JOURNAL OF CLINICAL MICROBIOLOGY, Feb. 1982, p. 243-248 0095-1137/82/020243-06$02.00/0 Vol. 15, No. 2 Improved Immunofluorescence Antigens for Detection of Immunoglobulin M Antibodies to Epstein-Barr Viral Capsid Antigen and Antibodies to Epstein-Barr Virus Nuclear Antigen DANA GALLO,`* KIRSTEN H. WALEN,2 AND JOHN L. RIGGS1 Viral and Rickettsial Disease Laboratory, California Department of Health Services, Berkeley, California 94704,1 and Children's Hospital Medical Center, Oakland, California 946092 Received 31 March 1981/Accepted 28 September 1981 Epstein-Barr viral capsid antigen and nuclear antigen produced by modified procedures were evaluated for use in measuring viral capsid antigen immunoglobulin M and Epstein-Barr virus nuclear antigen antibody responses in sera from patients with suspected Epstein-Barr virus infections. Viral capsid antigen production was stimulated with a phorbol ester, and the Epstein-Barr virus nuclear antigen cells were fixed in suspension to eliminate loss of antigen during the drying process. Both preparations proved to be sensitive and reliable. Epstein-Barr virus (EBV), the causative agent of infectious mononucleosis (9), stimulates heterophil antibodies in the sera of most patients during the acute stage of infection (4, 14). These antibodies agglutinate sheep and horse erythrocytes and lyse bovine erythrocytes in the presence of complement, and detection of these antibodies is the diagnostic test of choice. However, EBV-specific antibody assays, as well as assays for cytomegalovirus, Toxoplasma gondii, and adenovirus, which can also cause infectious mononucleosis-like symptoms, are often required for heterophil-negative cases (10, 12). Heterophil antibody is sometimes detectable for many months after infection (3), which can be a source of diagnostic confusion, requiring additional serological tests for EBV to clarify the role of this agent in such problem cases. Henle et al. (11) have shown that immunoglobulin G (IgG) and IgM antibodies to EBV viral capsid antigen (VCA) are stimulated in virtually 100% of EBV infections. These antibodies increase so rapidly that greater than or equal to fourfold IgG antibody rises to VCA can be demonstrated in only approximately 20% of the cases. VCA-IgM antibody is elevated during the acute stage of infection and drops to undetectable levels after convalescence (12), so presence of this class of antibody is suggestive of current or recent infection. Antibody to EBV nuclear antigen (EBNA) is stimulated late in the course of infection and is detected 3 weeks to 6 months after onset (8). Thus, the presence of VCA-IgM antibody and the absence of EBNA antibody in an acute serum is a reliable indicator of recent EBV infection; and, conversely, a VCA-IgM-negative and EBNA-positive serum reaction is interpreted as indicating past infection (12). Unfortunately, diagnostic tests for VCA-IgM and EBNA antibodies are available in very few laboratories because of inherent problems in producing reliable antigens. IgG antibody to VCA is readily determined by the indirect immunofluorescence test, employing an EBV-infected lymphoblast line that produces VCA (producer line) (7). However, a portion of these lymphoblast cells contain IgM and react with the anti-igm conjugate in the VCA-IgM antibody test (12). A specific reaction of IgM antibody in the serum of a patient with the VCA antigen is difficult to differentiate from the background fluorescence in the IgM-producing cells. EBNA antibody is detected by the sensitive anticomplement immunofluorescence (ACIF) test performed on cell smears of a nonproducer EBV-infected human lymphoblast line (15). These cells contain EBV antigens in the nucleus but produce no infectious virus. Dense nuclear fluorescence should be visible in virtually 100% of the cells with a serum possessing EBNA antibody. However, EBNA is unstable, and when the cells are applied to slides or cover slips, the antigen tends to diffuse out from the nucleus during the drying process. Extensive loss of antigen from the cells can result in falsenegative EBNA antibody tests. This problem can be partially alleviated by spreading the cells across glass cover slips to promote rapid drying of the cells. The preparations are then air dried for 1 to 2 h and either fixed in carbon tetrachloride for 15 min at room temperature (17) or in equal volumes of methanol and acetone at -20 C for 3 min (12). The authors of the latter 243
