Prevalence of human papillomavirus 16 and nucleolar organizer region counts in oral exfoliated cells from normal and malignant epithelia

Prevalence of human papillomavirus 16 and nucleolar organizer region counts in oral exfoliated cells from normal and malignant epithelia Er-Jia Mao, DDS, MS, PhD, a Seattle, Wash. DEPARTMENT OF ORAL PATHOLOGY, SCHOOL OF CLINICAL DENTISTRY, SHEFFIELD UNIVERSITY, U.K., AND COLLEGE OF STOMATOLOGY, WEST CHINA UNIVERSITY OF MEDICAL SCIENCES, CHENGDU, CHINA.

This study was designed to identify the prevalence of human papillomavirus 16 (HPV 16) in oral exfoliated cells from 26 patients with oral cancer and matched healthy volunteers with the use of polymerase chain reaction. In addition, the value of a silver staining technique for nucleolar organizer regions (AgNORs) was also investigated. HPV 16 was detected in 30.8% of the cancer lesions, 26.9% of the unaffected sites, and 15.4% of samples from normal mucosa. AgNOR counts on the same cases were analyzed. Although AgNOR counts are useful in distinguishing between normal and malignant oral exfoliated cells, they provided no additional prognostic information for oral cancer. However, when AgNOR counts were compared with HPV 16-positive and HPV 16-negative counts in cancer lesions, AgNOR counts were higher in HPV-positive lesions. These findings suggest that HPV 16 may play a role in tumor cell proliferation, but it is unlikely to play a significant role alone in the cause of oral cancer. Therefore evidence of HPV 16 infection in oral

malignant neoplasms should be cautiously interpreted. (ORALSURG ORAL MED ORAL PAYHOLORAL RADIOLENDOD 1995;80:320-9)

Many recent reports have suggested that human papillomaviruses (HPVs) play an important role in the development of neoplasms. HPVs have been detected with different techniques in anogenital lesions in as many as 84% of patients with cervical cancer. 1 In oral lesions many independent groups have identified HPV antigens and HPV-DNA in benign, premalignant, and malignant lesions of the oral cavity, which included oral squamous cell papilloma, 2,3 condyloma acuminatum, 3 focal epithelial hyperplasia, 4 verruca vulgaris, 5 oral lichen planus, 6 oral hairy leukoplakia, 7 smokeless tobacco keratoses, 8 odontogenic keratocyst, 9 ameloblastoma, 1~ oral leukoplakia, 6 oral epithelial dysplasia, 8 and squamous cell carcinoma.6, 8, 11 Additionally, both normal cervical epithelium and oral keratinocytes have successfully been immortalized and transformed by HPV 16 DNA in vitro. 12-14 Thus HPV appears to be one of the etiologic agents involved in development of not only benign but also premalignant and malignant epithelial lesions. Currently HPV has been reported to include more than 70 types and several subtypes. More than 12 types of HPV have been identified in oral lesions, inSupported by the British Council. aResearch Associate. Received for publication Oct. 26, 1994; returned for revision Nov. 21, 1994; accepted for publication Dec. 20, 1994. Copyright 9 1995 by Mosby-Year Book, Inc. 1079-2104/95155.00 + 0 7/14!63035

32O

cluding HPV 1, 2, 4, 6, 7, 11, 13, 16, 18, 30, 32, and 57. However, HPV 16 is one of the most common types to be associated with both oral and genital carcinomas.1, 11, 15 The role of HPVs in oral carcinogenesis still requires investigation for a number of reasons. Reports of HPV prevalence in patients with oral cancer varies from 0% to 100%. 16 This variability could be related to the sample size or the sensitivity of applied techniques. Some of the previous research did not include control samples or had a small number of con, trol samples compared with the number of case sampies.5,6, 8 Other difficulties in oral carcinogenesis studies include obtaining normal oral biopsy specimens from healthy individuals and even greater difficulty of obtaining multiple biopsies from the same individual. Additionally, the background information for the normal control groups often is not completed, particularly for related cofactors that are known oral carcinogens, for example, tobacco use and alcohol consumption. Nucleolar organizer regions (NORs) are the morphologic sites around which the nucleoli develop at the end of mitosis. 17 Hybridization techniques have shown that these regions represent the loops of DNA actively transcribing to ribosomal RNA and thus to ribosome and ultimately to protein. 18 NORs can be identified by means of the argyrophilia of their associated proteins (AgNORs) as nuclear dark dots. Many recent reports have suggested that the number of

