identical igm antibodies recognizing a glycine-alanine epitope are induced during acute infection...

Journal of Clinical Laboratory Analysis 4:456-464 (1990)

Identical IgM Antibodies Recognizing a Glycine-Alanine Epitope Are Induced During Acute

Infection With Epstein-Barr Virus and Cytomegalovirus Gary Rhodes,' Richard S. Smith,' Ruth E. Rubin,* John Vaughan,'

and Charles A. Horwitz3 'Department of Basic and Clinic Research, Scripps Clinic and Research Foundation, La Jolla, California;

'Johnson & Johnson Biotechnology Center, San Diego, California; 3Metropolitan-Mount Sinai Medical Center, Minneapolis, Minnesota

We studied antibody production in serial se rum samples from patients with acute Epstein-Barr virus (EBV) and cyiomegalovi- r u s (CMV) infections. Sera were analyzed both by enzyme-linked immunosorbent assay (ELISA) using a synthetic peptide (P62) derived from the glycine-alanine repeating region of the Epstein-Barr nuclear antigen (EBNA-1) and by immunoblotting. In prior studies, we have shown that patients with acute EBV infection make IgM antibodies that react with this peptide, that recognize a viral-specific protein (EBNA-l), and that bind with a number of proteins present in uninfected cells; however, antibody binding to these autoantigens was inhibited by the peptide. IgG antibodies reactive with the peptide

did not appear until months after the disease and were specific for the EBNA-1 protein. We now find that patients with acute CMV infection but not those with acute infections from a variety of other nonherpes organisms also produce IgM antibodies that recognize the EBV-derived peptide P62. These antibodies also appear to recognize the same cellular proteins as the EBV-induced IgM antibodies. The IgM antibodies appeared in all acutely infected CMV patients studied and occurred both in patients with previous EBV infections and in one patient studied who had not previously been exposed to EBV. It appears that infection with EBV or CMV can induce the synthesis of a very similar or identical set of IgM antibodies.

Key words: EBV, CMV, glycine-alanine repeating region

INTRODUCTION

It is well-established that Epstein-Barr virus (EBV) causes infectious mononucleosis (IM) in humans (1). A primary infection is confirmed serologically by the initial detection of antibody to the EBV-specific viral capsid antigen (VCA) followed by the later appearance (usually after 1-6 mo) of antibodies to the EBV encoded nuclear antigen (EBNA) (2); for a review, see de The (3) and Henle and Henle (4). EBNA-I , the first nuclear antigen to be recognized, has been identified as a 69,000-85,000 protein by the immunoblotting technique (5-7). The size of the EBNA-I protein is correlated with the variation of the length of the IR-3 region of EBV-DNA (6). The IR-3 region encodes a repeating glycine-alanine sequence that has been characterized to be the major epitope of the EBNA- 1 protein (8).

Synthetic peptides representing portions of this glycine- alanine region have been shown to be reactive with sera from patients with EBV-induced IM (9- 12). One synthetic peptide called P62 can be used in an enzyme-linked immunosorbent assay (ELISA) to distinguish, serologically, acute EBV-induced IM from the convalescent and recovery phases (9-12).

0 1990 Wiley-Liss, Inc.

Two characteristics of the antibody response to the EBNA- 1 protein are of particular interest. The IgM antibodies arising early and recognizing the glycine-alanine region also recognize more than a dozen other normal cellular proteins (13). IgM antibody binding to EBNA-1 and to the autoantigens can both be inhibited by the peptide P62. Further, antibody eluted from selected bands on Western blots and used to probe new, unstained strips detects the same normal cellular proteins. Thus, acute EBV infection induces the synthesis of numer- ous IgM autoantibodies that seem to share an epitope related to the glycine-alanine repeating region of EBNA- 1 .

Also, IgG antibodies to EBNA-I do not appear until several months after the disease. IgG anti-peptide P62 antibodies measured by the ELISA assay (9-12), anti-EBNA-1 antibodies measured by standard anticomplement immunofluorescence (ACIF) (14,15), and IgG antibodies to the

Received July 3, 1990; accepted July 10, 1990

Address reprint requests to Richard S . Smith, Ph.D., Johnson & Johnson Biotechnology Center, 4245 Sorrento Valley Blvd., San Diego, CA 92121.

