We have analyzed 114 serum specimens from women with cervical intraepithelial neoplasia (CIN) or carcinoma of the cervix uteri for the presence of serum ...
JOURNAL OF CLINICAL MICROBIOLOGY, Mar. 1990, p. 624-627
Vol. 28, No. 3
0095-1137/90/030624-04$02.00/0 Copyright © 1990, American Society for Microbiology
Serological Responses to Papillomavirus Group-Specific Antigens in Women with Neoplasia of the Cervix Uteri LENA DILLNER,' JORGE MORENO-LOPEZ,2 AND JOAKIM DILLNER1* Department of Virology, Karolinska Institute, SBL, S-105 21 Stockholm,' and Department of Veterinary
Microbiology (Virology), The Biomedical Centre, Uppsala,2 Sweden Received 2 November 1989/Accepted 1 December 1989
Certain types of human papillomaviruses have been linked to the development of carcinoma of the cervix uteri. We have analyzed 114 serum specimens from women with cervical intraepithelial neoplasia (CIN) or carcinoma of the cervix uteri for the presence of serum antibodies against purified, disrupted bovine papillomavirus (BPV). The titers of immunoglobulin A (IgA) antibodies against BPV were slightly elevated (P < 0.025) in the sera from CIN or cervical carcinoma patients compared with the titers of 139 serum specimens from sex- and age-matched healthy controls. In contrast, both the IgG and IgM serum antibody titers against BPV were significantly decreased for CIN and cervical carcinoma patients compared with those of healthy controls (P < 0.001 and P < 0.005, respectively). These results suggest that the difference between IgA and IgG or IgM antibodies to papillomavirus group-specific antigens may represent interesting serological parameters that could possibly be used in the epidemiologic study of women at risk for CIN.
There are more than 60 different types of human papillomaviruses (HPV) that are associated with several different HPV-associated diseases (13). Several types infect the genital tract (13). Certain types are linked to benign genital warts, condyloma acuminata, whereas others have been found in cervical intraepithelial neoplasia (CIN) and carcinoma of the cervix uteri (12, 18). HPV infection is usually diagnosed by DNA hybridization techniques (14). It is not possible to perform virus isolation from clinical samples, since the replication of HPV is critically dependent on the differentiation stage of the infected keratinocyte, and to date, it has not been possible to obtain papillomavirus (PV) replication in vitro (14). Type-specific serology for HPV can thus only be performed on HPV virions purified from human warts (15-17), a material which is not readily obtainable in sufficient amounts. Detection of HPV infection by typespecific serology is therefore not routinely used today, although type-specific assays with synthetic peptides (4) or fusion proteins expressed in bacteria (7, 9, 11) have been reported recently. The PVs also contain epitopes shared by many different types of PV: the so-called group-specific antigenic sites. Antisera reacting with both HPV and bovine PV (BPV) can be readily prepared (8). With antisera against disrupted BPV virions, HPV capsid antigen can be detected in tissue from patients with CIN and in condylomatous tissue (19). Additionally, some studies have indicated that immunoglobulin G (IgG) and IgM antibodies to purified BPV virions are elevated in sera from CIN patients (1, 2). We have previously reported the presence of IgA antibodies to BPV in cervical secretions (5), as well as a correlation between the presence of such antibodies in cervical secretions and the diagnosis of CIN (5). In order to further analyze the serum antibody response to the group-specific PV capsid antigens, we studied the levels of serum antibodies against BPV in a healthy population compared with the levels of serum antibodies in women with CIN or cervical carcinoma. A total of 139 control serum samples from healthy, adult women were collected. Of these, 62 serum samples were *
from
attending gynecology outpatient clinics, either Huddinge Hospital (Stockholm, Sweden; 39 women) or the University of Ferrara (Ferrara, Italy; 9 women). The women either had no pathological findings (48 women) or had condyloma but no histopathological evidence of CIN (14 women). A total of women
at Scripps Clinic (La Jolla, Calif.; 14 women),
