• Open Access

Prevalence of human papillomavirus in mobile tongue cancer with particular reference to young patients


To whom correspondence should be addressed. E-mail: ishikawa@jfcr.or.jp


The carcinogenetic role of human papillomavirus (HPV) in mobile tongue cancer remains unclear because of conflicting results reported in the literature. This disparity is likely to be due to variations in the samples and methods used. Furthermore, despite a tendency for increased prevalence of mobile tongue cancer in young adults, only a few reports specifically in young patients have been published. In the present study on 32 patients, including six in their 20s, we genotyped the prevalence of HPV using a highly sensitive detection tool in fresh-frozen samples from surgical specimens and a novel detection device with electrochemical DNA chip and loop-mediated isothermal amplification. In addition, we confirmed HPV prevalence by in situ hybridization and immunohistochemistry for the p16INK4a protein, regarded as a biomarker of HPV-associated cancers. The frequency of 13 genotypes of high-risk HPV was 0/32 (0%), which was further confirmed by in situ hybridization. Overexpression of p16INK4a protein was observed in six of the 32 patients (19%), with four (67%) also overexpressing p53. Because there is usually a lack of p53 overexpression in HPV-associated cancer, it is unlikely that p16INK4a protein overexpression is correlated with HPV infection. Consequently, it is unlikely that HPV infection plays an important role in mobile tongue carcinogenesis, in particular in young adults. In addition, our data suggest that the overexpression of p16INK4a protein is not an appropriate biomarker for HPV association in mobile tongue carcinogenesis. (Cancer Sci 2012; 103: 161–168)

Human papillomavirus (HPV) is a small DNA virus and more than 130 different types of HPV have been identified based on DNA sequence variations.(1,2) Human papillomavirus is divided into low-risk and high-risk types, depending on the carcinogenic power.(3) It has been conclusively established that more than 99% of uterine cervical cancers are associated with HPV infection.(2,4) Because HPV-associated cancers may be prevented with vaccination,(5–7) it is important to determine whether cancer of extracervical organs is associated with HPV infection.(8,9) Among the head and neck cancers, oropharyngeal cancer, particularly tonsillar cancer, and basal tongue cancer are known to be strongly associated with HPV infection.(3) However, the role of HPV in mobile tongue cancer (MTC) remains unclear because of the disparity of results published so far, which have reported HPV infection rates ranging from 0% to 100%.(10–20) This disparity may be due to differences in the types of samples used and the methods of detection. A recent study has reported an increasing incidence of MTC among young adults.(21) Thus far, alcohol consumption and heavy cigarette smoking have been suggested as causal factors for MTC.(11,13) However, it is difficult to explain the high incidence of MTC in young adults simply as a result of accumulated exposure to alcohol and smoking.(11,21,22) To date, there have been only a few reports on relationship between MTC (as well as tonsillar cancers) and HPV in young patients.(23)

The primary aim of the present study was to determine the prevalence of HPV in MTC using a comprehensive assay series. To that end, we screened rapidly frozen surgical specimens that were subsequently assayed with electrochemical DNA chips (EC chips) and the loop-mediated isothermal amplification (LAMP) method. We confirmed HPV infection using other methodologies, such as in situ hybridization (ISH) and immunohistochemistry (IHC) to investigate the overexpression of p16INK4a protein.(24) The secondary aim of the study was to elucidate whether overexpression of p16INK4a protein is an appropriate marker HPV infection.

Materials and Methods

Patients.  Thirty-two patients presenting with untreated MTC who subsequently underwent surgical resection at the Cancer Institute Hospital, Japanese Foundation for Cancer Research (JFCR), between January 2006 and December 2009 were enrolled in the study. All samples were dissected from a main viable part of each tumor and snap-frozen in liquid nitrogen, typically within 20 min of removal, and stored at −80°C until processing. Histological diagnosis and grading were performed using sections stained with H&E, made from formalin-fixed and paraffin-embedded tissues, based on the World Health Organization (WHO) classification.(25) Each tumor was staged according to the Cancer Staging Manual of American Joint Committee on Cancer TNM classification.(26) All samples were collected from patients who had provided informed consent, and the study was approved by the Institutional Review Board of the JFCR.

