Int. J . Cancer: 48, 682-688 (1991) 0 1991 Wiley-Liss, Inc.

w, -6:

Pubhcation of the International Union Against Cancer Pub1,cation de I'Unton Internationale Contre ie Cancer

OCCURRENCE OF ANTIBODIES TO L1, L2, E4 AND E7 GENE PRODUCTS OF HUMAN PAPILLOMAVIRUS TYPES 6b, 16 AND 18 AMONG CERVICAL CANCER PATIENTS AND CONTROLS Heinrich G. K&HEL',~,Masyar MONAZAHIAN', Kai SIEVERT',Michaela HOHNE',Christoph THOMSSEN~, Alexander TEICHMANN3, Peter h E N D T 3 and Reiner THOMSSEN University Center of Hygiene and Human Genetics, Dept. of Medical Microbiology, Kreuzbergring 57,0-3400 Gottingen; 2Frauenklinikder Technischen Universitat, Ismaninger Str. 22, 0-8000 Munich; and 3University Dept. of Gynecology and Obstetrics, Robert-Koch-Str, 40, 3400 Gottingen, Germany. Sera from I18 women of 33 t o over 90 years of age, with or without a history of cervical squamous-cell carcinoma, were examined for the presence of antibodies t o HPV-6b, HPV-I6 and HPV-18 LI, U,E4, and E7 gene products by the use of bacterially derived P-Gal fusion proteins and Western-blot analysis. Among the cervical cancer patients, 29/46 (63.0%) were positive for antibodies t o E4 and/or E7 of HPV- I 6 and/or E7 of HPV-18. In contrast, only 2 of 31 (6.5%) non-genital cancer patients and 4 of 41 (9.8%) healthy individuals were antibody-positive for HPV- I 6 E4 or E7. while antibodies t o the homologous proteins of HPV-18 could not be detected. Prevalence rates of antibodies t o the HPV- 16118 late proteins were 25/46 (54.3%) in the cervical carcinoma group, 1313 I (4 I.9%) among women with non-genital cancer types, and 18/41 (43.9%) among normal, healthy individuals. Antibodies t o HPVdb late gene products ranged between 6.5% and 12.2% in the different patient groups. Antibodies t o HPV-6b E4 and E7 were detected only once. By studying an additional control group of 207 women with a different age distribution, agedependence of antibodies to HPV gene products could be ruled out. Whereas antibodies t o late proteins may indicate that, regardless of clinical stage, HPV infections are widespread among the female population, the striking difference between the prevalence rates of antibodies to early proteins of HPV-I6 and HPV-I8 among cervical cancer patients and controls (p < 0.001) supports the idea of the involvement of these virus types in carcinogenesis of the cervix.

Epidemiologic studies have suggested that a sexually transmitted agent is involved in the development of cervical cancer. There is evidence that certain human papillomavirus types (HPV) may play an important role in the development of this disease. Nucleic acid hybridization has shown that HPV-6b and HPV-11 are the most prevalent HPV types in benign genital lesions (de Villiers et al., 1981; Gissmann et al., 1983); whereas HPV-16 and HPV-18 are the predominant types in severe cervical lesions and cervical carcinomas, as well as in metastasis and cell lines derived from cervical cancers (Boshardt et al., 1984; Crum et al., 1984; Durst et al., 1983; Lancaster et al., 1986; Schwarz et al., 1985). In addition, differences between cancerous and normal tissue concerning the physical state of HPV-16/18 DNA have also been revealed. Whereas in most invasive tumors the viral DNA is integrated within the host genome, in benign cervical lesions or normal cells the HPV genomes exist in episomal form (Durst et al., 1985). The fact that the HPV-DNA in cancer cells may be more than a silent passenger has been demonstrated by its biological activity. Viral transcripts and gene products from the early region such as E6 and E7 are necessary for cellular transformation (Schwarz et al., 1985; Shirasawa et al., 1987; Crook et al., 1989; Munger et al., 1989). Despite the knowledge accumulating about the events necessary for cellular transformation there is only limited information about the immune response directed against human papillomavirus infections. This is mainly due to the fact that