59 serum samples from healthy, female laboratory workers and 18 serum samples from women attending an annual health check-up were collected at the National Bacteriological Laboratories (Stockholm, Sweden). The patient group consisted of 114 women with untreated, histopathologically confirmed cervical neoplasia, of which 13 lesions were classified as CIN 1, 16 lesions were classified as CIN 2, 16 lesions were classified as CIN 3, and the remaining 69 lesions were classified as invasive cervical carcinoma. These sera were collected either at Scripps Clinic (41 serum samples), Huddinge Hospital (9 serum samples), University of Ferrara (32 serum samples), or Radiumhemmet, Karolinska Hospital (Stockholm, Sweden; 32 serum samples). Virus was isolated from bovine warts and purified by two cycles of CsCl equilibrium centrifugation as previously described (5). Our preparation of BPV virions was previously shown to consist of four proteins, all of which contained epitopes immunoreactive with human sera as shown by immunoblotting (5). The purified BPV was disrupted by five cycles of freezethawing, diluted to 3 Fg/ml in 10 mM carbonate buffer (pH 9.6), and then added to 50-,ul 96-well microdilution plates (Costar) and kept at room temperature overnight. After being washed once with phosphate-buffered saline-0.05% Tween 20 (PBS-T), the plates were blocked with 10% lamb serum in PBS (LS-PBS) and incubated for 60 min at 37°C. Human sera were diluted 1:20 in LS-PBS, added to the plates in duplicate wells, and allowed to react for 120 min at 37°C. After the plates were washed five times with PBS-T, a horseradish peroxidase-labeled monoclonal antibody against human IgA (Janssen) diluted 1:500 in LS-PBS was incubated on the plates for 120 min at 370C. The plates were then washed five times with PBS-T and developed with 20 mg of 2,2'-azino-di(3-ethylbenzthiazolinesulfonate) deammonium
Corresponding author. 624
VOL. 28, 1990
NOTES
625
1,5
1,0
0,5
.0 -0.
l'uni ...
.uu:L
a1 1
E
n
n
m
S-
M
M
OI
E
M
*a
i
-F
m
FIG. 1. Detection of IgA antibodies against PV capsid antigen in sera from patients with CIN or carcinoma of the cervix uteri (SCC). Sera were diluted 1:20 and were added to plates coated with BPV. IgA antibodies were detected with a horseradish peroxidase-labeled monoclonal antibody against human IgA. Each datum point represents the mean of duplicate absorbances for serum blanks subtracted from the mean absorbance of duplicates. Compared with the absorbances in the No CIN group, the absorbances in the CIN and carcinoma group (CIN 1, CIN 2, CIN 3, and SCC) are elevated (P < 0.025).
salt (ABTS) per ml diluted 1:50 in 0.1 M citrate buffer (pH 4) with 0.9% hydrogen peroxide. The A415s were recorded after 60 min. For detection of IgG, the plates were washed and blocked with LS-PBS as described above. Thereafter, a rabbit anti-human IgG-alkaline phosphatase conjugate (Dako) diluted 1:1,000 in LS-PBS was incubated on the plates for 120 min at 37°C. After the plates were washed five times with PBS-T, 1 mg of phosphatase substrate per ml (Sigma Chemical Co.) in 0.1 M diethanolamine buffer (pH 9.6)-i mM MgCi2 was added, and the plates were read at 405 nm after 90 min. For detection of IgM antibodies, the plates were washed and blocked as described above and were then incubated with an anti-human IgM-glucose oxidase conjugate (Sera-lab) at a dilution of 1:800 in 10% LS-PBS for 120 min. As a substrate, we used 0.36 mg of ABTS per ml, 2.4% glucose, and 8 ,ug of horseradish peroxidase per ml (Sigma) in 0.1 M phosphate buffer (pH 6.0). The plates were read at 415 nm after 60 min. Serum samples from two CIN patients with known reactivity were used as internal standards in all tests. Uncoated wells in duplicates served as negative controls. The single dilution optical density titers (3, 6, 10) were calculated as the mean of duplicate absorbances on uncoated wells subtracted from the mean of duplicate absorbances. For the statistical analysis, the different CIN groups and the cervical carcinoma group were combined to one single cervical neoplasia group, which was compared with the control group by a two-sided, nonparametric ranking test (Mann-Whitney test). The IgA antibody levels were increased in the CIN or carcinoma patient group, compared with the age- and sex-matched controls without known CIN (P < 0.025) (Fig. 1). In contrast, the levels of IgG antibody against PV were decreased in cervical neoplasia patients compared with those of controls (P < 0.001) (Fig. 2). There
FIG. 2. Detection of IgG antibodies to PV capsid antigen in serum from patients with CIN or carcinoma of the cervix uteri (SCC). The same sera as those used in Fig. 1 were tested for the presence of IgG antibodies against BPV with a rabbit anti-human IgG-alkaline phosphatase conjugate. Compared with the reactivities in the No CIN group, the reactivities in the CIN and carcinoma (SCC) group are decreased (P < 0.001). See legend to Fig. 1 for a definition of the data points.
