J. gen. Virol. (1985), 66, 1383-1395. Printed in Great Britain

Key words: arenaviruses/Junin virus/Tacaribe ~,irus/monoclonal antibodies

1383

Properties and Characterization of Monoclonal Antibodies to Tacaribe Virus

By C. R. H O W A R D , 1. H. L E W I C K I , 2 L. A L L I S O N , 1 M. S A L T E R 1 AND M. J. B U C H M E I E R 2

1 Department of Medical Microbiology, London School of Hygiene and Tropical Medicine, Keppel Street, London WCIE 7HT, U.K. and 2 Department of Immunology, Scripps Clinic

and Research Foundation, 10666 North Torrey Pines Road, La Jolla, California 92037, U.S.A.

(Accepted 4 April 1985)

SUMMARY

Monoclonal antibodies prepared against Tacaribe and Junin viruses have been used to define further the serological relationships between arenaviruses of the Tacaribe complex. A close relationship was found between these two viruses and the heterologous Amapari and Machupo viruses, with Pichinde virus and Parana virus being more distantly related. Among the antibodies specific for Tacaribe virus, five were found to react with viral antigens at the surface of infected cells and to neutralize virus infectivity in vitro. These five antibodies could be differentiated by competitive immunoassay as recognizing at least two antigenically distinct epitopes. The kinetics of reaction between antibody and virus were examined for all five neutralizing antibodies. One antibody (2.25.4) effectively neutralized all infectious virus. The remaining four directed against a second epitope gave significant persistent fractions which could be reduced by addition of complement, anti-mouse immunoglobulin, or antibody 2.25.4. Variants of Tacaribe virus resistant to neutralization by antibody 2.25.4 were obtained by growth in the presence of this antibody and neutralization kinetics were re- examined using the heterologous monoclonal neutralizing antibodies. Several different neutralization profiles were obtained, suggesting that point mutations resulted in conformational changes at topographically selected distinct epitopes recognized by the remaining antibodies.

INTRODUCTION

The Tacaribe complex of the family Arenaviridae presently contains eight viruses, all of which serologically cross-react to varying degrees with Tacaribe virus, the type member of the group (Wulffet al., 1978). Other members include Pichinde, Amapari and Parana viruses, all of which cause persistent infections in their natural rodent hosts but are non-pathogenic for man. Machupo and Junin viruses are also members of the complex, being the causative agents of Bolivian and Argentinian haemorrhagic fevers respectively.

Tacaribe virus in common with other arenaviruses contains large (L) and small (S) RNA genome segments (Vezza et al., 1978). The virus contains a major internal nucleocapsid protein (N) and has been reported to contain a single glycoprotein species in its outer envelope (Gard et al., 1977). This is in contrast to other arenaviruses so far characterized including lymphocytic choriomeningitis (LCM) virus, Pichinde and Junin viruses, all of which contain two glycoprotein moieties in the viral envelope (Ramos et al., 1972; Vezza et al., 1978; Grau et al., 1981; Buchmeier et al., 1978).

Extensive serological cross-reactivity among members of the Tacaribe complex is readily demonstrable by complement fixation techniques (Casals et al., 1975; Wulff et al., 1978) although the degree of relatedness between individual members has been difficult to assess using convalescent or hyperimmune antisera. However, Junin, Machupo, Amapari and Tacaribe viruses appear to be particularly closely related. All available evidence indicates that the

0000-6574 © 1985 SGM

1384 c . R. HOWARD AND OTHERS

complement - f ix ing ant igen is associated wi th the internal N polypept ide (Buchmeie r et al., 1977). The neut ra l iza t ion test is m u c h more specific, showing no cross-neutra l iza t ion be tween viruses of the Taca r ibe complex which are o therwise closely related by c o m p l e m e n t fixation. Fo r example , Jun in and M a c h u p o viruses are qui te d is t inct ( Johnson et al., 1973), a l though there is some ev idence to suggest that Jun in virus is weakly neutra l ized by Taca r ibe i m m u n e se rum (Henderson & Downs , 1965; Wei s senbache r et al., 1975/1976). Fur the rmore , ne i the r guinea- pigs nor marmose t monkeys inocula ted wi th Taca r ibe virus show signs o f disease but they are pro tec ted against chal lenge wi th the normal ly lethal Jun in virus, an observa t ion wh ich has led to the suggest ion that Taca r ibe virus m a y be a possible cand ida te vacc ine against Argen t ine haemor rhag ic fever (Weissenbacher et al., 1975/1976, 1982). One reason for this pro tec t ion m a y be that Taca r ibe virus pr imes an i m m u n e response to Jun in virus as a result o f the an t igen ic s imilar i ty be tween them (Coto et al., 1980).

