Human papillomavirus genome detection by in situ hybridization in fine-needle aspirates of metastatic lesions from head and neck squamous cell carcinomas




Patients with head and neck squamous cell carcinoma (HNSCC) often present with metastatic disease. The diagnosis of metastatic lesions usually is determined by fine-needle aspiration. Human papillomavirus (HPV) is now being considered as a causative agent in a subset of HNSCC. The objectives of this study were, first; to search for the presence of HPV DNA by in situ hybridization (ISH) in metastatic lesions from HNSCC using alcohol-fixed, archival, cytopathologic material; second, to characterize the cytologic features of HPV-positive metastatic lesions of HNSCC; and, third, to determine whether there is a correlation between the presence of HPV DNA and the origin of metastatic lesions.


The authors performed chromogenic ISH analysis for HPV DNA on fine-needle aspiration materials from metastatic lesions from 26 patients with HNSCC. Along with the ISH analysis, a detailed cytologic review was performed, and cytopathologic features were recorded. The HPV DNA status in metastatic lesion was correlated with cytopathologic features and primary tumor location.


The integration of HPV DNA was visualized microscopically on tumor cell nuclei in 15% of aspirates. The anatomic locations of the study samples were as follows: 16 lymph node aspirates (11 cervical lymph nodes and 5 lymph nodes at other sites other), 5 tracheostomy sites, and 5 miscellaneous sites located on the head and neck area. Cytologic review revealed 13 keratinized and 13 nonkeratinized metastatic tumors. HPV DNA was detected in four metastatic sites (three lymph nodes and one tracheostomy site). All HPV DNA-positive tumors were of the nonkeratinizing type (P < 0.05; Fisher exact test). The origins of HPV-positive tumors included two laryngeal sites, one nasopharyngeal site, and one oral cavity site.


The current findings showed that archival cytology slides can be used for HPV DNA detection with ISH. The results also showed that HPV DNA-containing HNSCC has distinctively nonkeratinizing cytologic features. The authors concluded that HPV DNA not only is involved in the initiation of tumoral processes but also plays an important role in the development of metastatic disease. Cancer (Cancer Cytopathol) 2005. © 2005 American Cancer Society.

About half a million patients worldwide are diagnosed with head and neck squamous carcinoma (HNSCC) annually.1 Most patients have metastatic disease at the time of diagnosis.2 Mortality rates among patients with advanced disease have not changed significantly despite improvements in therapeutic and diagnostic procedures over the years.

HNSCC has multifactorial etiology, and the most prominent etiologic factors are a history of tobacco and alcohol exposure. The risks of tumoral development with these agents are dose-related and time-dependent.3, 4 Along with tobacco and alcohol exposure, the identification of human papillomavirus (HPV) in a patient with HNSCC in 1985 suggested a possible role of viral etiology in the tumorigenesis of HNSCC.5, 6 Since then, many studies have demonstrated the presence of HPV, and HPV is now under consideration as a causative agent in a subset of HNSCCs.7–9 In a recent review, using polymerase chain reaction (PCR) analysis, HPV DNA was found in 34% of HNSCC tumors from all sites.8, 9

The objective of this study was to search for the presence of HPV DNA by in situ hybridization (ISH) in metastatic lesions from HNSCC using archival cytopathologic materials. Another objective was to characterize further the cytologic features of HPV-positive metastatic lesions in HNSCC. We also investigated the correlation between the presence of HPV DNA and the origin of metastatic lesions.


We retrieved cytologic materials of previously diagnosed patients with metastatic HNSCC from the files of the Cytopathology Department of Izmir Atatürk Research and Training Hospital in Turkey. In addition, we sought relevant clinical information. All cytology materials were ethanol-fixed smears that had been obtained with conventional fine-needle aspiration. We reviewed cytologic materials for adequacy, confirmation of diagnosis, and grading. At the same time, we selected a representative slide for each case for ISH analysis. We excluded materials if they were not sufficient to discard one slide. We used the Fisher exact test for statistical analysis. Primary tumors were not tested, because they were not available in the majority of patients.


