VIROLOGY

188,

143-l

5 1 (1992)

Monoclonal IgM Antibodies from Cytomegalovirus-Infected Mice Recognize the GlcNAc-Containing Receptor Determinant of Murine CMV as well as Neutralizing Anti-CMV IgG Antibodies RAJESWARI Department

of Neurology,

Reed Neurological

M. H. RAVINDRANATH’ Research

Received

Center.

September

School 23,

AND

of Medicine,

199 1; accepted

MICHAEL University January

C. GRAVES of California,

Los Angeles,

California

90024-1769

10, 1992

This study examines monoclonal antibodies derived from mice at different time points after infection with attenuated murine cytomegalovirus (MCMV). The antibodies obtained from mice at 3 weeks p.i. were of IgG type (designated V-antibodies) and several could neutralize the virus. Those obtained at 5 weeks p.i. were of the IgM class (designated R-antibodies), bound to uninfected (MEF, mouse embryo fibroblast) cells, and thereby blocked MCMV plaque formation. In ELISA, the IgM monoclonals (R-antibodies) bound to GalBl-3GlcNAc and GalBl-4GlcNAc, the receptordeterminants for MCMV. Similarly, these GlcNAc-containing residues blocked the binding of purified IgM monoclonal antibodies (MAbs) to MEF. The R- and V-series of antibodies showed mutual binding activities; for example, IgM MAb R-2D8 bound specifically to four (V-8C4, V-i C7, V-8C7, V-9C5) of six neutralizing IgG MAbs in ELISA. The same neutralizing IgG MAbs bound to antireceptor IgM antibodies in an immunoblot assay. This suggests that the IgM monoclonals directed against the known cell surface receptor determinant are anti-idiotypic antibodies against neutralizing antiviral IgG antibodies. The neutralizing antiviral IgG MAbs bound to 60- and 66-kDa MCMV polypeptides on Western blots, suggesting that these viral polypeptides may be important in MCMV binding to this receptor. The R-series might represent anti-idiotype antibodies capable of down-regulating antiviral V-antibodies and may also represent a mechanism for the induction of IgM autoantibodies directed at cell surface glycolipids present in autoimmune CMV-associated neuropathies. 0 1992- Academic Press, Inc.

INTRODUCTION

For example, influenza virus may bind to a single receptor determinant occurring in more than one species of glycoconjugates (Rogers et al., 1983; Paulson, 1985; Suzuki et al., 1987; Yamada et al., 1990). Recently, we have found that attenuated murine CMV (MCMV) infection of MEF (mouse embryo fibroblasts) involves virus binding to GlcNAc-containing oligosaccharide residues on the cell surface (Ravindranath and Graves, 1990). Free GlcNAc and GlcNAc-containing oligosaccharides prevented virus binding to MEF. Enzymatic removal of GlcNAc residues from the cell surface or blocking cell surface GlcNAc residues with specific lectins also prevented virus binding. The virus was also found to bind to a GlcNAc-containing glycolipid extracted from host cells, indicating that such a molecule could possibly function to carry a receptor determinant for MCMV (Ravindranath and Graves, 1990). Since the diseases of the peripheral nerve which are associated with CMV infection are also associated frequently with autoantibodies directed at carbohydrate structures similar to the MCMV receptor determinant, we wondered if the virus could induce antireceptor autoantibodies. In this paper we report experiments aimed at determining (1) if antireceptor antibodies are produced during the course of MCMV infection, and (2) if any of these antibodies are anti-idiotypic antibodies capable

Cytomegalovirus (CMV) infection is associated with autoimmune inflammatory diseases of peripheral nerve which occur sporadically and also in patients with AIDS (Alford and Britt, 1985). Several mechanisms have been proposed for virus induction of autoimmunity, including virus acting as adjuvant (Pearson, 1956; Weigle, 1980; Negoro et al., 1979) viral release of cryptic antigens (Weller et al., 1989), molecular mimicry or shared viral and self-epitopes (Bartholomaeus et a/., 1988) and anti-idiotypic antibodies, which recognize the host cell viral receptor (Erlanger et a/., 1986; Marriot et a/., 1987; Kauffman et al., 1983; Gaulton et al., 1985; Krah and Choppin, 1988). Many examples of antireceptor and anti-idiotype antibodies induced by viral infection are known (Krah and Choppin, 1988; McClintock et al., 1986; Powell et al., 1989; Beauclair and Khansari, 1990; Ertl, 1989; Rogers and Paulson, 1983). The study of antireceptor antibodies is greatly facilitated when the specific molecular structure of the receptor determinant is characterized. Specifically, it is important to distinguish the task of determining the structure of the receptor determinant from that of identifying the protein or other molecule which carries it. ’ To whom

reprint

requests

should

be addressed, 143

0042.6822/92

$3.00

CopyrIght 0 1992 by Academic Press. Inc All rlghts of reproduction in any form reserved

144

RAVINDRANATH

of recognizing neutralizing antiviral antibodies. We describe antiviral and antireceptor antibodies from MCMV-infected mice. To better study antibody specificities, we derived hybridomas from spleens of mice and screened for clones secreting antiviral and also antireceptor monoclonal antibodies (MAbs).

MATERIALS

AND METHODS

Virus The attenuated MCMV (the source of the virus is Dr. Jack Stevens: it has been widely refered to as the “Smith” strain) was grown, assayed, purified, and stored as previously described (Ravindranath and Graves, 1990; Howard et al., 1978).

