helpful in finding a link between the virus and tonsillar carcinogenesis. Squamous cell carcinoma (SCC) of the palatine tonsils represents ∼15–23% of all oral cavity SCC in the USA.2 The most frequently reported risk factors for oropharyngeal cancer are smoking and alcohol, however, transplant recipients or the spouses of cervical cancer patients have also been reported to be high risk group.1, 7, 8

In a recent overview of HPV and TC, 51% of the patients contained HPV DNA, and of these cases HPV-16 was the most frequently isolated type.1 It has also been shown that patients with HPV-positive head and neck cancer demonstrate better survival rates than those with HPV-negative cancers.9

We aimed to clarify whether HPV directly affects the oncogenesis and biologic behavior of TC by comparing the infection prevalence, physical status of the virus and clinicopathologic prognostic factors. To compare the effects of HPV integration in TC and chronic follicular tonsillitis (CFT), p16 overexpression and c-myc gene amplification were evaluated using immunohistochemistry and fluorescent in situ hybridization (FISH), respectively.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. References

Selection of tissue samples and DNA extraction

Samples from 52 patients were collected from the archived, paraffin-embedded tonsillar SCC registry of Yonsei University Medical College Department of Pathology and Head and Neck Oncology Division of Otorhinolaryngology, from the period between 1995 and 2005. A total of 61 patients who underwent tonsillectomy for CFT were also selected as the control group.

Ten-micrometer sections were cut from paraffin blocks and collected in 1.5-ml Eppendorf tubes for DNA extraction. To prevent cross-contamination, each block was cut after a thorough cleaning of the microtome blade. Paraffin-embedded samples were placed in xylene for 5 min and centrifuged at 14,000 rpm. DNA extraction was carried out using the QiaAmp DNA minikit (Qiagen, CA). The quality (ratio of 260/280 nm) and quantity (absorbance at 260 nm) of the isolated DNA were determined by optical density measurement. CasKi and SiHa cells were grown for ∼5 days in the appropriate medium (CasKi cells and SiHa cells, RPMI 1600 [Gibco-BRL, Grand Island, NY]). DNA isolation was performed with the QiaAmp DNA minikit, according to the protocol for cultured cells grown in a monolayer.

HPV genotyping

We used an HPV genotyping DNA chip (Biocore, Korea, Seoul) arrayed by multiple oligonucleotide probes of L1 sequence of 26 types of HPV, according to the manufacturer's protocol.10 Consensus PCR products of L1 were hybridized to the arrayed probes on the HPV chip, and HPV genotypes were identified using a fluorescence scanner (GenPix 4000B; Axon Instruments, CA) with a 532-nm laser for excitation of Cy3. The fluorescence intensity data of the specific probes were then printed out from an Excel spreadsheet.