Identical Antibodies in EBV and CMV Infection 457

EBNA- 1 protein measured by immunoblotting all appear con- currently. These IgG antibodies can also be inhibited by the peptide P62 but, in contrast to the IgM anti-peptide P62 antibodies, are exclusively specific for the EBNA-1 protein

In this article, we illustrate another interesting feature of anti-P62 antibodies. We show that acute-phase sera from patients with cytomegalovirus (CMV)-induced mononucleosis- like illnesses also have IgM antibodies reactive with the EBNA P62 peptide. Sequential sera from 9 of 10 such patients with acute CMV infections had serologic evidence of past EBV infection. The last patient was seronegative for EBV at the time of the CMV-induced illness. Immunoblotting of the serum from both acute EBV-infected and acute CMV- infected patients showed that the same set of IgM antibodies is induced by both viruses and that the antibody reacts with a number of normal cellular proteins present in uninfected cells. Thus, infections with CMV as well as with EBV induce the synthesis of cross-reacting IgM antibodies that recognize an epitope similar to the glycine-alanine region of the EBNA-1.

(5-8,13).

MATERIALS AND METHODS

Specimen Sources and Serology

The sera studies were done at the clinics of the Mount Sinai Hospital in Minneapolis, and the clinical features of the patients have been reported previously (16). Standard tests for EBV-related antibodies to VCA (IgM and IgG), early antigen (EA), and EBNA, as well as CMV macroglobulins (CMV-IgM), were performed by indirect immunofluorescent (IFA) procedures (16,17). Anti-CMV was also detected by complement fixation using the AD- 169 strain of virus (Mi- crobiological Associates, Bethesda, MD). Assays for anti- peptide antibodies by ELISA and antibody detection by immunoblotting were carried out in the Johnson & Johnson Biotechnology Center and in the Research Institute of Scripps Clinic, as described below.

Peptides

The peptide sequence for EBNA P62 (9,lO) is A-G-A-G- G-G-A-G-G-A-G-A-G-G-G-A-G-G-A-G-C. It was synthe- sized by the solid-phase method of Memfield (18) as modified by Houghten et al. (19). The EBNA P62 peptide was shown to be >90% pure by high-performance liquid chromatography (HPLC) analysis. The sequence of the P62 peptide was confirmed by analytical amino acid analysis (20).

ELISA

The EBNA P62 ELISA utilized in these studies has been described previously (1 1,12). Briefly, microtiter wells were coated with 50 p1 of 20 pg P62/ml in borate-buffered saline (0.05 M sodium borate, 0.15 M NaCI, and 0.001 M MgC12, pH 8.3) for 12 hr at 4°C. This solution was discarded and the

plates were blocked with 300 pl well of 10% normal goat serum in phosphate-buffed saline (PBS) (pH 7.3). After 90 min incubation at 37"C, the wells were emptied and dried for 60 min at 37°C. At this time, 200 p1 of 10% goat serum in PBS was added to the wells. A volume of 10 pl of patient serum was added to each well and mixed. The specimen was al- lowed to bind to the P62 peptide for l hr at room tempera- ture. After washing five times with 350 p1 of 0.05% Tween in PBS, a solution of 200 pl of horseradish peroxidase- conjugated mouse monoclonal antibody to human IgG or IgM (Ortho Diagnostic Systems, Raritan, NJ) was added to each well. Color was developed in the standard ELISA procedure, and the absorbance was determined at 490 nm.

Analysis of Data

The results of the ELISA with EBNA P62 were normal- ized as has been previously described ( 1 1 ,12). Cutoff values (mean OD + SD) have also been previously described (12). Absorbance of >0.3 for the IgM conjugate was considered to be positive for the diagnosis of IM. A ratio of ELISA values of IgG to IgM (G/M) greater than 0.7 was determined in a healthy population. Ratios less than 0.7 were considered to represent an acute infection due to EBV (1 1).

lmmunoblots

Cells were grown and harvested, and extracts were prepared as previously described (8,13). WI-L2 and K562 cells were obtained from the American Type Culture collection (Rockville, MD). The WI-L2 cells are EBV-positive B cells, while the K562 served as an EBV-negative control.

The CMV strain AD-169 was a gift from Dr. D. Richman of the University of California, San Diego. The virus was propagated in human fibrolastic line GM-2504, obtained from the Human Genetic Mutant Cell Repository (Camden, NJ). The cells were grown in DMEM medium containing 10% fe- tal calf serum and were infected just before confluency at a multiplicity >3. Cells were harvested after 9-12 days and washed twice with PBS. Extracts were prepared by adding 10 ml of DMEM (2% SDS, 2% 2-mercaptoethanol, 0.004% bromphenol blue, 40 mM Tris-HCL, pH 6.8, and 15% glyc- erol) containing 10 pgiml phenylmethylsulfonyl fluoride di- rectly to a 150 cm2 tissue culture flask containing the adherent infected cells. The flask was scraped, and the solution was removed, placed in a boiling water bath for 5 min, and stored at - 20°C.