However , the na ture of this cross-protect ion and the ant igen(s) invo lved are as yet uncharac ter ized . The d e v e l o p m e n t o f monoc lona l an t ibodies specific for these viruses wh ich was necessary for the analysis o f this an t igenic re la t ionship is r epor ted in this paper . These ant ibodies have also been used to clar ify fur ther the extent of an t igen ic cross-react ivi ty be tween member s o f the Taca r ibe complex , thereby ex tend ing previous results ob ta ined wi th monoc lona l ant ibodies against L C M and P ich inde viruses (Buchmeie r et al., 1980, 1981). Neu t r a l i z ing ant ibody against m a n y arenaviruses can only be de tec ted wi th difficulty (Chanas et al., 1980) and the f inding of monoc lona l an t ibodies wi th neut ra l iz ing proper t ies against Taca r ibe virus p r o m p t e d a detai led examina t i on o f these reagents and their use to p repare stable var ian ts resis tant to neutral izat ion.

METHODS

Virus and viral growth. Tacaribe (strain TRVL 11573) and Parana viruses were obtained from the Center for Disease Control, Atlanta, Ga., U.S.A. (courtesy of Drs K. Johnson and J. McCormick) as a suckling mouse brain suspension. Attenuated Junin virus (XJ-C 13 strain) was obtained from Dr G. Eddy, U.S. Army Research Institute of Infectious Diseases, Fort Derrick, Md., U.S.A.

When required, extracellular virus was purified from roller bottle cultures of infected BHK-2I cells and purified as described previously (Buchmeier et at., 1978).

Immunization andfi~sion. BALB/c mice selected at 4 to 6 weeks of age from the Scripps Clinic and Research Foundation breeding colony were primed by intraperitoneal injection of 103 p.f.u, of virus. Primed animals received a further 103 p.f.u. 1 week later. All animals immunized with Tacaribe virus contained circulating antibody detectable by immunofluorescence after 4 weeks. Mice showing the highest titres were selected and boosted by intraperitoneal injection of 0.3 ml of a 10% (v/v) saline extract of infected cells on three successive days prior to sacrifice.

Animals primed initially with two doses of Junin X J-C13 at 1000 p.f.u, per dose failed to produce a detectable antibody response. These animals were rested for 3 months and then further immunized, at first with 10 ~ p.f.u., and then with 105 p.f.u. 3, 6, 7 and 8 days later. Mice were sacrificed 24 h later. Serum from the Junin virus- inoculated mice used for fusion were confirmed retrospectively as having antibody to the virus.

Spleens selected for fusion were prepared as described by K6hler & Milstein (1975) and splenocytes fused with the non-secreting P3-X63-Ag8-6531 line of mouse plasmacytoma cells. Detection of hybrid cell colonies secreting antibody was accomplished by indirect immunofluorescence using both acetone-fixed infected L or Vero cell substrates and viable cell suspensions as previously described for the detection of LCM virus antibody (Buchmeier & Oldstone, 1978; Collins et al., 1982). Positive colonies were subsequently subcultured and cloned by limiting dilution as described previously (Buchmeier et al., 1981) and cloned cultures were grown on ascites using BALB/c mice primed with Pristane (Aldrich).

Virus assay and neutralization. Both Tacaribe and Junin viruses were quantified by plaque assay using Vero cells. The procedure was essentially as described by Mann et al. (1980). In some early experiments, CM-cellulose was used as a liquid overlay, but plaques were consistently of smaller diameter after 8 days of incubation than they were when agarose was used.

Aliquots of supernatant fluid from antibody-producing hybridoma cell cultures were initially screened for neutralization in vitro by measurement of virus titre reduction in the presence of a constant amount of antibody. Mixtures containing 0.1 ml of antibody-containing fluid and 0.1 ml of virus diluted in 10-fold increments were made directly in 24-well tissue culture plates and left in the dark for 30 min at 37 °C. Vero ceils (5 × l0 s) in 1 ml growth medium [Medium 199 containing 5~ (v/v) foetal calf serum and antibiotics] were added to each well and plates further incubated at 37 °C for 3 h. The medium was then withdrawn and a CM-cellulose overlay added. The

Tacar ibe virus monoc lona l an t ibod ies 1385

extent of virus neutralization was recorded as the log~ 0 reduction in virus titre observed after plaque visualization 7 days later. Mouse ascites containing polyclonal antibodies to Tacaribe virus or hamster immune serum to Junin virus were used as positive control reagents where appropriate (kindly provided by Dr K. Johnson).