Real-time PCR

The copy numbers of the HPV E2 and E6 ORFs were assessed using a TaqMan-based 5′-exonuclease quantitative real-time PCR assay based on the DNA amplification of a 76-bp sequence of the E2 ORF and an 81-bp sequence of the E6 ORF in the presence of HPV-16 E2- and E6-specific hybridization probes, respectively.11 The primers and probe for the E2 assay were designed to recognize the E2 hinge region of the E2 ORF, which is most often deleted upon HPV-16 integration in cervical carcinomas.12, 13 For each specimen, identical amounts of DNA were quantified for the E6 and E2 sequence of HPV-16. Each specimen was assayed 3 times. The PCR amplification was performed in a 25-μl volume containing 1× iQ SuperMix (BioRad, Hercules, CA), 200 nM E2 and E6 specific primers (Table I), 100 nM dual-labeled (5′Hex and 3′BHQ2) E2 and (5′FAM and 3′BHQ1) E6 fluorogenic hybridization probe and 200 ng of the genomic DNA template. All experiments were performed using the real-time iCycler™ PCR platform (BioRad, Hercules, CA). Two standard curves were obtained by amplification of a dilution series of the HPV viral copy number using CaSki (American Type Culture Collection, Manassas, VA) cell line genomic DNA, known to have 600 copies/genome equivalent (6.6 pg of DNA/genome) and were included in each experiment.14, 15 There was a linear relationship between the threshold cycle values plotted against the log of the copy number over the entire range of dilutions. The amplification ramp included 2 hold programs of 2 min at 50°C and 10 min at 95°C, followed by a 2-step PCR cycle with a melting step for 15 sec at 95°C and an annealing step for 1 min at 60°C and a total of 45 cycles. The ratio of E2 to E6 copy numbers was calculated to determine the physical status of the HPV-16 viral gene. HPV-16 in pure episomal form was expected to have equivalent copy numbers of E2 and E6 genes (i.e., E2/E6 ratio = 1), whereas preferential disruption of E2 upon viral integration should result in fewer E2 gene copies than E6 genes. This means that an E2/E6 ratio of less than 1 would indicate the presence of both the integrated and episomal forms, while a ratio of 0 would indicate the presence of an integrated form only. The copy number of the integrated E6 gene was calculated by subtracting the copy number of E2 (episomal) from the total copy number of E6 (episomal and integrated). The ratio of E2 to integrated E6 genes represents the amount of the episomal form in relation to the integrated form. Values less than 1 indicate the predominance of the integrated form. DNA extracted from the cervical carcinoma cell line SiHa, known to harbor a pure, integrated form of the HPV-16 gene and to disrupt the E2 and E4 ORFs, was used as the control for E2 (negative) and E6 (positive) amplification.16, 17 The relative viral load can be estimated by calculating the ratio of copies of E6 to the SiHa cell E6.

Table I. Primers and Probes Used for Real-Time PCR
NameSequenceTm (°C)
Probe 16E25′-(Hex)-CACCCCGCCGCGACCCATA-(BHQ2)-3′70

Tissue microarray

Recipient blocks were made from purified agar in 3.8 × 2.2 × 0.5-cm3 frames. Holes of 2 mm were made in the recipient blocks using a core needle, and the agar cores were discarded. The paraffin donor blocks were prepared after a thorough evaluation of the hematoxylin–eosin-stained slides. Every 2 adjacent areas of invasive carcinoma from the matching donor blocks were transplanted into the recipient blocks using a 2-mm core needle. The array for the cohort of patients with adjacent normal areas was constructed from paraffin-embedded, formalin-fixed tissue blocks. The recipient blocks were also framed in the mold that was used for the conventional paraffin block. Paraffin was then added to the frame. Consecutive 4-μm-thick sections were cut from the recipient blocks using an adhesive-coated slide system (Instrumedics, NJ).


Four-micrometer sections were placed on silane-coated slides, deparaffinized, immersed in phosphate-buffered saline (PBS) containing 0.3% (v/v) hydrogen peroxide and then processed in a microwave oven (10 mmol/l sodium citrate buffer, pH 6.5, for 15 min at 700 W). After blocking with 1% (w/v) bovine serum albumin in PBS containing 0.05% (v/v) Tween-20 for 30 min, the slides were incubated overnight at 4°C with p16, survivin, HIF-1α, skp-1, cyclin A and cyclin B1. Immunoperoxidase staining was performed using the streptavidin–biotin peroxidase complex method (LSAB universal kit; Dako, Carpinteria, CA). For negative controls, the antibodies were replaced with equivalent amounts of the subtype-matched normal mouse IgG. The final reaction product was visualized by the addition of 0.03% (w/v) of 3, 3′-diaminobenzidine tetrachloride for 5–20 min. Strong nuclear staining, except in the cases of p16 and survivin with cytoplasmic diffusion, was considered positive. Immunostaining was graded and scored as follows: 0 (no staining), 1+ (weak and diffuse or strong and focal staining), 2+ (strong and diffuse staining). In addition, a 3-tiered grading system was used for p16 which included 1+ (weak focal staining), 2+ (weak and diffuse or strong and focal staining) and 3+ (strong and diffuse staining).