detecting antibodies against papillomaviruses has been difficult and not well standardized. The main obstacle in this context has been the lack of purified viral antigen. This drawback may be overcome by recombinant DNA technology (Jenison et al., 1988; Jochmus-Kudielka et al., 1989). An understanding of the immune response to human papillomaviruses may be of importance for the possible prevention of cervical carcinoma by vaccination, provided that a viral infection turns out to be essential for the development of this disease. Furthermore, antibodies to HPV gene products may be useful for diagnostic purposes, especially if differences between the antibody pattern of cervical cancer patients and that of normal individuals may represent prognostic parameters for the progressive or regressive behavior of cervical lesions. The goal of the present study was therefore to examine whether differences exist between the humoral immune responses against human genital papillomaviruses in cervical cancer patients and in controls with non-genital carcinoma types, and individuals who have never suffered from any type of cancer. For this purpose, HPV gene products were expressed in Escherichia coli as @-Gal fusion proteins and employed in western-blot analysis to investigate the pattern of antibodies to the HPV-16 and HPV-18 late gene products L1 and L2, and the early proteins E4 and E7. These proteins have been identified in the case of HPV-16 as the main targets recognized by the human humoral immune system (Kochel et al., 1991). Sera were also tested against the homologous gene products of HPV-6b in order to exclude the possibility that the antibodies detected to HPV-16/ 18 gene. products resulted from crossreactions with gene products of the ''harmless'' HPV type. This report will show that antibodies to E4 andlor E7 of HPV- 16/18 correlate strongly with cervical carcinoma, whereas antibodies to late gene products indicate similar high prevalence rates of HPV infections among female individuals independent of clinical diagnosis. PATIENTS AND METHODS

Human sera One set of sera was collected in 1987 from 95 women of age 60 years and older (mean age 70.2 years) being treated at the outpatient clinic of the University Women's Hospital of Gottingen, for their routine Pap-smear. This age group was selected because it contained a substantial number of women with a history of carcinoma. Twenty-three women had a history of cervical cancer. From 4 of the patients, the sera were drawn at the time when the cancer was discovered. In 9 women

4To whom correspondence and reprint requests should be sent. Received: February 22, 1991 and in revised form April 5 , 1991

683

ANTIBODIES TO HPV-6b. -16 AND -18 L1, L2, E4 AND E7

the cancer was detected and treated in 1986, i.e. 1 year before the sera were taken. In 4 women the cancer was diagnosed between 1980 and 1984, and in 6 women cancer detection and treatment dated back to more than 10 years before serum sampling. Thirty-one women had non-genital cancers. Eleven of them had cancer of the uterine corpus, 11 had breast cancer, 8 ovarian cancer, and 1 gastric cancer. Most of these cancers were detected in 1986 or 1987. Two cases dated back to 1977. The remaining 41 women had never developed cancerous diseases. To ensure that the antibody pattern observed among the control groups did not depend on the high age of the patients, further sera collected from 207 cancer-free women 20 to 59 years of age (mean age 43.1 years) at the above-mentioned institution were included in the study. In order to exclude the possibility that the HPV-16/18 antibody pattern among the cervical cancer patients from Gottingen may reflect a regional peculiarity, we also examined sera from 23 cervical cancer patients, obtained from the Women’s Hospital of the Technical University in Munich, Germany. These sera were taken during the years 1986 to 1989 at the time when the cancer was discovered and treated. The ages of these patients ranged from 33 to 83 years (mean age 57.6 years). All sera were stored frozen at - 20°C until they were processed. HPV DNA and expression plasmids HPV-6b, HPV-16 and HPV-18 genomes cloned in pBR322 were kindly provided by Dr. H. zur Hausen, German Cancer Research Center (Heidelberg). We subcloned parts of the viral genomes in pUC vectors ourselves, to facilitate the isolation of the individual open reading frames. The expression plasmid pSS20 (Scholtissek and Grosse, 1988) was obtained from Dr. F. Grosse, MPI for Experimental Medicine (Gottingen). Cloning of HPV DNA fragments Generally, it was attempted to insert restriction fragments of the HPV-6b, HPV-16 and HPV-18 DNA into the expression plasmids covering as much as possible of the nucleotide sequence of each individual open reading frame to be expressed. The expression plasmids used in this study contained the following HPV genome fragments (the numbering for HPV-6b, HPV-16 and HPV-18 derives from Schwarz et al. (1983), Seedorf et al. (1985), and Cole and Danos (1987), respectively): HPV-66: L 1 (Sau96VSau961, nucleotides 5695-7203), L2 (BamHUPstI, nucleotides 4722-5408), E4 (PflMVPflMI, nucleotides 3264-3793) E7 (NsiVNsiI, nucleotides 530-1640); HPV-16: L1 (DdeYSphI, nucleotides 5692-7463), L2 (AccIi BamHI nucleotides 4138-6150), E4 (SspURsaI, nucleotides 3332-3616) and E7 (MboIiNcoI, nucleotides 622-863); HPV18: L1 (AccI/AccI, nucleotides 5529-7766), L2 (OxaNI/ OxaNI, nucleotides 4340-5799), E4 (RsaURsaI, nucleotides 3433-3725) and E7 (SspI/SspI, nucleotides 580-1756). The fragments were ligated to the suitably digested expression vectors. In some cases the termini of the vector and HPV DNA fragments had to be modified by nuclease digestion or by filling in with Klenow fragment polymerase from E . coli. Recombinant plasmids were introduced into E . coli JM109 according to Hanahan (1983). The correct insertion of the HPV genes was confirmed by restriction digestion and by nucleotide sequence analysis. Protein expression and Western blot analysis Bacteria containing the pSS20 expression plasmids were grown at 37°C until the mid-log phase of bacterial growth, and plasmid-dependent protein expression was induced by addition of 2 m~ isopropyl-6-D-galactopyranoside(IPTG). Incubation

