8 76

0 1992

The Japanese Society of Pathology

Detection of Human Papiliomavirus DNA in lnvasive Cervical Cancers by the Polymerase Chain Reaction and Its Clinical Significance

Kenji Kashiwabara and Takashi Nakajima

In order to detect human papillomavirus (HPV) DNA in invasive cervical cancers, three different polymerase chain reactions to amplify different subgenomic fragments of HPV DNA were carried out on DNA extracted from 93 formalin-fixed and paraffin-embedded tumor tissues. This study detected HPV DNA in 5 4 cases (58.1%), which broke down to HPV 1 6 in 3 9 (41.9%) cases, HPV 18 in six (6.4%), HPV 5 2 in three, HPV 3 3 in one and unclassified HPV type in the remainder. Histopathologically, squamous cell carcinomas frequently contained HPV 16, whereas, HPV 18 was present in adenocarcinoma, adenosquamous cell carcinoma and small cell carcinoma of the cervix. Clinicopathological study revealed that HPV 16 and 18 DNA found were more frequently than other HPV subtypes in premenopausal patients. Moreover, HPV 18 DNA-positive cancers had a relatively high recurrence rate. These results indicate that cervical cancers might be clinically influenced by the difference in subtypes of the infecting HPV. Acta Pathol Jpn 4 2 : 876-883, 1992. Key words : Cervical cancers, Human papillomavirus, Polymerase c ha in reaction, Clinicopat hologicaI study

DNA tumor viruses are thought to contribute to the development of 20% of the world's cancers(1). In particular, HPVs (human papillomaviruses) are implicated as major etiological agents of cervical cancers (2). A large bulk of HPV studies on cervical carcinogenesis has revealed that various HPV DNAs are present not only in almost all invasive cancers of the uterine cervix but also in its various precancerous lesions (3, 4). Up to the present, many HPVs of more than 66 subtypes have Received August 3, 1992. Accepted for publication October 14, 1992. Second Department of Pathology, Gunma University School of Medicine, Gunma. Mailing address: Takashi Nakajima, Second Department of Pathology, Gunma University school of Medicine, 3-39-22, Showa-machi, Maebashi, Gunma 371, Japan.

been isolated from various lesions of the uterine cervix, larynx and skin, of which about 20 subtypes are found in various cervical lesions including cancers (5). Many investigators have reported that the prevalence of HPV in cervical cancers varies according to the differences in geographic distribution, detect ion met hods and materials used (6-14). Southern blot analysis detected various types of HPV in 15% to 92% of patients with cervical cancer outside Japan (8-10) and from 39% to 60% among Japanese patients(11-13). The recently developed polymerase chain reaction (PCR) technique is more rapid, sensitive and convenient than other methods including Southern blot hybridization (14). In a comparative study of the sensitivity of Southern blot hybridization and PCR, the rate of detection of HPV DNA in cervical cancers by the PCR method was about 20% higher than that obtained by Southern blot hybridization (13). Moreover, PCR can be applied not only for a small amount of DNA material but also degraded DNA extracted from formalin-fixed and paraffin-embedded material (15-18). This method is suitable for HPV detection in a retrospective study using pathological archival tumor materials. In this study using formalin-fixed and paraffin-embedded materials, the presence of HPV DNA in invasive cervical cancers was confirmed by three different PCR systems. Moreover, the HPV status in invasive cervical cancers was retrospectively evaluated in co m pa riso n with severa I clinicopat hological features.

MATERIALS AND METHODS Cancer patients and t u m o r specimens Tumor specimens of 9 3 invasive cervical cancer patients were obtained from the pathology files of the Gunma University and its affiliated hospitals from July 1 9 7 8 to December 1990. The tumor tissues were fixed

877

Acta Pathologica Japonica 42 (12): 1992

in formalin and embedded in paraffin for routine pathological diagnosis. They were obtained from total hysterectomies of 91 patients and only biopsied materials were available for the other two patients. Pathological and clinical information was obtained by reviewing the hematoxylin and eosin-stained slides prepared for routine surgical pathological examination and clinical records, respectively.