We performed chromogenic ISH with a probe that contained DNA from HPV subtype 6 (HPV-6), HPV-11, HPV-16, HPV-18, HPV-31, HPV-33, HPV-35, HPV-45, HPV-51, and HPV-52 (Dako Corporation). The ISH protocol was as follows: Alcohol-fixed, archival cytology slides were destained and rinsed in tap water for 5 minutes and rinsed twice for 3 minutes in distilled water. Heat-induced retrieval of slides was performed in a microwave oven with 0.01 M citrate buffer, pH 6.0, at 500 W for 5 minutes. After retrieval, slides were allowed to cool for 30 minutes and then were rinsed in 3 changes of distilled water. Slides were digested with 5 μg/mL of proteinase K in 50 mM Tris-Cl, pH 7.5, for 5 minutes at room temperature and washed extensively with distilled water. Endogenous peroxidase was quenched with 3% H2O2 for 10 minutes, and slides were washed with distilled water. Twenty microliters of biotin-labeled HPV probe solutions (DAKO Corporation) were applied to individual sections and coverslipped. Slides were denatured at 95 °C for 5 minutes, and hybridization was carried out in a humidified box at 37 °C for 90 minutes. After hybridization, the slides were washed 3 times in Tris-buffered saline (TBS) with 0.05% Tween 20 for 5 minutes each. Posthybridization stringency wash was in 0.1×sodium saline citrate/0.1% sodium dodecyl sulfate at 52 °C for 10 minutes followed by a rinse in TBS-Triton X-100. Detection of hybridized probe was performed by chromogenic signal detection (DAKO® 5-bromo-4-chloro-3-inodolyphosphate/nitroblue tetrazolium [BCIP/NBT] Substrate System; DAKO Corporation). Specimens were covered with BCIP/NBT solution and incubated for 30 minutes. Slides were rinsed gently with distilled water and coverslipped with glycergel. Tissue sections that were positive for HPV wide spectrum (HPV control slides; DAKO Corporation) were used as positive controls. Biotin-labeled plasmid probes served as negative controls.


We were able to identify 26 patients with sufficient cytologic material from the cytopathology archives. All materials were smeared fine-needle aspirations from metastatic lesions of HNSCC that were obtained by the same cytopathologist with a 27-gauge needle. All slides were fixed in alcohol. There were 16 aspirates from lymph nodes, 5 aspirates from tracheostomy sites, and 5 aspirates from miscellaneously located lesions. The origins of tumors were as follows: 9 laryngeal sites, 6 nasopharyngeal sites, and 11 oral cavity sites. Twelve patients were smokers. Fourteen patients were females. There were 4 patients with Stage IV HNSCC, and the remaining patients had Stage III NHSCC. The results are depicted in Table 1.

Table 1. Cytologic Grading and Chromogenic In Situ Hybridization Results from 26 Patients with Metastatic Head and Neck Squamous Cell Carcinoma
FNA siteOrigin of primary tumorHPV DNA statusGrade
  1. HPV: human papillomavirus; WD: well differentiated; PD: poorly differentiated; NK: nonkeratinizing; NOS: not otherwise specified.

Cervical lymph node   
 9Oral cavity, NOSNegativeNK
 10Oral cavity, NOSPositiveNK
 11Oral cavity, NOSNegativeNK
Jugulodigastric lymph node   
Tracheostomy site   
Other sites   
 Occipital lymph nodeOral cavity, NOSNegativeNK
 Supraclavicular lymph nodeOral cavity, NOSNegativeNK
 Buccal mucosa, recurrence siteOral cavity, NOSNegativePD
 Jaw, soft-tissue massGingivalNegativePD
 Sternoidocleidomastoid muscle massOral cavity, NOSNegativeNK
 Tongue, recurrence siteTongueNegativeNK
 Tonsillectomy site, massTonsilNegativePD