was then continued for another 2 hr. Bacteria were collected by centrifugation at 3000g. Lysis of bacteria and isolation of fusion proteins were carried out according to Jenison et al. (1988). For Western blot analysis, approximately 1 pg of each of the fusion proteins was applied to SDS-PAGE (Mighty Small, Hoefer, San Francisco, CA). Proteins were then transferred to Immobilon P (Millipore, Munich) or nitrocellulose membranes (BA 85, Schleicher and Schull, Dassel) by means of a semi-dry blot-apparatus. Sera and filters were pre-treated as described by Jenison et al. (1988). Prior to incubation, sera were finally diluted to 1: 100 in milk buffer and incubated with the filters at 4°C overnight. After washing the filters, bound human antibodies were visualized with 1:2000-diluted alkaline-phosphatase-labelled anti-human IgG, IgA and IgM (Dakopatts, Hamburg) from rabbits. Blots were evaluated by visual inspection and by comparison with the reactivities of control sera against the respective fusion proteins and against P-Gal. Sera were regarded as positive when signals could be clearly distinguished from background (examples for weak reactivities are given in Fig. la, panels 2 and 5). To evaluate the significance of the differences between the results from the patient groups examined, the odds ratio (OR) and Chi-square values were calculated. RESULTS

Antibody reactions to HPV proteins and controls of specijicity Figure la shows reactivities of different sera against HPV16 E4 or E7. The related Coomassie-stained fusion proteins separated on a 7% SDS polyacrylamide gel are demonstrated in panel 1. Panel 2 shows a strong reactivity and panel 3 a weak reactivity of antibodies to E4. Panels 4, 5, and 6, on the other hand, depict strong, medium and weak activities, respectively, against E7. These blots illustrate that the immune reactions observed are directed against specific fusion proteins and not against @-Gal. If these reactions were directed against P-Gal or the collagenlinker, then positive signals should be visible in all lanes. Nevertheless, antibody reactions to HPV gene products had to fulfill additional criteria before they were finally considered positive. They had (1) to react also against the same gene products expressed by means of the so-called pEX vectors. Furthermore, signals had to disappear if positive sera were (2) preabsorbed with lysates from bacteria expressing the respective HPV protein. Another important aid in ensuring specificity of the reaction was (3) offered by the expression vector pSS20. The fusion proteins expressed by this plasmid contain the recognition sequence for collagenase from human pro-a-collagen separating P-Gal and the viral moiety. After collagenase digestion of the respective fusion proteins, positive signals had to be shifted to a position referring to the molecular weight of the viral component as calculated from the nucleotide sequence of its corresponding gene. Exceptions to this rule, however, were L2 and E7, which were apparently larger than expected. Examples of the specificity test by collagenase digestion are given in Figure 1b. The left panel shows reactivities against E4 prior to (U) or after collagenase digestion ( C ) . While the blot carrying number 214 displays the reactivity against HPV-6b E4 of patient 214, number 2 and 3 correspond to the same numbers as in Figure la. The positions of the fusion protein (E4/P) and the collagenase digestion product (E4c) are indicated. The right panel of Figure l b represents the signals of the anti-E7-positive sera from Figure la before and after collagenase digestion. Blot U shows the reactivity of the same serum

684

FIGUREI

KOCHEL ET AL.