Neutralizing antibodies against Tacaribe virus were examined further in order to quantify the rate of reaction. Neutralization kinetics were measured by incubating dilutions of heat-inactivated ascites containing antibody with approximately 5 x l0 s p.f.u, of Tacaribe virus in equal volumes. All dilutions were pre-warmed at 30 °C and aliquots of each mixture were withdrawn at various times after mixing. Virus-antibody reactions were stopped by an immediate 1:100 dilution into Medium 199 containing 2% (v/v) foetal calf serum held on ice; residual virus infectivity was quantified on Vero cell monolayers.

Selection o f virus variants. Approximately 107 p.f.u, of freshly cloned Tacaribe virus was mixed with a 1:500 final dilution of monoclonal antibody 2.25.4 for 15 min at 37 °C. The mixture was then plated on Vero cells at high dilution under an agar overlay containing the same dilution of antibody. The plates were examined by neutral red staining 7 days later and well-separated plaques representing virus that had escaped neutralization were recloned three times in the presence of antibody and stocks grown in BHK-21 cells for further analysis. Variants were selected that retained reactivity in kinetic neutralization experiments with a mouse hyperimmune antiserum but were resistant to antibody 2.25.4.

Immune precipitation. The specificity of monoclonal antibodies against Tacaribe virus was investigated by immune precipitation of [3SS]methionine-labelled viral proteins from cell extracts as described previously (Buchmeier et aL, 1981 ; Collins et al., 1982). The resulting immune complexes were removed by addition of formalin-fixed Staphylococcus aureus and analysed on 10.5% SDS-poiyacrylamide gels (Buchmeier et al., I978).

Solid-phase competition radioimmunoassay. Purified Tacaribe virus was bound to the surface of flexible polyvinyl microtitre plates (Cooke Engineering) by inoculation overnight at room temperature in 0.1 M-bicarbonate buffer pH 9.0. Approximately 1 gg of viral protein per well was previously determined as optimal for detection of specific antibody. The wells were then washed and treated with 0.25 % bovine serum albumin (BSA~2 % foetal calf serum in PBS to block remaining sites. Individual monoclonal antibodies were purified by Protein A affinity chromatography (Ey et al., 1978) and radiolabelled with Bolton & Hunter reagent (Amersham) to a specific activity of 0-1 to 0-3 ktCi/~tg.

Competition experiments were performed by adding purified monoclonal antibody IgG diluted fivefold in 0.25% BSA PBS directly into coated wells and leaving for 30 rain at room temperature. A fixed amount of radiolabelled antibody was then added to all wells and incubation continued for 3 h at 37 °C. The wells were then washed three times with 0.1% Tween 20-PBS, dried, and counted by gamma spectroscopy.

RESULTS

Production of monoclonal antibodies A total of 21 cell clones secreting antibodies to Tacaribe virus were obtained from a total of

112 hybrid cultures after fusion of immune spleen cells with P3-X63-Ag8-6531 plasmacytoma cells. Positive cultures were identified by immunofluorescence analysis of tissue culture fluids using fixed, infected L-cell substrates. No reactions were recorded using uninfected cell substrates as negative controls. Of the 21 cell cultures, three were dispensed with on grounds of poor antibody production. The remaining 18 were cloned, re-checked for antibody production and hybridoma cells injected into Pristane-treated mice for the production of ascites. Of these, the majority secreted antibody of the IgG2a subclass (12), and four produced antibody of IgG type 1. The remaining two produced IgG2b and IgM antibodies respectively (Table 1). In one case (Tacaribe antibody 2.14.D4) it was necessary to re-clone the culture to establish a satisfactory cell line.

Following the immunization protocol outlined in Methods, monoclonal antibodies to the attenuated X J-El 3 Junin virus were produced in seven of a total of 120 cultures. All monoclonal antibodies to Junin were of the IgG type, three of subclass 2a, three of subclass 1 and one of subclass 2b (Table 2).

Properties of monoclonal antibodies The initial screening of cultures by immunofluorescence segregated the antibodies to

Tacaribe virus into three groups. The first of these produced both a fine granular cytoplasmic fluorescence and a reaction at the plasma membrane (antibodies 2.2. I to 2.82.2, Table 1 a). All five antibodies in this group produced a positive immunofluorescence reaction at the surface of unfixed infected cells. The second group of antibodies (2.1.2 to 2.43.1, Table 1 a) gave positive

1386 C. R. H O W A R D A N D O T H E R S

Table 1. Immunofluorescence analysis of monoclonal antibodies to Tacaribe virus

(a) Ant ibodies reac t ive agains t homologous subst ra tes only

Cytop lasmic P l a sma m e m b r a n e Ref. no. Ig subclass f luorescence* fluorescence*