The extracts were loaded onto a 13-cm-wide slot of a 7.5% acrylamide gel, transferred to nitrocellulose, and cut into 0.5-cm strips. These strips were processed as previously described (7,8,13).

Synthetic peptide inhibitions were performed by diluting the sera 150 in a solution of 50 mM sodium borate, 100 mM sodium chloride, and 3% powdered milk, pH 8.3, and adding free peptide to a final concentration of 400 pg/ml to in-

458 Rhodes et al.

hibit IgM antibody and of 20-50 p,g/ml to inhibit IgG antibody. This solution was incubated overnight at 4°C and then reacted for 1 hr with the nitrocellulose strips at 4°C.

RESULTS

in the extract, as previously described (13). The IgM antibody level to most of these proteins increased in the first two serum samples, peaked at the third, and declined after that. This is the same temporal sequence that was observed with the peptide ELISA (Fig. 1).

2 -

E c 0 z ' - 0 9'

Sequential Sera From a Patient With EBV-Induced IM Patient D. B. had a typical episode of heterophil-positive

EBV-IM followed by an uneventful clinical recovery. During the acute phase of illness, he developed high titers of antibodies to VCA (IgM, >1:80; VCA-IgG, 1:1280) without ac- companying antibodies to anti-EBNA (<1:2). By 174 days after onset, IgM anti-VCA antibodies were no longer detectable, while anti-EBNA antibodies evolved between 76 and 1 1 1 days. Antibodies to CMV were not detected (<1:8) in serial sera by complement fixation.

We have previously shown that IgM antibodies that recognize peptide P62 are detected in an ELISA assay during the acute phase of EBV-IM (9- 12). These observations were confirmed when sequential serum samples from patient D.B., with EBV-induced IM, were analyzed for immunoreactivity by the EBNA P62 ELISA (Fig. 1). There was an early rise of IgM antibodies in the acute phase of the illness (January- March 1985) that corresponded in time to an early rise of IgM and IgG anti-VCA (not shown). Initially, IgG anti-EBNA P62 was only marginally positive (>0.200 unit), rose sig- nificantly at 2 mo, and reached a plateau at 6 mo after onset of illness. The IgG/lgM ratio exceeded 0.7 (1 1) at 4 mo (May), when the patient was in the late convalescent phase of illness. At this time, IgG anti-EBNA was first detectable by immunofluorescence, and the immunofluorescent anti- IgM-VCA titer had fallen to a level that was barely detectable.

The rise and fall of IgM class-specific antibodies during and following the acute phase of the illness was studied by immunoblotting experiments using the sequential sera to probe an EBV-transformed B cell line (Fig. 2). In addition to EBNA- I , the IgM antibodies present in the patient's sera recognized a series of eight major and several more minor bands

8 1 2 5 7 1 1 5 8 1 2 5 7 1 1 5 N 5 9 1 6 1 7 1 5 9 1 6 1 7 1 0

4 7 4 4 7 4 R M

Days Fig. 2. Analysis of sequential sera from patient D.B. by immunoblotting. Cell extracts were prepared from EBV-immortalized Wi-L2 cells. A nor- I I I I I

The IgG antibodies to EBNA- 1 (77 kDa) in the patient appeared late on the immunoblots (Fig. 2), an observation consistent with the EBNA P62 ELISA data. Figure 2 shows a strong IgG response to the 77-kDa EBNA band from a VCA-positive, EBNA-positive individual as a control. IgG anti-EBNA- 1 antibodies were first detected in the blots at 5 1 days after onset of the disease (fourth time point), the same time that IgG antibodies to the peptide started to rise in the ELISA (Fig. 1 ) .

Sequential Sera From a Patient With CMV-Induced IM

Sequential serum samples from a second patient, D.S., were also analyzed by the EBNA P62 ELISA and Western

A

95-

7 7-

66-

45-

29-

B


during both diseases. The IgM antibody levels detected in the blots were at their highest levels during the first and the initial part of the second infection and declined thereafter. This is similar to the results seen in the antipeptide P62 ELISA assay. IgG antibody to EBNA-1 appeared only at the last time point, at the weak band next to a normal EBNA-1 + control in the right lane (Fig. 4).