Cytopathologic Findings

Metastatic carcinomas were classified broadly based on the presence of keratinization. Keratinizing-type tumors were classified further as either well differentiated or poorly differentiated. The nonkeratinizing tumors consisted of large polygonal cells with prominent and overlapping pleomorphic nuclei. Cytoplasm mainly was basophilic (Fig. 1). Ten of 16 lymph node aspirates as well as 2 miscellaneously located lesions and 1 tracheostomy site lesion corresponded to this type. Keratinizing squamous carcinoma cells were large and sometimes had bizarre shapes, and the cytoplasm mainly was orangophilic (Fig. 2). Keratinizing HNSCCs are termed well differentiated if pearl formation or relatively low nuclei/cytoplasm ratio has been observed. Two lymph node aspirates and four tracheostomy site lesions were well differentiated keratinizing tumors. Four lymph node aspirates and three miscellaneously located lesions were in the poorly differentiated group (Table 2).

Figure 1.

In this fine-needle aspirate from metastatic head and neck squamous cell carcinoma, the tumor is of the nonkeratinizing type and is composed of pleomorphic nuclei with relatively scant cytoplasm. There is no evidence of keratin production.

Figure 2.

In this photomicrograph of well differentiated, metastatic head and neck squamous cell carcinoma, the cells are bizarre in shape, and the cytoplasm mainly is orangophilic.

Table 2. Distribution of Metastatic Head and Neck Squamous Cell Carcinoma Subtypes by Anatomic Site
Metastatic siteNonkeratinizing HNSCCKeratinizing HNSCC
Well differentiatedPoorly differentiated
  • HNSCC: head and neck squamous cell carcinoma.

  • a

    Recurrence sites and soft-tissue masses.

Lymph node1024
Tracheostomy scar140
Miscellaneous sitesa203

HPV Study

With ISH analysis, HPV DNA was visualized as dot-like signals inside the nuclei of tumor cells (Fig. 3). The signal intensity varied from one inconspicuous dot to many conspicuous dots in single tumor nucleus (Fig. 4). This punctuate pattern of hybridization has been correlated with viral DNA integration, and the number of dots were correlated with copy numbers of HPV DNA.10

Figure 3.

Using in situ hybridization, human papillomavirus DNA was detected in this sample of metastatic head and neck squamous cell carcinoma as black dots in the nuclei of tumor cells. The signal, which is represented as dot-like punctuation, is not present outside the nuclear membrane or background.

Figure 4.

The dot-like products in the nuclei of this metastatic head and neck squamous cell carcinoma represent viral integration. The signal intensity varies from one inconspicuous dot to many conspicuous dots in a single tumor nucleus. The number of dots is related to the viral load (see Samama et al., 200210).

Among all of the lymph node aspirates, 3 aspirates (19%) were positive for HPV DNA (2 cervical lymph node aspirates and 1 jugulodigastric lymph node aspirate). One of 5 tracheostomy site lesions displayed signal (20%). None of the miscellaneously located lesions were positive for HPV DNA. Two metastatic tumors from laryngeal primary tumors (22%), 1 of 6 metastatic lesions from nasopharyngeal carcinomas (17%) and 1 metastatic lesion from an oral cavity primary tumor were positive for HPV (9%). The identification of HPV DNA according to primary tumor site is shown in Table 3.

Table 3. The Origin of Metastatic Tumors and Human Papillomavirus DNA Detection
OriginHPV DNA status: No. of tumors (%)
  1. HPV: human papillomavirus.

Nasopharyngeal1 (17)5
Laryngeal2 (22)7
Oral cavity1 (19)10
Total4 (15)22

Statistical Analysis

When tumors were stratified by cytologic grade, HPV DNA-positive tumors were of the nonkeratinizing type (Table 4). There was no significant correlation (P = 0.9) between the presence of HPV in metastatic lesions and the origin of the primary tumor (whether oral or and nonoral). No correlations were found between HPV status and gender, tobacco use, or and disease stage.

Table 4. Cytologic Grade and Human Papillomavirus Status in Metastatic Head and Neck Squamous Cell Carcinomaa
Cytologic gradeHPV status: No. of tumors
  • HPV: human papillomavirus.

  • a

    P = 0.04 (Fisher exact test).