Tumor DNA extraction DNA was extracted from paraffin blocks according to Goelz' method with some modifications (19). Briefly, two or three 1 0 - p m thick sections were cut from a paraffin block and placed in Eppendorf tubes. The sections were deparaffinized with xylene and rehydrated in a series of decreasing ethanol concentrations to sterile distilled water. Then they were digested in 7 0 0 p l of digestion mixture (400 p g of proteinase K [Boehringer Mannheim-Yamanouchi, Tokyo, Japan] per ml, 10% SDS and x 2 0 SSC) for 48 h at 48°C. After tissue digestion, the DNA was extracted three times with phenol-chloroform and precipitated with cold ethanol. The extracted DNA was dissolved in 10 m M Tris-HCI (pH 7.4) containing 1 mM of EDTA and adjusted to a final DNA concentration of 100 pg/ml, then stored at -20% until use as a template DNA.

PCR for HPV detection Five pairs of oligonucleotide primers were used to amplify different open reading frames (ORFs) of HPV genomes by PCR ; consensus primers (designated C1 and C2) to amplify a part of L1 ORF of many HPV types (LLPCR)(20) and type-specific primers to amplify a part of E6 ORF (H1 and H2 for HPV 16 and H1 and H3 for HPV 18) (E6-PCR) (14) and to amplify a part of E7 ORF (P1 and P2 for HPV 16 and P1 and P3 for HPV 18) (E7-PCR). A pair of primers to amplify the 99 base

pairs of the c-K-ras 2 gene including codons 1 2 and 1 3 were used to check the validity of the template DNA (21). Each primer was prepared by DNA synthesizer (Applied Biosystems, Foster city, CA, USA). The sequence of each primer and the sizes of the PCR products are detailed in Table 1. Ten microliters of extracted DNA solution, i.e. 1 p g of DNA for each assay, were added to a PCR reaction mixture. Besides template DNA or HPV plasmid DNA as positive controls, the reaction mixture consisted of 2 0 0 p M of each dNTP, 50 mM of KCI, 3 m M of MgCI, 10 mM of Tris-HCI (pH 8.3), 0.01% gelatin, 100 pM of each oligonucleotide primer and 2.5 units of Taq polymerase (Perkin-Elmer/Cetus, Norwalk, CT, USA). The total volume of reaction mixture was loop1 in the reaction tube, which was sealed with loop1 of light mineral oil (Sigma Chemical Co., St. Louis, MO, USA). The PCR was carried out in a DNA Thermal Cycler (Perkin Elmer/ Cetus) for 3 1 cycles (2-min denaturation at 94"C, 3-min annealing at 55°C and 2-min extension at 72°C) using type specific primers, for 40 cycles (1.5-min denaturation at 94"C, 2-min annealing at 50°C and 2-min extension at 72°C) using consensus primers and for 40 cycles (1.3-min denaturation a t 9 4 C , 4-min annealing at 50°C and 1.7-min extension at 74°C) to amplify the c-K-ras gene. After the PCR, l o p 1 of the reaction mixture was electrophoresed on a 3% agarose gel and the presence or absence of a specific band of PCR products was revealed by staining with 0.5 pg of ethidium bromide per ml. Detection of HPV subtype The L1-PCR products were analyzed for HPV s u b genomic typing by testing them for the presence of restriction fragment length polymorphs (RFLP). About 2 0 p1 of PCR product was digested with various restriction enzymes (Rsa 1, Dde I and Hae 111) and analyzed by 3% agarose gel electrophoresis, following Yoshikawa's method (20).

Table 1. Detailes of Various Primers Used in this Study and the Size of their PCR Products HPVs to be detected

E6-PCR

E7-PCRa Ll-PCR

~-

a

HPV 16 HPV 18 HPV 16 HPV 18 various HPVs

Primers H1 and H2 H1 and H3 P1 and P2 P1 and P3 C 1 and C2

PCR product 110bp 110bp 92 bp 154 bp 230-253 bp

The sequence of primers for E7-PCR is described below. P1 (a common 5' primer): HPV 16; (562-586), HPV 1 8 ; (590-61 4) ATGCATGGAGCTACAGCTACATTGC. P2 (3' primer for HPV 16) (635) TACTCGTTAATTTACTGTCG (654). P 3 ( 3 primer for HPV 18) (724) TTAGTAGTTGTAAATGGTCG (743)

RESULTS The sensitivity of the PCR used in this study was Table 2.