immunoblotting. The clinical and serological data from this previously healthy patient have been published previously as case 63 in Horwitz et al. (16). D.S. had an initial active, primary cytomegalovirus infection (CMV-IM), with recovery, followed 4 mo later by a primary EBV infection. D.S. initially demonstrated IgM anti-CMV by immunofluorescence (1:256), a fourfold rise in anti-CMV by complement fixation (CMV-CF), and no antibody responses to any EBV-related antigens. During his second illness, beginning 124 days after the onset of the first, a primary EBV infection occurred, characterized by IgM anti-VCA (1:320) and the initial ab- sence (<1:2) but later appearance of antibodies to EBNA by ACIF. The IgM antibodies to VCA disappeared with serial testing. During this second illness, IgM anti-CMV was not detectable, while complement-fixing anti-CMV was stable at 1 : 128. Thus, by classical serology, patient DS had an acute CMV-induced, heterophil negative illness. Four months later, he became ill again, with acute EBV-induced IM.

The results of the antipeptide ELISA studies are shown in Figure 3 . Sera in the acute CMV phase of the illness showed a pronounced IgM response to the P62 peptide detected by ELISA, just as patient D.B. had had during his acute EBV infection. The IgM antipeptide signal was high for the three serum samples obtained during the first illness and was also elevated in the initial sample of the second illness. We do not know if the antibody level decreased between the two illnesses. IgG antibodies to P62 were not detected until the last serum sample, which was obtained a year after the second illness.

Samples from this case were analyzed by immunoblotting on extracts of EBV-transformed B cells (Fig. 4). IgM antibodies to proteins of molecular weights 82, 80-77, 69, 62, and 58 kDa were found during both infections. These proteins are all encoded by the host (13). Note that the IgM antibodies recognized antigens of the same molecular sizes

2

E c 0

S ' s' 0

0

W =lgG/lgM =IgM

o =&G W

Onset of CMUIIM Onset of €BU/lM ///

/ /

/ /

- Jan" Dee

1981 1982

Fig. 3. Anti-peptide P62 ELISA analysis of sequential sera from patient D.S. with a CMV infection (first illness) followed by an EBV infection (second illness).

Sera From 10 Patients With Acute CMV-Induced IM

We examined sera from 10 additional patients with CMV- induced IM by ELlSA and immunoblot . The acute-phase sera of the patients demonstrated significant titers (> 1 :32) of IgM anti-CMV (IF) and anti-CMV (CF) in all cases. Fourfold ti-

9s-

77- 66-

4s-

29-

9 1 2 4 1 1 4 9 1 2 4 1 1 4 N 2 9 4 2 4 7 2 9 4 2 4 7 0

9 0 1 9 01 R M

Fig. 4. Analysis of sequential sera from patient D.S. by immunoblotting. Cell extracts were from EBV-immortalized Wi-L2 cells. (A) IgM antibodies were detected in serum samples, and (B) IgG responses were seen from the same sequential samples. Serum from a normal VCA-positive individual was also used, indicating the 77-kDa EBNA-I protein. Day 9 indicates the first sample (September), taken during the CMV illness (Fig. 3), and day 129 is the second sample (January), taken during the EBV illness.

460 Rhodes et al.

ter rises or falls of the complement-fixation antibodies were noted in six serially sampled patients. All 10 patients demonstrated serological evidence of past EBV infection, with moderately constant levels of IgG anti-VCA and anti-EBNA (>1:10) and no antibodies to IgM anti-VCA. Some case data from these patients were published previously (16).

The 10 patients were considered to have active CMV infections on the basis of their CMV serology (16), and each had had a previous EBV infection. All 10 patients showed a pronounced elevation in IgM anti-peptide P62 antibodies by ELISA during the acute phase of CMV disease. Representa- tive anti-peptide P62 data for two of these patients are shown in Figure 5 . The elevated IgM antibodies persisted for at least 1 mo following the acute illness before returning to the base- line level. In contrast, IgG anti-peptide P62 was present initially (consistent with the patients being positive for VCA and EBNA-1) and showed minimal change during the course of the disease. By immunoblotting (data not shown), all of the patients showed a transient rise in IgM antibodies to a number of normal tissue antigens during the acute phase of the disease, similar to that seen in Figures 2 and 4. Little, if any, change could be detected in the IgG anti-EBNA- 1 antibody levels. Thus, all of the patients with acute CMV-induced mononucleosis-like illness showed a pronounced transient increase in IgM anti-peptide P62 antibodies, with minimal changes in the IgG antipeptide level.

Previously, we showed that the ratio of IgG to IgM is less than 0.7 during the acute phase of EBV-induced IM and increases during convalescence to a value greater than 0.7 (1 1,12), where it apparently persists indefinitely. The ratio has been advocated as a useful diagnostic aid for infectious mononucleosis ( 12). When we analyzed the CMV-induced

IM, 8 of 10 patients had sufficient rises in their IgM anti- peptide P62 levels to give ratios below 0.7 during the acute illness despite elevated levels of IgG anti-peptide P62. In the two other two patients, one of whom was S.G. (Fig. 5B), the ratio remained high.