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY

Mao

321

Volume 80, Number 3

AgNORs per nucleus is related to cellular proliferation and differentiation. 19,2~ This finding suggests that AgNORs can be used as an aid in diagnosing borderline lesions and grading malignant lesions as well as an indicator or prognostic factor in various malignant neoplasms. The purpose of this investigation was to assess the prevalence of HPV 16 in oral exfoliated cells from patients with oral cancer with the use of the polymerase chain reaction (PCR) technique. Lesions and unaffected sites from patients with oral cancer were compared with samples from healthy individuals. Despite the fact that exfoliative cytologic evaluation has been successfully used for identifying dysplasia or malignant change in anogenital lesions, some oral pathologists still doubt its value in the diagnosis of oral malignancy because of reported high false-negative rates. 21 To further investigate whether the A g N O R method can play a role in the diagnosis and prognosis of oral squamous cell carcinoma, exfoliated cells from oral cancer lesions and normal oral mucosa were analyzed by the A g N O R method. As part of this study it was necessary to establish suitable fixation and counting methods for the A g N O R stain in exfoliated cells. Finally, the correlation between HPV 16 infection and A g N O R counts in oral exfoliated cells from both healthy persons and patients with oral cancer was assessed.

MATERIAL AND METHODS Study subjects Cases. Specimens from 26 patients with oral cancer who were admitted to hospitals in the vicinity of Sheffield, U.K., were analyzed. None of these patients had previously undergone any cancer treatment. Seventeen men and nine women ranging in age from 44 to 84 years were enrolled in testing. In addition to specimens obtained from patients' oral cancer lesions, specimens were collected from matching sites on their unaffected sides as internal control samples. With the use of clinical tumor-nodes-metastasis classification, 22 the 26 patients were divided into four groups: three in Stage I, four in stage II, 11 in stage III, and eight in stage IV. Eight of the 26 patients had at least one metastasis in neck lymph nodes. Control group. Twenty-six healthy volunteers were selected from staff and patients at the Charles Clifford Dental Hospital in Sheffield as the control group. These volunteers were matched for site, age, and sex with each patient with cancer. All were nonsmokers, drank less alcohol than average (less than a social drinker), and had no clinical lesions involving their oral mucosa.

Table I. E6 primers and probe of HPV 16 genome Sequences (5' ad 3') Primer 1 TCA AAA GCC ACT GTG TCC TG Primer 2 CGT GTT CTT GAT GAT CTG CA Probe GAC AAA AAG CAA AGA TTC CAT AAT ATA AGG GGT CGG TGG A

Genomic location

421-440

540-521

460-499

Oral exfoliated cells Four cytologic samples were taken from each healthy individual. All samples w e r e taken from the right side of the mouth: the buccal mucosa in the region opposite the first molar of both jaws, the alveolar mucosa in the buccal region of the upper first molar, the lateral border of the tongue in the region near the lower third molar, and the floor of mouth adjacent to the lower third molar. For each patient with squamous cell carcinoma, a sample was also taken from the lesion in addition to samples from the same mucosal sites as those of the healthy subjects. Sample collection Samples were obtained with a cytobrush (Colgate Medical Ltd. Berkshire, U.K.). The brush was dipped in phosphate-buffered saline solution before the sample was taken and was then held against the mucosa of the collection site and rotated for 10 full turns. After this procedure was done, the cells were removed by agitation of the brush in a microcentrifuge tube containing 1 ml phosphate-buffered saline solution until the cells thoroughly dispersed into the buffer. The exfoliated cells were then pelleted at 1000g centrifugation for 10 minutes at room temperature. Each pellet was washed three times with phosphate-buffered saline solution. Each sample pellet was divided: one half for detection of HPV 16 by PCR and one half for A g N O R staining. The pellets were stored at - 2 0 ~ C until needed.

DNA preparation Cellular D N A was extracted from exfoliated cells with a standard proteinase K and phenol/chloroform extraction techniquefl 3 The samples were incubated in 1/10 volume of sodium dodecyl sulfate (SDS) and 1/100 volume of proteinase K overnight at 37 ~ C. Then an equal volume of phenol/chloroform was added to extract the DNA. Finally, the aqueous phase was precipitated in two volumes of ethanol, 1/10 vol-

322

Mao

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY

September 1995

Fig. 1. Oral exfoliated cells fixed by 95% alcohol and stained by AgNOR technique. Numerous black dots (AgNORs) within nuclei are clearly defined (arrows), (Original magnification x405.)

ume of 3 mol/L NaOAc, and 1/1000 volume of transfer R N A (10 mg/ml) overnight at - 2 0 ~ C. The sample was centrifuged, redissolved in H20, and stored a t - 2 0 ~ C.