Histological sub-classification.  As described previously,(27,28) squamous cell carcinoma (SCC) of MTC was divided into three groups based on histological features: (i) non-keratinizing (NK) SCC; (ii) keratinizing (K) SCC; and (iii) hybrid SCC. This division is thought to be particularly useful for the analysis of HPV-related SCC because HPV-related tonsillar cancer may derive from the cryptal epithelium, having a more basaloid nature, whereas non-related tonsillar cancer may originate from the surface epithelium.(29,30) Briefly, NK SCC is defined as forming sheets, nests, or trabeculae with pushing borders and forms comedo-type necrosis. Tumors have ovoid to spindled hyperchromatic cells that lack prominent nucleoli and have indistinct cell borders. In the present study, K SCC was defined as that composed entirely of mature squamous cells with keratinization and intercellular bridges without areas exhibiting NK SCC morphology. Hybrid SCC was defined as SCC having definitive areas with NK SCC morphology but exhibiting squamous maturation in >10% of the tumor. We also used conventional differentiation grading of SCC.

Immunohistochemistry.  Formalin-fixed, paraffin-embedded (FFPE) tumor-rich blocks were prepared for p53 and p16INK4a IHC. After deparaffinization, the sections were submerged in either sodium citrate buffer or Tris–EDTA buffer for heat-induced epitope retrieval at 97°C for 40 min. The immunostaining was performed using the EnVision + dextran polymer kit (Dako, Tokyo, Japan) and the CINtec histology kit (MTM Laboratories, Heidelberg, Germany) using a Dako Autostainer (Dako, Glostrup, Denmark). A positive control consisting of an invasive cervical cancer was included. Results were evaluated by one pathologist (YI). “Positive cells” were defined as those exhibiting strong nuclear and cytoplasmic staining for p16INK4a and strong nuclear staining for p53, and we classified tumors in terms of the percentage positive cell as follows: 0, negative; 1+, 1–25% positive cells; 2+, 26–50% positive cells; 3+, 51–75% positive cells; and 4+, >75% positive cells.

In situ hybridization.  Additional sections were prepared and deparaffinized for HPV DNA detection using ISH. The ISH was performed using a wide-spectrum HPV biotinylated DNA probe (Dako) for common HPV types and the GenPoint amplification system (Dako) according to the manufacturer’s instructions. This wide-spectrum probe targets the genomic DNA of HPV types 6, 11, 16, 18, 31, 33, 35, 39, 45, 51, and 52. In each series, a positive control sample, consisting of a tonsillar cancer already identified as HPV positive, was included. Punctate nuclear staining (brown nuclear dots) was considered as positive.

Detection of HPV DNA.  Fresh-frozen blocks were added to individual tubes in addition to 200 μL virus standard buffer (comprising 10 mL of 5 M NaCl, 10 mL of 1 M Tris–HCl (pH 7.5), 5 mL of 0.1 M EDTA-2Na (pH 7.5), and 975 mL milliQ water). Each sample was homogenized, followed by the addition of 700 μL virus standard buffer, 100 μL of 10% SDS, and 25 μL of 25 mg/mL proteinase K solution. Samples were incubated overnight at 37°C. After digestion, DNA purification was performed by phenol–chloroform extraction and ethanol precipitation. The quantity and quality of the DNA were assessed by spectrophotometry and electrophoresis on a 1% agarose gel. The DNA solutions were stored at 4°C.