Infection

GRAVES

says for many months. In protocol No. 2, the receptorblocking assay, we tested serum for the ability to react with the virus receptor on the cell surface and thus block plaque formation by MCMV. The pooled and diluted sera (1:8) were directly layed over MEF and the monolayer was incubated at 25” for 1 hr.’ After gentle and repeated washing (X4), the virus (with known PFU) was overlayed on the monolayers and the reduction in PFU was determined using the standard plaque assay. The percentage inhibition of infectivity of sera was measured as (PFU with control sera - PFU with infected sera) x ,oo, (PFU with control sera) The results are plotted as the percentage infectivity on different days postinfection

of mice

Inbred Swiss Webster (male; 4 weeks old) and BALB/c (female; 4-6 weeks old) mice were purchased from Charles River and maintained under conventional conditions at the Reed Neurological Research Center vivarium. Swiss Webster mice were inoculated i.p. with 3 x lo5 plaque forming units (PFU) of purified attenuated MCMV or saline. Control mice were injected with saline on Day 0. For antibody studies, one control and two infected mice were sacrificed and bled by cardiac puncture on alternate days from Days 0 to 44. After serum collection the mice were eliminated from the experiment. The serum was stored at -20”. For hybridoma production, BALB/c mice were inoculated i.p. with 5 X 1O5 PFU of the virus.

Assessment of inhibition from infected mice

AND

of viral infectivity

by sera

For this purpose, a group of 70 mice were inoculated with virus and bled at designated times p.i. Control sera were obtained from a similar group of saline-injected mice which were bled at the same times. Previously described protocols (Ravindranath and Graves, 1990) were used for the plaque-reduction assay and are designed to test serum reaction either with virus or with the cell surface receptor. In protocol No. 1, virus neutralization, 10 ~1 of virus stock of known dilution was mixed and incubated with 100 ~1 of pooled and diluted sera (1:4) for 1 hr at 4”. After incubation, the virus-serum suspension was diluted to 1 ml and 200 ~1 of this suspension was overlaid per monolayer and the plaque assay was carried out as previously described (Graves et a/., 1978; Howard et al., 1978). Under the conditions of this assay, we did not detect neutralizing antiviral antibody after Day 21 p.i., probably due to the low sensitivity of the assay, as antiviral antibody can be detected by immunofluorescence as-

inhibition of

Assessment of monoclonal antibodies for virus neutralization, and receptor-blocking activity Plaque assay and plaque-reduction assay (Ravindranath and Graves, 1990; Graves et a/., 1978) were done as indicated above, but the MAb to be tested for neutralizing activity was incubated with 10 ~1 of virus stock for 1 hr at 4”, and then inoculated on MEF. Controls included virus without antibody and antibody without virus. The R-series antibodies were also tested for virus-neutralizing activity. The receptor-blocking assay was done also as previously described by overlaying MEF monolayers with serially diluted MAb. After incubation at 25” for 1 hr and repeated (4X) washing, a standard viral inoculum was added and the reduction in PFU determined (Ravindranath and Graves, 1990). Neutralizing antibodies (viral (V)-antibodies) were also tested in this assay for antireceptor activity. The percentage inhibition of infectivity of monoclonal antibodies was measured as (PFU of the virus - PFU with monoclonal antibody) (PFU) of the virus)

Production

of monoclonal

x ,oo,

antibodies

Spleens were harvested from infected BALB/c female mice at Days 2 1 and 36 p.i. A single-cell suspension was prepared and fused to myeloma cell line SP2/ 0 Ag14 following a conventional procedure (Coligan et 2 IgM overlay and incubation were carried out at 25” because the IgM behaved like cryoglobulins and formed precipitates at 4”. 3 Sera of control mice per se showed a range of 25 to 40% reduction in PFU when compared with wells containing a test batch of virus alone. In all our studies on sera. the percentage inhibition of infectivity of infected sera was determined in comparison with the control sera as shown In the formula.

MCMV ANTIVIRAL AND ANTIRECEPTOR ANTIBODIES

al., 1991). After HAT medium selection, hybridomas in the culture wells were periodically screened for antibody against purified MEF-CMV and also against the synthetic virus receptor analogues, GalBl-3GlcNAc and GalBl-4GlcNAc (a gift from Dr. James C. Paulson, Cytel, La Jolla, San Diego, CA) in an ELISA (Coligan et a/., 1991). Antibody-producing hybridoma cultures were cloned by limiting dilution in 96-well plates containing 100 ~1of feeder cells (5 X 1O* BALB/c peritoneal wash cells/well) in RPMI 1640 with 10% FCS, 100 U/ml of penicillin, and 100 pg/ml of streptomycin. After selection, antibody-secreting clones were isolated and propagated in RPMI with 10% FCS and antibiotics, and supernatants were concentrated to one-tenth volume with an Amicon filter (cutoff size 10 X 105). Purification of IgG MAbs The individual antibody-secreting clones (3 X 1O6 cells) were inoculated i.p. into BALB/c mice primed with 0.5 ml Pristane (1,2,6,10,14tetramethyl pentadecane, Sigma). Ascites fluid was collected from 8 to 16 days after implantation of cells. After centrifugation, the supernatant was stored at -20” until further use. IgG antibodies were purified from ascites (Temponi et a/., 1989) using caprylic acid (25 PI/ml of sample), centrifugation (10 X 1O3g for 30 min), and 45% ammonium sulfate precipitation. The IgG pellet was collected by centrifugation (5000 g for 15 min), dialyzed against PBS (pH 7.00) and stored as aliquots at -70”. IgG concentration was determined by titrating against standard mouse IgG in an ELISA assay.

145

diamine, 7.7 mg/lO ml, in citrate buffer, pH 5.00) was added and incubation continued for 10 min. The color reaction was arrested with 6 N H,SO,. The color formed was measured with a v,,,,, kinetic microplate reader (Molecular Devices) at 490 nm.