2 .2 .1 G 2 a + + 2 . 1 2 . 5 G I + + 2 . 1 4 . D 4 G 2 a + + 2 . 2 5 . 4 G 2 a + + 2 . 8 2 . 2 G 2 a + + 2 . 1 . 2 G 2 a + - 2 . 7 . 2 G1 + - 2 . 2 9 . 9 G 2 b + - 2 . 3 0 . 1 0 G 2 a + - 2 .31 .1 G 2 a + - 2 .69 .1 G2a + - 2 .84 .11 M + - 2 .43 .1 G I + -

(b) Ant ibodies reac t ive agains t heterologous substra tes

Cross - reac t ions t

Ref. no. Ig subclass Taca r ibe Junin M a c h u p o P ich inde A m a p a r i P a r a n a

2 . 1 6 . 2 G 2 a 320 + ~ - :~ - 10 10 2 . 4 8 . 3 G 2 a 40000 4680 500 3200 10000 - 2 . 5 2 . 2 G 2 a 160000 + - - 80000 - 2 . 7 4 . 3 G 2 a 320000 1620 - - 80000 - 2. 100.3 G 1 32000 + 500 - - -

* Expe r imen t s were pe r fo rmed using e i ther acetone-fixed cells (cytoplasmic fluorescence) or unfixed ( m e m b r a n e fluorescence) infected L-cell cultures (Buchmeie r et al., 1981).

"~ Expressed as reciprocal of endpoint t i tres ob ta ined wi th ascites fluids. :~ + , Posi t ive (not t i t ra ted) ; - , negat ive .

Table 2. Cross-reactions by immunofluorescence with monoclonal antibodies to Junin virus

Cross-react ions*

Ref. no. Ig subclass Junin M a c h u p o P ich inde T a c a r i b e A m a p a r i P a r a n a

3 .6 G 2 b + + - + - - 3 .46 G 2 a + + - + - - 3 .67 G 2 a + + + + - - 3 .69 G I + + - + - - 3 .70 G I + + - + - - 3 .88 G 2 a + + - + - - 3 .95 G1 + + - + - -

* + , Posi t ive; - , negat ive . All substrates were acetone-fixed infected mouse L-cell cul tures (Buchmeie r et al., 198l).

reactions only with acetone-fixed substrates infected with Tacaribe virus. The remaining antibodies were characterized by reactions with acetone-fixed substrates infected with heterologous arenaviruses in the absence of surface fluorescence (Table 1 b). All in this group reacted with Junin and four with Amapari viral antigens. Among these five antibodies, two also reacted with Machupo viruses and one additionally recognized Pichinde antigens (antibody 2.48.3). Parana virus-infected cells were only recognized by antibody 2.16.2 which failed to react with either Machupo or Pichinde virus. Representative antibodies from the second and third group reacted with the internal N polypeptide of Tacaribe virus as assessed by analysis of immune complexes obtained by reaction of antibody with radiolabelled cell extracts (Fig. I a, c, d,e).

Tacar~e v~usmonoclonal antibodies

~) (~ (3 (d) (d (f) ~) (h)

1387

N

G

N

G

Fig. 1. SDS-PAGE analysis of immunoprecipitates formed by the reaction of infected cell extracts with Tacaribe virus monoclonal antibodies giving cytoplasmic immunofluorescent reactions on acetone-fixed cells [(a, c, d, e) antibodies prefixed 2.2, 2.16, 2.74 and 2.7 respectively in Table 1] or surface fluorescence on unfixed cells and neutralization of virus infectivity [(b, J) antibodies prefixed 2.14 and 2.2 respectively]. Both N and G polypeptides were precipitated by mouse hyperimmune serum (g). No precipitation occurred with normal mouse serum (h). (a) to (d) and (e) to (h) represent two separate experiments.

All of the monoclonal antibodies to Tacaribe virus which gave positive reactions at the surface of unfixed infected cells immunoprecipitated the single envelope glycoprotein (G) of this virus (e.g. Fig. 1 b,J). However, these immunoprecipitates also contained the nucleocapsid (N) protein, a finding in common with previous studies of antibodies to arenavirus glycoproteins (Kiley et al., 1981). Attempts to confirm the specificity of these antibodies by Western blotting techniques were unsuccessful, suggesting they are directed against conformation-dependent epitopes. As a control, a polyclonal immune ascitic fluid positive for neutralizing antibodies to Tacaribe virus precipitated both structural proteins (Fig. lg).