Fluorescent in situ hybridization

Two-color FISH was done on 3.5-μm consecutive sections from the same TMA paraffin blocks. Before hybridization, the sections were deparaffinized, air-dried and dehydrated in 100% ethanol after incubation at 56°C for 24 h. TMA slides were treated in a wash buffer (Vysis, Downers Grove, IL) for 3 min after treatment with 0.2 N HCl for 20 min. The pretreatment solution (Vysis) at 80°C was applied for 30 min, and the slides were then washed with purified water. Slides were treated with wash buffer twice for 5 min each. Protease solution (Vysis) was applied to the immersed slides at 37°C for 10 min, and the slides were washed with a wash buffer and air-dried. Slides were fixed in 10% buffered formalin for 10 min and were washed with a wash buffer. Slides were immersed in denaturation solution (Vysis) for 5 min at 72°C, followed by serial dehydration with 70, 85 and 100% ethanol. For c-myc hybridization, 20 μl mixed LSI c-myc and CEP8 probes (Vysis) were applied, and a coverslip was placed. After overnight hybridization at 37°C in a humidified chamber, the slides were washed with a posthybridization wash buffer (Vysis) at 42°C for 2 min. Nuclei were counterstained with 20 μl 4, 6-diamino-2-phenylindole (Vysis). For EGFR hybridization, 20 μl LSI EGFR/CEP 7 probes (Vysis) were applied, and a coverslip was placed. After overnight hybridization at 37°C in a humidified chamber, the slides were washed with a 72°C posthybridization wash buffer for 2 min. Nuclei were counterstained with 20 μl 4, 6-diamino-2-phenylindole.

The c-myc and EGFR gene copy numbers in the tumor cells were estimated in ∼200 nuclei in relation with centromere (CEP) 8 and CEP 7. Hybridization signals were enumerated as the ratio of orange signals (c-myc and EGFR) to green signals (CEP8 and CEP 7) in morphologically intact and nonoverlapping nuclei. At least a 3-fold increase of the c-myc signals over the CEP8 signals in the tumor cells was considered the criterion for gene amplification. The same criterion was applied to the EGFR amplification analysis using the LSI EGFR probe together with the CEP 7 probe. Positive signals were categorized as follows: 1 for FISH-negative but with low genomic gain less than 4 copies of the gene in >40% of cells, 2 for FISH-positive with high level of polysomy in more than 4 copies of the gene in >40% of cells or gene amplification. Gene amplification was defined by the presence of tight gene clusters more than 15 copies of the gene per cell in more than 10% of analyzed cells.

Statistical analysis

The relationship between HPV status and clinicopathological parameters (TNM stage, recurrence, survival, origin of cancer, depth of invasion) or molecular factors (p16, cyclin A, cyclin B1, skp-1) were analyzed using cross-tabulations and Fisher's exact test with SAS software, version 9.1 (SAS Institute, Cary, NC).


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. References

HPV prevalence in TC and CFT

HPV was detected in 38/52 (73.1%) of the TC patients, and of the HPV-positive tumors, 34 were HPV-16 positive (87.2%). The remaining 4 samples were infected by non-16 high-risk types such as HPV-18, 33, 35, 58. There were no patients with multiple infections. Among the 69 CFT specimens, HPV was detected in 8 samples (11.6%): 3 were found to have HPV-16 and the rest were infected by HPV-58 and low-risk types (HPV-6, 11 or 84). TCs were found to be significantly associated with HPV infection (p < 0.0001 by Chi square).