ELISA of MAb (R-antibodies) for binding synthetic receptor determinants

to

An ELISA was performed using Galpl-3GlcNAcO(CH,),COOCH, and GalPl-4GlcNAcO(CH,),COOCH, as antigens. Antigen coating was done by adding a suspension of 100 ~1of each glycoconjugate in ethanol (1 pg/well) to microtiter plates which were then dried in vacua. The antigen-coated plates were treated with 1% gelatin at 37” for 1 hr to block nonspecific absorption. MAbs to be tested were serially diluted, added to wells, and incubated for 1 hr at 37”. After washing five times with 0.1% gelatin in PBS, and incubation with peroxidase-coupled goat anti-mouse IgG (1: 1000) or IgM (1:2000) or IgG/lgM (1:5000, from Jackson Immunoresearch Labs), as described previously, color was developed with o-phenylenediamine in citrate buffer and the OD was determined at 490 nm.

ELISA of antireceptor MAb (R-antibodies) neutralizing antiviral MAb V-antibodies

binding to

IgM was separated by gel filtration on Sephacryl 300, eluted with PBS (pH 7.00) concentrated to onetenth volume with an Amicon filter (cutoff size 10 X 105), aliquoted, and finally stored at -70” (William and Chase, 1967). The concentration of IgM was measured by titrating against standard mouse IgM in an ELISA.

Six neutralizing monoclonals (V-8C4, V-l C7, V-9C5, V-l OE9, V-8C6, V-8C7) were diluted (1 :16) in PBS, and 50 ~1was added to microtiter wells for overnight incubation at 4”. After the supernatant was removed, wells were blocked with 1% gelatin in PBS for 1 hr at 37”. One hundred microliters of antireceptor MAbs (R-2D8, R-9Ell) at 1:lOO dilution was added and incubated for 1 hr at 37”. The binding affinity was assessed using peroxidase-coupled goat anti-mouse IgM (1:2000). Wells untreated with antireceptor antibodies and those treated with unrelated IgGs were used as controls. Neutralizing antibodies without antireceptor antibodies but with peroxidase-coupled IgM gave OD values ranging from 0.080 to 0.140.

ELISA of antiviral

Western

IgM MAb purification

MAbs (V-antibodies)

Microtiter plates were coated with purified MCMV (1 pg/well in PBS) and incubated at 4” overnight. After PBS was removed, the wells were blocked with 1% gelatin for 1 hr at 37”. Appropriate dilutions of MAbs were added and incubated for 1 hr at 37”. The plates were washed with 0.19/o gelatin in PBS five times before addition of the peroxidase-coupled goat antimouse IgG (1 :lOOO) or IgM (1:2000) (Sigma, St. Louis, MO) or IgG/IgM (1:SOOO) (Jackson lmmunoresearch Lab). After 1 hr of incubation at 37”, the plates were washed five times, and the substrate (orthophenylene-

immunoblotting

Proteins from purified virus were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis using 8.5% resolving and 3.50/a stacking gels (Laemmli, 1970). Viral proteins were electrotransferred to nitrocellulose paper by the method of Towbin et al. (1979) using 25 V for 15 hr. Protein transfer was assessed by staining the nitrocellulose strips with 0.1% Fast Green in 10% acetiuc acid in 40% methanol and the strips were compared with Coomassie bluestained protein bands. Replicas were treated with antiviral MAbs at a dilution of 1: 100 in sealer bags for 1 hr

146

RAVINDRANATH

at 37”. The binding of the antibodies was assessed by incubating the strips with ‘251-labeled goat anti-mouse IgG (1 X lo5 cpm) or peroxidase-coupled anti-mouse IgG (1:lOOO) for 1 hr at 37”. Peroxidase activity was measured with diaminobenzedine in 0.05 M Tris-HCI. Radioactive antibody binding proteins were identified after autoradiography as described earlier (Ravindranath and Graves, 1990). Direct binding

GRAVES

j jJJ\dl,;, 0

4

6

12

assay

Binding of neutralizing IgG MAbs to antireceptor IgM MAbs was also determined with a nitrocellulose immunodot assay. Purified R-2D8 and R-9El 1 IgM antireceptor MAbs were diluted 1:20 and 10 ~1 (20 ng) was spotted on nitrocellulose strips. After blocking with 3% gelatin, the strips were overlaid with different IgG neutralizing antiviral MAbs (1 :lOO) for 1 hr at 37”. After washing extensively, the strips were incubated in lz51labeled goat anti-mouse IgG (1 X lo5 cpm) for 1 hr, washed, and autoradiographed as described earlier. Strips containing antireceptor IgM but no IgG-neutralizing antibodies were used as controls. The bound radioactivity of treated and untreated spots was measured in a gamma counter. RESULTS MCMV-infected mice produce early antiviral antibody and late antireceptor antibody The objective of the first series of experiments was to look for antireceptor antibodies in pooled polyclonal

16

20

24

26

32

36

40

44

Days Post- infection

radioimmunoassay

In this assay, binding of antireceptor IgM to MEF and blocking of this binding with GlcNAc-containing synthetic receptor determinants were measured. Synthetic oligosaccharides (100 pg of either Galfll3GlcNAc or Gal/31 -4GlcNAc in ethanol) were coated in Eppendorf tubes and dried in vacua. Purified antireceptor IgM MAb (R2D8) diluted at 1: 1600 in 300 ~1 PBS (with 1 m/W CaCI,, 0.05% BSA, pH 7.2) was added and incubated for 1 hr at 4” or 25”. The MAb incubated without carbohydrates was a control. Treated and untreated MAbs (100 ~1) were added to MEF cell suspension (5 X 1O4cells) and incubated under constant agitation for 1 hr at 4” or 25”. The cells were extensively washed (4X) with PBS with 0.1% BSA and then incubated with 1251-labeled goat anti-mouse IgM (60-80 X 1 03/cpm) for 1 hr at 4” or 25”. The cells incubated with radioiodinated anti-mouse IgM were maintained as control and all the values were corrected against this control. The cells were washed three times and the radioactivity in the cell pellet was measured in a gamma counter. lmmunodot