A limited number of monoclonals to Junin virus strain XJ-C 13 were also prepared to examine the extent of reciprocal cross-reactions by immunofluorescence. All seven clones showed positive reactions against fixed cell substrates infected with the same Junin virus strain. There was no evidence of antibody binding at the cell surface and all showed cross-reactions against other members of the Tacaribe virus complex (Table 2). All gave positive reactions with Tacaribe virus-infected substrates, confirming the close relationship between these viruses. Additionally, all recognized Machupo virus antigens. In contrast to the monoclonal antibodies to Tacaribe virus, no cross-reactions were seen against Amapari or Parana viruses and only one antibody recognized Pichinde virus antigens. Titration analysis using substrates infected with virulent strains of Junin virus showed that these antibodies gave varying titres of reactivity compared to the X J-C13 strain (C. R. Howard & J. I. Maiztegui, unpublished observations).

1388 ¢. R. HOWARD AND OTHERS

100 I I I I . ~ I i A

~a) , ~ t ~ n (b)

"0 " . 5 0 - - -

g

g

100 I I I ~ - ° T-,,A I I • ~ • • • •

• , . _ •

5 0 - -~ • • -

O I I I I I I 0 1000 100 10 1 1000 100 10 1 0 - 1

IgG/well (ng)

Fig. 2. Solid-phase competition radioimmunoassays using monoclonal antibodies with neutralization properties against Tacaribe virus. Four purified immunoglobulins were 125I-labelled with Bolton & Hunter reagent and each ligand was examined for binding to purified virus in the presence of increasing amounts of heterologous antibodies. Labelled ligands used in the four experiments were (a) 2.2.1, (b) 2.12.5, (c) 2.14. D4 and (d) 2.82.2. Competing antibodies prepared as purified IgG are represented as follows: I , 2.25.1; O, 2.2.1; A, 2.12.5; I-], 2.14.D4; O, 2.82.2.

Competition radioimmunoassay

All five monoclonal antibodies to the Tacaribe virus glycoprotein were radiolabeUed and used as radioactive probes in a solid-phase radioimmunoassay containing whole virus bound to the solid phase as described in Methods. Each antibody was tested in turn for cross-competition with the remaining four antibodies (Fig. 2). Complete cross-reaction was seen between antibodies 2.2.1, 2.82.2, 2.12.5 and 2.14. D4, these being the same groups of antibodies which produced a high level of non-neutralized virus (Fig. 4). Antibody 2.25.1, from a clone similar in properties to 2.25.4, did not cross-compete with any of these antibodies. Attempts to perform the reciprocal experiment whereby antibody 2.25.1 was used as the radioactive probe proved unsuccessful as this antibody lost its ability to bind to virus after radiolabelling using either the Bolton & Hunter reagent, lactoperoxidase, chloramine-T or Iodogen or a variety of other labelling procedures. Hence, a labelled probe was unavailable. Unlabelled 2.25.1 IgG used in competition experiments was positive by immunofluorescence, confirming that reactivity of this antibody had not been lost on purification.

Neutralization o f Tacaribe virus with monoclonal antibodies

Aliquots of supernatant fluid from antibody-producing hybridoma cell cultures were screened for neutralization in vitro by measurement of virus titre reduction in the presence of a constant amount of antibody. All five antibodies in the group previously found to react with the surface of Tacaribe virus-infected cells (see Table 1) reduced viral infectivity (Table 3). In contrast, no significant levels of neutralization were found against the heterologous arenaviruses Pichinde and Amapari (data not shown). Although some neutralization was obtained against Junin virus XJ-C13 (up to 0.3 log10 p.f.u, of total plaque number), this was deemed to be not significantly

Tacaribe virus monoclonal antibodies

Table 3. Neutralization of inJectivity by monoclonal antibodies to Tacaribe virus

Specificity Envelope glycoprotein

Nucleocapsid protein

Virus neutralized* A

f - -

Antibody no. Tacaribe Junin 2.2.1 2.87 0.23 2.12.2"1" 1.95 0-16 2.14. D4 1.83 0.02 2.25.4 2.84 0.27 2.82.2 1-84 0.34 2.16.2 0.10 ND~ 2.48.3 0.18 0.17 2.52.2 0.33 ND 2.74.3 0.20 ND 2. 100.1 0'25 ND 2.1.5 0 0"24 2.7.2 0"16 ND 2.29.9 0"14 ND 2.30.10 0' 19 ND 2.31.1 0"19 ND

Control serum Mouse anti-Tacaribe 2.39 > 1.0 Mouse anti-Junin 0 > 3.0

* Reduction in log~0 titre compared to control. ~" Clone similar in properties to 2.12.5 (Table 1). :~ NO, Not done.

1389

different from the maximum values obtained against Tacaribe virus using antibodies specific for the nucleocapsid. The one-way cross-neutralization observed using polyclonal mouse antisera as positive control was noteworthy : anti-Junin sera failed to neutralize Tacaribe virus but anti-Tacaribe immune ascites neutralized both Junin and the homologous virus, as previously reported by Henderson & Downs (1965).