Human papillomavirus detection and genotyping were performed using an HPV DNA genotyping system. The DNA extracted from each sample was incubated at 95°C for 5 min and then cooled immediately on ice. A 1-μL aliquot of the DNA was added to each of the six LAMP tubes, containing 24 μL LAMP reagent and specific primers for the different types of HPV. The tubes were incubated at 65°C for 90 min, then at 80°C for 5 min, and the cooled immediately on ice. A 10-μL aliquot of the amplified products from each tube was mixed with 6 μL hybridization buffer (saline citrate solution), and 50 μL mixed LAMP product was applied to an EC chip, which uses novel current detection technology (CLINICHIP HPV; Sekisuimedical, Tokyo, Japan). The EC chip was placed in Genelyzer GLH-2C601 (Toshiba Hokuto Electronics, Asahikawa, Japan), which is an instrument designed to measure electrochemical signals from the EC chip. A positive control was included in each series, consisting of an invasive cervical cancer. The results were evaluated automatically. The principle of measurement was as follows. Thirteen DNA types (type 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68) of high-risk HPV were amplified using the LAMP method with 13 primers that amplify HPV DNA in a type-specific manner. This HPV DNA genotyping assay targets the L1 open reading frames, conserved among most of the HPV genomes. The DNA probes for the target genes are fixed on the electrodes of the EC chip. When HPV DNA extracted from samples is amplified and introduced to the EC chip, it reacts and binds only with the probe with the complementary sequence. Subsequently, when an intercalator (i.e. labeling reagent for the double-strand structure) is added, it only combines with the DNA that has reacted and current flows only between the bound intercalator and the EC chip electrode. The sequence of the sample DNA is determined by detecting this current. The criteria used to evaluate the samples were as follows. Positive samples were scored where the current level of the test specimen was higher than that of the average for the negative control by ≥10 nA. Negative samples were scored where the difference in current levels between the test specimen and the average of the negative control was within 10 nA. Invalid samples were scored when the average current level of the negative control was >45 nA. All genotypes were detected at a 100% positive at HPV DNA 250 copies per 1 μL. The HPV genotypes determined by this HPV DNA genotyping system for 244 test clinical samples were compared with direct PCR sequencing using specific primers for 13 genotypes. The concordance rate for HPV genotyping was 95.5% (233/244).

Statistical analysis.  Statistical analyses for a correlation between age groups, younger (20–39 years) and older patients (40–90 years), or immunoreactivity and the clinical outcome were performed using Fisher’s exact test. Differences in disease-free survival and overall survival between younger (20–39 years) and older patients (40–90 years) were evaluated using the Kaplan–Meier method, starting from the date of surgery. Survival curves was compared using the log-rank test. < 0.05 was considered significant.


Table 1 lists the characteristics of the 32 patients in the present study. Twenty-nine patients (91%) were male and three (9%) were female. The age of the patients ranged between 21 and 90 years, with a median age 55 years. Ten patients (31%), including six who were in their 20s, were younger than 40 years. The mean follow-up time was 26.1 months, with a maximum of 53 months. One of 32 patients was diagnosed histologically as having spindle cell carcinoma, whereas the others were all diagnosed with standard SCC. The patient with spindle cell carcinoma was a 55-year-old male smoker who also had the heaviest alcohol consumption per day among the 32 patients. The spindle cell carcinoma recurred in 10 months and the patient died 23 months after surgery. Most patients had well-differentiated tumors, but three had moderately (26-, 44-, and 76-year-old men) and two had poorly (31- and 38-year-old men) differentiated tumors. Although the three patients with moderately differentiated tumors had no recurrence, tumors in the two patients with poorly differentiated tumors recurred: one within 7 months, with the patient dying 30 months after surgery, and the other within 8 months. This second patient is still alive 40 months after surgery, although at a terminal stage.

Table 1.   Patient and tumor characteristics
  1. *Excluding two patients in whom regional lymph nodes could not be assessed. Alcohol consumption categorized as “yes” indicated daily, whereas that categorized as “incidental” indicated social drinkers. SCC, squamous cell carcinoma.