-

(a) Immune reactions against P-Gal fusion proteins of HPV-16 E4 (panels 2 and 3, pats. M61 and M45,resp.) and E7 (panels

4,5, and 6 , pats. 107, 379 and M39, resp.). A corresponding Coomassie-stained gel is depicted in panel 1. (b) Immune reactions of antibodies

to HPV-16 E4 (left panel) and HPV-16 E7 (right panel) against P-Gal fusion proteins prior to and after collagenase digestion (U, undigested; C, collagenase-digested; E4c and E7c, collagenase-released viral moieties). ( c ) Reactivities of human serum containing antibodies directed against HPV-18 E7 (a, pat. M52), and of serum reactive against both HPV-16 E4 and E7 (b, pat. 381). Positions of molecular weight markers, fusion proteins (F, E4/P, E7/P) and unfused P-Gal (P) are marked.

as in Figure la (panel 4) against the undigested E7/P-Gal fusion protein (E7/P), while the following blots exhibit the reactivities of the same sera used for blots 4, 5 and 6 against collagenase-released E7 (C, E7c). Further examples of immune reactions against HPV gene products are shown in Figure lc. Panel (a) represents an immune reactivity against HPV-18 E7 while in panel (b) reactivities against HPV-16 E4 and E7 are to be seen. Figure 2, on the other hand, shows 6 juxtaposed examples of reactions of antibodies against the @-Gal fusion protein of HPV-16 L2 and the collagenase-released viral gene product. The Coomassie-stained proteins are shown in panel 1. The positions of the fusion protein (L2/p), P-Gal and the collagenase-released viral moiety (L2c) are indicated on the left of the Figure. The positions of molecular weight reference markers are shown on the right. After collagenase digestion, a signal appeared at about 70 kDa which was also clearly visible at weaker reactivities (panel 6) or at high background (panel 7).

This deviation of L2 from its expected molecular weight of about 50 kDa has already been described and appears to be an intrinsic property of this protein (Komly et al., 1986; Kochel et al., 1991). Occurrence of HPV antibodies in patients with non-genital carcinoma types and in cancer-free individuals Table I shows the results from the examination of women with non-genital cancers (A) and of cancer-free individuals (B). Positive reactions are indicated by “plus” signs. Both groups exhibited a rather similar pattern of anti-HPV reactivity which, in the case of HPV-16/18, was mainly directed against the minor nucleocapsid protein L2. In group A, 13/31 (41.9%) women were positive for anti-12, antibodies to HPV-18 L1 being detected in one of them. In group B , 18/41 (43.9%) of the sera showed antiL2 without the appearance of antill. Antibodies to E4 and/or E7 of HPV-16 were identified in 2/31 (6.5%) sera from group A and in 4/41 (9.8%) from group

685

ANTIBODIES T O HPV-6b, -16 A N D -18 L1, L2, E4 A N D E7

FIGURE2 - Examples of immune reactions against P-Gal fusion proteins of the HPV-16 L2 gene product blotted onto nitrocellulose membranes (1, Coomassie-blue-stainedproteins; 2-7, reactions of antibodies to L2 present in different patient sera, i.e. from the left to the right pats. 282, 71, M52, M45, 349, 386; U, undigested; C, collagenase-digested). Positions of the fusion protein (L2/p), of free P-Gal, and collagenase-released L2 (L2c) are indicated on the left of the figure. Molecular weights of marker proteins are depicted on the right.

B. Antibodies to early proteins of HPV-18 were not detected. One woman belonging to group A showed, besides antibodies to HPV-16 E7, also those directed against HPV-6b E4 and E7, which may have resulted from a double infection with both HPV-6b and HPV-16. Antibodies to late proteins of HPV-6b were present in 3/31 (9.7%) of the sera from group A and in 5/41 (12.2%) of those from group B. Obviously, no difference in the immune response to HPV gene products existed between these 2 control groups.

TABLE I - PAlTERN OF ANTI-HPV POSITIVITY AMONG PATIENTS WITH OTHER CANCER TYPES (N = 31 A) AND HEALTHY SUBJECTS (N = 41,'B)'.*.' Patient number

HPV-6b E4

E7

L1

HPV-16 L2

E4

E7

L1

HPV-18 L2

E4

E7

L1

L2

Antibodies to HPV gene products in cervical cancer patients As depicted in Table 11, among women with a history of cervical carcinoma (A = Gottingen, B = Munich) the pattern

of anti-HPV reactivity appeared to be more complex than that of the groups described above. While the prevalence rates of anti-L-positive women did not deviate from those of the abovementioned groups to any appreciable extent (A = 12/23, 52.2%; B = 13/23, 56.5%), the frequency of women positive for antibodies to E4 andlor E7 of HPV-16/18 was substantially increased (A = 14/23, 60.9%; B = 15/23, 65.2%). The different antibody types occurred either singly or in various combinations. Although both groups differed in their age-distribution and in their regional origin, the similarity in the immune response to the different HPV proteins was striking. However, a minor difference could be observed regarding the immune response to HPV-18 E7. Whereas in the Gottingen group only one woman exhibited antibodies to this protein, this number was increased to 6 in the group from Munich. This may be attributed to a regional fluctuation in the prevalence of the different genital HPV types. In none of the cervical carcinoma patients could antibodies to E4 and E7 of HPV-6b be detected. Thus, the possibility of immunological cross-reactions of antibodies to HPV-16/18 E4 and E7 against the homologous gene products of HPV-6b could be excluded. Antibodies to HPV-16/18 gene products in normal patients of other age groups