Viral physical status investigation

Both validated assays of the real-time amplification systems for E2 and E6 ORFs using serially-diluted HPV-16 plasmid DNA showed similar amplification efficiencies, as reflected by the almost identical slopes of the amplification curves (Fig. 1). The results of the physical states and the copy numbers of HPV-16 E2 and E6 are summarized in Table II. Integrated E6 was calculated by subtracting the values of E2 from those of E6. The ratio of E2 to integrated E6 represented the amounts of the episomal form in relation to the integrated form. A value of less than 1.0 indicated a predominance of the integrated form.

thumbnail image

Figure 1. Comparison of PCR amplification efficiencies for HPV-16 E2 and E6. A 5-point 10-fold series of Caski cell line genomic DNA (2 × 10−2 to 2 × 102 ng) with amplification efficiencies was found to be very similar for the 2 reactions.

Download figure to PowerPoint

Table II. Physical Status of HPV-16 Infection in Tonsillar Cancer and Tonsillitis Patients
Dx/originHPV-16 E6 (copies)HPV-16 (copies/cell)E2/E6 ratioE2/integrated E6Physical status
  1. TC, tonsillar cancer; CFT, chronic follicular tonsillitis; C, crypt origin; S, surface origin; M, mixed crypt and surface origin; Epi, episomal.

TC-1/C476100.290.030.03Mixed (>integrated)
TC-2/M391700.240.330.50Mixed (>integrated)
TC-5/C4200002.580.050.06Mixed (>integrated)
TC-6/C347500.210.450.82Mixed (>integrated)
TC-7/C89830.060.340.52Mixed (>integrated)
TC-8/M143700.090.450.83Mixed (>integrated)
TC-9/M109.40.0010.370.58Mixed (>integrated)
TC-11/C210000.130.180.22Mixed (>integrated)
TC-12/M64720.040.390.64Mixed (>integrated)
TC-13/C5555003.410.410.70Mixed (>integrated)
TC-16/C58920.040.070.07Mixed (>integrated)
TC-18/C11230.010.060.06Mixed (>integrated)
TC-19/C135400.080.080.09Mixed (>integrated)
TC-20/C147100.090.050.06Mixed (>integrated)
TC-21/C259000.160.050.05Mixed (>integrated)
TC-22/C227400.140.070.08Mixed (>integrated)
TC-23/C2054001.260.490.96Mixed (>integrated)
TC-24/C47640.030.230.30Mixed (>integrated)
TC-25/C561000.340.360.57Mixed (>integrated)
TC-26/C40610.020.270.37Mixed (>integrated)
TC-27/M110600.070.340.53Mixed (>integrated)
TC-28/C415900.260.130.15Mixed (>integrated)
TC-29/C2411001.480.280.40Mixed (>integrated)
TC-30/C3945002.430.340.51Mixed (>integrated)
TC-31/S749.10.0050.651.84Mixed (>episomal)
TC-32/C121300.070.280.38Mixed (>integrated)
TC-33/C1007000.620.330.50Mixed (>integrated)

When only the episomal form is present, equivalent copy numbers of E2 and E6 should be detected. This was found in 2 samples (5.9%) of 34 HPV-16-positive TCs. Our data showed HPV-16 integration in 94.1% of 34 HPV-16-positive TCs. These included 14 specimens (41.2%) without detectable E2 sequences, which indicated the complete integration of viral genes into the host genome, and also included 18 specimens (52.9%) with an E2/E6 ratio between 0 and 1, which were indicative of the presence of mixed, integrated and episomal forms. Among the 31 HPV-16 nonepisomal status samples, 30 cases showed a predominance of the viral integrated form, with a ratio of E2 to integrated E6 of less than 1.

Three samples of HPV-16-positive CFT patients showed purely episomal forms. Ominous clinical findings including death, T-stage 4, N-stage 3, M-stage 1 and recurrence were observed more frequently in the HPV-16 of nonepisomal (integrated and mixed form) physical state, but these findings were not statistically significant except for the N-stage (p < 0.0001 by Fisher's exact test). There was no significant difference in survival between the HPV-negative group and the HPV-16-positive group (survival rate of the HPV-negative group vs. HPV-16 episomal group vs. HPV nonepisomal group: 61.5% vs. 100% vs. 75%).