Direct Comparison of Antigens

Since similar antigens were recognized by IgM antibodies generated during acute EBV and CMV infections (Figs. 2 and 4), we immunoblotted sera from patients with these diseases side by side on the same cell extract for IgM antibodies. The cells used were from the K562 cell line, which contains no EBV or CMV sequences. Thus, any antigens detected would be normal cellular proteins.

The results of this experiment are shown in Figure 6A. Lanes 1 , 2, 9, and 10 were probed with sera from EBV-IM patients while the rest were from patients with acute CMV infection. The major antigen bands at 92, 82-77, 69, 62, and 58 kDa were recognized by the sera of patients infected with either virus. Some less intense bands were also seen in both patient groups, while others were found in only one or a few sera. These same sera, when developed for IgG antibodies in an immunoblot with the same K562 extract, were negative (not shown). Thus, acute infection by either virus apparently induces IgM but not IgG autoantibodies to a common set of proteins.

The same sera were also blotted on an extract of fibro- blasts productively infected with CMV in order to detect anti-CMV-specific antibodies. The immunoblots were developed for IgG antibody (Fig. 6B). All CMV patients had IgG antibodies to a prominent 50-kDa antigen and to another at

B 'r - c

I 1 1

I 1 I I

Oec May "Marl i p r Jun May Nov

1976 1977 1978 1974 1976 1973

Fig. 5. Anti-peptide P62 ELISA analysis of sequential serum samples from two patients with CMV-IM who had previous EBV infections. A: Patient L.S.; B: patient S.G.


K 562 Extract A CMV Infected

Cell Extract

IgM Detected IgG Detected

Number Serum 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 910

-77kD

-66kD

-45kD

--u E N CMV EBV

-77kD -66kD

-45kD

~

EBV CMV EBV infection

infection

Fig. 6. Immunoblot analysis of four sera from acute EBV-IM and six acute CMV-IM patients. The lanes containing the sera of each patient group are marked at the bottom. A: Sera were irnrnunoblotted on extracts of the K562

(EBV- and CMV-negative) cells. IgM antibodies to several cellular antigens were detected. B: The same sera were blotted on extracts from late CMV-infected cells. IgG antibodies were detected. All sera were used at 1/20 dilutions,

64 kDa. Little, if any, reaction was seen with the sera from the EBV-infected patients. We conclude that acute infection by CMV or EBV induces IgM autoantibodies reactive with a common set of autoantigens. In contrast, the IgG antibodies present during acute infection are specific for the virus causing the disease and have no reactivity with host components (1 3).

Inhibition of Antibody Binding With Peptide P62

We have previously demonstrated that peptide P62 will inhibit the binding to autoantigens of a number of IgM antibodies induced by EBV (13). This is again demonstrated in Figure 7, where two EBV sera were blotted on an extract of EBV-positive B cells. Lane 1 contains EBV serum from patient D.B. 15 days after onset of disease and lane 5 contains another acute EBV-IM serum. Lanes 2 and 6 have the corresponding sera containing 400 pg/ml of peptide P62. These experiments were done at a higher serum dilution than shown in Figure 2, so that only the strongest of the bands in Figure 2 are easily visible. The peptide inhibited antibody binding to bands at 82-77 kDa, which includes EBNA-1 and one or more cellular proteins (Rumpold and Rhodes, manuscript in preparation), and at 58-62 kDa.

Lane 3 was probed with serum from D.S. taken 12 days after the onset of his CMV illness. Lane 4 was probed with the same serum with added P62. Again, the 82-77-kDa bands and some proteins in the range 58-62-kDa are inhibited by the P62 peptide.

We also performed inhibition studies of the IgG antibodies seen in blotting experiments. Lane 7 contains serum from patient D.B. 514 days after disease onset, lane 9 is patient D.S. 417 days after onset of his CMV illness, and lane 11 is from a healthy adult with a long past EBV infection. Lanes 8, 10, and 12 contain the corresponding sera with 50 pg/ml of P62. Antibody binding to the EBNA-1 band at 77-kDa was inhibited by the peptide in all cases.

We conclude from these studies that CMV infection pro- duces IgM antibodies that also recognize the P62 synthetic peptide, which represents a major epitope in the EBV-encoded EBNA- 1 antigen. These CMV-induced antibodies also bind a series of cellular autoantigens that are identical in mobility to the autoantigens produced during EBV infection. Presurn- ably, EBNA-1, an as-yet unidentified CMV-encoded antigen and several human autoantigens share an epitope specificity identified by the P62 peptide.