Because of the relatively small numbers of normal individuals in the above patient groups, and in order to ascertain that the antibody pattern observed was not influenced by the age of the patients, sera from 207 additional women belonging to different age groups were examined for antibodies to L2, E4 and E7 of HPV-16/18. Thirty-eight of the women were 20 to 30 years of age (mean age 24.5 years), 43 were 3 1 to 40 (mean age 35.2 years), 62 were 41 to 50 (mean age 45.1 years), and

~~

~

'Fifteen patients in group A were antibody-negative.-'Sixteen patients in group B were antib~dy-negative.-~Positivereactions are indicated by crosses.

64 were 51 to 59 years of age (mean age 55.0 years). The results of this examination are shown in Table 111. It follows from these data that obviously no difference exists in the prev-

686

KOCHEL ET A L .

TABLE I1 - PATTERN OF ANTI-HPV POSITIVITY AMONG CERVICAL CANCER PATIENTS FROM GOTTINGEN (N = 23. A) AND MUNICH (N = 23. B)’.’

alence rates of antibodies to early or late proteins between women older than 60 years and other age groups. A summary of all results is given in Table 111. A total of 325 sera was examined for the presence of antibodies to different HPV types and gene products. With regard to the antibodies against HPV-16 and HPV- 18, a striking difference between the occurrence of antibodies to E4 and E7 in cervical cancer patients and cancer-free subjects could be observed (OR = 33.5, p < 0.001). Furthermore, antibodies to early proteins were observed among patients with non-genital cancer types to the same extent as in cancer-free individuals. These data therefore indicate that antibodies to early proteins are correlated with cervical cancer, which may be a further indication of a connection between HPV- 16/18 infection and the development of cervical cancer. DISCUSSION

The aim of the present study was to determine whether differences in the humoral immune response against the most important genital HPV types (6b, 16, 18) exist between cervical carcinoma cases and controls suffering from other carcinoma types or who have never developed cancer. In all groups examined, antibodies directed against early and against late proteins could be identified. If only antibodies to late proteins had been investigated, then the HPV- 16/18 prevalence rates would apparently have been the same in cervical cancer cases and controls which, in turn,