PCR A single pair o f E6 primers was selected for PCR amplification according to the study b y Maitland et al. 24 (Table I). A standard PCR amplification was performed. 25 The reaction mixture consisted of a 100 lal reaction volume containing 1 ~tg genomic DNA, 10 rnmol/L Tris-HC1 pH 8.4, 2.5 mmol/L MgC12, 50 mmol/L KC1, 0.01% gelatin, 100 mmol/L deoxyribose nucleoside triphosphates, 250 ng of each primer, and 2.5 U Taq polymerase. D N A from SiHa ceils (5 to 50 copies per cell) was used as a positive control for HPV 16. To avoid a false-positive reaction caused by contamination, D N A from K562 cells (which are known to have no H P V - D N A ) and two water blanks (one added during the reagent preparation and one at the same time as the samples) were incorporated as negative control samples. PCR reaction mixtures were spun briefly and heated in an automated thermal cycler (Perkin Elmer Corporation, Norwalk, Conn.) for 40 cycles of amplification. The PCR cycles consisted of 95 ~ C for 1 minute, 37 ~ C for 2 minutes, and 72 ~ C for 3 minutes. When PCR amplification was completed, 10 pl PCR products was subjected to

electrophoresis in a 2% agarose gel. In the gel the D N A was visualized under ultraviolet light after staining with ethidium bromide. Southern blotting of the amplified products was done by a standard method with a D N A probe labeled by [~/_32p] with T4 polynucleotide kinase. The specific oligonucleotide probe used was the internal portion of the amplified sequence (Table I). The hybridized filters were washed in 2 x SSC, 0.5% sodium dodecyl sulfate (SDS) at 42 ~ C and dried. The membranes were then autoradiographed at - 7 0 ~ C for 5 days. All negative samples were further tested by [~-globin analysis to ensure that they were all amplificable by PCR.

AgNOR technique Fixation method. The exfoliated cells were extruded onto a clean, dry, glass slide with a Cytospin (Shandon Scientific Lit., Runcorn, Cheshire, U.K.). Because no previous report of oral exfoliated cells stained by the A g N O R method was known by the author, a suitable fixative had to be determined. The visualization of AgNORs is thought to be dependent on the type of fixation used, and certain fixatives are highly deleterious to the A g N O R reaction. 26 A pilot study examined three commonly used fixation methods: 95% ethanol, 10% neutral-buffered formalin,

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY

Mao

Volume 80, Number 3

Fig. 2. Oral exfoliated cells fixed by 10% neutral-buffered formalin and stained by AgNOR technique. Black intranuclear dots (AgNORs) are visible but not as clearly defined as in Fig. 1. (arrows). (Original magnification x405.)

Fig. 3. Oral exfoliated cells fixed by air drying and stained by AgNOR technique. Black dots (AgNORs) are visible but not as separate individual dots (arrows). (Original magnification x405.)

323

324

Mao

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY

September 1995

Fig. 4. Southern blot hybridization with HPV 16specific probe. The results of the Southern transfer hybridization with a 32p-labeled HPV 16 internal oligonucleotide probe confirm those obtained by PCR. SiHa DNA was positive, and sample lanes 1, 3, and 4 were also positive; whereas negative control samples (K562 and blank water) remained negative. and air drying. Smears were collected from the buccal mucosa of three volunteers, and three separate smears were taken from each person; one was fixed in 95% ethanol, another in 10% neutral-buffered formalin, and the third smear was air-dried. The minim u m fixation time for each smear was 4 hours. A modified A g N O R staining method (described later) was then performed on these smears. The results showed that more AgNORs are clearly visible after fixation in 95% ethanol than with the other two fixation methods. However, AgNORs were visualized after all fixation methods (Figs. 1, 2, and 3). Therefore 95% ethanol was used as the fixation method in this study. AgNOR staining method. The A g N O R staining method was based on the original method designed by Ploton et al., 27 as modified by Chungpanich 2s for paraffin-embedded oral mucosa. The staining was done in the dark at 4 ~ C for 1 hour. The staining mixture was freshly prepared and consisted of one volume of 2% gelatin in 1% formic acid solution and two volumes of 50% aqueous silver nitrate solution. T h e stained slides were washed with distilled water and dehydrated through increasing concentrations of ethanol, cleared in xylene, and mounted in Piccolyte (Kodak, London, U.K.). No counterstaining was needed.

Counting procedure All slides were examined by light microscopy by a x l 0 0 0 oil immersion objective with an automatic counter (Wild M501, Heerbrugg, Switzerland) to prevent duplicate counting. In this investigation nucleolar clusters of AgNORs were counted as a single A g N O R dot, as were the more readily discernible single extranucleolar dots.

Unlike biopsy sections, A g N O R dots in oral exfoliated cells were occasionally covered by artifacts such as oral bacteria or debris. The microscopic fields for analysis were selected by random number tables; any of these fields that contained cells affected by the artifacts or had overlapping nuclei were disregarded, and the next random number field was counted. The resulting data were analyzed by one-way analysis of variance and the student t test on the Minitab (Minitab Inc., State College, Pa.) statistical computer package.

Nuclei count To determine the suitable counting number for oral exfoliated cells, three samples were randomly selected from each of the three sample groups. Cumulative means were obtained for each individual sample by counting up to 100 nuclei. To evaluate whether the results of AgNOR counting were reproducible and reliable, 15 randomly selected samples were counted twice by the same person. The repeat counts for each sample varied by less than 9%.

