INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Epstein-Barr virus (EBV) humoral immunology has played a major role in studies dealing with a relationship between this virus and nasopharyngeal carcinoma (NPC) (12, 13, 24). Detection of antibodies to the EBV viral capsid antigen (VCA) and EBV early antigen (EA) in sera by indirect immunofluorescence (IF) assays was one of the earliest tests developed. To date, the IF assays still serve as the "gold standard" of EBV serodiagnosis (10, 11, 13). These tests showed the importance of antibodies directed against some of the serologically defined EBV antigens in the diagnosis of EBV-associated diseases. They also help in the clinical management of patients with EBV-associated malignancies. Diagnostically relevant antibodies that have been identified by a number of investigators over the years are immunoglobulin G (IgG) and IgA antibodies directed against EA and VCA. The IgA-EA test, which is routinely used in many laboratories throughout the world, is one of the more specific EBV-associated NPC diagnostic tests available. Moreover, detection of anti-IgA antibodies by IF is suitable for the identification of patients with occult NPC, and the identification of populations at high risk for the development of this cancer (3, 12, 19, 25, 30, 31). However, the IF assays are time-consuming, not suitable for automatic handling, and difficult to standardize because of the variability of antigen-producing cells as well as the subjective reading of results. This makes their application in mass screening of populations not convenient. Some of the technical difficulties associated with IF have been overcome by the development of specific enzyme-linked immunosorbent assays (ELISAs), which are easily automated, quick to perform, and thus well suited for mass screening programs involving populations at high risk for NPC.

The advent of monoclonal antibody technology, gene cloning, and expression of foreign proteins in cells and organisms has greatly facilitated our understanding of the profile of distinct antibodies to individual EBV polypeptides in patients with NPC and other EBV-related diseases (7, 8, 21). It is therefore worthwhile to address the question of whether antibodies against individual polypeptides might be more sensitive and specific than antibodies directed against the whole complex for diagnosing and monitoring patients with NPC. EA and nuclear antigen (EBNA)-specific ELISAs based on recombinant antigens have been successfully used in EBV diagnosis (6, 17). In contrast, the VCA IF assay serologically defines antigens that are more difficult to replace by recombinant proteins. This is related to the complexity of the VCA protein family and the lack of a complete definition of proteins within the VCA complex (18).

In this study, we report results of IgG and IgA responses of NPC patients to the recombinant antigens p54 (BMRF1) and p138 (BALF2) and p18 (BFRF3) and p23 (BLRF2) of the EA and VCA complexes, respectively. Seroreactivity to p72 (BKRF1), representing the EBNA complex, was analyzed concomitantly. We then compared the sensitivity and specificity of these antigens alone or in combination with IgG antibodies directed to the EBV transactivator protein ZEBRA (BZLF1), an immediate-early protein responsible for the switch between EBV latency and replication. The study focused on NPC, especially in young patients who show a high frequency of serological nonresponders by IF (i.e., IgA-EA and -VCA) (2).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Patients and controls. One hundred sera were collected from patients with histologically proven NPC, admitted to the two public oncology centers of Morocco: the National Institute of Oncology in Rabat and the Oncology Center in Casablanca. Young patients were defined as those between 10 and 30 years old, and adult patients were defined as those over 30 years old. The control group consisted of 77 healthy individuals and 28 patients with cancers other than NPC (larynx, tongue carcinoma, and Hodgkin's disease). Healthy individuals consisted of randomly selected blood donors, laboratory staff, and healthy parents of patients awaiting renal transplantation. Healthy controls were matched for sex and age.

EBV-specific serology. All sera were titrated for IgG and IgA antibodies to VCA and EA by using the IF assay (14, 15). The P3HR-1 cell line was used to detect VCA antibodies, while Raji cells treated with 12-O-tetradecanoyl-phorbol-13-acetate (40 ng/ml) and n-butyrate (17 mM/ml) were used for EA. Titers represent the highest serum dilutions at which clear IF staining of the positive target cells was observed.

Detection of IgG and IgA antibodies to the EBV-VCA-p18+23, EA-p54+138, and EBNA1-p72 recombinant proteins by ELISA. The following recombinant antigens were used within an anti-EBV ELISA system (Biotest AG, Frankfurt am Main, Germany).

Detection of IgG antibodies to the EBV-ZEBRA protein by ELISA. Recombinant ZEBRA protein was purified from E. coli transformed by the recombinant pET3c plasmid containing the BZLF1 gene under the control of the T7 RNA polymerase promoter. The anti-ZEBRA titer of a serum was determined with an ELISA involving plates coated with 30 ng of purified protein per well. An EBV-negative serum and an NPC serum were used as negative and positive controls, respectively, in each microtiter assay (22). The titer of the positive control was set at 9,600. The tests were repeated when a substantial deviation from this value was observed. The EBV-ZEBRA-specific ELISA titers were defined by the amount of anti-ZEBRA antibody in the serum measured by the endpoint dilution method as previously described (29). The endpoint dilution was defined as the intersection between the dilution curve of a given serum and the negative/positive cutoff curve.