would have obscured the connection between an HPV infection and cervical cancer. However, we could demonstrate that the frequencies of antibodies to the early proteins E4 and E7 of HPV- 16 and E7 of HPV- 18 among women with cervical cancer were significantly higher than in women with non-genital carcinoma types and normal, healthy individuals. Furthermore, antibody reactivities against E4 and E7 of HPV-6b were not detected in cervical cancer patients. Thus, the assumption of an etiologic connection between infection with HPV-I6 or - 18 and the occurrence of cancer of the uterine cervix is additionally supported by the occurrence of antibodies to early gene products in women suffering from this disease. Due to the lack of cancer biopsies and cervical smears, we were unfortunately not able to demonstrate a coincidence between HPV-16/18 DNA and the occurrence of antibodies to HPV- 16/18 gene products. From hybridization data, however, HPV-16 DNA can be expected in up to 50% of cervical cancers (Gissmann et al., 1984; Lorincz et al., 1986; Reid et al., 1987). The frequency with which antibodies to early gene products of HPV-16 among cervical cancer patients could be detected was 52.1%. This value agrees well with that found by nucleic acid hybridization. Moreover, with regard to the antibodies to late HPV-16 proteins, the frequency of antibodypositive individuals in this patient group increased to 71.7% (33/46). Therefore, in the case of cervical cancer, serology seems to be at least as sensitive as nucleic acid hybridization. The occurrence of antibodies to early gene products most probably reflects the gene activity of HPV genomes in cervical cancer cells. In such cells, E7 is the most abundant HPV-16 gene product (Smotkin and Wettstein, 1986). Its expression, together with that of E6, appears to be inevitable for cellular transformation and the maintenance of the malignant phenotype (Munger et al., 1989). However, in contrast to antibodies to HPV-16 E7, those directed against E6 were detected in only one of the cervical cancer patients (not shown in Tables). This low frequency of antibodies to E6 may be due to the low amount of this protein expressed in cancer cells, or to a low antigenicity. On the other hand, it cannot be excluded that the missing 30 aminoterminal amino-acids within the E6/P-Gal fusion protein employed by us may contain the antigenic determinant of E6. This would also lead to a low number of E6-positive individuals. Antibodies to E4 were also frequently detected among cervical cancer patients. Although this protein is a major protein constituent in HPV-I-induced skin warts (Doorbar et al., 1986), its function has still not been elucidated. Whether it contributes to malignant progression of HPV- 16-infected cells has never been demonstrated but may be worth examining because of the elevated frequency of anti-E4-positive individuals among cervical cancer patients. Jochmus-Kudielka et al. (1989) have also demonstrated an association of antibodies to HPV-16 E7, but not of antibodies to E4, with cervical cancer. The latter may be explained by the fact that these authors used a truncated HPV-16 E4 gene product, whereas the one used in our experiments was complete. These authors also reported that they have found anti-E4 in up to 30% of girls aged 1 to 10 years. When we examined sera from 46 1- to 10-year-old girls, none of them was positive for this antibody (Kochel et al., data not shown). Because, on the other hand, we detected higher frequencies of a n t i 4 3 in cervical cancer patients than Jochmus-Kudielka et al. (1989) (30.4% vs. 15.9%), no explanation for the above-mentioned discrepancy between their results and ours concerning anti-E4 positivity in young girls can be offered. Furthermore, the study of Jochmus-Kudielka et al. (1989) did not include antibodies to late proteins. If only the antibodies to early proteins are taken into consideration, the main impression gained from their study is that healthy individuals are infected with HPV-16 to a much lesser extent than cervical

ANTIBODIES TO HPV-6b, -16 AND -18 L1, L2, E4 AND E7

687

TABLE I11 - PREVALENCE RATES OF ANTIBODIES TO EARLY AND LATE PROTEINS OF HPV-16 AND HPV-18 IN CERVICAL CANCER PATIENTS AND CONTROLS Antibody positive (70)

Diagnosis

Anti-E pos. + / - anti-L

Anti-L pos. + / - anti-E

Anti-HPV- 16118 total

Cervical cancer Other cancers No cancer ( 2 6 0 years) No cancer (other age-groups) No cancer total

63.0% (29/46) 6.5% (2/31) 9.8% (4141) 3.9% (81207) 4.8% (121248)

54.3% (25/46) 41.9% (13/31) 43.9% (18/41) 41.5% (86/207) 41.9% (1041248)

80.4% (37/46) 45.1% (14/31) 51.2% (21/41) 42.5% (88/207) 43.9% (109/248)

+ / - indicates that a patient positive for a given antibody may be negative or also positive for the other antibody type. carcinoma patients. However, in the present study, it is shown that antibodies to late proteins were found in similar high frequencies in all groups examined irrespective of clinical diagnosis, suggesting an HPV-16/18 infection rate of more than 40% in the normal female population. It can therefore be deduced from these data that an HPV- 16/18 infection per se may not represent a sufficient single etiological component for cervical cancer development. The role of possible co-factors in cervical carcinogenesis has already been discussed by zur Hausen (1982). At first glance the high prevalence of HPV- 16/18 infections in normal individuals, as determined by serology, appears rather surprising. However, de Villiers et al. (1987) have already claimed that, bearing in mind the fluctuations in virus production occurring in HPV-infected cells, the actual HPV prevalence rate may range between 30% and 40% in adults with normal cervical cytology. This value would exceed the prevalence rates revealed by these authors by nucleic acid hybridization by a factor of 2 to 3. Our data, which were based on serology, therefore agree well with these proposed values. From the cervical cancer patients, 22/46 (47.8%) were found to be positive for antibodies to late proteins of HPV-16, 14 of these 22 (63.6%) being also positive for antibodies to HPV-16 early gene products. In comparison, 25/72 (34.7%) of the control patients aged 2 6 0 years were positive for antibodies to HPV-16 late proteins, but only 2 of them (8.0%) simultaneously exhibited antibodies to early proteins. Since antibodies to late proteins are possibly induced as a consequence of a primary HPV infection, it is feasible that antibodies to early