RESULTS Prevalence of HPV type 16 All oral smears f r o m healthy individuals and patients with oral cancer were tested by PCR for the presence of HPV 16 DNA. A typical result of a Southern blot is illustrated in Fig. 4. Of the 26 samples from oral cancer lesions, eight (30.8%) were positive for H P V 16 by PCR, whereas unaffected sites and healthy control sites were positive in seven (26.9%) and four (15.4%) samples, respectively. Comparison of the HPV 16 D N A infection rates in these three groups showed t h a t they did not differ

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 80, Number 3 9

Mao

32,,$

~

Unaffected sites

8

', Cancer Lesions

X Healthy Mucosae

7

6 5

I

I

4 3 X

2

~..~

1

]

0

I

I

I

5

I

I

I

I

t

10

~

I

I

I

I

15

I

I

I

I

I

I

I

20

I

I

I

25

I

I

I

I

I

I

1

I

30

I

I

35

I

[

i

I

I

40

I

I

I

I

~

45

I

I

I

I

I

50

Number of Nuclei Fig. 5. Number of nuclei and cumulative means of AgNORs from oral exfoliated cells. significantly (p > 0.05), although the infection rate in the cancer lesions was higher than in the other two groups.

AgNOR count In all the exfoliated cell preparations, AgNORs were visible as distinctly black dots within yellowbrown nuclei (Fig. 1). Fig. 5 illustrates the three groups of mean values; these became stable after 40 nuclei had been counted in cancer lesions or after nuclei had been counted 45 in unaffected sites and healthy mucosa. These results indicated that a count of 100 nuclei is sufficient for oral exfoliated cells. The mean A g N O R counts per nucleus for each group were as follows: cancer lesions, 4.69 ___0.72; unaffected sites, 2.99 ___0.43; and healthy mucosa, 2.44 -+ 0.37. These results are graphically displayed in Fig. 6; individual points represent the mean count per sample in each group. Statistical analyses of the A g N O R counts showed that the mean counts for the cancer lesions were significantly higher than those for the unaffected sites of the same patients and for healthy mucosa (/9 < 0.05), whereas no difference was seen between the latter two groups (p > 0.05). Comparison of A g N O R counts in oral cancer lesions showed no statistically significant differences between the clinical stages of cancers or between the presence or absence o f metastases (Table II). HPV infections and AgNORs counts In all three groups (cancer lesions, unaffected sites, and healthy mucosae) the mean A g N O R counts from

HPV 16-positive samples were higher than those of HPV 16-negative samples (Table III). However, only in cancer lesions were the A g N O R counts significantly higher in the HPV-positive samples than in the HPV-negative samples (p < 0.01). DISCUSSION Oral exfoliated cells Exfoliative cytologic evaluation is one of the most important diagnostic tools used for the detection of genital lesions, and it is currently widely used to screen various diseases in the gynecology field. The advantages of oral exfoliative cytologic evaluation in viral detection studies have been previously described. 29 This procedure is noninvasive and painless, and when it is compared with a biopsy section, more evaluable oral epithelial cells are likely to be obtained in oral smears. To my knowledge this is the first report of the use of A g N O R staining technique for oral exfoliative cells.

HPV-16 infection in oral mucosa Determining the prevalence of HPV 16 infection in the normal population is very important in establishing a link between H P V 16 infection and cancer. A few recent studies involving the PCR technique have claimed that HPV 16 has been found not only in malignant neoplasms but also in histologically normal tissues. Maitland's group 15 identified H P V - D N A related to HPV 16 in 41% of normal oral biopsy specimens. More recently, oral exfoliated cells from healthy adults and preschool children were examined for HPV 6b and HPV 16 with the PCR technique. 3~ In adults HPV 6b and HPV 16 were present in 17%

326

Mao

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY

September 1995 Table II. The relationships between clinical features and A g N O R counts Clinical features Site of lesions Buccal Alveolar Tongue FOM Sex Women Men Age <49 yrs 50-59 yrs 60-69 yrs >70 yrs Size of tumor <4 cm >4 cm Grade of tumor Stage I Stage// Stage III Stage IV Metastasis + -

No. of cases

AgNOR counts

Table III. The correlation between HPV 16 infection and A g N O R counts AgNOR counts

p Value

2 9 11 4

3.39 4.38 4.90 4.60

-+ 1.44 - 0.62 + 0.59 - 0.33

>0.05

9 17

4.55 _+ 0.44 4.62 -+ 0.69

>0.05

1 8 7 10

4.05 4.48 4.81 4.75

-+ 0.67 -+ 0.76 -+ 0.50 -+ 0.59

>0.05

10 16

4.62 -+ 0.52 4.58 -+ 0.73

>0.05

3 4 11 8

4.12 5.18 4.56 4.87

-+ 0.83 -+ 0.88 _+ 0.73 -+ 0.86

>0.05

8 18

4.51 -+ 0.54 4.65 + 0.69

>0.05

FOM, Floor of mouth.