AND

FIG. 1. Plaque-reduction assay of polyclonal sera forvirus-neutralizing and receptor-blocking activities obtained at the indicated days p.i. from MCMV-infected Swiss Webster mice. Neutralizing antibodies were assessed by incubating virus inocula with antisera (dilution 1:4) prior to plaque assay on monolayers of MEF. Receptor-blocking antibodies were assessed after coating MEF monolayers with antisera (16 dilution) and washing, prior to adding virus inoculum. The control values were used to calculate the percentage of inhibition, as described under Materials and Methods (see also footnotes 2 and 3). Virus-neutralizing antibodies (line) were detected early, at Days 7, 15, and 21, and receptor-blocking antibodies (bars) were detected later, at Days 26, 33, and 37.

sera of CMV-infected mice, which were bled on alternate days from Day 0 to Day 44 p.i. The results are presented in Fig. 1. Sera from Days 2 to 22 p.i. had detectable virus-neutralizing activity, but not receptorblocking activity (Fig. 1). However, sera collected at later times, up to Day 41 p.i., did have measurable activity in the receptor-blocking assay (Fig. 1). The lack of measurable virus neutralization by these sera may be due to the lower sensitivity of the plaque reduction assay, compared to immunofluorescence assays, or to the presence of anti-idiotype antibodies which neutralize the virus-neutralizing antibodies. To further characterize the antiviral (V) and antireceptor (R) antibodies formed in the course of MCMV infection, we prepared monoclonal antibodies from mice at the early (21 days) and late (36 days) times p.i. Monoclonal

antiviral

(V) antibodies

Table 1 shows the frequencies of hybridomas produced from spleen of mice on Days 21 and 36 postinfection. The antibodies secreted by all hybridomas were analyzed. Six of 80 hybridomas developed from spleen cells obtained from mice on Day 2 1 p.i. showed neutralizing activity. Since they produced antiviral antibodies, they are designated as V-series. Ten of 18 hybridomas developed from spleen cells obtained from mice on Day 36. These hybridomas secreted antireceptor antibodies and hence are designated as R-series. All except one clone (V-5G4) secreted the IgG class of antibodies. The antibody activities were assessed after a month of plating and antibody secretion

MCMV TABLE

ANTIVIRAL

AND

1

FREQUENCIES OF HYBRIDOMA PRODUCING VIRUS-NEUTRALIZING MONOCLONAL (V-SERIES) AND ANTIRECEPTOR MONOCLONAL ANTIBODIES (RSERIES) Number containing Treatments Plating (10 plates) Hybridoma: total Hybridoma: neutralizing Hybridoma: antireceptor Gall -3GlcNAc Gall -4GlcNAc

antibodres antibodies

of wells hybridomas

Day 21 p.i.

Day 36 p.1.

960 80 6

960 18 0

0 0

10 10

increased steadily during the first 10 weeks after plating and remained steady thereafter. The results in Table 2 show that the neutralizing antiviral IgG MAbs bound to MEF-CMV in ELISA as well as inhibited infectivity by binding to virus in a PFU assay. There is no correlation between the ELISA OD and plaque inhibition, which is not unexpected in view of the differences in the nature and sensitivity of the assays; for example, an antibody with high ELISA might bind to the virus in a way which is not effective in neutralizing infectivity. The virus binding affinity and inhibitory potency of these six IgGs (V-8C4, V-lC7, V-9C5, V-l OE9, V-8C6, and V-8C7) did not change after subcloning or storing the hybridomas at -70” or with time.

Production of ascites and purification characterization of neutralizing MAbs

and

Ascites fluid was collected from mice inoculated with hybridoma clones V-l OE9 and V-8C6. Ascites of V-lOE9 was purified and the IgG concentration was 1.25 mg/ml. Upon testing neutralization activity in the plaque-reduction assay, it was noted that 1.5 pg of the purified antibody reduced viral PFU by 85% under the conditions of this assay. As shown in Table 2, all the IgG MAbs which bind to virus in ELISA also can neutralize virus. These antibodies may block infectivity of the virus by binding either to the receptor binding domain (RBD) on the virus or possibly to other proteins in the close proximity of the RBD. The idiotype of the neutralizing antibodies directed specifically toward the RBD may resemble the receptor determinant, and the antireceptor antibodies may bind to the idiotype of such neutralizing antibodies (Gaulton et al., 1985). Therefore, one way to further study the viral receptor binding domain is to obtain antireceptor MAbs and to determine if they recognize any of the neutralizing antiviral MAbs.

ANTIRECEPTOR

ANTIBODIES

147

Production and characterization antireceptor (R) antibodies

of monoclonal

Because Day 36 p.i. mice produced polyclonal antireceptor sera, they were used for developing hybridomas. The antibodies secreted by these hybridomas (designated R-series) were screened in ELISA against synthetic oligosaccharides simulating glycosyl receptor determinants of attenuated MCMV, namely Gal/31 3GlcNAc and GalPI -4GlcNAc (Table 3). Interestingly, all of the MAbs positive in the screening test were of the IgM class and there was no class switching during growth in culture for more than 4 months. The receptor-blocking activity of these MAbs on the plaque reduction assay is shown in Table 3.

Purification IgM MAbs

and characterization

of antireceptor

Ascites fluids were collected from mice inoculated with hybridomas (R-2D8 and R-9Ell). IgM was purified from the ascites fluid obtained with R-2D8 clones. The purified IgM was serially diluted and tested against Galbl-3GlcNAc and GalPl-4GlcNAc in an ELISA. The purified IgM bound to both oligosaccharides equally. In the plaque-reduction assay, it was noted that 0.8 pg of purified R-2D8 IgM reduced PFU by 80%.