Sex (n)
Age (years)
 Median (range)55 (21–90)
Tumor histology (n)
 SCC/spindle cell carcinoma31/1
Histological grade of SCC (n)
 Well/moderately/poorly differentiated26/3/2
Pathological stage* (n)
 I+II/III+IV 8/22
Alcohol consumption (n)
Cigarette exposure (n)
 Yes/no/former smoker12/12/8

Regarding pathologic stages, eight patients (27%) were in early Stages I/II, 22 (73%) were in advanced Stages III/IV, and two were unstaged because their lymph nodes were not examined. Of the eight patients in early Stages I/II, three (38%) were younger and five (62%) were older than 40 years of age. Two (67%) of the three younger patients in early Stages I/II had local recurrence, whereas none (0%) of the five older patients had any recurrence (= 0.107). Of the 22 patients with advanced Stages III/IV, seven (32%) were younger and 15 (68%) were older than 40 years of age. The percentage of patients with an unfavorable outcome was 72% (5/7) in the younger age group and 54% (7/13) in the older age group (= 0.392). The remaining two patients in whom staging was not performed died from other diseases. Therefore, there were no significant differences in the prognosis between younger (<40 years) and older (>40 years) patients.

Of the 32 patients, 12 (38%) were daily smokers, eight (24%) were past smokers, and 12 (38%) were life-time non-smokers. The mean cumulative smoking cigarette volume was 31.6 pack-years for the 12 daily smokers and 11.9 for the eight past-smokers. Only 11 of the 32 patients (34%) in the present study were drinkers. In addition, seven of the 32 patients (22%) were neither smokers nor consumers of alcohol and one of these seven was in his 20s.

Table 2 gives the clinicopathological features and results of IHC (p16INK4a and p53), ISH, HPV DNA detection and genotyping analyses. The 32 tumors were classified histologically as one spindle cell carcinoma and 31 SCCs. Of the 31 SCCs, 21 (68%) were classified as K SCC (Fig. 1a), none (0%) was found to be an NK SCC, and 10 (32%) were determined to be hybrid SCC (Fig. 1b). Overexpression of p16INK4a was noted in six of 32 patients (19%; Fig. 2a). The grading of the p16INK4a-positive tumors was as follows: two were 1+, one was 2+, one was 3+, and two were 4+. Two of three tumors (67%) with strong reactivity (3+ or 4+) were in patients who were aged in their 20s. Only one of six patients (17%) with overexpression of p16INK4a experienced local recurrence and died, whereas 14 of 26 patients (54%) without p16INK4a overexpression died, indicating a prognostic value of p16INK4a, although the difference failed to reach statistical significance in the current sample set (= 0.116). Overexpression of p53 was noted in 19 of 32 tumors (59%; Fig. 2b). Specifically, 13 tumors were not stained, four were weakly stained (three 1+ tumors and one in 2+ tumor), and 15 were strongly stained (five 3+ tumors and ten 4+ tumors). Eight of 10 patients (80%) who were <40 years of age were found to overexpress p53. Eleven of the 15 patients (73%) with strong p53 overexpression died of MTC, one from another disease; in contrast of the 17 patients with no or only weak p53 expression, only four (24%) died of MTC, one from another disease. All other patients survived. Therefore, patients with strong p53 overexpression had a significantly less favorable prognosis than those with weak or no p53 expression (< 0.01). With ISH using the wide-spectrum probes against carcinogenic HPV DNA, none of the samples in our series showed the punctate pattern, correlated with viral DNA integration (Fig. 2c). Consequently, although we used two carcinogenic HPV DNA detection methods (i.e. ISH and the novel HPV DNA genotyping system) and they worked well for a positive control case with tonsillar and cervical cancer (Figs 2d and 3b), all samples of MTC were negative (Figs 2c and 3a).

Table 2.   Clinicopathologic features, results of immunohistochemistry (p16INK4aand p53 status), ISH, human papillomavirus DNA detection, and clinical outcome
Patient no.Age (years)SexpTNHistologic features*p16INK4A expression†p53 expression†ISHGS‡Outcome
  1. *Tumors were classified into three types histologically: (i) keratinizing squamous cell carcinoma (K SCC); (ii) non-keratinizing squamous cell carcinoma (NK SCC); and (iii) mainly NK SCC but with areas of squamous maturation (Hybrid SCC). †The percentage of positive-stained cells were scored as follows: 0, negative; 1+, 1–25% positive cells; 2+, 26–50% positive cells; 3+, 51–75% positive cells; and 4+, >75% positive cells. Note, p16INK4A-positive cells were those in which both the nucleus and cytoplasm were stained, whereas p53-positive cells exhibited nuclear staining only. ‡The genotyping system (GS) used in the present study is a novel tool to detect 13 types of high-risk human papillomavirus (HPV) strains using electrochemical technology and loop-mediated isothermal amplification (LAMP). DOD; death from other disease. IHC; in situ hybridization, Spindle CC; spindle cell carcinoma.

