VIROLOGY

187, 508-5 14 (1992)

Antibodies to HPV-16 E6 and E7 Proteins as Markers for HPV-I 6-Associated Invasive Cervical Cancer MARTIN MULLER,* RAPHAEL P. VISCIDl,t YEPING SUN,t ELOISA GUERRERO,t PETER M. HILL,t FARIDA SHAH,t F. XAVIER BOSCH,+ NUBIA MUtiOZ,+ LUTZ GISSMANN,* AND KEERTI V. SHAH+’ *Deutsches

Krebsforschungszentrum, Heidelberg, Germany; t The Johns Hopkins Medical Institutions, and #The International Agency for Research on Cancer, Lyon, France Received September

17, 1991; accepted

December

Baltimore,

Maryland

2 1205;

4, 1991

Transforming proteins E6 and E7 of human papillomaviruses (HPVs) are consistently expressed in HPV-associated cervical cancers. In ELBA with four HPV-16 E6-E7 peptides, patients with HPV-16-associated invasive cervical cancer (group 1) had a greater seroreactivity than all other groups, which included patients with HPV-16-associated cervical intraepithelial neoplasia, invasive cervical cancer patients without HPVs, and unaffected controls. A larger proportion of group 1 sera, as compared to sera of all other groups, was reactive with at least one peptide (49% vs 17-27%), and with two or more peptides (22% vs O-6%). A clear difference between group 1 and all other groups was also found for high ELISA absorbance values to at least one peptide (22% vs O-8%). This high seroreactivity of group 1 sera was confirmed by a radioimmunoprecipitation assay with in vitro transcribed and translated HPV-16 E7 protein. Sera from 50% of group 1 but only 3% of controls were reactive in this test. Antibodies to HPV-16 E6 and E7 proteins appear to be virus-specific and disease state-specific markers of HPV-associated cervical cancer. o issz Academic PRESS, IN.

INTRODUCTION

study of cervical cancer for reactivity with HPV-16 E6 and E7 peptides, and with in vitro translated full-length HPV-16 E7 protein, and report here that serum reactivity to epitopes on E6 and E7 proteins is a marker of HPV-16-associated invasive cervical cancer but not of HPV-16-associated CIN lesions or of invasive cervical cancer not associated with HPV-16. Furthermore, seroreactivity with the translated E7 protein was a more discriminating marker for invasive cancer than that with synthetic E7 peptides.

Human papillomavirus (HPV) genomic sequences are recovered from a large majority of preinvasive and invasive cervical cancers, and HPV-16 has been recognized to be the predominant HPV type in these tumors in studies all over the world. HPV-I 6 genome is present in about 50% of cervical cancers and is often integrated into the cellular DNA (DOrst et al., 1985; Fuchs et al., 1989; Cullen et al., 1991). E6 and E7 proteins, the transforming proteins of oncogenic HPVs, are consistently expressed in HPV-associated cervical cancer (Schwarz et al., 1985; Smotkin and Wettstein, 1986; Androphy et al., 1987; Baker et al., 1987; Banks et al., 1987). In previous studies, we reported that serum antibodies to HPV-16 E7 fusion protein were detected in 20.5% of invasive cervical cancer cases but in only 1.4-3.8% of control subjects (Jochmus-Kudielka et al,, 1989). Other investigators have also noted the higher serum reactivity of cervical cancer cases to HPV-16 E7 peptides and protein (Mann et al., 1990; Krchnak eta/., 1990; Dillner, 1990). In many of the above studies the interpretation of serologic data was difficult because the serum donors were not fully characterized virologitally or epidemiologically. We have now tested sera from a subset of the participants of a case-control ’ To whom correspondence dressed. 0042-6822/92

and reprint requests

$3.00

Copyright 0 1992 by Academic Press. Inc. All rights of reproduction I” any form reserved.