462 Rhodes et al.

WiL2 Extract

DB2 DS2 101 088 DS7 0 6

IgM IgG Peptide P62

- + - + - + - + - + I +

1 2 3 4 5 6 7 8 9 1 1 1 0 1 2

Fig. 7. Inhibition of IgM and IgG antibodies from acute and convalescent specimens from patient DB and DS. Lanes 1 and 2 contain serum from patient D.B. obtained 15 days post-onset of disease (pod), lanes 3 and 4 are from patient D.S. 12 days pod, lanes 5 and 6 are from an acute IM patient 4 days pod, lanes 7 and 8 are from patient D.B. 514 days pod, lanes 10 and 11 are from patient D.S. 471 days pod, and lanes 11 and 12 are from a normal VCA-positive, EBNA-1-positive donor with a long-past EBV infection. Serum dilutions are 1/50 for the first six lanes and 11500 for the last six. The odd-numbered lanes contain only diluted serum, while the even- numbered lanes contain peptide P62 as well. The concentration of P62 is 400 p g h l in lanes 2, 4, and 6 and 50 pg/ml in lanes 8, 10, and 12. IgM antibodies are detected in the first six lanes and IgG in the remainder.

DISCUSSION EBNA-1 contains a region of over 200 amino acids that

include only glycine and alanine (6,21). Peptides made from

these sequences are immunoreactive with sera from VCA- positive individuals, and this region contains the major antigenic determinant of EBNA-1 (8-10,13,22). We have shown that almost all patients with acute infectious mononucleosis due to EBV have IgM antibodies to the synthetic glycine- alanine peptide P62 (9-12). Conversely, patients with a variety of other acute infections do not (12).

In this study, we examined sequential sera from 12 patients with either EBV or CMV infection for antibody to EBNA P62 and by immunoblots. One patient, D.B., had a typical case of primary EBV infection as manifested by a classical immune response to the VCA and EBNA antigens. D.B. had a standard response to the EBNA P62 (Fig. l), including an acute-phase IgM antibody to the peptide followed by a delayed IgG response to P62 as seen in the majority of acute cases of EBV-induced IM (9- 12). Immunoblotting experiments confirmed the cor- relation between the rise of IgM antibodies to several proteins in the Wi-L2 extract with anti-peptide P62 responses in the ELISA. Specificity of this response was demonstrated by inhibition experiments with the P62 peptide (Fig. 7).

We studied sequential sera from patient D.S., who was first infected with CMV and then several months later by EBV. During the primary infection with CMV, there were high levels of IgM antibody to the P62 peptide in this patient and also IgM antibodies reactive with at least six autoantigens. These autoantigens were of the same apparent molecular weights as those we previously reported to be recognized by IgM antibodies during acute EBV infection (1 3 ) . Further, as had been true in the EBV sera, the antibodies binding to these common antigens were inhibited by the P62 peptide. Thus, IgM antibodies of the same or a very similar set of specifici- ties to those produced during acute EBV infection were also produced during this patient’s infection with CMV.

We previously described a simple ELISA assay for EBV- induced IM that uses peptide P62 as the solid-phase antigen (9,11,12). The assay takes advantage of the fact that individuals with an initial EBV infection have a high TgM and low IgG signal, i.e., the ratio of IgG to IgM is less than 0.7 (1 1,12). During recovery, this ratio increases and becomes greater than 0.7, where it remains for life ( 1 1,12) (Fig. 1). The peptide P62 ELISA is remarkably accurate in diagnosing EBV infectious mononucleosis. The assay was able to identify greater than 98% of more than 200 clinical samples (12 and unpublished results) including a number of specimens that were negative in the standard heterophil assay (1 I , 12). The assay was also positive in some IM patients before the heterophile assay was scored as positive (1 1). The use- fulness of the EBNA- 1 glycine-alanine peptide to diagnose infectious mononucleosis has been confirmed by others (23).