and 23% of samples, respectively; the findings in preschool children were 24% and 19%, respectively. HPV 16 has also been identified in six of 10 histologically normal samples of buccal mucosa by Adler-Storthz et al.7 A similar finding was reported by Lawton et al., 31 who detected HPV in 60% of oral samples from 60 individuals; in that study HPV 16 was the most prevalent genotype detected. These findings indicate that HPV, particularly type 16, could be present in normal oral mucosa as an asymptomatic or latent form. Some recent publications have shown that HPV 16 is present in a higher percentage of patients with oral cancer compared with healthy control groups, u Additionally, the E6 gene of HPV 16 has been reported to be capable of transforming and immortalizing normal oral keratinocytes in vitro.13, 14 However, the reported prevalence of H P V 16 in oral cancer varies among different research groups. The possible etiol o g i c association between HPV 16 and oral cancer still needs to be further defined. This study found an H P V 16 infection rate of 30.8% among patients with oral cancer. In contrast, 26.9% of unaffected sites from the same group of patients were HPV-positive. No significant difference in HPV 16 positivity was seen among the cancer lesions, their

HPV 16+ HPV 16p Value

Cancer lesions

Unaffected sides

Health mucosae

6.03 _+ 0.39 3.82 _+ 0.59 <0.01

2.82 -+ 0.51 2.41 -+ 0.64 >0.05

2.77 _+ 0.44 2.15 -+ 0.48 >0.05

unaffected sites, and the matched healthy control lesions (15.4%). In comparing different sites of the oral cavity, no significant correlation could be found between HPV 16 infection and the specific sites of oral mucosa. These results suggest that there appears to be no clear evidence showing an association between HPV 16 infection and oral cancer. This observation is in agreement with a recent publication that showed that HPV 16 D N A was present in a high percentage of normal epithelial tissue. 15 Thus if HPV 16 D N A is involved in oral carcinogenesis, the virus alone may not play a significant role in the cause of oral cancer. H P V - D N A has been found either integrated in the host chromosome or in a n extrachromosomal location. 32 The physical state of viral D N A in tumor cells is important, because it is closely a s s o c i a t e d with risk factors for the development of carcinomas. 32 Previous reports showed that HPV 16 in oral lesions has been found to be integrated frequently in the late gene region. There also appears to be a deletion in the late gene region of HPV 16 D N A sequence, which is considered to be unique to oral lesions. 24 The specific location of viral integration appears to be very important in the regulation of HPV gene expression.

AgNOR counts in oral exfoliated cells In this study exfoliated cells were spread onto the glass slides by cytospin. These samples were thicker than sections, making it difficult to examine individual intranucleolar dots. However, the advantage of using exfoliated cells for A g N O R counting is that the whole cell can be examined, reducing the possibility of underestimating the A g N O R counts per nucleus. The risk of obscuring some AgNORs by superimposition and coalescence is minimal. Exfoliated cells were found to show distinct AgNOR dots with 95% alcohol used as a fixative. In addition, 95% alcohol considerably reduced the nonspecific background staining and made the A g N O R dots more distinguishable. Cells fixed in 10% neutralbuffered formalin stained less clearly, and cells fixed

Mao

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 80, Number 3

7

327

AgNORs per nucleus

i

i

I

I

f

Healthy Mucosae

Unaffected Sites

Cancer Lesions

Fig. 6. Distribution of mean AgNOR counts from three groups. by air drying gave the least suitable results. Previous reports are consistent with this finding. 26 Two techniques are commonly used for counting A g N O R dots. Both methods count all silver-stained structures in the nucleus (intranucleolar and extranucleolar). In the first method dots that lie in groups as a cluster (almost aggregated o r partly desegregated nucleoli) are considered as one structure (dot) and are counted with the extranucleolar dots. This method has been considered to be simple and reproducible and to have fewer interobserver differences. 18 The second method is used when AgNORs can be seen separately within a cluster. The number of dots in the cluster is estimated and counted together with the smaller single AgNORs seen outside the nucleolus. In this study it was initially found to be difficult to count the tiny AgNORs within the nucleolus of exfoliated cells. Thus the first counting method was chosen for the A g N O R study with oral exfoliated cells. AgNORs represent the loops of D N A that contain the ribosomal RNA gene. It has been suggested that the A g N O R counts in nuclei may reflect cell proliferative activity or the degree of malignancy of a tumor.19, 20 The close association between the number of AgNORs and the state of activation of cells has also been determined by flow cytometric analysis. 33 Many types of tissues from different body sites have been examined by the A g N O R technique; these sites include: cervix, stomach, lungs, esophagus, intestine, liver, breast, colorectum, and lymphoid tissue. 18 Resuits from these studies have shown that AgNORs may be useful in diagnostic evaluation. With the use of formalin-fixed and paraffin-embedded tissues, Cabrini et al. 34 reported that A g N O R counts in oral squamous cell carcinoma were significantly higher than were those in oral papilloma and normal oral mucosa.