Antibodies against Tacaribe virus found positive for homologous neutralization were examined further in order to quantify the rate of reaction between antibody and homologous virus. Kinetic neutralization showed that the five antibodies differed according to the level of infectivity remaining after 20 min of incubation. At high dilutions, antibody 2.25.4 effectively neutralized all infectious virus within the limit of sensitivity of the assay system; neutralization curves demonstrated an initial lag period followed by linear neutralization of virus with time (Fig. 3). 'Least squares' linear regression analysis of data obtained in the linear region of each neutralization curve showed highly significant values for correlation coefficients (P = <0.001) and extrapolation of values calculated for slope towards time zero gave intercept values greater than 1 (2, 1.6 and 1.95 at dilutions 1/16000, 1/32000 and 1/64000 respectively). This result is indicative of more than one molecule of antibody being required for neutralization, as single-hit theory assumes a linear response from the time of mixing virus and antibody and a zero time intercept value of 1 (Della-Porta & Westaway, 1978). The remaining four neutralizing antibodies gave significant non-neutralized fractions after an initial linear response (Fig. 4). This residual virus infectivity was decreased by addition of either antibody 2.25.4 at 20 min (Fig. 5b) or by the addition of either anti-mouse IgG or complement to the virus-antibody mixtures at time 0 (Fig. 5 c, d). Given the monoclonal nature of the antibodies it is unlikely that the persistent fraction was due to steric hindrance from non-neutralizing antibodies. Although the possibility of a heterogeneous population of virus particles cannot be excluded, all virus stocks were repeatedly recloned during the course of these experiments. Cloned virus stock grown from residual virus after reaction with antibody 2.82.2 gave a similar fraction on re- exposure to the same antibody.

Antibody 2.82.2 was examined in combination with antibody 2.25.4. The reduction in titre produced by the two antibodies was found to be equivalent to the sum of the reduction in

1390 C . R . HOWARD AND OTHERS

o

0.1

0.01

I I I I

I I I I

5 10 15 20

Time (rain) Fig. 3. Kinetic neutralization of Tacaribe virus with antibody 2.25.4 at 30 °C. Dilutions of antibody examined were 1/16000 (O), 1/32000 (O) and 1/64000 (Z~). Reduction in infectivity was plotted as the log~o value of residual virus (V) divided by titre at time 0 (Vo).

o

o

I l I I I I I I I (a) (b)

all w

I I I I I (a9

I I

0.1 I I I I I (c)

1

0 . 1

I

0 4 8 12 16 20 0 4 8 12 16 20

Time (min)

Fig. 4. Kinetic neutralization of Tacaribe virus with antibodies directed against the surface glycoprotein but producing a large resistant fraction: (a) 2.2.1, (b) 2.12.5, (c) 2.82.2 and (d) 2.14. D4. Twofold dilutions (O to O to A) of the antibody examined started from (a) 1/4000, (b) 1/1000, (c) 1/500 and (d) 1/100. Residual infectivity was calculated as in Fig. 3.

infect ivi ty produced by each an t ibody tested singly (Fig. 6). The pers is tent non-neu t ra l i zed fract ion obta ined with an t ibody 2 .82.2 a lone is thus unl ikely to be due to virus aggregat ion and provides further evidence of two dis t inc t epi topes p laying a role in the neut ra l iza t ion of Tacar ibe virus.

Tacaribe virus monoclonal antibodies 1391

>

0 " 1 B

I I I I I I I I I I I

I I I I I I I

0.1

I I (c)

I I I I I

• v

4 k -

I I (d)

I I I I I I I I I I I I I I

0 10 20 30 40 50 60 0 10 20 30 40 50 60 Time (min)

Fig. 5. Neutralization kinetics of Tacaribe virus. Effect of e i t h e r h e t e r o l o g o u s monoclonal antibody 2.25.4 added at 20 min (b), complement (c) or anti-mouse IgG (d) on the residual fraction of v i r u s infectivity remaining after reaction with antibody 2.82.2. (a) No addition; antibody alone.

e~

_.> > , , . .J

etl) o

O.1

0.01

I I I I

O

I 2O

I I I 0 5 10 15

Time (min)

Fig. 6. Neutralization of T a c a r i b e v i r u s w i t h antibodies 2.25.4 (1/3200 final dilution) and 2.82.2 (1/500 final dilution) added simultaneously at time 0 (A) compared w i t h n e u t r a l i z a t i o n by antibodies s e p a r a t e l y (O, 2.82.2; 0 , 2.25.4; final dilutions as above). Extrapolation of the plot obtained w i t h 2.25.4 a l o n e i n d i c a t e s d o u b l e - h i t k i n e t i c s .