We examined sera from 10 patients with CMV mononucleosis-like illness, all of whom had had EBV infections in the distant past. In every case, we found large increases in IgM anti-peptide P62 antibodies (assayed by ELISA and blotting) during the acute CMV infection. It should therefore be pos-

identical Antibodies in EBV and CMV Infection 463

Richman for providing the CMV virus and Dr. M. Lotz for providing the cells and help in their growth. We are grateful to Drs. M. Lotz and S. Fong for many helpful comments. We thank Rickey Jayne Richardson for typing this manuscript. We also thank the late Dr. Werner Henle for performing the EBV serology to diagnose many of the patients reported in this study. This is publication 5 149 BCR of the Department of Basic and Clinical Research of Scripps Clinic.

sible to develop an ELISA assay that will diagnose acute EBV and CMV infections. The assay as we presently use it should be useful for the detection of acute CMV illness in individuals with no previous EBV infection. However, the ratio method is not as accurate in diagnosing CMV mononucleosis- like illness when there has been previous EBV exposure, be- cause the initial level of IgG anti-P62 antibodies in patients with past EBV infections can vary by at least an order of mag- nitude. In our limited sample of 10 acute CMV sera with prior EBV, we found two that were scored as negative by the ratio method, even though both sera had high levels of IgM anti-peptide P62. A modification of the assay will therefore be necessary to increase its accuracy for this subset of patients.

Two mechanisms can be proposed to account for the pres- ence of similar or identical IgM antibodies during infection by the two different viruses. One possibility is that CMV encodes a protein or proteins with sequences similar to the glycine-alanine repeat in EBNA-I . Another is that CMV infection of individuals with past EBV infection reactivates the latent EBV. The antibody response in this latter case would be truly to EBV, not to CMV.

Several lines of evidence favor the former hypothesis. Pa- tient D.S., who had a CMV infection without previous exposure to EBV, exhibited significant IgM antibody to the P62 peptide during the CMV phase of his illness. Although these data are limited to one patient at present, they do imply that CMV itself is producing the antigenic stimulation.

A candidate CMV protein that could serve to induce the IgM antibodies has been described (24-26). CMV-infected cells contain a 2.2-kb mRNA, which has an open reading frame encoding a glycine-alanine region 17 amino acids long (15 glycines, 2 alanines) (24,25). Patients with acute CMV (or EBV) infections produce IgM antibodies that recognize synthetic peptide with this sequence (Smith, unpublished). The RNA transcribed from this region is apparently differen- tially spliced and present in four different CMV proteins (26). The peptide P62 inhibits binding of acute CMV sera to at least three CMV-encoded proteins (Rhodes, manuscript in preparation). These proteins are the same apparent molecular weight as three of the four proteins identified by Wright et al. (26). There are also four other open reading frames in CMV, which contain glycine-rich sequences, but it is not known if these proteins are immunologically active (27). In any event, it seems highly likely that the two viruses encode proteins with very similar antigenic epitopes.

ACKNOWLEDGMENTS

This work was supported in part by NIH Grants AR2 1 175 to J.H.V. and RROO833 to the General Clinical Research Cen- ter of Scripps Clinic and by the Johnson &Johnson Biotech- nology Center. G.R. was supported in part by Grant AR33489 from the NIH , a stipend from Johnson & Johnson, Inc . , and a grant from the Kieckhefer Foundation. We thank Dr. D.

REFERENCES

1. Niederman JC, McCollum RW, Henle G, Henle W: Infectious mononucleosis. Clinical manifestations in relation to Epstein-Barr virus antibodies. JAMA 203:205-209, 1968.

2. Henle WG, Henle E, Honvitz CA: Epstein-Bar virus specific diagnostic tests in infectious mononucleosis. Hum Path01 5 5 5 1-565, 1974.

3. de The G: Epidemiology of Epstein-Barr virus and associated disease in man. In The Herpesviruses (Vol. I). Plenum Press, New York, 1982,

4 . Henle W, Henle G: Immunology of Epstein-Barr virus. In The Herpes- virus (Vol. 1) . Plenum Press, New York, 1982, pp 209-252.

5. Stmad BC, Schusler TC, Hopkins RE Neubauer RH, Rabin H: Identi- fication of an Epstein-Barr virus nuclear antigen by fluoroimmunoelec- trophoresis and radioimmunoelectrophoresis. J ViroZ38:996- 1004, 1981.

6. Hennessey K, Kieff E: One of two Epstein-Barr virus nuclear antigens contains a glycine-alanine copolymer domain. Proc Nur Acad Sci USA

7. Billings PB, Hoch SO, White PJ, Carson DH, Vaughan JH: Antibodies to the Epstein-Barr virus nuclear antigen and to rheumatoid arthritis nuclear antigen identify the same polypeptide. Proc Nar Acad Sci USA 807 104-7 108, 1983.

8. Rumpold H, Rhodes GH, Bloch PL, Carson DA, Vaughan JH: The glycine-alanine repeating region is the major epitope of the Epstein- Barrnuclearantigen-lC EBNA-I. JImmuno/138:593-599, 1987.