In this study of oral cancer lesions, the mean AgN O R count in oral exfoliated cells was 4.69 • 0.72. The means of the unaffected sites and matched healthy mucosa were 2.99 ___0.43 and 2.44 • 0.37, respectively. In cancer lesions the mean A g N O R count was significantly higher than in those patients' unaffected sites or in healthy mucosa, but no significant difference was seen between the latter two groups. This finding suggests that the mean A g N O R counts may help distinguish normal from malignant oral exfoliated ceils. This finding is useful because the false-negative rate for cytologic examination alone could be as high as 31%. 21 The A g N O R method may well be of some use in the early diagnosis of malignant change in oral exfoliated cells, particularly for epidemiologic studies. Although considerable overlap was seen between each category (Fig. 6), the resuits are encouraging. Even though the A g N O R technique is labor-intensive and to some extent subjective, it could be an important tool for distinguishing borderline lesions. Initial reports indicated that the number of A g N O R counts could be used as a possible indicator for malignant neoplasms, for example, esophageal and nonsmall cell lung cancersY, 36 The prognostic value of A g N O R counts in oral squamous cell carcinomas 37 and salivary gland tumors 38 has also been considered by a number of researchers. Sano et al. 37 demonstrated that oral cancer lesions in a group of patients with a poor prognosis had a higher pooled mean A g N O R number (9.74 ___ 1.72) than did those in a group with a good prognosis (6.39 • 1.67). Their studies suggested that oral c a n c e r lesions with high A g N O R counts may be more aggressive and therefore have a poorer prognosis. However, several other studies have shown that A g N O R counting is only of limited value in grading tumors and that it is inferior

328

Mao

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY

September 1995 to classic prognostic methods such as histologic grading, D N A flow cytometry, mitotic activity measurements, and immunohistochemistry. 28' 33, 39 This investigation did not establish any correlation between the means of A g N O R counts and the clinical stages of tumors. No statistical differences were found in A g N O R numbers between the clinical features analyzed, which included lesions from different sites (buccal, alveolar, tongue, and floor of mouth), the four age groups, sex, and the presence or absence of metastases. This result is apparently due to the counting method. This is partially supported by Chungpanich's investigation, 28 which showed that only the number of AgNORs per cluster was found to be a parameter for predicting the malignant potential of oral epithelial dysplasia. The number of AgNORs represents the NOR-associated protein, and the activity of this molecule is related to protein synthesis. It is possible that the increase in A g N O R counts in superficially located malignant cells is not related to an increase in malignant potential. Therefore it may be that the mean A g N O R number simply does not have any prognostic value for malignant oral exfoliated cells. Furthermore the small number of samples examined in this study may have made it difficult to separate different clinical groups.

Correlation between HPV-16 and AgNOR counts In this study A g N O R counts from oral exfoliated cells positive for HPV 16 were significantly higher than those that were HPV 16-negative (Table III). However, this difference was limited only to the cancer group. This finding indicates that the numbers of A g N O R were correlated not only with cell proliferation (e.g., high A g N O R counts in malignant cells) but also with H P V infection in oral mucosa. The association between HPV infection and A g N O R counts was also supported by a previous report. 4~ Genova et al. 4~ claimed that there were different patterns Of A g N O R distribution between HPV-infected lesions with dysplasia and without dysplasia. They suggested that A g N O R counts were diagnostically useful in the evaluation of borderline lesions. Another study analyzing a series of HPV-positive cervical cancers showed that the ratio of large/small AgNORs and the total number of AgNORs proved to be independent prognostic predictors. 41 However, whether H P V infection affects the means of AgNORs in the nuclei of tumor ceils still remains unclear. Research involving a larger sample group and including other HPV types may be required to confirm these associations. In summary, this investigation has shown that collection of oral exfoliated cells by the cytobrush is a practical method for obtaining samples in viral epi-