proteins, which are assumed to be involved in viral replication and cellular transformation, may concomitantly occur. Therefore, one would expect anti-E and anti-L in at least similar frequencies in HPV-infected normal individuals as well as in cancer patients. This is, as shown above, obviously not the case. This suggests that, in most instances, early proteins are only transiently expressed, in amounts not sufficient for antibody production. Thus, the generation of antibodies to these proteins may require a continued expression of early HPV gene products upon a primary HPV infection. The permanent expression of viral oncoproteins such as E7 and E6 may then lead to the malignant progression of HPV-induced precancerous lesions. Another pathway for the generation of antibodies to early proteins may be the reactivation of latent, persistent viruscarrier states. Such a reactivation is, for example, well known from nasopharyngeal carcinoma which is associated with a latent, persistent Epstein-Barr virus (EBV) infection. Since there are some similarities in the antibody pattern of EBV (Henle and Henle, 1979) and HPV, the reactivation of HPV genomes after long-lasting latent periods may also represent an initial event in the development of cervical cancer. If anti-E antibodies appeared at very early stages of malignant transformation or at recurrences after cancer treatment, then their detection may be of great prognostic and diagnostic value. This remains to be investigated. Future serological studies which will involve methods easier to handle than Westernblot analysis, may thereby help to elucidate the role of HPV infections in cervical carcinogenesis.

REFERENCES

BOSHARDT, M . , GISSMANN,L . , IKENBERG, H . , KLEINHEINZ, A., SCHEURLEN, W. and ZUR HAUSEN,H . , A new type of papillomavirus DNA, its presence in genital cancer biopsies and in cell lines derived from cervical carcinomas. EMBO J . , 3, 1151-1157 (1984). COLE,S.T. and DANOS,O . , Nucleotide sequence and comparative analysis of the human papillomavirus type 18 genome. Phylogeny of papillomaviruses and repeated structure of the E6 and E7 gene products. J . mol. Biol.. 193, 59%608 (1987). CROOK,T., MORGENSTERN, J.P., CRAWFORD, L. and BANKS,L . , Continued expression of HPV-16 E7 protein is required for maintenance of the transformed phenotype of cells co-transformed by HPV-16 plus EJ-rus. EMBO J . , 8, 513-519 (1989). CRUM,C.P., IKENBERG, H . , RICHART, R.M. and GISSMANN, L . , Human papillomavirus type 16 and early cervical neoplasia. N. Engl. J . Med., 310, 880-883 (1984). DE VILLIERS, E.M., GISSMANN, L. and ZUR HAUSEN,H . , Molecular cloning of viral DNA from genital warts. J. Virol., 40,932-935 (1981). DE VILLIERS, E.M., SCHNEIDER, A., MIKLAW, H., PAPENDICK, U., WAGNER, D . , WESCH,H . , WAHRENDORF, J . and ZUR HAUSEN,H., Human papillomavirus infections in women with and without abnormal cervical cytology. Lancer, 11, 703-706 (1987). DOORBAR,J . , CAMPBELL, D . , GRAND,R.J.A. and GALLIMORE, P.H., Identification of the human papilloma virus- la E4 gene products. EMBO J . , 5, 355-362 (1986). DORST,M., GISSMANN, H . , IKENBERG, H. and ZUR HAUSEN,H . , A pap-

illomavirus DNA from a cervical carcinoma and its prevalence in cancer biopsy samples from different geographic regions. Proc. nut. Acad. Sci. ( W ~ s h . )80, , 3812-3815 (1983). DURST,M., KLEINHEINZ, A , , HOTZ,M. and GISSMANN, L., The physical state of human papillomavirus type 16 DNA in benign and malignant genital tumours. J . gen. Virol., 66, 1515-1522 (1985). GISSMANN, L., BOSHARDT, M . , DORST,M . , IKENBERG, H., WAGNER,D. and ZUR HAUSEN,H . , Presence of human papillomavirus in genital tumours. J. invesr. Dermar., 83, 26s-28s (1984). GISSMANN, L., WOLNIK,L., IKENBERG, H . , KOLDOVSKY, U . , SCHNORCH, H.G. and ZUR HAUSEN,H . , Human papillomavirus types 6 and 1 1 DNA sequences in genital and laryngeal papillomas and in some cervical cancers. Proc. nar. Acad. Sci. (Wash.), 80, 560-563 (1983). HANAHAN,D . , Studies on transformation of Escherichia coli with plasmids. J. mol. Biol., 166, 557-580 (1983). HENLE,W. and HENLE,G., Seroepidemiology of the virus. In: M.A. Epstein and B.G. Achong (eds.), The Epstein-Burr virus, pp. 61-78, Springer, Berlin (1979). JENISON, S . A . , FIRZLAFF, J.M., LANGENBERG, A. and GALLOWAY, D.A., Identification of immunoreactive antigens of human papillomavirus virus 6b by using Escherichia coli-expressed fusion proteins. J . Virol., 62, 2115-2132 (1988). JOCHMUS-KUDIELKA, I . , SCHNEIDER, A , , BRAUN,R . , KIMMIG, R., KOLWVSKY, U., SCHNEWEIS, K.E., SEEDORF,K . and GISSMANN, L., Antibodies against the human papillomavirus type 16 early proteins in human