244 GALLO, WALEN, AND RIGGS procedure state that, when this method is used, the cover slip preparations usually contain some satisfactory areas, even when leakage of the antigens has occurred. However, we have found that the whole smear must be carefully screened before it can be determined that a serum lacks EBNA antibody, since the amount of antigen in the cells varies due to leakage. Thus, EBNA substrates are rather laborious to prepare, and reading the test is time consuming. A tumor-promoter agent (TPA), 12-O-tetradecanoyl-phorbol-13-acetate, has been shown to stimulate the number of VCA-producing cells from 5 to 80% in a producer line (18). We postulated that, if the number of VCA-producing cells was this much greater than the number of IgM-producing cells (usually 5 to 10% of the cell population), VCA-IgM antibody could then be reliably detected. In cell biology studies, cells are often fixed in suspension to stabilize cellular components. An adaptation of this method was used to prepare EBNA smears to determine if loss of EBNA could be prevented. The present study evaluated these two antigen preparations for use in measuring VCA-IgM and EBNA antibody levels in sera from patients with suspected EBV infections. With rare exceptions, heterophil tests were either negative or had not been performed. MATERIALS AND METHODS VCA antigen preparation. The producer lymphoblast line, P3HR-1, was stored in liquid nitrogen until needed, to prevent loss of VCA production owing to extensive passage. Each stored ampoule contained cells from one 7-day-old bottle culture in 1 ml of growth medium with 10o dimethyl sulfoxide. For use, the cells from one ampoule were propagated in a 4- ounce (120-ml) prescription bottle containing 20 ml of Dulbecco modified Eagle medium (high glucose) with 10%o fetal bovine serum, 0.09% glutamine, 200 U of penicillin and 200 U of streptomycin per ml, and 3 p.g of amphotericin B per ml. Sodium bicarbonate was added as needed to maintain a ph of 7.2. At 7 days, 6 x 101 cells (78% viability by trypan blue stain) were transferred to each of three 4-ounce (120-ml) bottles containing growth medium with 20 ng of TPA (Consolidated Midland Corp., Brewster, N.Y.) per ml. A stock solution of TPA was prepared by reconstituting 10 mg in 1 ml of dimethyl sulfoxide; this was stored in 0.1-ml amounts at -20 C. A sample was reused provided that no precipitate formed when the solution was thawed. In 24 h, many of the cells in the TPA-treated cultures had attached to the glass. These cells were pipetted into the fluids at 3 days, and slides were prepared at 5 days. VCA slide preparation. The cell suspension from the three bottle cultures of P3HR-1 cells was centrifuged at 450 x g for 5 min, the supernatant fluids were removed, and the cell pellet was reconstituted in 3 ml of 0.01 M phosphate-buffered saline, ph 7.2, containing 2% fetal bovine serum. Three cell spots, each approximately 3 mm in diameter, were applied to a J. CLIN. MICROBIOL. slide, the cell density was checked microscopically, and the concentration was adjusted, if necessary, to give an almost confluent cell smear. Eight cell spots were made on each of 200 glass slides (gold seal; Clay Adams, Parsippany, N.J.). The slides were air dried, fixed in acetone at room temperature for 10 min, and stored at -60 C. EBNA antigen preparation. The nonproducer Raji lymphoblast cells were also stored in liquid nitrogen until needed, maintained in the above growth medium without TPA, and split 1:3 at 7 days. The antigen was fixed at 14 days. EBNA slide preparation. Ten-milliliter samples of Raji cell suspension from the three bottle cultures were centrifuged at 450 x g for 5 min. Immediately after the fluid was removed from a sample, 1 ml of a freshly prepared mixture of equal volumes of methanol and acetone was added dropwise to the cell pellet while the cells were gently agitated in a Vortex mixer. The fixed cells were pooled and centrifuged, and the supematant fluids were removed. The cell pellet was reconstituted in 3 ml of methanol-acetone, and the density was adjusted so that the cells in a 3-mm spot were almost confluent. The cell suspension was stored in a tightly stoppered tube at 4 C. Eight spots per slide were made on the day that EBNA antibody tests were performed, and the slides were air dried at room temperature for 30 min before use. Immunofluorescence procedures. A 1:8 dilution of each test serum was absorbed with equal volumes of glutaraldehyde cross-linked human IgG for 1 h at 35 C to remove rheumatoid factor before being tested for the presence of VCA-IgM antibody (1). The IgM (6), IgG (2), and ACIF (13) tests were performed as previously described. Sera screened for EBNA antibody were tested in duplicate by the indirect immunofluorescence method on EBNA slides to detect antinuclear antibody, which would mask EBNA staining. The cell smears were not allowed to dry during the ACIF procedure for EBNA antibody determinations. Excess moisture around the smears was aspirated, and the appropriate reagent was applied immediately. RESULTS Approximately 70% of the TPA-stimulated P3HR-1 cells fluoresced when reacted with VCA-IgM- or VCA-IgG-positive sera (Fig. 1A). Five percent of the cells reacted to varying degrees of brightness with the anti-igm conjugate alone (Fig. 1B). Because of this marked contrast between the numbers of VCA- and IgM-producing cells, even low levels of IgM antibody could be easily detected. In the ACIF test for EBNA antibody, 100% of the cells (except the dead ones) in every smear tested showed dense nuclear fluorescence with positive sera (Fig. IC). Variability in staining intensity within a smear, which would indicate leakage of the antigen, was not observed. This observation is in contrast to the uneven staining patterns encountered with the previous method of preparing EBNA slides (Fig. 1D). Either no fluorescence or faint diffuse staining was visible with negative sera. The reliability of this antigen
VOL. 15, 1982 IMPROVED VCA-IgM AND EBNA ANTIGENS 245 Downloaded from http://jcm.asm.org/ FIG. 1. TPA-stimulated P3HR-1 cells stained with a VCA-IgM-positive serum and an anti-human IgMspecific conjugate (A) and with the anti-human IgM conjugate (B); Raji cells fixed in suspension with methanolacetone and stained by ACIF with an EBNA-positive serum (C); and uneven EBNA staining pattern in Raji cells prepared by spreading the cells on a slide and fixing in methanol-acetone (D). Magnification, x440. greatly facilitated the reading of the EBNA antibody assays. The improved antigen preparations were evaluated in tests on 122 sera from patients with recent and past EBV infections. Twenty patients possessed antibody patterns consistent with recent EBV infection. The ages of the patients ranged from 1 to 62 years (Table 1). Fifteen patients showed VCA-IgM antibody and no antibody to EBNA; both IgM and EBNA antibodies were present in later specimens taken from patients 3, 5, 6, 11, and 18. Neither VCA-IgM nor EBNA antibodies were demonstrable in the single, early serum from patient 4. A rise in IgG antibody to VCA and EBNA was demonstrated between the first and second serum specimens submitted on patient 10, but no IgM antibody was present. No IgM antibody could be detected L:,--.z,-- in patient 18 until 49 days after onset. A serum sample (which was not available for this study) taken on day 1 was positive for heterophil antibody. Three weeks later, the patient presented with a very sore throat and severe leukocytopenia. Haemophilus influenzae was isolated from a throat culture. He was hospitalized, and antibiotics and prednisone were administered. A severe leukocytopenia and no VCA-IgM or EBNA antibody were present at 31 days. The 49-day sample, taken after convalescence, possessed antibodies to both antigens, and the leukocyte count was normal. EBNA antibody was present in two patients (5-2 and 14) earlier than 21 days after onset. The patients examined in this study, with the exception of patient 18, were heterophil negative, and all were diagnostic problem cases. V-i D on June 17, 2018 by guest
246 GALLO, WALEN, AND RIGGS J. CLIN. MICROBIOL. TABLE 1. Recent EBV infections as indicated by VCA-IgG, VCA-IgM, and EBNA antibody results Titer Patient no. Age (yr) Days after onset VCA-IgG VCA-IgM EBNA 1 23 1.32-512 <2 2 1 1.512 16 <2 3-1 19 2.512.512 <2 3-2 33.512 128 4 4a 3 6 128 <8 <2 5-1 19 7 32 32 <2 5-2 19 32.512 32 6-1 18 8.32.512 <2 6-2 23.32.512 4 7 25 10.32 128 <2 8 40 11.32.512 <2 9-1 15 11.32 128 <2 9-2 25.32 8 <2 10-1 8 11 <8 <8 <2 10-2 27.32 <8 16 11-1 20 12.128.512 <2 11-2 26.128.512 4 12 8 15.32 128 <2 13 8 15.32 32 <2 14 26 19.32 16.256 15-1 22 19.512 128 <2 15-2 47.512 32 <2 16 34 20 512 128 <2 17 7 25.32 32 64 18-1 15 31 128 <8 <2 18-2 49 128 32 4 19 62 31.512 8 <2 20 5 NAb.512 128 <2 a Results on patient 4 were interpreted as infection at some time. b NA, Not available. Antibody patterns failed to implicate EBV as the causative agent of the current illness in 94 patients. Representative results indicating past EBV infections are shown in Table 2. Titers on patient 27, a 5-month-old infant, were interpreted as possibly being due to maternal antibody. DISCUSSION