GlcNAc-containing oligosaccharides absorb out binding activity of antireceptor IgM MAb to MEF in a radioimmunoassay We tested binding of the antireceptor IgM MAbs to a suspension of MEF at 4” and 25”, after preabsorbing

TABLE

2

VIRUS BINDING (AS MEASURED BY ELISA)AND INHIBITION OF INFECTIVITY (AFTER BINDING TO VIRUS) BY IgG MONOCLONAL ANTIVIRAL (MEF-CMV) ANTIBODIES’ SECRETED BY DIFFERENT HYBRIDOMAS (SOURCE: SPLEEN DAY 2 1 P.I.) Antiviral IgG-secreting hybridoma ID V-8C4 v-lC7 v-9c5 V-l OE9 v-8C6 V-8C7 Others (n = 74)

ELISA OD (range n = 3) 0.46-0.75 1.31-2.20 1.96-2.10 1 .l O-2.00 1.68-2.00 1.08->2.00 0.00-0.10

Percentage inhibition PFU (range n = 3) 68-70 60-64 64-76 72-84 60-68 50-72 0

‘Antibodies were not diluted for ELISA. Fifty microliters of media from wells containing hybridoma was used directly for ELISA. Values were corrected for background absorbance (~0.1) of culture medium. For PFU inhibition, 50 ~1 of media was used after diluting fourfold.

148

RAVINDRANATH TABLE

TABLE

ELISA OD range (n = 3) ID

R-2D8 R-9Ell RlB8 R-5C9 R-8G7 R-2D5 R-lG9 R-9E7 R-l B9 R-8G5 Others

Galpl-3GLcNAc 0.44-0.86 0.70-l .15 0.54-l .03 0.48-l .13 0.49-0.86 0.67-0.85 0.46-0.73 0.53-0.80 0.28-0.60 0.36-0.56 0.02-0.08

Percentage inhibition PFU range (n = 3)

GalPl-4GlcNAc 0.47-0.64 0.60-0.78 0.36-0.85 0.54-0.72 0.48-0.64 0.51-0.65 0.43-0.62 0.47-0.68 0.28-0.59 0.41-0.63 0.02-0.08

70-75 50-55 45-50 45-50 50-55 45-50 25-30 50-55 45-50 1 O-25 0

the antibodies with Gal01 -3GlcNAc or GalPI -4GlcNAc residues (Fig. 2). The experiment carried out at 4” shows that Galpl-3GlcNAc absorbed 100% of the MEF binding activity of MAb R-2D8, whereas Galpl-4GlcNAc reduced the binding by 700/o, under the conditions of this assay. These results indicate that the relative receptor specificity of IgM MAb R-2D8 is

CPM x 10-3

0

MEF*(Gall-4GlcNAc+Ab)

4c

25 c Incubation

temperature

FIG. 2. Blocking of binding affinity of antireceptor MAb R-2D8 to MEF with GlcNAc-containing synthetic receptor determinants, measured using 1251-coupled anti-mouse IgM. Values corrected against controls (MEF + radioiodinated anti-mouse which included MEF and anti-mouse IgM. R-2D8 binding carried out at 4” and 25”. Each value represents the mean of cates and the mean is less than 2 SD.

4

BINDING OF ANTIRECEPTOR IgM MAbs (R-2D8; R-9Ell) TO NEUTRALIZING ANTIVIRAL IgG MAbs IN AN ELISA Antiviral antibody (IgG)-secreting hybridoma ID V-lOE9 v-8C6 V-8C4 v-9c5 v-lC7 V-8C7

ELISA

OD (mean

(IgM) as were IgM), was tripli-

+ SD) (n = 8)

R2D8 0.220 0.420 0.380 0.420 0.340 0.410

?I k f f f 5

R-9Ell 0.029 0.234 0.130 0.086 0.056 0.034

Note. Values are mean + SD corrected bancy. Coated antigens are IgG antibodies antibodies were diluted 1 :lOO. For details ods.

a Antibodies were not diluted for ELISA. Fifty microliters of media from wells containing hybridoma were used directly for ELISA. The values were corrected for background absorbancy (~0.1) of culture medium. For PFU inhibition 50 ~1 of media was diluted fo fourfold, before use.

6,

GRAVES

3

CMV RECEPTOR DETERMINANT BINDING (AS MEASURED BY ELISA) AND INHIBITION OF INFECTIVITY (AFTER BINDING TO TARGET CELLS) BY IgM MONOCLONAL ANTIRECEPTOR ANTIBODIES~ SECRETED BY DIFFERENT HYBRIDOMAS (SOURCE: SPLEEN DAY 36 P.I.)

Antireceptor IgM-secreting hybridoma

AND

0.420 0.580 0.980 0.790 0.700 0.760

f f f + 2 t

0.071 0.285 0.096 0.136 0.060 0.088

for background absordiluted (1 :16) in PBS. IgM see Materials and Meth-

similar to that of attenuated MCMV for binding to these synthetic receptor determinants (see Ravindranath and Graves, 1990).