Viral loading investigation

The viral loading number was determined in reference with the SiHa cell line, which was variable (Table II). Most of the results showed relatively low copy numbers. When compared with a ratio of more than 1 and less than 1 in SiHa, patients with higher copy numbers of HPV tended to survive, but the number was not significant (p = 0.0721). There was no significant correlation with the physical status of HPV.

Pathologic correlation according to HPV infection

To demonstrate that HPV-associated tonsillar SCC originates from the cryptal epithelium while non-HPV-related SCC emerges from the surface epithelium, we classified the cases into (i) invaginating tumors arising from the crypt and (ii) fungating or verrucous tumors directly arising from the surface, on macroscopic and microscopic reviews (Fig. 2). TCs of crypt showed a tendency for centrifugal growth (Fig. 3a), multinodularity (Fig. 3b) and occasional invasion of the surface epithelium in a pagetoid manner (Figs. 3c and 3d). The different phenotypes of TCs correlated significantly with the HPV infection status (p < 0.0001 by Fisher's exact test). The differentiation of the SCC was relatively poor in the HPV-positive group (p = 0.0106 by Fisher's exact test).

thumbnail image

Figure 2. Microscopic differences between tonsillar cancer of surface origin and crypt origin. (a1) Verrucous or papillary changes are restricted to the surface (labeled as V). (a2) Carcinomatous changes totally involve the whole tonsillar surface epithelium (Sr). (b1) Papillary changes are restricted to the surface adjacent to the crypt (labeled as Cr) and densely packed lymphoid cells (labeled as Ly). Notice a normal-looking surface epithelium (Sr) in contrast to (a1). (b2) A multilobulated pattern (arrowheads) of growth is characteristic of TC arising from the crypt. [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

thumbnail image

Figure 3. Microscopic findings to suggest a crypt origin for tonsillar cancer. (a) Early carcinoma (labeled with an asterisk) develops from the squamous epithelium lining the crypt (labeled as Cr). (b) Lobular arrangement (arrowheads) provides additional evidence of cryptal origin. Notice well-developed keratinizing pearls (asterisk). (c) Underlying nested cancer tissues invade the tonsillar surface (labeled as Sr). Carcinoma in situ was recognized predominantly in the crypt epithelium (labeled as Cr) but spared in the surface epithelium (Sr). (d) Higher magnification of carcinoma in situ along the cryptal epithelium. [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

Molecular comparison according to HPV infection

On analysis of p16, skp1, survivin, cyclin A, cyclin B1, HIF-1α by immunohistochemistry (Fig. 4), and EGFR and c-myc by FISH (Fig. 5), positive expression rates were as follows in decreasing order: 44 cases (84.62%) for cyclin B1, 43 cases (82.69%) for cyclin A, 37 cases (71.2%) for p16, 37 cases (71.2%) for survivin, 30 cases (57.69%) for skp1, 25 cases (48.08%) for HIF-1α, 13 cases (25%) for c-myc gene amplification and 7 cases (13.46%) for EGFR gene amplification.

thumbnail image

Figure 4. Immunohistochemical results from tissue microarray. (a1, a2) Negative staining for p16 and skp1. (b1) Grade 1 staining (diffuse weak) for skp1. (b2) Grade 2 staining (focal strong) for p16. (c1) Grade 2 staining (diffuse strong) for skp1. (c2) Grade 3 staining (diffuse strong) for p16. [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

thumbnail image

Figure 5. Fluorescent in situ hybridization (FISH). FISH shows increased c-myc gene copy numbers in tonsillar cancer (orange = c-myc; green = CEP8 control). [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

Compared with HPV-negative TC, HPV-positive TC showed a strong association with p16 overexpression (p < 0.0001) and with EGFR amplification (p = 0.0478). Moreover, HPV-16 integration status was strongly associated with c-myc amplification (p = 0.034), and HIF-1α expression (p = 0.022).