The TPA-stimulated VCA antigen preparation provided a reliable substrate for VCA-IgM antibody determinations. Commercially available VCA slides that we have tested (Gull Laboratories, Inc., Salt Lake City, Utah; Litton Bionetics, Inc., Kensington, Md.) were not suitable for IgM antibody studies, owing to comparable levels of VCA- and IgM-producing cells in the slide preparations. Fixation of Raji cells in suspension prevented the loss of EBNA antigen during the drying of the cell smears. Storage of these slides at -60 C resulted in some loss of antigenicity, so fixed cells were held at 4 C, and smears were prepared when needed. Because the cell suspension was stored in the methanol-acetone mixture at the appropriate cell density for preparation of smears, the number of slides necessary for the day's run could be rapidly made. EBNA cell suspensions have been stored at 4 C for 6 weeks with no loss of antigenicity. An attempt to fix a cell pellet from three Raji bottle cultures resulted in clumping of the cells, presumably due to uneven fixation. A homogeneous mixture was obtained when the fixation procedure was applied to cells pelleted from 10-ml samples of cell suspension. Although the ACIF test requires an additional step to that needed for the indirect immunofluorescence test, it is not a difficult procedure. In our experience, drying of the cell smears between the addition of reagents may result in a loss in sensitivity and should be avoided. Substituting guinea pig serum for human serum eliminates the problem of locating an EBV-negative serum source for complement. After the working dilution of a new lot of freshly reconstituted guinea pig complement is determined, the complement can be dispensed in appropriate working volumes and stored at -60 C, and the amount needed for the test can be thawed immediately before use. A commercial anti-guinea pig C3 conjugate (N. L. Cappel Laboratories, Inc., Cochranville, Pa.) has given excellent results. Demonstration of the presence or absence of antibody to the various antigens was sufficient
VOL. 15, 1982 IMPROVED VCA-IgM AND EBNA ANTIGENS 247 TABLE 2. Past EBV infections as indicated by VCA-IgG, VCA-IgM, and EBNA antibody results Titer Patient no. Age (yr) Days after onset VCA-IgG VCA-IgM EBNA 21 27 2 128 <8.2 22 9 3 128 <8.2 23 9 4 232 <8.2 24 18 5 232 <8.2 25 22 6.32 <8-2 26 23 6-32 <8.2 27 5 months 6.32 <8-8 28 10 6 232 <8 28 29 33 9 32 <8.2 30 18 9 232 <8 28 31 14 9 232 <8 28 32 12 10 128 <8.8 33 19 11 128 <8.2 34 52 13.32 <8.2 35 24 13-32 <8.2 36 25 14.32 <8.2 37 12 16.32 <8.2 38 16 20.32 <8.2 39 NAa 21.32 <8.2 40 23 21.32 <8.2 a NA, Not available. for interpretation in the majority of cases, and determination of serum endpoint titers was usually not necessary. Presence of VCA-IgG antibody and lack of VCA-IgM and EBNA antibody in the serum of patient 4 was interpreted as infection at some time. The unusual delay in the appearance of IgM antibody seen in patient 18 would suggest that a later serum specimen should be tested when a pattern of VCA-IgGpositive, VCA-IgM- and EBNA-negative is present in an acute-phase serum sample. This result may indicate infection occurring as long as 6 months previously, since EBNA antibody production can be delayed. As Henle et al. have stressed (12), VCA-IgM tests or EBNA tests alone are not ideal for determining recent and past EBV infections. VCA-IgM antibody was not detected in specimens from patients 4, 10-2, and 18-1. Also, because EBV becomes latent in the host after primary infection, it is possible that another disease state might activate the virus and stimulate VCA-IgM antibody production (16). VCA- IgM antibody has been demonstrated in current cytomegalovirus infections (5), and we have detected low levels of VCA-IgM antibody in some cancer patients (data not shown). Absence of EBNA antibody and presence of VCA-IgG antibody usually indicates infection within the last 6 months. Presence of EBNA antibody in a specimen taken less than 3 weeks after onset is considered to be indicative of past infection. However, results on specimens from patients 5-2 and 14 would have been interpreted as indicating past infections if only the EBNA test had been performed. These reactions may reflect inaccurate onset dates rather than unusually early EBNA antibody responses. The combinations of VCA-IgM antibody and lack of EBNA antibody, or EBNA antibody and lack of IgM antibody, are much more reliable indicators of recent and past EBV infections, respectively, than is either test alone. Negative or