was also a decrease in the IgM reactivity to PV in the cervical neoplasia patient group compared with that of the controls (P < 0.005) (Fig. 3). When the differences between the IgA and the IgG responses (IgA-IgG) were compared for the group without known CIN and the CIN or carcinoma
1.GF
0,5 a un ee»»
a..Mm_ _M
ee
.00 ...0 a
n
nz
CH
_::::O
.
n
FIG. 3. Detection of IgM antibodies to PV capsid antigen in sera from patients with CIN or carcinoma of the cervix uteri (SCC). The same sera as those used in Fig. 1 were also tested for the presence of IgM antibodies against PV, with an anti-humàn IgM glucose oxidase conjugate. Compared with the antibody levels in the patients without CIN (No CIN), the antibody levels in the CIN and carcinoma group are decreased (P < 0.005). See legend to Fig. 1 for a definition of the data points.
626
J. CLIN. MICROBIOL.
NOTES
It is clear that there are many problems involved in PV serology, e.g., distinguishing between cutaneous and mucous membrane infections, between genital tract and other mucous membrane infections, between infections with malignancy-associated and benign types of HPV, and finally between latent infections and clinically important, active infections. The PV serology for different immunoglobulin classes described here has several distinct advantages. It is likely that the immunoglobulin class of the response could reflect the extent of the mucous membrane involvement of the infection. At least theoretically, cutaneous warts should give rise mainly to IgG antibodies, although some IgA antibodies are also likely to be elicited. It is therefore possible that measurement of the IgA-IgG difference could be more specific for mucous membrane PV infection than just measuring the IgA response. It is also likely that serology could distinguish between latent and active infection. By analogy with the situation for other viruses, IgA and IgM antibodies should not persist for long after the active infection has been resolved, whereas IgG antibodies may persist long after the primary infection has been resolved. It is possible that consecutive testing of the same patient could provide information as to whether a PV infection is chronically active or has resolved. Distinction between different types of PV cannot be made by the enzyme-linked immunosorbent assay described herein, but since the number of different types of PV is very large and increasing, there is clearly also a need for a group-specific test like ours.
1.0E
0.51-
~ ~ ~ ~
~
*-
FIG. 4. The
(IgA-IgG)
to
difference
PV
~
~
as were
~
~
~
~
~
~
~
cu:
.
between
capsid antigen
absorbance values
:
..
*33
the
IgA and IgG
responses
in relation to disease. The
recorded in Fig.
and 2
were
same
plotted
as
IgG subtracted from IgA. The absorbances in the CIN and carcinoma (SCC) group are significantly elevated (P < 0.0001). See to Fig. for a definition of the data points.
legend
group, a striking elevation for the CIN or carcinoma patient group was noted (P < 0.0001) (Fig. 4). We thus demonstrate here that there is an increase of serum IgA antibodies to PV in patients with CIN or carcinoma of the cervix uteri, although it is smaller than what we previously found for cervical secretions (5). Contrary to previous studies (1, 2), ours was not able to detect any significant increase in IgG or IgM antibodies against the PV group-specific capsid antigens in sera from CIN patients. We actually found the IgG and IgM antibodies unexpectedly to be significantly decreased. Whereas Baird (1, 2) used BPV virions disrupted by treatment with sodium dodecyl sulfate, we used a treatment of multiple freeze-thawing cycles. In our studies, the presence of sodium dodecyl sulfate impaired the binding of antigen to enzyme-linked immunosorbent assay plates, and we have therefore not been able to use the same treatment as that used by Baird (1, 2). The discrepancy between our results and those previously reported could thus possibly reflect the difference in methodology used. It is noteworthy, however, that some studies that used HPV purified from human warts found the IgG antibody levels of HPV to be decreased among patients with persistent or multiple cutaneous wart infections (16, 17). Conceivably, HPV-induced CIN could also be associated with immunosuppression of the antibody response to HPV. In the case of the IgA antibodies, it is possible that an IgA-stimulating mucous such an
membrane involvement of the lesion may
immunosuppression.