demiologic and histologic studies. Although the prevalence of HPV 16 in oral cancer lesions was higher than that in matched control lesions, no statistical difference was seen (p > 0.05). Whether HPV 16 is an etiologic agent in oral cancer or is simply a passenger in oral mucosa or whether it has to combine with some specific cofactors such as smoking and alcohol consumption in oral carcinogenesis requires further evaluation. The high prevalence of HPV 16 in normal oral mucosa indicates that HPV 16 infection is not a unique factor linked with transformation of normal epithelium to squamous cell carcinoma. Thus the presence of H P V 16 in oral malignant neoplasm should be interpreted with caution. This study shows that the fixation method is important in studies of this nature and is directly associated with the quality of visualization of the A g N O R dots. The optimal method for oral exfoliated cells is 95% alcohol fixation. Although the mean count of AgNORs per nucleus in oral cancer lesions was significantly higher than those in control lesions, the diagnostic usefulness in an individual patient is still doubtful, because considerable overlap was seen. In oral cancer lesions no relationship was demonstrated between A g N O R counts and clinical features: lesion site, sex, age, size of tumor, tumor clinical stage, and metastases. A g N O R counts from oral cancer lesions that were HPV 16-positive were statistically higher than those that were HPV-negative. This result suggests that HPV 16 may play a role in tumor cell proliferation. However, HPV 16 D N A alone is unlikely to play a significant part in development of oral squamous cell carcinoma. The author thanks Prof. C. J. Smith, head of the Department of Oral Pathology, for his excellent supervision, invaluable advice, and constructive criticism during this study; Mr. P. G. McAndrew (consultant in oral and maxillofacial surgery), who provided access to patients with oral cancer; Mr. D. Thompson, for his photographic assistance; and S. Mahan and D. Teale, for their help in preparation of this manuscript. REFERENCES 1. zur Hausen H, Schneider A. The role of papillomaviruses in anogenital cancer. In: Salzman NP, Howley PM, eds. The papillomaviruses. New York: Plenum, 1987:245-63. 2. Eversole LR, Laipis PJ. Oral squamous papillomas: detection ofHPV DNA by in situ hybridization. ORAL SURGORALMED ORAL PATHOL 1988;65:545-50. 3. Miller CS, White DK. In situ hybridization analysis of human papillomavirus in orofacial lesions using a consensus biotinylated probe. Am J Dematopathol 1993;15:256-9. 4. Padayachee A, Van Wyk CW. Human papillomavirus (HPV) DNA in focal epithelial hyperplasia by in situ hybridization. J Oral Pathol Med 1991;20:210-4.

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 80, Number 3 5. Premoli-de-Perc0co G, Galindo I, Ramirez JL, Pert'one M, Rivera H. Detection of human papillomavirus-related oral vermca vulgaris among Venezuelans. J Oral Pathol Med 1993 ;22:113 -6. 6. Kashima HK, Kutcher M, Kessis T, Levin LS, de Villiers EM, Shah K. Human papillomavirus in squamous cell carcinoma, leukoplakia, lichen planus, and clinically normal epithelium of the oral cavity. Ann Otol Rhinol Laryngol 1990;99:5561. 7. Adler-Storthz K, Ficarra G, Woods KV, Gaglioti D, DiPietro M, Shillitoe EJ. Prevalence of Epstein-Barr virus and human papillomavirus in oral mucosa of HIV-infected patients. J Oral Pathot Med 1992;21:164-70. 8. Greet RO, Eversole LR, Crosby LK. Detection of human papillomavirus-genomic DNA in oral epithelial dysplasias, oral smokeless tobacco-associated leukoplakias, and epithelial malignancies. J Oral Maxillofac Surg 1990;48:1201-5. 9. Cox M, Eveson J, Scully C. Human papillomavirus type 16 DNA in an odontogenic keratocyst. J Oral Pathol Med 1991; 20:143-5. 10. Van Heerden WFP, van Rensburg EJ, Raubenheimer EJ, Venter EH. Detection of human papillomavirus DNA in an ameloblastoma using the in situ hybridization technique. J Oral Pathol Med 1993;22:109-12. 11. Chang F, Syrj~nen S, Nuutinen J, K~irj~iJ, Syrj~nen K. Detection of human papillomavirus (HPV) DNA in oral squamous cell carcinomas by in situ hybridization and polymerase chain reaction. Arch Dermatol Res 1990;282:493-7. 12. Wemess BA, Miinger K, Howley PM. The role of the human papillomavirus oncoproteins in transformation and carcinogenic progression. In: DeVita VT, Hellman S, Rosenberg SA, eds. Important advances in oncology. Philadelphia: JB Lippincott, 1991:2-18. 13. Park N-H, Min B-M, Li S-L, Huang MZ, Cherick HM, Doniger J. Immortalization of normal human oral keratinocytes with type 16 human papillomavirus. Carcinogenesis 1991; 12: 1627-31. 14. Oda D, Garrett L, Disteche CM. Chromosomal abnormalities in HPV 16 immortalized cultured oral epithelial cells. J Dent Res 1994;73:360. 15. Maitland NJ, Cox MF, Lynas C, Prime SS, Meanwell CA, Scully C. Detection of human papillomavirus DNA in biopsies of human oral tissue. Br J Cancer 1987;56:245-50. 16. Chang F, Syrj~inen S, Kellokoski J, Syrj~tnen K. Human papillomavirus (HPV) infections and their associations with oral disease. J Oral Pathol Med 1991;20:305-17. 17. Babu KA, Verma RS. Structural and functional aspects of nucleolar organizer regions (NORs) of human chromosomes. Int Rev Cytol 1985;94:151-76. 18. Crocker J. Nucleolar organiser regions. In: Underwood JCE, ed. Current topics in pathology: pathology of the nucleus. Berlin: Springer Verlag, 1990:91-149. 19. Quinn CM, Wright NA. The clinical assessment of proliferation and growth in human tumours: evaluation of methods and applications as prognostic variables. J Pathol 1990; 160:93 102. 20. Hall PA, Levison DA, Woods AL, et al. Proliferating cell nuclear antigen (PCNA) immunolocalisation in paraffin sections: an index of cell proliferation with evidence of deregulated expression in some neoplasms. J Pathol 1990; 162:28594. 21. Folsom TC, White CP, Bromer L, Canby HF, Garrington GE. Oral exfoliative study: review of the literature and report of a three-year study. ORAL SURG ORAL MED ORAL PATHOL 1972;33:61-71. 22. Hermanek P, Sobin LH. UICC International Union Against Cancer. TNM classification of malignant tumours. 4th ed. Berlin: Springer Verlag, 1987:13-8. 23. Kawasaki ES. Sample preparation from blood, cells, and other fluids. In: Innis MA, Gelfand DH, Sninsky JI, White TJ, eds. PCR protocols: a guide to methods and applications. San Diego: Academic Press, 1990:146-52.