MATERIALS AND METHODS Serum donors The serum donors were among the subjects of a study of cervical cancer in Spain and Colombia in which incident cases of invasive cervical carcinoma and of cervical intraepithelial neoplasia (CIN) grades l-3 were compared for behavioral and virological characteristics with their respective controls (MuAoz et al., 1990). The disease status was confirmed by a panel of pathologists. Exfoliated cervical cells were tested for HPV by ViraPap, Southern hybridization and polymerase chain reaction techniques for the invasive cases and controls and by ViraPap, and Southern hybridization for CIN cases and controls (Guerrero et al., 1990). HPV-positive specimens were diagnosed specifically for HPV-6, 1 1, 16, 18, 31, 33, and 35, or as containing an unknown type.

should be ad-

508

ANTIBODIES

Enzyme immunoassay (ELISA)

with synthetic

509

TO HPV-16 EARLY PROTEINS

peptides

Four synthetic peptides representing two epitopes on E6 (E6 aal -23; Met-His-Gln-Lys-Arg-Thr-Ala-MetPhe-Gin-Asp-Pro-Gln-Glu-Arg-Pro-Arg-Lys-Leu-Pro -Gin-Leu-Cys and E6 aa8-37;Met-Phe-Gln-Asp-Pro -Gin-Glu-Arg-Pro-Arg-Lys-Leu-Pro-Gln-Leu-Cys-Thr -Glu-Leu-Gln-Thr-Thr-Ile-His-Asp-lle-lle-Leu-Glu-Cys) and two on E7(E7aa6-35;NH,-Pro-Thr-Leu-His-GluTyr-Met-Leu-Asp-Leu-Gln-Pro-Glu-Thr-Thr-Asp-Leu -Tyr-Cys-Tyr-Glu-Gln-Leu-Asn-Asp-Ser-Ser-Glu-GluGlu-OH and E7 aa29-52; NH,-Asn-Asp-Ser-Ser-GluGiu-Glu-Asp-Glu-lle-Asp-Gly-Pro-Ala-Gly-Gln-Ala-GluPro-Asp-Arg-Ala-His-Tyr-OH), identified in a previous investigation (MOller et al., 1990), were prepared for use in ELISA with human sera. Automated peptide synthesis was performed with a Milligen/Biosearch model 9050 peptide synthesizer according to protocols specified by the manufacturer. The purity of the peptide was verified by analytical reverse phase HPLC using a Vydac C,, column. Wells of microtiter plates (Immunol II, Dynatech Laboratories, Arlington, VA) were coated with 10 fig/ml of E7 aa6-38 or of 30 pg/ml of E7 aa2952 in phosphate-buffered saline, pH 7.2, or with 25 pg/ml of E6 aal -23 or 10 pg/ml of E6 aa8-37 in 0.06 M carbonate buffer, pH 9.6. Serum specimens were tested in duplicate at a 1:25 dilution. The assay was completed with a goat anti-human IgG conjugated to horseradish peroxidase (Zymed Laboratories, San Francisco, CA) and ABTS substrate solution. For each serum, the mean reactivity of buffer wells was subtracted from the mean reactivity of wells coated with peptide to calculate a net absorbance value. Negative net absorbance values were scored as 0. Individual sera were scored as antibody-positive or antibodynegative for each peptide, using a cut-off absorbance value which was based on the distribution of absorbance values of the control sera, excluding the outliers. The means and standard deviations of absorbance values of control sera were calculated (separately for groups 4 and 6) and sera with values greater than a mean + 3 SD were excluded. The means and standard deviations were then recalculated and additional sera excluded, if necessary, by the same criteria. This process was repeated until none of the remaining sera were excluded, and the final mean + three standard deviations was taken as the cut-off value. In vitro transcription-translation and radioimmunoprecipitation assay A pGEM-1 plasmid containing the E7 ORF of HPV-16 (Dyson et al., 1989) (gift of Dr. Peter Howley) was lin-

earized with Pstl. One microgram of plasmid DNA was transcribed in vitro with 10 units of T7 RNA polymerase in the presence of cap analog (5’7meGpppG) according to the manufacturer’s instructions (Promega, Madison, WI). One microgram of RNA was used to program a micrococcal-nuclease-treated rabbit reticulocyte lysate (Promega) in the presence of [35S]methionine. Immune precipitations were performed by incubating 10 ~1 of lysate and 10 ~1 of serum in a 200 &volume containing 10 mMTris-HCI, pH 8, 140 mrJl NaCI, 0.025% NaN,, 0.1% Nonidet P-40 (modified RIPA buffer) for 16 hr at 4”. Then 80 ~1 of washed Protein A Sepharose (Pharmacia LKB, Piscataway, NY) suspended in modified RIPA bufferwas added and the mixture was rocked for 2 hr at 4”. After washing twice with 300 ~1 of modified RIPA buffer, once with modified RIPA buffer without detergent, and once with 10 mMTris-HCI, pH 6.8, the beads were resuspended in 30 yl of SDS gel sample buffer and heated to 80” for 15 min. The immunoprecipitated proteins were resolved on a 15% SDSpolyacrylamide gel and detected by fluorography with Amplify’“. Statistical