688

KOCHEL ET AL.

sera: correlation of anti-E7 reactivity with cervical cancer. J . nat. Cancer Znst., 81, 1698-1704 (1989). KBCHEL,H.G., SIEVERT,K., MONAZAHIAN, M . , MITTELSTADTA,, TEICHMANN, A. and THOMSSEN, R., Antibodies to huDETERDING, man papillomavirus type-16 in human sera as revealed by the use of prokaryotically expressed viral gene products. Virology, 182 (1991) (In press). KOMLY,c.A., BREITBURD, F., CROISSANT, 0 . and STREECK, R.E., The L2 open reading frame of human papillomavirus l a encodes a minor strucma1 protein c a w i n g type-specific antigens. J . Viral., 60, 813-816 (1986).

LANCASTER, w.D., cASTELLANO, p , , sANTOS,c , , D ~ G , , KUR~ MAN, R.J. and JENSON,A.B., Human papillomavirus deoxyribonucleic acid in cervical carcinoma from primary and metastatic sites. Amer. J . Obsret. Gynecol., 154, 115-1 19 (1986). LBRINCZ,A.T., TEMPLE,G.F., PATTERSON, J.A., JENSON,A.B., KURMAN,R.J. and LANCASTER, W.D., Correlation of cellular atypia and human papillomavirus deoxyribonucleic acid sequences in exfoliated cells of the uterine cervix. Obsret. Gynecol., 68, 508-512 (1986). MONGER,K., PHELPS,W.C., BUBB,V., HOWLEY,P.M. and SCHLEGEL, R., The E6 and E7 genes of the human papillomavirus type 16 together are necessary and sufficient for transformation of primary human keratinocytes. J . Virol., 63, 44174421 (1989). REID,R . , GREENBERG, M., JENSON,A.B., HUSAIN,M., WILLETT,J., C.R., SHERMAN, A.I., PHIBBS, DAOUD,Y., TEMPLE,G . , STANHOPE, G.D. and LORINCZ,A.T., Sexually transmitted papillomavirus infections. I. The anatomic distribution and pathologic grade of neoplastic lesions

associated with different viral types. Arner. J . Obstet. Gynecol., 156, 212-222 (1987). SCHOLTISSEK, S. and GROSSE,F., A plasmid vector system for the expression of a triprotein consisting of P-galactosidase, a collagenase recognition site and a foreign gene product. Gene, 62, 55-64 (1988).

SCHWARZ, E., DORST,M., DEMANKOWSKI, C., LATTERMANN, O., ZECH, R., WOLFSPERGER, E., SUHAI,S. and ZUR HAUSEN,H., DNA sequence and genome organization of genital human papillomavirus type 6b. EMBO J . , 2, 2341-2348 (1983). SCHWARZ,E., FREESE,u.K., GISSMANN, L., MAYER,w., ROGCENBUCK, A. and ZUR HAUSEN,H., Structure and transcription of B., STREMLAU, human ~ papillomavirus ~ ~ sequences , in cervical carcinoma cells. Narure (Land.), 3147 111-114 (1985). SEEWRF,K . , KRAMMER, M., DORST,M.,SuHA1, s, and R~JWEKAMP, W., Human papillomavirus type 16 DNA sequence. Virology, 145, 181lg5 SHIRASAWA, H., TOMITA,Y., SEKIYA,S., TAKAMIZAWA, S. and SIMIZU, B., Integration and transcription of HPV type 16 and 18 sequences in cell lines derived from cervical carcinomas. J . gen. Virol., 68, 583-591 (1987). SMOTKIN,D. and WETTSTEIN, F.O., Transcription of human papillomavirus type 16 early genes in a cervical cancer and a cervical cancer-derived cell line and identification of the E7 protein. Proc. nat. Acad. Sci. (Wash.), 83, 4 6 8 M 6 8 4 (1986). ZUR HAUSEN,H., Human genital cancer: synergism between two virus infections or synergism between a virus infection and initiating events. Lancer, 11, 1370-1372 (1982).