Determination of cutoff values. The cutoff values were calculated as the mean optical density (OD) of negative controls plus 3 standard deviations (SD) for IgA-EBNA-p72, 2 SD for IgA-EA-p54+138, and 1 SD for IgG-EA-p54+138 and IgA-VCA-p18+23.

Clinical sensitivity and specificity. Sensitivity of the tests was defined as the percentage of NPC individuals detected positive, and specificity was defined as the percentage of negatives scored among healthy individuals.

Statistical analysis. The Mantel-Haenszel ² and Student's tests were used to compare the significance of differences of the trend in seropositivity rates and the mean titers and mean OD values of various antibodies between NPC patients and controls, between young and adult NPC patients, and between young controls and older controls. Pearson's correlation analysis was performed to determine correlations among different types of antibodies as detected by IF and ELISA tests.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Immunoreactivity of sera from NPC patients, healthy donors, and patients with other tumors to VCA-p18+23, EA-p54+138, and EBNA-p72 antigens. One hundred sera from NPC patients, 77 sera from healthy donors, and 28 sera from patients with tumors other than NPC were tested for their IgG and IgA immunoreactivities to EA, VCA, and EBNA recombinant antigens (Table 1). IgG antibodies specific for EA-p54+138 were detected in 86% of NPC sera, compared with only 10% in healthy donors and 22% in patients with other tumors (P1 < 0.001, P2 < 0.001). IgA antibodies against EA-p54+138 were detected in 70% of NPC sera, in only 3% of healthy donor sera, and in 11% of sera from patients with other tumors (P1 < 0.001, P2 < 0.001). A significant difference was also observed in the mean OD value of both IgG and IgA antibodies against EA-p54+138 between NPC patients and healthy donors (1.246 versus 0.573, P1 = 0.00031; 1.518 versus 0.521, P1 = 0.0003) and between NPC patients and patients with other tumors (1.246 versus 0.533, P2 = 0.00041; 1.518 versus 0.485, P2 = 0.0004).

Comparison between ELISA and IF for the detection of IgG and IgA to VCA, EA, and EBNA. As illustrated in Table 3, 64% of NPC sera negative for IgA-EA by IF were positive for IgA-EA-p54+138, and 7% that were negative for IgA-EA-p54+138 were positive for IgA-EA by IF. A positive correlation was observed between the mean OD value of IgA-EA-p54+138 antibodies as detected by ELISA and the mean titer of IgA-EA as detected by IF (RR = 0.229, P < 0.01). Of NPC sera negative for IgG-EA by IF, 75% were positive for IgG-EA-p54+138, and only 9% of sera that were negative for IgG-EA-p54+138 were positive for IgG-EA by IF. Our results also showed that only 24% of sera negative for IgA-VCA by IF presented IgA to VCA-p18+23, and 61% of sera negative by ELISA were positive by IF.

Sensitivity complementation of recombinant antigens EA-p54+138 and ZEBRA protein. Taking into account the sensitivity and specificity of all of the recombinant proteins tested, IgA-EA-p54+138 proved more sensitive and specific than both of the other antigens (i.e., IgA-VCA-p18+23 and IgA-EBNA-p72) and the IF assay for detection of NPC (Table 4). To compare the sensitivity and specificity of IgA-EA-p54+138 alone or in combination with IgG-ZEBRA for detection of NPC, data were considered only for those sera examined by all tests. The serological analysis of IgG antibodies against the EBV-ZEBRA protein showed that these antibodies were detected in more than 90% of NPC patients irrespective of their age, compared to only 3% found in healthy donors. The combination of IgA-EA-p54+138 with IgG-ZEBRA improved the sensitivity to 95%. However, the improvement in sensitivity did not significantly affect specificity (97%).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, the seroreactivities to the recombinant proteins of EBV-encoded antigens expressed in different phases of virus infection (i.e., EBNA-p72 in the latent phase, EA-p54+138 in the early phase, and ZEBRA in the immediate-early phase of active replication, as well as VCA-p18+23 in the late phase) were examined in NPC patients. IgG and IgA antibodies against EA-p54+138 polypeptides showed a high sensitivity for detecting NPC (86% for IgG and 70% for IgA), in contrast to IgA-VCA-p18+23 (30%) and IgA-EBNA-p72 (42%). Considering the low reactivity observed with IgG and IgA-EA-p54+138 with sera from patients with tumors other than NPC, these markers show a high specificity for detecting NPC (78% for IgG and 89% for IgA). In fact, of patients with tumors other than NPC, only 22% were positive for IgG-EA-p54+138 and 11% were positive for IgA-EA-p54+138. Furthermore, these markers can detect a substantial portion of NPC cases negative by IF (75% for IgG and 64% for IgA). This may be due to the sensitivity of the ELISA test, a method that can detect antibodies to particular EBV proteins, such as anti-p138 antibodies, which are not detected by IF with chemically induced Raji cells.