In the current study, we detected HPV DNA in 22% of laryngeal metastases and in 12% of metastases from oral cavity-nasopharyngeal primary sites. Previous studies have shown the prevalence of variations of HPV DNA in HNSCC originating from different anatomic locations. Of these, the most common locations are palatine tonsil and oropharynx.7, 11 The rates of HPV DNA expression reported have varied among studies, and the overall expression HPV DNA has been reported between 7% and 36% in different studies evaluating HNSCC.7, 11, 12 HPV DNA was identified in 7–26% of laryngeal primary tumors and in 2.9–4.0% of oral cavity primary tumors.13, 14 The current results in NHSCC metastases correspond to those prior investigations that evaluated primary tumors.

It has been reported that HPV-positive squamous cell carcinomas have distinct histopathologic features.15–18 Among these features, basal cell morphology is especially of interest. It has been reported that HPV-positive tonsillar carcinomas have nonkeratinizing basal cell morphology.17 Likewise, in the genital tract, a distinct basaloid pattern for HPV DNA-positive squamous cell carcinomas has been reported.15, 16, 18 In our cytologic evaluation, we were unable to find a cytologic feature that corresponded to basaloid histopathology. However, our results showed that, although all HPV DNA-positive tumors were of the nonkeratinizing type, none of the keratinizing metastatic tumors contained HPV DNA. This difference was significant in our statistical analysis. The absence of keratinization in HPV DNA-containing tumors suggests that this mucosatrophic virus either preferably infects nonkeratinizing squamous cells or integrates to the host DNA, inhibiting keratinization.19

Current studies have shown that HPV-positive oral-pharyngeal carcinomas are distinct entities.9, 16 The diagnosis of these distinct entities relies on HPV DNA detection. We believe that HPV DNA detection may increase the diagnostic ability of cytologic evaluation in patients who present with masses of the head and neck area, especially in specimens that contain limited numbers of cells.

We used chromogenic ISH in our study for the detection of HPV DNA. Eighteen percent of all tested HNSCC samples were HPV-positive according to the ISH results.7 Higher rates of detection were identified with Southern blot and PCR analyses. Except for in situ PCR, ISH it is the only method by which the HPV genome can be observed in topographic relation with pathologic lesions. Compared with PCR, ISH is inexpensive and is feasible for routine use. In addition, the complicated laboratory infrastructure that is mandatory for PCR is not required for ISH.20–24

In the current study, we were not able to find a correlation between the primary tumor sites and HPV DNA status of metastatic sites; however, recent studies have shown that HPV has a predilection for lymphoepithelial structures in certain anatomic regions, especially the tonsils and Waldeyer ring.19 Begum et al. showed that detection of HPV-16 in metastatic cervical lymph nodes is correlated with an oropharyngeal origin.25

The biologic behavior of HPV-positive HNSCCs is subject to debate. In the study by Paz et al., patients who had HPV-positive tumors presented with later stage disease compared with patients who had HPV-negative tumors.26 Although HPV-positive tumors seem to spread more rapidly than HPV-negative tumors, overall survival is almost the same or better than in patients with HPV-negative tumors. Moreover, Gillison et al. found that the risk of dying from disease was reduced (59%) in patients with HPV-positive HNSCC after correcting for age, metastasis, and alcohol consumption.9

Although tobacco and alcohol are chemical agents that contact the upper aerodigestive mucosa directly, the mechanism by which this mucosatrophic carcinogenetic virus infects the mucosa of the upper aerodigestive track has not been well established. Some studies demonstrated a significant correlation between a sexual behavior, such as young age of activity and multiple sexual partners; the risk is independent of tobacco products and alcohol use.27, 28 In addition, it was found that HPV infection of the oral region was rare in preadolescent children prior to sexual debut.29

In conclusion, we demonstrated the presence of HPV in metastatic lesions of HNSCC. We also showed that HPV DNA-containing tumors have distinctively nonkeratinizing cytologic features and alcohol-fixed archival cytologic material can be used for HPV DNA detection by ISH.