 121MpT2N2bHybrid SCC4+4+Death
 225MpT1N0K SCC4+Death
 326MpT3N2bK SCC3+4+No recurrence
 427MpT3N1K SCC3+Death
 528MpT2N0K SCC3+Death
 628MpT4aN2bK SCC4+No recurrence
 731MpT2N2bHybrid SCCDeath
 833MpT4aN2bHybrid SCC4+Death
 935MpT2N2bHybrid SCC4+Death
1038MpT1N0Hybrid SCCRecurrence, alive
1144MpT4aN0Hybrid SCCNo recurrence
1247MpT2N2K SCCNo recurrence
1348MpT2N0K SCC2+1+No recurrence
1449FpT3N1Hybrid SCC4+No recurrence
1551MpT3N0K SCCNo recurrence
1655MpT3N0Spindle CC4+Death
1755MpT2N0K SCCNo recurrence
1857MpT1N0K SCCNo recurrence
1959MpT2N0K SCC1+1+No recurrence
2060FpT3N2cHybrid SCC3+Death
2164MpT2N2bK SCC3+DOD
2268MpT2N2bK SCC2+DOD
2369MpT3N0K SCCNo recurrence
2471MpT2NXK SCC4+No recurrence
2572MpT2N1K SCC1+No recurrence
2676MpT2N0Hybrid SCCNo recurrence
2777MpT4aN2bHybrid SCC4+Death
2878MpT3N2bK SCC3+Death
2979MpT2N2bK SCC1+Death
3080FpT3N2bK SCC4+Death
3185MpT3N2bK SCCDeath
3290MpT2NXK SCCDeath
Figure 1.

 Histological features (H&E stain). (a) Squamous cell carcinoma (SCC) composed entirely of keratinizing malignant cells and (b) SCC composed mainly of non-keratinizing malignant cells but also exhibiting partly formed squamous maturation (hybrid SCC). Scale lines, 1.00 mm (a); 400 μm (b).

Figure 2.

 (a–c) Immunohistochemical staining for p16INK4A (a) and p53 (b), and in situ hybridization (ISH) findings (c) for Patient 1 (see Table 2). (d) Tonsillar cancer positive control for ISH (brown indicates nuclear staining). (a) Strong nuclear and cytoplasmic staining was seen for p16INK4A (i.e. >75% positive cells) and (b) strong nuclear staining was seen for p53 (also >75% positive cells). (c) The tumor was negative by ISH. Scale lines, 200 μm (a,b); 100 μm (c,d).

Figure 3.

 Presence of high-risk human papillomavirus (HPV), as determined using a novel HPV DNA genotyping system with an electrochemical DNA chip and loop-mediated isothermal amplification (LAMP). Lane numbers correspond to HPV genotypes. PC, positive control; NC, negative control. (a) Patient 1 (see Table 2) was negative for all HPV genotypes. (b) A positive control sample (HPV genotype 18, cervical cancer, DNA extracted from paraffin-embedded block).

The 3-year disease-free and overall survival rates were 48% and 50%, respectively. The prognosis for younger patients (20–39 years) was less unfavorable than that for older (40–90 years) patients in terms of both disease-free and overall survival. A significant difference was found between these two age groups in terms of disease-free survival (= 0.027), but not for overall survival (= 0.102; Fig. 4).

Figure 4.

 Kaplan–Meier survival curves for (a) disease-free survival and (b) overall survival in younger (20–39 years) and older (40–90 years) patients. Both disease-free and overall survival analyses indicated a tendency for a less favorable prognosis in younger compared with older patients. However, the difference only reached statistical significance in the case of disease-free survival (log-rank test).