Detectability of HPV DNA by E6-, E7- and L1-PCRs (total 93 cases)

L1 positive Both E6 and E7 positive E6 sositive only E7 positive only Negative by both Total

41 1 2 3 47 (50.5%)

L1 negative 7

0 0 39 46

8 78

1

HPV DNA Subtypes and Cervical Cancers (Kashiwabara and Nakajima)

2

3

4

5

6

7

8

9

1

0

- 1357 - 603 - 31 0 -

194

- 118 bp I

3

3

4

5

1

6 603

Figure 1. Agarose gel electrophoresis of PCR products and its restriction enzyme clevage pattern of HPV 16- and 18-positive cases. Lane 1 : E6-PCR product for HPV 16, lane 2 : E7-PCR product for HPV 16, lane 3 : Ll-PCR product, l a n e 4 and 5 : Rsa I and Dde I clevage pattern of Ll-PCR product of HPV 16, respectively. l a n e 6 : E6-PCR product for HPV 18, l a n e 7 : E7PCR product for HPV 18, l a n e 8 : L1-PCR product of HPV 18, lane9 and 10: Rsa I and Dde I clevage pattern of L1-PCR of HPV 18, respectively. DNA size marker: q5 X174-Hae / / I digest.

2

3

4

5

-

603 -

310

-

310-

194118-

194 118

-

bP

Figure 2. Agarose gel electrophoresis of HPV 33-positive case. Lane 1 : E6-PCR product for HPV 16, lane 2 : E7-PCR product for HPV 16, lane 3 : L1-PCR product, lane 4, 5 and 6 : Rsa I, Hae / I / and Dde I clevage pattern of L1-PCR product, respectively.

Figure 3. Agarose gel electrophoresis of HPV 52-positjve case. Lane 1 : no band of E6-PCR product for HPV 16, lane 2 : L1-PCR product, lane3.4, 5 : Rsa I, Dde I and Hae I11 clevage pattern of L l -PCR product, respectively.

determined by the serial dilution of HPV 16 and 18 plasmid DNA. With each set of primers, the PCR could detect l 0 0 f g of HPV plasmid DNA per sample, which corresponded to lo-' copies of HPV DNA per cell (data not shown). As a control, when the c-K-ras gene was amplified, the reaction product yielded a specific 99-bp band in all samples (data not shown). Therefore, all extracted tumor tissue DNAs were considered to be suitable for PCR as a template.

The results of various PCRs are shown in Table 2. The L1-PCR showed an HPV DNA band in 47 (50.5%) of the 93 cervical cancer cases. Also, the E6- and E7PCR detected a positive band in 49 (45 for HPV 16 and 4 for HPV 18 DNA) and 50 (44 for HPV 16 DNA and six for HPV 18 DNA) cases, respectively. HPV DNA was detected by both E6- and E7-PCR in 48 cases. On the other hand, L1-PCR failed to reveal HPV DNA in seven of the 48 cases.

Acta Pathologica Japonica 4 2 (12): 1 9 9 2

1

2

3

4

5

8 79

6

357 -

603 -

310194118-

.--

--

Figure 4. Agarose gel electrophoresis of unclassified HPVpositive case. Lane 1 : E6-PCR product for HPV 16, lane 2 : E7-PCR product for HPV 16, lane 3: L1-PCR product, lane4, 5 and 6 : RSJ I, Dde I and Hae / / I cleavage pattern of L1-PCR product, respectively. Figure 5. Microscopic appearances of cervical squamous cell carcinomas with HPV 3 3 (a) and 5 2 (b). HE. stain. Table 3. HPV Subtypes and Histological Classification of Cervical Cancers

Squarnous cell carcinoma large cell type keratinizing non-keratinizing small cell type v erruco us type Adnosquamous carcinoma Adenocarcinoma Small cell carcinoma Total

33 5 2 U*

N o’of Cases

16

18

(66)

35

0

1

(29) (35) ( 1)

15 20

0 0 0 0 1 3

1 1 1 0 2 1 0 0 0

t 1) ( 5)

0 0 0

4 (20) ( 2) 0 2 9 3 39(42%) 6(6%)

0 0 0 0 1

3

0 0 0 0 3

2

0 0 3 0 5

* U : unclassified HPV type.