Antireceptor IgM MAbs R-2D8 and R-9Ell some neutralizing antiviral IgG

bind to

Previous reports on other viruses indicated that antireceptor antibodies occurring late after viral infection may represent the anti-idiotypic second antibody directed at the idiotype of the first (neutralizing antiviral) antibody. Therefore it would be of interest to determine if our antireceptor antibodies bind to any of the antiviral MAbs. We devised two kinds of experiments to detect such binding, based on the observation that the antiviral MAbs are IgG and the antireceptor MAbs are IgM in class. In the first experiment, the neutralizing IgG MAbs were used as antigens in ELISA for antireceptor IgM MAbs. The results (Table 4) reveal that of the six neutralizing IgG MAbs (V-series antibodies) tested, two (V-8C6, V-l OE9) were poorly recognized by both antireceptor antibodies (R-2D8 and R-9El l), and four were recognized by R-9El 1 with varying activity (V-8C4 > V-9C5 > V-8C7 > V-l C7). Binding of V-series antibodies to R-2D8 cannot be compared with the binding of these antibodies with R-9El 1. Among various V-series antibodies, V-9C5, V-8C6, and V-8C7 showed better binding with R-2D8 in ELISA. In order to further confirm the interaction between the two series of antibodies, another series of experiments was carried out. In the second experiment (immunodot assay), antireceptor IgM MAbs, R-2D8 and R-9El 1, were used as antigens (20 ng) on nitrocellulose strips and overlaid with neutralizing IgG MAbs (1: 100). The purpose of using dot spot analysis instead of ELISA is to directly visualize the binding of IgG antibodies to the antireceptor IgM

MCMV 2500 IgG binding I

ANTIVIRAL

AND

ANTIRECEPTOR

ANTIBODIES

149

45-kDa proteins, suggesting that these antibodies recognize an epitope common to other proteins.

kpm)

may

2000

DISCUSSION

1600

1000

600

:

0I v-6C4

V-lC7

Neutralizing

v-m7

V-&Z6

Monoclonal

V-lOE9

IgGs

v-six

(1:lOO)

FIG. 3. Binding of neutralizing IgG MAbs (V-8C4, V-lC7, V-8C7, V-8C6, V-lOE9, and V-9C5) to antireceptor IgM MAbs (R-2D8 and R-9El 1) on nitrocellulose. IgG binding was measured with 1z51-coupled anti-mouse IgG. Values were corrected against controls, which included radiolabeled antibody binding to nitrocellulose and unrelated IgM MAbs. Values are means of triplicates and the mean is less than 2 SD.

monoclonals. The results (Fig. 3) confirmed the strong mutual reactivity of the neutralizing antibodies with the antireceptor MAbs. As noted earlier, V-8C6 and V-l OE9 did not bind to antireceptor MAbs, and V-8C4, V-l C7, and V8C7 showed high binding for R-2D8 and R-9El 1 on nitrocellulose strips. Since monoclonal IgMs did not bind to unrelated IgG antibodies (data not shown), any nonspecific binding of IgM to IgG antibodies is excluded. High binding of V-9C5 with R-2D8 and R-9El 1 in ELISA and minimal binding of the same antibody with antireceptor antibodies in a nitrocellulose assay could be due to an inherent discrepancy between the two assays. Neutralizing antiviral kDa virion proteins

antibodies

Recently, we have obtained evidence showing that the receptor determinant of murine CMV is a GlcNAccontaining oligosaccharide (Ravindranath and Graves, 1990). Removal of the GlcNAc residues from the target cell surface or blocking the same with GlcNAc-specific lectins prevented MCMV binding to MEF. The antireceptor antibodies described in this study are similar to the GlcNAc-binding lectins (Ravindranath and Graves, 1990) in that they recognized GlcNAc-containing oligosaccharides in an ELISA and their binding to target cells was blocked by the GlcNAc-containing oligosaccharides in a radioimmunoassay. These antibodies were of the IgM class, which is not unexpected, since carbohydrate and glycolipid binding antibodies are commonly IgM (Zenita et al., 1990). The relatively late appearance of the IgM antireceptor antibodies, after the peak of neutralizing antiviral IgG, suggested an anti-idiotype mechanism. If they were induced by another mechanism, such as virus acting as adjuvant, or viral cytolysis releasing cryptic antigens, one might have expected the peak response to be concurrent with that of antiviral antibodies. Antiidiotype antibodies would be expected to appear later than and might down-regulate antiviral-neutralizing an-

bind to 60- and 66-

Since V-8C4 and V-l C7 avidly recognize antireceptor IgM MAbs R-2D8 and R-9E11, they may specifically recognize the virion receptor binding domain. To identify the MCMV polypeptide which carries the receptor binding domain, Westerns were stained with MAbs V-8C4, V-8C7, V-l C7, V-9C5, V-8C6, and V-l OE9. The results presented in Fig. 4 reveal that all the neutralizing IgGs bound to viral proteins. The irrelevant IgG antibody as a control did not stain with any protein. None of the V-antibodies bound to protein fractions extracted from uninfected MEF cells (data not shown). V-8C4 and V-l C7, the MAbs avidly recognized by antireceptor IgMs, bound to 66- and 60-kDa proteins only, suggesting that these polypeptides may contain the receptor binding domain. The neutralizing IgGs not recognized by antireceptor IgMs bound to three or more protein fractions which include 200-, 66-, and

ff6976660-

C -

45-

-

FIG. 4. Binding of neutralizing IgG MAbs (V-8C4, V-8C7, V-lC7, V-9C5, V-8C6, V-l OE9) to Western blots of purified MEF-CMV proteins The first two lanes are electropherograms of standard (Std) and viral proteins from purified attenuated MEF-CMV stained with Coomassie blue. The antIvIral MAbs bound to fractions with molecular weights of 66 and 60 kDa. Some of the antibodies (VlOE9 and V-lC7) bound to a fraction with a molecular weight of 45 kDa. Nonantiviral IgG antibodies drd not stain any of the viral proteins. They were not included in the ftgure due to poor quality and background. The neutralizing antibodies did not stain uninfected cell proteins.