It is important to determine whether extracervical cancers are associated with HPV because these cancers may be preventable by HPV vaccination. Of the extracervical sites, tumors in the oral cavity, oropharyx, esophagus, penis and anus are possibly associated with HPV because of its route of infection(31–33) and, indeed, tumors in these sites exhibit morphological similarities.(16,30) Of the head and neck cancers, oropharyngeal cancer is related to HPV, as is tonsillar cancer in particular, with detection rates of HPV16, HPV18, and/or HPV33 reported in the range of 21–100%.(19,34–36) Interestingly, HPV-positive tumors have been reported to respond better to chemoradiotherapy than HPV-negative tumors and survival is longer in patients with HPV-positive tumors than those with -negative tumors.(23,37–41) Although the importance of carcinogenic HPV infection is well recognized for diagnostic and therapeutic strategies in tonsillar cancer, the prevalence of HPV in MTC (adjacent to tonsillar cancer) remains uncertain. In the present study, we evaluated the prevalence of carcinogenic HPV infection and investigated the correlation between p16INK4a protein expression and HPV. We found that there was no evidence of carcinogenic HPV infection in MTC and that the p16INK4a protein was not an appropriate marker of HPV infection.

Although the possibility that HPV is relevant to MTC has been suggested, there are no reliable prevalence data supporting this assertion. This may be due mainly to the diverse target groups or methodologies used by different studies. In fact, some previous studies have classified MTC as an oral cancers,(14,42) which is too broad a classification to allow causal agents to be specified. Furthermore, different studies have used different experimental procedures: ISH, Southern blot hybridization, IHC and PCR have been used for detection; brush, biopsy, and surgery have been used for sampling; and fresh-frozen or FFPE tissue blocks have been used for storage.(43,44) Fresh-frozen materials seem to provide higher detection rates than FFPE samples, probably because of target DNA degeneration following formalin fixation; furthermore, PCR methods appear to be more sensitive than others.(45) Using FFPE samples may result in the possible contamination of carcinogenic HPV DNA at the time of sectioning.(13) Therefore, in the present study, we chose to use fresh-frozen materials from surgical specimens to avoid both DNA degeneration and contamination. In addition, the LAMP method used in the present study is reportedly equivalent to, or even more sensitive than, PCR.(46,47) We also used ISH for additional confirmation of HPV DNA. Thus, we tried to achieve higher sensitivity and higher accuracy than in previous studies.

In addition to differences in the methodologies used in previous studies, the wide range in the prevalence of HPV reported in different studies may be ascribed to variations in the nature of HPV-associated cancers. For example, a geographic heterogeneity of HPV-associated cancers has been reported. In India and Southeast Asia, oral cancer is a predominant malignancy, accounting for up to 50% of all cancers.(10) Similarly, HPV prevalence in oropharyngeal cancer is significantly higher in North America and Asia than in Europe.(8) Another factor that may impact on the apparent prevalence of HPV is the time period/s examined. For example, Okinawa, a subtropical island in south-west Japan, the prevalence of HPV in SCC of the lung we found to decrease significantly over a relatively short period of time, from 68% in 1995, to 35% in 1996, 23% in 1997, and 24% in 1998.(48) In the present study, all patients examined were from mainland Japan, a region with a lower prevalence of HPV infection than Okinawa. Based on the findings of the previous studies discussed above, our results of no HPV infection in MTC may be explained, in part, by the facts that all our patients were from a region with low HPV infection and that the present study was performed using recent cases (2006–2009).

Characteristic histopathology of HPV-related cancers has been reported, with differentiation grades appearing to differ depending on the site. For example, HPV-related SCC in the lung are mostly well differentiated,(48) half of all uterine cervical cancers are moderately differentiated,(49,50) and tonsillar SCC are poorly differentiated.(29,30) Unfortunately, the relatively low number of cases in the present study means that we did not have a sufficiently wide enough spectrum of differentiation (e.g. we only had two cases of poorly differentiated tumors and no NK SCC cases) and so we cannot draw any conclusions about the histopathology of HPV-related MTC. However, it should be kept in mind that the subclassification of SCC used in the present study was derived originally from 89 tonsillar cancers and that there is a basic histological difference between the mobile tongue and the tonsil. In fact, the mobile tongue is covered by keratinized squamous epithelium, whereas the tonsil is coated with stratified, non-keratinized squamous epithelium. Thus, a subclassification system based on squamous cell maturation (i.e. keratinization) may not be suitable for classifying MTC.