The L1-PCR product was subjected to HPV subtyping by RFLP analysis. Of 4 7 cases in which HPV DNA was amplified by L1-PCR, 3 2 and six cases showed the typical RFLP pattern of HPV 16 and 18 DNA, respectively (Fig. I). These results with these 38 cases were completely identical to the results of E6- and/or E7PCR. However, in one case, which was positive for HPV 16 DNA by E6-PCR but negative by E7-PCR, the DNA was classified as HPV 33 DNA by RFLP analysis (Fig. 2). HPV 52 DNA, which was not detected by both E6- and E7-PCR, was detected in three cases by RFLP analysis of L1-PCR product (Fig. 3). In the other five cases, which were positive for HPV 16 DNA by both E6- and

E7-PCRs, the HPV subtype remained to be determined HPV subtype because the L1 regions were not digested by Rsa I or Dae I (Fig. 4). They were typed as unclassified HPV DNA in this study. Other HPV subtypes such as HPV 6, 11, 31 and 5 8 were not detected in this study. The subtyping of seven cases in which HPV 16 DNA was detected by both E6- and E7-PCR but not by L1PCR, depended on the results of type specific PCR. Overall, various HPV DNAs were detected in 5 4 (58.1%) of 93 cases of invasive cervical cancer and HPV 16 and 18 DNA were detected in 39 and six cases, respectively (Table 3). No co-amplification of either HPV 16 or 18 DNA was seen in any case studied. As shown in Table3, squamous cell carcinomas contained mainly HPV 16 DNA followed by HPV 5 2 and unclassified HPV DNA, but no HPV 18 DNA. A single HPV 33-positive squamous cell carcinoma was histologically subclassified to keratinizing large cell type (Fig. 5). HPV 18 DNA was present in various other histological types of cancers. In adenocarcinoma, HPV 16, 18 and unclassified HPV DNA were detected. Three of five unclassified HPV-positive cancers showed various histology of adenocarcinoma (Fig. 6). HPV 18 DNA was detected in both small cell carcinomas, which were immunohistochemically shown to have chromogranin granules in the cytoplasm (Fig. 7).

880

HPV DNA Subtypes and Cervical Cancers (Kashiwabara and Nakajima)

Figure 6. Microscopic appearances of three cervical adenocarcinomas with unclassified HPV type. Other two cancers with unclassified HPV type show the histology of squamous cell carcinoma, non keratinizing large cell type. H.E. stain.

Table 4. The Time of Onset of Clinical Cervical C a n cers in this Study HPV Subtypes 16 18 33 52 U" Undetected Total (75)

Premenopausal 18 (62.1%) 2 ( 6.9%) 0

0 1 ( 3.4%) 8 (27.6%) . ,-, 29 cases

Postmenopausal 17 (37.0%) 1 ( 2.2%) 1 ( 2.2%) 3 ( 6.5%) 4 ( 8.7%) 20 (43.5%) , ,-, 46 cases

* U: unclassified HPV type.

Figure 7. Microscopic appearance of a cervical small cell c a r cinoma with HPV 18. This tumor have some features similar to malignant carcinoid tumor (a: H.E. stain) and many tumor cells contain chromogranin in the cytoplasm (b: chromogranin immunohistochemistry counterstained with hematoxylin).

The mean age of all patients in this study was 52.0% 12.4 years (HPV-positive patients; 50.81-12.4; HPVnegative patients; 52.7k12.4). Among the HPV DNA-positive patients, the mean ages of those with HPV 16, 18, 33, 52 and unclassified HPV DNA were 49.0% 12.8, 44.5k13.0, 63, 60.0k2.6 and 56.2k6.3 years, respectively. These differences were not statistically significant. Patient information such as menarche and menopause, marital history, number of pregnancies and deliveries was available from 75 patients. The mean age of menarche in cervical cancer patients with HPV DNA is almost identical with that without HPV DNA. In Table 4, the time of onset of clinical cervical cancer is classified as premenopausal or postmenopausal, and is compared with the HPV status. Detection of HPV DNA was more frequent in cervical cancer patients with premenopausal onset than in those with postmenopausal onset. M o r e over, there was a tendency for HPV 16 and 18 DNA to

Acta Pathologica Japonica 42 (12) : 1992 Table 5. Interrelation between HPV DNA Subtypes and Tumor Recurrence in Cervical Cancer Patients (79 Cases).

HPV Subtypes Tumor Recurrence a

16 1 8 - 3 3 526 U* Undetected 4/35a)3/5 0 / 1 1 / 3 015 7/30

The numerator/denominator means cervical cancer with recurrenceltotal number of cases. U : unclassified HPV type.

be present in the younger group and for other HPV DNAs to occur in the older group (Fisher's exact probability test, p