Selection of Tacaribe virus variants

In order to analyse further the nature of neutralization reactions with monoclonal antibody, Tacaribe virus was mixed with antibody 2.25.4 for 15 min and the mixture plated on Vero cells at high dilution. Variants were selected on the basis of retention of reactivity with a mouse hyperimmune antiserum but continuing resistance to 2.25.4 (Table 4). Five of these variants were selected for further study (nos. l, 3, 5, 6 and 7) using heterologous antibodies. All variants

1392 c . R . HOWARD AND OTHERS

>__. o

e ~ o

I a ( a ) ' ' '

~a

0.1

I I l

0 5 10 15

Time (min)

i

20

o 0.1

(b) I ~ f I 4 -

O O - - >__.

0.1

( C ) I I I I

I I I I I I I

0 10 20 30 5 10 15 20

Time (min) Time (min)

Fig. 7. Reaction of Tacaribe virus variants selected for resistance to antibody 2.25.4 from Tacaribe virus stock 3V3B (O). (a) All variants showed similar kinetics using antibody 2.2.1 (1/2000) although only results obtained with V5 (O) and V7 (•) are shown for clarity. (b) Variants V3, V5 and V7 (A, Eli) were not neutralized by antibody 2.14.D4 (1/100) whereas variants V6 (A) and VI (O) were. (c) Variants V5 (O) and V7 (Z~) showed enhanced neutralization with antibody 2.12.5 (1/2000).

Tab le 4. Neutralization of Tacaribe virus variants selected by antibody 2.25.4

Virus + immune Virus diluent ascites to Tacaribe Virus + monoclonal

Variant no. only virus* antibody 2.25.4*

1 4.37~ 2.90 (1.47)+ + 4.36 (0.01) 2 5.94 4.25 (1.69) 5.57 (0.37) 3 5.95 4.31 (1.64) 5.60 (0.35) 4 4-12 2.66 (1-46) 3.99 (0.13) 5 5.87 4-66 (1.21) 5-92 (0.00) 6 5.48 4-12 (1,36) 5.44 (0-04) 7 4.18 2.68 (1.50) 4.16 (0.02)

Control 6.62 3.92 (2.70) < 1,30 ( > 5.32)

* Immune ascites and monoclonal antibody used at final dilutions of 1 : 10 and 1:500 respectively. Mixtures of virus and antibody incubated for 15 min at 37 °C prior to assay for residual infectivity as described in Methods.

] Logm infectivity titre. ++ Reduction in log~o infectivity titre compared to reaction with diluent only.

reac ted wi th an t ibody 2.2.1 with s imilar kinet ics to those ob ta ined with the or iginal virus (Fig. 7a). However , the var iants were clearly di f ferent ia ted into two groups by the comple te absence o f neut ra l iza t ion o f var iants 3, 5 and 7 by an t ibody 2 .14 .D4 (Fig. 7b) whereas a large non- neutra l ized f rac t ion was still present wi th var ian ts 1 and 6. Immunof luorescence analysis o f cell cultures infected with var ian ts 3, 5 and 7 conf i rmed a loss of react iv i ty for an t ibody 2 .14 .D4 (data not shown). Both var iants 5 and 7 were neutra l ized by monoc lona l an t ibody 2.12.5 to a m u c h greater degree as compared to the parental virus, wi th a l inear decrease being observed over the whole 20 min per iod of observa t ion (Fig. 7c). Ex tended incuba t ion up to 60 min showed the presence of a resistant f rac t ion at the same level as seen at 20 rain after mix ing the virus wi th ant ibody.

Tacaribe virus monoclonal antibodies 1393

D I S C U S S I O N

Antibodies directed against determinants on the viral nucleocapsid of Pichinde virus have been found to recognize one or more heterologous members of the Tacaribe complex, confirming the particularly close relationship between the arenaviruses Pichinde, Tamiami and Parana (Buchmeier et al., 1981). One of these antibodies was found additionally to recognize Old World arenaviruses (Buchmeier et al., 1980) in addition to recognizing both the Tacaribe and Junin virus-infected cell substrates (C. R. Howard & M. J. Buchmeier, unpublished observations).