Sexual behavior may be another factor affecting the prevalence of HPV. A higher number of lifetime sexual partners and engaging in oral sex are principal risk factors for exposure to HPV. In fact, several studies have reported increased risks of oral and oropharyngeal cancers among individuals with a high number of sexual partners.(9,42,43,51,52) Unfortunately, we cannot comment on the relationship between HPV infection in MTC in terms of sexual behavior in the present study because there are no data available. In terms of the impact of sexual behavior, notably that oral sex is a predominant cause of increased rates of MTC, a possible drawback of the present study is that only three of the 32 patients were women. There may be a greater correlation between MTC and HPV in women than in men because the issue of oral sex may be more relevant in the case of women.

Overexpression (1+ to 4+) of p16INK4a protein was observed in six of 32 patients (19%) of MTC. Integration of HPV to host cells leads to increased expression of oncogenes E6 and E7. The HPV E6 and E7 proteins bind and functionally inhibit p53 and retinoblastoma protein (pRb), respectively, and the functional inactivation of pRb by E7 leads to upregulation of p16INK4a protein as a result of the loss of negative feedback control.(53,54) Thus, both overexpression of p16INK4a protein and lack of expression of p53 are usually evidence of an HPV-associated cancer.(28,55,56) However, in the present study, four of six tumors (67%) with p16INK4a overexpression also exhibited p53 overexpression. This suggests that the p16INK4a overexpression in these tumors bears little relation to HPV. It is true that overexpression of p16INK4a protein is a biomarker for cervical or tonsillar cancer arising from carcinogenic HPV infection,(30,33,57,58) but this is not the case for MTC.

The incidence of MTC among young adults is reported to be increasing, but this cannot be explained by ordinary carcinogens for oral cancer, such as smoking and alcohol consumption, because young people must have less cumulative exposure to carcinogens than older people.(21,59,60) It has been suggested that the increased HPV infection in industrialized countries was due, in part, to changing sexual behavior and an increase in oral sex.(61,62) This may be a cause of increased MTC, but are only few studies have investigated MTC in young patients, especially those in their 20s (Table 3). In the present study there were no HPV-positive cases in the six patients aged in their 20s or in the 10 young adults (20–39 years old). On the basis of these results, it is unlikely that HPV infection is a cause of increased MTC among the young. Thus, the cause of the increasing incidence of MTC among young people remains unclear.

Table 3.   Human papillomavirus prevalence in mobile tongue cancer among young patients (<45 years of age)
ReferencesDetection techniqueAge (years)Detection rateSample storage
  1. *The genotyping system (GS) used in the present study is a novel tool to detect 13 types of high-risk human papillomavirus (HPV) strains using electrochemical technology and loop-mediated isothermal amplification (LAMP). FF, fresh-frozen; PE, paraffin-embedded; ISH, in situ hybridization; PCR, polymerase chain reaction.

Cruz et al.(16)PCR37, 392/2FF
Liang et al.(11)PCR<451/8FF
O’Regan et al.(20)PCR31, 33, 33, 37, 39, 39HPV16 DNA 2/6 HPV16 E6/E7 mRNA 0/6FF+PE
Cao et al.(18)PCR31, 38, 40, 42, 443/5PE
Premoli-de-Percoco et al.(15)PCR29, 31, 32, 35, 405/5FF
Siebers et al.(14)ISH, PCR23, 370/2PE
Present studyISH, GS*21, 25, 26, 27, 28, 28, 31, 33, 35, 38, 440/11FF


The authors thank Ms Reimi Asaka (Division of Pathology, JFCR Cancer Institute) for technical assistance. This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan (18014027), from the Japan Society for the Promotion of Science (20390105), grants from the Ministry of Health, Labour and Welfare (19–12), the Japan Science and Technology Agency, and the Smoking Research Foundation.

Disclosure Statement

The authors have no conflicts of interest.