The associations between all tested molecules are summarized in Table III.

Table III. Association Between all Tested Molecules in Tonsillar Cancer
 p16SurvivinSkp1Cyclin ACyclin B1HIF-1aEGFRc-myc
  1. Each value means p-value. S, statistically significant in Qui-square or Fisher's exact test (p < 0.05); NS, statistically not significant in Qui-square or Fisher's exact test (p > 0.05).

p16 0.02190.60140.78980.08091.00000.01080.7157
SurvivinS 0.16040.45240.06080.29620.15930.6552
Skp1NSNS 0.00060.39160.91500.58300.0109
Cyclin ANSNSS 0.01630.15350.49860.1377
Cyclin B1NSNSNSS 0.16310.80450.2037
HIF-1aNSNSNSNSNS 0.56140.0484

Among all tested molecules and clinicopathologic parameters, p16 overexpression was significantly associated with a higher T-stage (p = 0.0127) and EGFR amplification (p = 0.0108), and it was also inversely associated with survivin expression (p = 0.0219). Skp-1 expression was strongly associated with the tumor recurrence (p = 0.0233) and cyclin A expression (p = 0.0006), but inversely associated with c-myc amplification (p = 0.0109). HIF-1α expression revealed a strong relationship with recurrence (p = 0.0266) and c-myc amplification (p = 0.0484). EGFR gene amplification and c-myc gene amplification were also strongly associated with each other (p = 0.0132).

The positive associations of p16 overexpression with c-myc gene amplification and HPV infection with histological grade are summarized in Table IV. p16 overexpression was strongly associated with HPV infection (p < 0.0001) and poor histological grade (p = 0.0426) of cervical cancer. In particular, the c-myc gene amplification was significantly associated with physical status of HPV (mixed versus integrated; p = 0.034), and with poor histological grade (p = 0.036). EGFR amplification was inversely associated with HPV infection (p = 0.0478) but positively with well differentiation (p = 0.0342).

Table IV. Association of p16 and c-myc with Clinicopathologic Parameters
 HPV (n = 52)Physical status (n = 34)Grade (n = 52)
Neg (n = 14)Pos (n = 38)Episomal (n = 3)Mixed (n = 18)Integrated (n = 13)Well (n = 16)Mod (n = 26)Poor (n = 10)
 0 (n = 15)11 (21.15)4 (7.69)1 (2.94)1 (2.94)2 (5.88)9 (17.31)3 (5.77)1 (1.92)
 1 (n = 7)3 (5.77)4 (7.69)0 (0)1 (2.94)1 (2.94)3 (5.77)2 (3.85)2 (3.85)
 2 (n = 7)0 (0)7 (13.46)1 (2.94)4 (11.76)0 (0)1 (1.92)5 (9.62)1 (1.92)
 3 (n = 23)0 (0)23 (44.23)1 (2.94)12 (35.29)10 (29.41)3 (5.77)16 (30.77)6 (11.54)
  p-valuep < 0.0001p = 0.2192p = 0.0426
 0 (n = 39)10 (19.23)29 (55.77)1 (2.94)16 (47.06)8 (23.53)11 (21.15)23 (44.23)5 (9.62)
 1 (n = 12)4 (7.69)8 (15.38)2 (5.88)1 (2.94)5 (14.71)5 (9.62)3 (5.77)4 (7.69)
 2 (n = 1)0 (0)1 (1.92)0 (0)1 (2.94)0 (0)0 (0)0 (0)1 (1.92)
  p-valuep = 0.7916p = 0.0340p = 0.0360
 0 (n = 45)10 (19.23)35 (67.31)3 (8.82)17 (50.0)12 (35.29)11 (21.15)24 (46.15)10 (19.23)
 1 (n = 4)3 (5.77)1 (1.92)0 (0)0 (0)1 (2.94)3 (5.77)1 (1.92)0 (0)
 2 (n = 3)1 (1.92)2 (3.85)0 (0)1 (2.94)0 (0)2 (3.84)1 (1.92)0 (0)
  p-valuep = 0.0478p = 0.7273p = 0.0342

Skp-1 and HIF-1α expression were strongly associated with the tumor recurrence (p = 0.0233 and p = 0.0266, respectively).