overcome
We thank P. Bistoletti, D. DiLuca, N. Einhorn, G. Hall, I. Jones, E. Norrby, J. Robb, B. Wahren, and J. Willems for providing sera; C. Eklund for excellent technical assistance; and C. A. Petterson for help with illustrations. This work was supported by grants from the Medscand Foundation for Medical Research, by the Swedish Cancer Society, by the Swedish Medical Research Council, and by the Swedish Council for Forestry and Agricultural Research. LITERATURE CITED 1. Baird, P. J. 1983. Serological evidence for the association of papillomavirus and cervical neoplasia. Lancet ii:17-18. 2. Baird, P. J. 1985. Part Il: the role of human papilloma and other viruses. Clin. Obstet. Gynecol. 12:19-32. 3. Blomberg, J., I. Nilsson, and M. Andersson. 1983. Viral antibody screening system that uses a standardized single dilution immunoglobulin G enzyme immunoassay with multiple antigens. J. Clin. Microbiol. 17:1081-1091. 4. Dillner, J., L. Dillner, J. Robb, J. Willems, I. Jones, W. Lancaster, R. Smith, and R. Lerner. 1989. A synthetic peptide defines a serologic IgA response to a human papillomavirusencoded nuclear antigen expressed in virus-carrying cervical neoplasia. Proc. Natl. Acad. Sci. USA 86:3838-3841. 5. Dillner, L., Z. Bekassy, N. Jonsson, J. Moreno-Lopez, and J. Blomberg. 1989. Detection of IgA antibodies against human papillomavirus in cervical secretions from patients with cervical intraepithelial neoplasia. Int. J. Cancer 43:36-40. 6. Granstrom, M., I. G. Julander, S. Â. Hedstrom, and R. MolIby. 1983. Enzyme-linked immunosorbent assay for antibodies against teichoic acid in patients with staphylococcal infections. J. Clin. Microbiol. 17:640-646. 7. Jenison, S. A., J. M. Firzlaff, A. Langenberg, and D. A. Galloway. 1988. Identification of immunoreactive antigens of human papillomavirus type 6b by using Escherichia coli-expressed fusion proteins. J. Virol. 62:2115-2123. 8. Jenson, B., J. D. Rosenthal, C. Olsson, F. Pass, W. D. Lancaster, and K. Shah. 1980. Immunologic relatedness of papillomaviruses from different species. J. Natl. Cancer Inst. 64:495-498. 9. Komly, C. A., F. Breitburd, O. Croissant, and R. E. Streeck. 1986. The L2 open reading frame of human papillomavirus type la encodes a minor structural protein carrying type-specific
VOL. 28, 1990 antigens. J. Virol. 60:813-816. 10. Kurstak, E. 1985. Progress in enzyme immunoassays: production of reagents, experimental design and interpretation. Bull. W.H.O. 63:793-811. 11. Li, C.-C. H., K. V. Shah, A. Seth, and R. V. Gilden. 1987. Identification of the human papillomavirus type 6b Li open reading frame protein in condylomas and corresponding antibodies in human sera. J. Virol. 61:2684-2690. 12. Lorincz, A. T., G. F. Temple, R. J. Kurman, A. B. Jenson, and W. D. Lancaster. 1987. Oncogenic association of specific human papillomavirus types with cervical neoplasia. J. Natl. Cancer Inst. 79:671-677. 13. Pfister, H. 1984. Biology and biochemistry of papillomaviruses. Rev. Physiol. Biochem. Pharmacol. 99:111-181. 14. Pfister, H. 1987. Human papillomaviruses and genital cancer. Adv. Cancer Res. 48:113-147. 15. Pfister, H., and H. Zur Hausen. 1978. Seroepidemiological
NOTES
16. 17.
18.
19.
627
studies of human papillomavirus (HPV-1) infections. Int. J. Cancer 21:161-165. Pyrhonen, S., and E. Johansson. 1975. Regression of warts. An immunological study. Lancet i:592-595. Pyrhonen, S., and K. Penttinen. 1972. Wart-virus antibodies and the prognosis of wart disease. Lancet ii:1330-1332. Reid, R., M. Greenberg, A. B. Jenson, M. Husain, J. Willett, Y. Daoud, G. Temple, C. R. Stanhope, A. I. Sherman, G. D. Phibbs, and A. T. Lorincz. 1987. Sexually transmitted papillomaviral infections I. The anatomic distribution and pathologic grade of neoplastic lesions associated with different viral types. Am. J. Obstet. Gynecol. 156:212-222. Walker, P. G., A. Singer, J. L. Dyson, K. V. Shah, A. To, and D. V. Coleman. 1983. The prevalence of human papillomavirus antigen in patients with cervical intraepithelial neoplasia. Br. J. Cancer 48:99-101.