9 . Rhodes GR, Houghten A, Carson DA, J Valbracht, Vaughan JH: Map- ping the antigenic region of Epstein-Barr nuclear antigen using synthetic peptides. In Herpesvirus, R Rapp, ed. Alan R. Liss, New York,

0. Rhodes G, Carson DA, Valbracht J , Houghten R, Vaughan JH: Human immune response to synthetic peptides from the Epstein-Barr nuclear antigen. JImmunol134:211-216, 1985.

1 . Smith RS, Rhodes G, Vaughan JH, Honvitz CA, Geltosky JE, Whalley AS: A synthetic peptide for detecting antibodies to Epstein-Barr virus nuclear antigen in sera from patients with Infectious Mononucleosis. J Infect Dis 154:885-889, 1986.

2. Geltosky JE, Smith RS, Whalley AS, Rhodes G: Use of a synthetic peptide-based ELISA for the diagnosis of lnfectious Mononucleosis and other diseases. J Clin Lab Anal 1:153-162, 1987.

13. Rhodes G, Rumpold H, Kurkel P, Patrick KM, Carson DA, Vaughan JJ: Autoantibodies in infectious mononucleosis have specificity for the glycine-alanine repeating region of the Epstein-Barr virus nuclear antigen. JExpMed 165:1026-1040, 1987.

14. Reedman BM, Klein G: Cellular localization of an Epstein-Barr virus EBV-associated complement-fixing antigen in producer and non- producer lymphoblastoid cell lines. Int J Cancer 1 1:499-520, 1973.

15. Dillner J, Stemas L, Kallin B, Alexander H , Klein G, Lerner R: Anti- bodies against a synthetic peptide identify the Epstein-Bar virus determined nuclear antigen. Proc Nat Acad Sci USA 82:4652-4656, 1984.

16. Honvitz CA, Henle W, Henle G , et al: Clinical and laboratory evaluation of cytomegalovirus-induced mononucleosis in previously healthy individuals. Medicine 65:124-133, 1986.

17. Henle W, Henle GE, Honvitz CH: Epstein-Barr virus specific diagnostic tests in Infectious Mononucleosis. Human Pulho/5:55 1-565, 1974.

pp 25-104.

805665-5669, 1983.

1984, pp 487-496.

464 Rhodes et al.

18. Menifield RB: Solid phase peptide I. The synthesis of a tetrapeptide. J Am Chem Soc 85:2149-2154, 1963.

19. Houghten RA, Chang WC, Li CH: Human B-endorphin. Synthesis and characterization of analogs iodinated and tritiated at tyrosine residues 1 and27. IntJPeprProteinRes 16:311-320, 1980.

20. Smith RS, Naso RB, Rosen J, et al: Antibody to a synthetic oligopeptide in subjects at risk for human immunodeficiency virus infection. J Clin Microbiol25: 1498- 1504, 1987.

21. Baer R, Bankier AT, Biggin MD, et al: DNA sequence and expression of the 895-8 Epstein-Barr virus genome. Nafure 310:207-211, 1984.

22. Thiele G, Bicak M, Grierson H, Lai P, Purtilo D: Antibody reactivity to a synthetic peptide (P62) of the Epstein-Barr nuclear antigen in sera of patients with X-linked lymphoproliferative syndrome. J Immunol Methods 100:249-257, 1987.

23. Linde A, Kallin B, Dillner J, et al: Evaluation of enzyme-linked im-

munosorbent assay with two synthetic peptides of Epstein-Ban virus for diagnosis of infectious mononucleosis. JInfect Dis 161:903-909, 1990.

24. Shaw SB, Rasmussen RD, McDonough SH, Staprans SI, Vacquier JP, Spector DH: Cell-related sequences in the DNA genome of human cytomegalovirus strain Ad169. J Virol55:843-848, 1985.

25. Staprans SI, Spector DH: 2.2 Kilobase class of early transcripts encoded by cell-related sequences in human cytomegalovirus strain AD169. J Virol57:591-602, 1986.

26. Wright DA, Staprans SI, Spector DH Four phosphoproteins with common amino termini are encoded by human cytomegalovirus AD 169. J Virol6:331-340, 1988.

27. Chee MS, Bankier AT, Beck S, et al: Analysis of the protein-coding content of the sequence of human cytomegalovirus strain AD 169. In Current Topics in Microbiology and Immunology, JM McDougall, ed. Springer-Verlag, Berlin-Heidelberg, 1990, pp 125- 169.

identical igm antibodies recognizing a glycine-alanine epitope are induced during acute infection...

Documents