150

RAVINDRANATH

tibodies. Possibly for the same reason the antiviral antibodies could not be detected from Day 21 onward. Furthermore, the two antireceptor IgM MAbs (R2D8, R-9Ell) recognize four (V-8C4, V-lC7, V-9C5, V-8C7) of six neutralizing IgG MAbs in ELISA. Three (V-8C4, V-lC7, V-8C7) of these four neutralizing IgG MAbs bound avidly to the antireceptor IgM MAbs in an immunodot assay. These observations suggest that the antireceptor antibodies are anti-idiotypic antibodies to the neutralizing antibodies. The neutralizing antiviral antibodies which bind to the antireceptor antibody may specifically recognize that portion of the virus responsible for cell binding. This antibody might be used to identify the virion protein which carries the receptor binding specificity. Our experiments with Western blots of viral protein electropherograms show that these MAbs bind to viral polypeptides of 60 and 66 kDa. More work is needed to determine if these viral polypeptides carry receptor binding specificity. Binding of herpesviruses to cells is complex and probably involves several interactions of viral proteins to cellular receptors, some for initial recognition and binding, and others for fusion and penetration of virus into cells. Loh and co-workers (Loh and Qualtiere, 1988; Loh eT al., 1988) reported that MCMV has a major surface glycoprotein which is initially synthesized as a 128-kDa glycoprotein and then is processed with intrachain disulfide bonding and proteolytic cleavage to form gp150 on mature virions. In the presence of reducing agents, gpl50 can be resolved into a number of bands of reported molecular weights of 150, 105, and 52 kDa. We do not know if bands identified by our MAbs correspond to any of those identified by Loh et a/. (1988). It is likely that different viral proteins may serve receptor binding functions, and others may be needed for viral penetration into the cell. The relationship among 66-, 60-, and 45-kDa proteins deserves further investigation. Since antiviral monoclonals that bound to anti-Ids also bound to 60and 66-kDa proteins, it is suggested that these proteins may contribute to the receptor binding domain. Since we have used a reducing gel, it is highly likely that these polypeptides may be subunits of a large receptor binding protein. These speculations need further investigation. There is no doubt that the antiviral antibodies will be excellent reagents for identifying these proteins produced in an in vitro system using viral genome. The finding of anti-idiotyic, antireceptor antibodies following MCMV infection might explain how this virus can induce autoantibodies. The anti-idiotype antibody cross-reacts with a self-antigen, the GlcNAc-containing glycoconjugate on the target cell surface. Our results favor a viral etiology of inflammatory demyelinat-

AND

GRAVES

ing polyneuropathies, amyotrophic lateral sclerosis, and other diseases of motor neurons, wherein patients have autoantibodies directed against carbohydrate residues of glycoconjugates (Drachman and Kuncl, 1980). Antibodies to sialyl glycoconjugates are found in these patients (Salazar-Grueso et al., 1989). Since human CMV infection is closely linked to autoimmune neuropathies (Bishopric et a/., 1985), it is possible that the mechanism of sensitization involves the formation of anti-idiotypic antibodies in response to viral infection, just as we have demonstrated in murine CMV infection. In conclusion it may be stated that the results presented in this study have implications for diverse fields from regulation of antiviral antibody production to virus-receptor interaction.

ACKNOWLEDGMENTS We thank Dr. Mepur H. Ravindranath for valuable discussions, critical comments, and suggestions. Synthetic GlcNAc-containing oligosaccharides were kindly provided by Dr. James C. Paulson. This work was supported by a Public Health Service grant and Grant A I 25650 from the National Institutes of Health.

REFERENCES ALFORD, C. A., and BRITT, W. J. (1985). Cytomegalovirus. In “Virology” (B. Fields, Ed.), pp. 629-659. Raven Press, New York. BARTHOLOMAEUS, W. N., O’DONOGHUE, H., FOTI, D., LAUSON, C. M., SHELLAM, G. R., and REED, W. D. (1988). Multiple autoantibodies following cytomegalovirus infection: Virus distribution and specificity of antibodies. Immunology 64, 397-405. BEAUCLAIR, K. D., and KHANSARI, D. N. (1990). Protection of mice against Bruce/la abortus by immunization with polyclonal anti-idiotype antibodies. immunology 180, 208-220. BISHOPRIC, G., BRUNER, J., and BUTLER. J. (1985). Guillain-Barre syndrome with cytomegalovirus infection of peripheral nerves. Arch. Patho/. Lab. Med. 109, 1106-l 108. COLIGAN, J. E., KRUISBEEK, A., MARQUITES, D. H., and SHEVACH, E. M. (1991). “Current protocols in Immunology.” Wiley, New York. DRACHMAN, D. B.. and KUNCL, R. W. (1980). Amyotrophic lateral sclerosis: An unconventional autoimmune disease? Ann. Neural. 26, 269-274. ERLANGER, B. F., CLEVELAND, W. L., WASSERMAN, N. H., Ku, H. H., HILL, B. L., SARANGARAJAN, R., RAIAGOPALAN, R., CAYANIS, E., EDELMAN, I. S., and PENN, A. S. (1986). Auto-anti-idiotypy: A basis for autoimmunity and a strategy for anti-receptor antIbodies. Immunol. Rev. 94, 3-37. ERTL, H. C. J. (1989). Anti-idiotypic antibodies as vaccines to a murine Sendai virus infection. Viral immunol. 2, 247-254. GAULTON, G. N., Co, M. S., ROYER, H. D., and GREENE, M. I. (1985). Anti-idiotypic antibodies as probes of cell surface receptors. Mol. Cell. Biochem. 65, 5-21. GRAVES, M. C., SILVER, S. M., and CHOPPIN, P. (1978). Measles virus polypeptide synthesis in infected cells. virology 88, 254-263. HOWARD, R. J.. MAT~SSON. D. M., SEIDEL, M. V., and BALFOUR. H. H. (1978). Cell mediated immunity to murine cytomegalovirus. /. Infect. Dis. 138, 597-604. KAUFFMAN, R. S., NOSEWORTHY, J. H., NEPOM, 1. T., FINBERG, R., FIELDS, B. N.. and GREEN, M. I. (1983). Cell receptors for the mam-