positive reactions to both antigens should be interpreted with caution (patients 4, 14, and 17). Patients presenting with lymphatic disease who are heterophil negative and for whom other infections have been ruled out are diagnostic problems for the clinician, especially when lymphoma is a possibility. It is highly desirable that additional, reliable serological tests for EBV be available for studying these problem cases. Tests for measuring VCA-IgM and EBNA antibodies are within the capability of diagnostic microbiology laboratories, and the two antigens evaluated in this study were shown to be sensitive and reliable. It is hoped that similar reagents will soon become available through commercial sources. ACKNOWLEDGMENT We thank Nathalie J. Schmidt for her valuable discussions and review of the manuscript. LITERATURE CITED 1. Cremer, N. E., and J. L. Riggs. 1979. Immunoglobulin classes and viral diagnosis, p. 191-208. In E. H. Lennette and N. J. Schmidt (ed.), Diagnostic procedures for viral, rickettsial and chlamydial infections, 5th ed. American Public Health Association, Inc., Washington, D.C. 2. Emmons, R. W., and J. L. Riggs. 1977. Application of immunofluorescence to diagnosis of viral infections, p. 1-
248 GALLO, WALEN, AND RIGGS 28. In K. Maramorosch and H. Koprowski (ed.), Methods in virology, vol. VI. Academic Press, Inc., New York. 3. Evans, A. S., J. C. Niederman, L. C. Cenabre, B. West, and V. A. Richards. 1975. A prospective evaluation of heterophile and Epstein-Barr virus-specific IgM antibody tests in clinical and subclinical infectious mononucleosis: specificity and sensitivity of the tests and persistence of antibody. J. Infect. Dis. 132:546-554. 4. Fleisher, G., E. T. Lennette, G. Henle, and W. Henle. 1979. Incidence of heterophil antibody responses in children with infectious mononucleosis. J. Pediatr. 94:723-728. 5. Forsgren, M., and A. Demissie. 1980. IgM responses to EBV/CMV in cytomegalovirus and Epstein-Barr infections. Proceedings of the International Conference on Human Herpesviruses. The Woodruff Medical Center, Emory University, Atlanta, Ga. 6. Gallo, D., J. L. Riggs, J. Schachter, and R. W. Emmons. 1981. Multiple-antigen slide test for detection of immunoglobulin M antibodies in newborn and infant sera by immunofluorescence. J. Clin. Microbiol. 13:631-636. 7. Henle, G., and W. Henle. 1966. Immunofluorescence in cells derived from Burkitt's lymphoma. J. Bacteriol. 91:1248-1256. 8. Henle, G., W. Henle, and C. A. Horowitz. 1974. Antibodies to Epstein-Barr virus-associated nuclear antigen in infectious mononucleosis. J. Infect. Dis. 130:231-239. 9. Henle, W., and G. Henle. 1972. Epstein-Barr virus: the cause of infectious mononucleosis, p. 269-274. In I. M. Biggs, G. de The, and L. N. Payne (ed.), Oncogenesis and herpesviruses. International Agency for Research on Cancer. Lyon, France. J. CLIN. MICROBIOL. 10. Henle, W., and G. Henle. 1978. The immunological approach to study of possibly virus-induced human malignancies using the Epstein-Barr virus as an example. Prog. Exp. Tumor Res. 21:19-48. 11. Henle, W., G. Henle, and C. A. Horowitz. 1974. Epstein- Barr virus-specific diagnostic tests in infectious mononucleosis. Hum. Pathol. 5:551-565. 12. Henle, W., G. Henle, and C. A. Horowitz. 1979. Infectious mononucleosis and Epstein-Barr virus-associated malignancies, p. 441-470. In E. H. Lennette and N. J. Schmidt (ed.), Diagnostic procedures for viral, rickettsial and chlamydial infections, 5th ed. American Public Health Association, Inc., Washington, D.C. 13. Kettering, J. D., N. J. Schmidt, D. Gallo, and E. H. Lennette. 1977. Anticomplement immunofluorescence test for antibodies to human cytomegalovirus. J. Clin. Microbiol. 6:627-632. 14. Paul, J. R., and W. W. Bunnell. 1932. The presence of heterophile antibodies in infectious mononucleosis. Am. J. Med. Sci. 183:80-104. 15. Reedman, B. M., and G. Klein. 1973. Cellular localization of an Epstein-Barr virus (EBV)-associated complementfixing antigen in producer and non-producer lymphoblastoid cell lines. Int. J. Cancer 11:499-520. 16. Sumaya, C. V. 1977. Endogenous reactivation of Epstein- Barr virus infections. J. Infect. Dis. 135:374-379. 17. Suzuki, M., and Y. Hinuma. 1974. Evaluation of Epstein- Barr virus-associated nuclear antigen with various human cell lines. Int. J. Cancer 14:753-761. 18. zur Hausen, H., F. J. O'Neill, and U. K. Freese. 1978. Persisting oncogenic herpesvirus induced by the tumour promoter TPA. Nature (London) 272:373-375. Downloaded from http://jcm.asm.org/ on June 17, 2018 by guest