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. References

In a recent epidemiologic study, TC showed the strongest association with HPV with an overall detection rate of 51%, and with HPV-16 being the most prevalent type (84%) in 432 TC patients.1 In addition, the low-risk HPV type 6/11 DNA has been detected in 3% of the HPV-positive carcinomas, along with the occasional TC associated with HPV-5, 12, 31, 35 and 59.

In the present study, HPV was detected in 38/52 (73.1%) of the TC patients, and of the HPV-positive tumors, 34 were HPV-16-positive (87.2%). The remaining 4 samples were infected by non-16 high-risk types, such as HPV-18, 33, 35 and 58. None of the TC patients were infected with low-risk type HPV and none showed multiple HPV infections.

Eight and a half percent (17 of 200) of the tonsillitis samples contained HPV DNA in a previous study and were demonstrated to contain either type 16 (12 samples) or 6/11 HPV (5 samples).1 In this study, HPV was detected in 11.6%, of which 3 patients were infected with HPV-16, and the rest by HPV-58 and other low-risk types such as HPV-6, 11 or 84. Thus, there was a significant association of TC with HPV infection when compared with CFT cases.

The exact mechanism of HPV infection in nongenital regions remains uncertain, but the easy access to the tonsillar crypts and the favorable microenvironmental factors of the crypts may be causes of the high prevalence of HPV in nongenital regions. HPV DNA is detected predominantly in the epithelium lining the tonsillar crypt by in situ hybridization.18 In the present study, it has been speculated that HPV-associated tonsillar SCC originates from the cryptal epithelium, whereas non-HPV-related SCC emerges from the surface epithelium. We observed that HPV infection status was related to the characteristic gross and microscopic features of the tumors. HPV-positive TCs showed a tendency for invaginated (inverted), multilobulated and centrifugally expanding growth, whereas HPV-negative TCs tended to demonstrate fungating (polypoid) or flat and tentacular growth. HPV-positive tumors tended to be poorly differentiated with basaloid morphology.

The physical state of HPV in SCC has been analyzed with ease and reproducibility using the state-of-the-art technology of real-time PCR, but it has seldom been applied to TC. Our data showed HPV-16 integration in 94.1% of HPV-16 positive TCs, including 41.2% with complete integration and 52.9% that were indicative of the presence of mixed integrated and episomal forms. In contrast, 3 samples of HPV-16-positive CFT showed purely episomal forms. A recent paper regarding the physical status of HPV using restriction enzyme digestion, ligation and inverse PCR methods reported that all detected forms of HPV from 11 samples of 22 TCs (50%) were in episomal form.19 Venuti et al. found that 20% of HPV-16 existed in the integrated form in laryngeal carcinomas.20 However, most of the SCCs in head and neck cancers have indicated that HPV-DNA is more frequently integrated. Koskinen et al. reported that 65% of the HPV-DNA are in integrated or mixed form in 23 head and neck cancers.21 In other sites such as the esophagus, only 8.6% (3/35) of the HPV-16-positive specimens harbored the episomal form exclusively, whereas the remaining 91.4% contained either the integrated form only (5.7%) or a mixture of the episomal and integrated forms of the viral molecules (85.7%).22 In a large series of HPV-16-positive carcinoma patients, multiple deletions were found with the most common being in the region of the E2 ORF corresponding to the protein “hinge” region.12 With the real-time PCR method, HPV-16 has been shown to be integrated in CIN lesions, and a heavy load of integrated HPV-16 is closely associated with rapid progression in CINs.11 The high frequency of partial or total integration status of HPV-16 in TC in the present study together with many previous reports is similar to the situation in cervical cancer, in which viral integration can be detected in as many as 63–100% of cases and is frequently detected together with the episomal form.23, 24