Mao

329

24. Maitland NJ, Bromidge T, Cox MF, Crane IJ, Prime SS, Scully C. Detection of human papillomavirus genes in human oral tissue biopsies and cultures by polymerase chain reaction. Br J Cancer 1989;59:698-703. 25. Ting Y, Manos MM. Detection and typing of genital human papillomaviruses. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ, eds. PCR protocols: a guide to methods and applications. San Diego: Academic Press, 1990:356-7. 26. Smith PJ, Skilbeck NQ, Harrison A, Crocker J. The effect of a series of fixatives on the AgNOR technique. J Pathol 1988; 155:109-12. 27. Ploton D, Menager M, Jeannesson P, Himber G, Pigeon F, Adnet J-J. Improvement in the staining and in the visualization of the argyrophilic proteins of the nucleolar organizer region at the optical level. Histochem 1986;18:5-14. 28. Chungpanich S. The diagnostic and prognostic potential of nucleolar organizer regions in oral epithelial dysplasia. Sheffield, England: University of Sheffield; 1989. Thesis. 29. Mao E-J, Smith CJ. Detection of Epstein-Barr virus (EBV) DNA by the polymerase chain reaction (PCR) in oral smears from healthy individuals and patients with squamous cell carcinoma. J Oral Pathol Med 1993;22:12-7. 30. Jenison SA, Yu X-P, Valentine JM. Evidence of prevalent genital-type human papillomavims infections in adults and children. J Infect Dis 1990;162:60-9. 31. Lawton GM, Thomas SJ, Schonrock J, Monsour FN, Frazer !H. Human papillomaviruses in normal oral mucosa: a comparison of methods for sample collection. J Oral Pathol Med 1992;21:265-9. 32. Howley PM. Role of the human papillomaviruses in human cancer. Cancer Res 1991, 51 (suppl):5019S-22S. 33. Kahn MA, Mincer HH, Dockter ME, Hermann-Petrin JM. Comparing flow cytometric analysis and nucleolar organizer region enumeration in archival oral premalignant lesions. J Oral Pathol Med 1993;22:257-62. 34. Cabrini RL, Schwint AlE, Mendez A, Femopase F, Lanfranchi H, Itoiz ME. Morphometric study of nucleolar organizer regions in human oral normal mucosa, papilloma and squamous cell carcinoma. J Oral Pathol Med 1992;21: 275-9. 35. Morita M, Kuwano H, Matsnda H, Moriguchi S, Suginachi K. Prognostic significance of argyrophilic nucleolar organizer regions in esophageal carcinoma. Cancer Res 1991 ;51:533941. 36. Kaneko S, Ishida T, Sugio K, Yokoyama H, Sugimachi K. Nucleolar organizer regions as a prognostic indicator for stage I non-small cell lung cancer. Cancer Res 1991 ;51:4008-11. 37. Sano K, Takahashi H, Fujita S, et al. Prognostic implication of silver-binding nucleolar organizer regions (AgNORs) in oral squamous cell carcinoma. J Oral Pathol Med 1991;20: 53-6. 38. Matsumura K, Sasaki K, Tsuji T, Shinozaki F, The nucleolar organizer regions associated protein (Ag-NORs) in salivary gland tumors. Int J Oral Maxillofac Surg 1989; 18:76-8. 39. Eskelinen MJ, Lipponen PK, Collan Y, Syrj~len KJ. The role of nucleolar organiser regions as prognostic factors in breast cancer. Eur J Cancer 1991;27:989-92. 40. Genova G, Guddo F, Vita C, et al. Argyrophilic nucleoproteins of the cervical epithelium in HPV infection and intraepithelial neoplasia. Pathologica 1991;83:461-6. 41. Tosi P, Cintorino M, Santopietro R, et al. Prognostic factors in invasive cervical carcinomas associated with human papillomavirus (HPV): quantitative data and cytokeratin expression. Pathol Res Pract t992;188:866-73.

Reprint requests: Er-Jia Mao Program in Cancer Biology, C1-015 Fred Hutchinson Cancer Research Center 1124 Columbia St. Seattle, WA 98104