MCMV

ANTIVIRAL

AND ANTIRECEPTOR

malian rheovirus. 1 1. Monoclonal anti-idiotyotypic antibody blocks viral binding to cells. 1. Immunol. 131, 2539-2541. KRAH, D. L., and CHOPPIN, P. W. (1988). Mice immunized with measles virus develop antibodies to a cell surface receptor for binding virus. J. Viol. 62, 1565-l 572. ~.AEMMLI. U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227, 680-685. LOH, L. C., BALACHANDRAN,N., and QUALTIERE, L. F. (1988). Characterization of a major virion envelope glycoprotein complex of murine cytomegalovirus and its immunological cross-reactivity with human cytomegalovirus. Virology 166, 206-216. LOH, L. C.. and QUALTIERE, L. F. (1988). A neutralizing monoclonal antibody recognizes an 87K envelope glycoprotein on the murine cytomegalovirus virion. virology 162, 498-502. MARRIOT, S. J., ROEDER, D. J., and CONSIGLI, R. A. (1987). Antl-idiotypit antibodies to a polyomavirus monoclonal antibody recognize cell surface components of mouse kidney cells and prevent polyomavlrus infection. 1. Viral. 61 1 2747-2753. MCCLINTOCK, P. R., PRABHAKAR, B. S., and NOTKINS, A. L. (1986). Anti-idiotypic antrbodies to monoclonal antibodies that neutralize coxsackie virus B4 do not recognize viral receptors. Virology 150, 352-360. NEGORO, S., TAKASHIMA, T., FUJIWARA,H., and TSUYRIGUCHI,I. (1979). Regulatory mechanisms of autoantibody production in mice to bromelin treated isologus red blood cells. Immunology 36, 257264. PAULSON,I. C. (1985). Interactions of animal viruses with cell surface receptors. In “The Receptors II” (M. Conn, Ed.), pp. 131-219. Academic Press, New York. PEARSON, C. M. (1956). Development of arthritis, periarthritis and penortitis in rats given adjuvant. Proc. Sot. Exp. B/o/. Med. 91, 95-98. POWELL,T. J., SPANN, R., NGUYENDUC. M., and LAMON, E. W. (1989). Induction of effective immunity to moloney murine sarcoma virus using monoclonal anti-idlotypic antibody as immunogen. J. Immunol. 142, 1318-1324. RAVINDRANATH,M. H. R., and GRAVES. M. C. (1990). Attenuated munne cytomegalovirus binds to N-acetylglucosamine and shift to virulence may involve recognition of sialic acids. J. Viral. 64, 54305440. ROGERS,G. N., and PAULSON, 1. C. (1983). Receptor determinants of human and animal influenza virus isolates: Differences in receptor

ANTIBODIES

151

specificity of the H3 hemagglutinin based on species of origin. Virology 127, 361-373. ROGERS,G. N., PAULSON, J. C., DANIELS, R. S., SKEHEL,J. J., WILSON, I. A., and WILEY, D. C. (1983). Single amino acid substitutions in the influenza hemagglutinin change the specificity of the receptor binding. Nature (London) 304, 76-78. SALAZAR-GRUESO,E. F., ROUTBORT,M. J., MARTIN, J., DAWSON, G., and Roos, R. P. (1989). Polyclonal IgM anti-GM1 ganglioside antibody in patients with motor neuron disease and variants. Ann. Neural. 27, 558-563. SUZUKI, Y., NAGAO, Y., KATO, H., MATSUMOTO, M. K., NEROME, K., NAKAJIMA, K.. and NOBUSAWA, E. (1987). Human influenza A virus hemagglutinin distinguishes sialyloligosaccharides in membrane associated gangliosides as its receptor which mediates the adsorption and fusion process of virus infection. Specificity for oligosaccharides and sialtc acids and the sequence to which sialic acid is attached. J. Viol. Chem. 262, 17.057-17,061. TEMPONI, M., KAGESHITA,T., PEROSA, F., ONO, R., OKADA, H., and FERRONE.S. (1989). Purification of munne IgG monoclonal antibodies by precipitation with caprylic acid: Comparison with other methods of purification. Hybridoma 8, 85595. TOWBIN, H., STAEHELIN, T., and GORDON, J. (1979). Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc. Nat/. Acad. SC;. USA 76,4350-4354. WEIGLE, W. 0. (1980). Analysis of autoimmunity through experimental models of thyrotditis and allergic encepalomyelitis. Adv. lmmunol. 30, 159-273. WELLER, A. H., SIMSON, K., HERZUM, M., HOUTEN, N. V., and HUBER, S. A. (1989). Coxsackievirus-B3-induced myocarditis: Virus receptor antibodies modulate myocarditis. J. Immunol. 143, 18431850. WILLIAM, C., and CHASE, M. (1967). “Methods in Immunology and Immunochemistry,” Vol. 1, pp. 678. Academic Press. New York. YAMADA, M.. ZURBRIGGEN,A., and FUJINAMI, R. S. (1990). Monoclonal antibody to Theiler’s murine encephalomyelitis virus defines a determinant on myelin and oligodendrocytes, and augments demyelinatlon in experimental allergic encephalomyelitis. J. fxp. Med. 171, 1893-1907. ZENITA, K., HIRASHIMA, K., SHIGETA,K., HIFUIIWA,N., TAKADA, A., HASHIMOTO, K., FUJIMOTO.E., YAGO, K., and KANNAGI, R. (1990). Northern hybridization analysis of VH gene expression in murine monoclonal antibodies directed to cancer-associated ganglioside antigens having various sialic acid linkages. J. Immunol. 144, 4442-4451.