When compared between the episomal and nonepisomal (integrated and mixed forms) viral physical status, the most ominous clinical findings were restricted to the nonepisomal HPV-16 status, but were not statistically significant except for the N-stage.

Upon analysis of FISH, c-myc amplification was seen in 25% and strongly associated with the HPV-16 integration physical status. The overexpression of c-myc proto-oncogene is a common event in epithelial tumors, and in cervical cancers in particular.25 A recent study showed that even low levels of oncogene amplification with 3–7 copies of c-myc were found more often in HPV-infected tumors.26 In addition, c-myc amplification could be detected in a low percentage of preinvasive lesions and in a significantly high proportion of invasive ones, demonstrating that c-myc amplification takes place in the premalignant stage and can contribute to the progression to more compromised lesions.26 c-myc alteration may be associated with viral integration in genital cancer, including the activation of a proto-oncogene via a mechanism of insertional mutagenesis and/or gene amplification.27

With specific regards to EGFR in HPV-positive tumors, HPV-16 E5 protein can activate EGFR, which in turn can initiate diverse biochemical events that ultimately render the overexpression of a variety of proto-oncogenes.28

EGFR was reported to be overexpressed in ∼32% of TCs performed by immunohistochemistry.29 In the present study, EGFR gene amplification using FISH was seen in 13% of TC, which was strongly associated with p16 overexpression. However, we failed to find any association with clinical parameters. Discordance between IHC and FISH results in EGFR is interpreted as at least 2 different mechanisms of EGFR overexpression: those that involved gene amplification and those that did not.30

If pRb acts as a negative regulator of p16 expression at the transcriptional level, then disrupted retinoblastoma (pRb) by the HPV-16 integration pathway may induce upregulation of p16.31 In addition, 71.2% of TC exhibited strong and diffuse nuclear staining for p16, whereas there were no p16-expressing cells in CFT. In the present study, p16 overexpression was strongly associated with HPV infection status (p = <0.0001) and a high T stage (p = 0.0127). p16 overexpression was found in 56% of 34 TCs and was correlated with the presence of HPV-DNA and increased disease-free survival.32 In addition, p16 overexpression was reported to be correlated with HPV status and is useful for predicting radiotherapeutic response and prognosis in TC patients.33

The SCF complex (skp1/Cul 1/Rbx 1/F-box protein) mediating the ubiquitination of cyclin E and p27kip1 promotes S-entry.34 Although S-entry promotion caused by proteolysis of p27 can have an effect on S- or G2-phase cyclin and related molecules, there are no documents about an association between skp1 and cyclins in further steps. Skp-1 was strongly associated with a high recurrence and cyclin A overexpression, but was inversely associated with c-myc amplification in TC.

In a recent paper about survivin, 37% of TC was highly expressed and was associated with tumor progression.35 However, the present study revealed a higher frequency (71.2%) of survivin expression and that the overexpression of survivin was not significant by any parameters.

HIF-1α overexpression was intimately related to cancer recurrence in the present study, although the value of this phenomenon as prognostic predictors in head and neck cancer is controversial.36, 37

This study on the prevalence and mechanisms of HPV infection on palatine tonsillar SCC and CFT in a large population revealed a significant difference in HPV prevalence between TC and CFT. Most HPV infections were HPV-16 and were physically integrated or mixed with the human genome, and they showed aberrant genetic molecules directly related with HPV integration such as the overexpression of p16 and the amplification of c-myc.


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. References