methods

Distributions were compared by Mann-Whitney U test and frequencies by x2 test, using Fisher’s exact probability where necessary. RESULTS Serum samples were obtained from 449 subjects of a case-control study of cervical cancer in Spain and Columbia. For the present study, subjects were chosen from the Colombia-Spain study on the basis of disease status, virologic diagnoses, and serum availability, as follows: group 1 (n - loo), invasive cases with HPV-16 (Inv-HPV-16); group 2 (n = 15), invasive cases with other HPVs (Inv-other HPVs) [five with HPV18, four with HPV-33, and two each with HPV-31, 35, and undetermined type]; group 3 (n = 49), CIN cases (mostly CIN-3) with HPV-16 (CIN-HPV-16). Group 4 controls (n = 177) (Inv-Control) matched the age distribution and country of residence of groups 1, 2, and 3 and group 6 controls (n = 49) (GIN-Control) were individually matched for CIN cases in group 5. The 449 sera were tested in enzyme immunoassays for reactivity to four HPV-16 E6 and E7 peptides. When the distributions of absorbance values of cases (groups 1, 2, 3, and 5) were compared with those of their respective controls (groups 4 and 6) significant differences in absorbance values were found for three of the four E6 and E7 peptides in the comparisons of group 1 cases with group 4 controls. The difference

MijLLER

510

0

0

I

i 1 IllV HPV 16 (N=lOO)

2 Inv Other HPV (N=15)

3 IIW No HPV (N=62) Study

4 Inv Control (N=177)

5 CIN HPV 16 (N=49)

6 CIN Control (N=49)

Groups

FIG. 1. Distribution of absorbance values of sera to peptides E7 aa6-35 and E7 aa29-52. The summary statistics of each distribution are displayed in the box plot. The length of the box corresponds to the interquartile range, with the upper boundary of the box representing the 75th, and the lower boundary the 25th percentile. The horizontal solid line in the box represents the median value. The 90th percentile is shown by the small bar at the end of the line extending upward from the box plot. Each outlier absorbance value is shown individually by an open circle. In addition to the median value in the box plot, the mean absorbance value is shown with a broken line which may lie inside or outside the box.

was most marked for peptide E7 aa6-35 (Fig. 1). The median absorbance value (and 25-75% interquartile range) of peptide E7 aa6-35 with group 1 sera was 0.089 (0.482) as compared to the corresponding value of 0.009 (0.069) for group 4 sera; this difference was very highly significant (P < 0.00001). The distribution of absorbance values to peptide E7 aa6-35 in group 1 was also significantly different from that in group 2 (P < 0.05) group 3 (P < 0.05) group 5 (P < O.OOl), and group 6 (P < 0.0003). Less pronounced but highly significant differences between group 1 and

ET AL.

group 4 sera were also seen for reactivity to peptide E7 aa29-52 (P < 0.00001) (Fig. 1) and to peptide E6 aa837 (P < 0.001) (data not shown). The distribution of absorbance values to peptide E6 aalin group 1 sera was not significantly different from that in group 4 sera (P = 0.14). Cases of groups 2 (Inv-other HPVs), 3 (Inv-no HPVs), and 5 (CIN-16) had reactivities to E6 and E7 peptides which were not significantly different from those of the corresponding controls. The differences in the reactivities of the two control groups 4 and 6 were also not statistically significant. The percentages of sera in the six groups with antibodies to individual peptides and to selected combinations of peptide are shown in Table 1. In case-control comparisons, significant differences were found in antibody prevalences to each of the four E6 and E7 peptides and for several combinations of E6-E7 peptides (Table 1). High antibody prevalences to the peptides were found exclusively in group 1 (Inv-HPV-16). The percentage of sera with antibodies to individual peptides in group 1 ranged from a high of 37% for peptide E7 aa6-35 to a low of 10% for peptide E6 aa8-37; the corresponding percentages in group 4 were 9 and 1% (P values of