Our results on the IgG immunoresponse to VCA polypeptides agree with those of earlier studies showing a powerful IgG reaction with the small VCA proteins p18 and p23 (26-28). In fact, these two recombinant proteins (i.e., p18 and p23) have been recently described as having a high sensitivity for detection of EBV infection (18). VCA-p18-p23 proved to be a useful diagnostic ELISA antigen, yielding high IgG sensitivity among primary infections and in detecting previous infections (18). In our study, IgG-VCA-p18+23 antibodies were detected in more than 97% of members of the three groups studied (i.e., NPC patients, healthy donors, and patients with other tumors). However, the mean titers of these antibodies were significantly higher in the NPC group than in healthy donors and patients with other tumors. Several earlier investigations showed a significant decline in anti-EBV VCA antibody titers after radiotherapy, while a persistently elevated or rising IgG or IgA antibody titer could signal either a locally recurrent or systemic disease (12, 20). The detection of IgG-VCA-p18+23 antibodies at high titers in the NPC group, reported in the present study, may be considered an important finding. Thus, more investigations should be undertaken to evaluate this marker in the posttherapeutic surveillance of this malignancy.

The VCA complex protein, the antigen most commonly used for serological diagnosis, is composed of more than 30 different proteins. However, the relevant target molecules for IgA antibodies in NPC patients are still poorly defined (1, 4, 27). Our results show that only 30% of NPC patients were positive for IgA-VCA-p18+23 antibodies compared with 6% of healthy individuals. Comparison of the immunoreactivities of NPC sera with this fusion protein between young and older NPC patients showed that only 19% of young patients were positive for IgA-VCA-p18+23 compared with 52% of older patients. The results observed in adult NPC patients are in agreement with those of van Grunsven et al., who reported that 61% of adult NPC patients from Hong Kong presented IgA antibodies to VCA-p18 (27). The detection of IgA-VCA-p18+23 in 30% of our overall NPC population, as opposed to 61% of adult NPC patients from Hong Kong, may be explained by the fact that 25% of our NPC patients are young. These young patients generally show a low reactivity to EBV-specific IgA antibodies (2, 5). Moreover, the observation that IgA-VCA-p18+25 antibodies were detected at almost the same percentage in patients with NPC (30%) and in those with tumors other than NPC (26%) indicates that this marker is not specific to NPC. This marker also proved to be less sensitive than the classical IF assay for detecting IgA-VCA antibodies. The discordance between these two assays could be attributed to individual differences in the immune responses to various epitopes. Furthermore, in the IF assay, several possible epitopes of the whole protein may be accessible, including structural epitopes not detected by the ELISA test. It should also be noted that in NPC sera, the IgA and the IgG immunoreactivities are not necessarily against the same epitope. This may be related to aspects of viral antigen processing and presentation and the different immunobiological responses between local IgA immunity in the nasopharynx and B-cell reactivity (23).

IgG-EBNA-p72 antibodies have been described as having a high diagnostic significance, as being indicative of past infection, and as typically being negative among persons infected recently (16). However, little is known about IgA-EBNA-p72 immunoreactivity in NPC sera. Our results indicate that IgA responses to EBNA-p72 show a high specificity for detection of NPC in both young (100%) and older (97%) NPC patients. In contrast, this marker showed a low sensitivity for detecting NPC in the overall NPC group. This sensitivity was particularly low in young NPC patients (35%) compared with older ones (61%). This may be a feature of these recombinant antigens, which probably do not hold the epitopes that are necessary for adequate T-cell help in the induction of the IgA response. It may also be due to the generally accepted age-related changes in cellular immune function that result in decreased effectiveness to control latent EBV (9).

Taking into account the sensitivity and specificity of all of the recombinant proteins tested, IgA-p54+138 proved more sensitive (70%) and specific (97%) than the other antigens and than the IF assay for detecting NPC. The highest specificity was obtained by deciding a cutoff in favor of specificity rather than sensitivity. It is noteworthy, however, that the combination of IgA-EA-p54+138 with IgG-ZEBRA improved the sensitivity for detection of NPC to 95%. This improvement in sensitivity did not affect specificity (97%).

The availability of recombinant antigens will facilitate detailed studies of the pattern of antibody response, which may result in the development of useful serological markers to guide the treatment of NPC. Furthermore, the development of less expensive diagnostic tests would then facilitate population screening in countries with a high prevalence of NPC. Since early detection improves cure rates, such a test would be beneficial to individuals at an early stage of the disease.