In the present study, five monoclonal antibodies prepared against Tacaribe virus were found to recognize heterologous substrates (Table 1). All five reacted with Junin virus antigens and four with Amapari virus-infected cell substrates, providing further evidence of the particularly close relationship between Tacaribe virus and these two arenaviruses. Antibody 2.48.3 recognized all substrates prepared from New World arenaviruses except Parana virus (Table I b), illustrating that this reagent recognizes an epitope common to many viruses in the complex. However, no reactions were recorded with this or any of the remaining antibodies against LCM or Lassa viruses, confirming the specificity of these reagents for members of the Tacaribe complex. The value of reagents capable of recognizing epitopes common to pathogenic members of the group is illustrated by the finding that all monoclonal antibodies cross-reactive for Junin virus substrates also recognized wild-type Junin virus-infected cells of either human or rodent origin (C. R. Howard & J. I. Maiztegui, unpublished observations). Seven additional monoclonal antibodies specific for Junin virus showed again the close relationship between Tacaribe and Junin viruses (Table 2). Additionally, all antibodies to Junin virus reacted equally well with Machupo virus antigen. Collectively, these results suggest that several common antigenic determinants may exist on the internal nucleocapsid structures of these viruses. Cross- reactivity was not seen for Junin virus using the antibodies against the surface glycoprotein of Tacaribe virus by either immunofluorescence or neutralization tests. Additional antibodies may therefore be required to reveal the epitope responsible for the priming of the cross-neutralizing antibody response described in guinea-pigs by Coto et al. (1980). Alternatively, this priming effect may be due to another region of the Tacaribe virus which is non-immunogenic in the mouse, although the one-way cross-neutralization results obtained by the use of immune mouse ascites fluid (Table 3) would not support this.

The detection of neutralizing antibodies in hyperimmune or convalescent serum is particularly difficult for certain arenaviruses. For example, Chanas et al. (1980) have shown that reaction of Pichinde virus with antiserum forms virion-antibody complexes that remain fully infectious. In contrast, both Junin and Tacaribe virus antibodies may manifest neutralization activity in vitro and the availability of monoclonal antibodies which reduce viral infectivity at high titre is of potential value for the study of arenavirus neutralization. Of particular intererest are the five monoclonal antibodies that effectively neutralized the infectivity of Tacaribe virus (Table 3). These antibodies were deemed to be specific for the external glycoprotein moiety of the virion as (i) all neutralized virus infectivity, (ii) all five reacted with the surface of viable infected cells, (iii) the 40000 to 44000 mol. wt. glycoprotein appears to be the major peptide associated with the outer viral envelope and (iv) viral glycoprotein was immunoprecipitated from radiolabelled cell extracts by these antibodies. Although it is conceivable that other, hitherto minor components may be present in the outer envelope of Tacaribe bearing the relevant antigens for induction of neutralizing antibody, no other envelope protein has been identified (Gard et al., 1977).

The ubiquitous presence of N, the major nucleocapsid protein, in such precipitates (Fig. 1) may arise by a particularly tight bonding between these two proteins which is not broken during preparation for SDS-PAGE and/or non-specific affinity for Protein A molecules on the surface of the fixed S. aureus cells used for recovery of immune complexes. This latter problem has been encountered in similar analyses of monoclonal antibodies prepared against Lassa and Mopeia arenaviruses (Kiley et al., 1981; Buchmeier & Oldstone, 1981). At least two epitopes were defined in the present study as playing a role in Tacaribe virus neutralization, although our findings do not exclude the possibility of other antigenic determinants that remained undefined

1394 c . R . HOWARD AND OTHERS

using the monoclonal antibodies available. In the present study, antibody 2.25.4 to Tacaribe virus was found to recognize a distinct epitope compared to the remaining four glycoprotein antibodies as shown by competition radioimmunoassay (Fig. 2) and by the reaction of virus with this antibody being unaffected by prior addition by one of the remaining glycoprotein antibodies (Fig. 5). The large persistent fraction produced by the remaining antibodies may indicate incomplete neutralization resulting from an equilibrium between free and bound antibody molecules at a point prior to complete neutralization (Volk et al., 1982). Accurate measurements of antibody affinity are needed to correlate the extent of binding with the size of the non- neutralized fraction.

Several theories are available to explain the mechanism of virus neutralization but the process by which specific antibody reduces virus infectivity remains unclear. Considerable data have been provided by others to suggest a single-hit process whereby the infectivity of a virion is lost after interacting with a single antibody molecule, although it has been suggested that inactivation of certain viruses may follow multi-hit kinetics (Della-Porta & Westaway, 1978). In either case, neutralization is presumed to occur by antibody-mediated steric hindrance as the result of blocking of relevant sites on viral protein required for interaction with uninfected cell surfaces, the efficacy of which may vary according to the avidity or isotype of the antibody (Lafferty, 1963). The number of relevant sites has been defined for several viruses using monoclonal antibodies. For example, it has been estimated that at least three and perhaps as many as nine distinct antigenic determinants on the 80000 mol. wt. glycoprotein may be involved in the neutralization of rhabdoviruses (Coulon et al., 1983; Lafon et al., 1983; Volk et al., 1982). Such findings are in contrast with the report of Massey & Schochetman (1981), however, who found only a single antigenic determinant on mouse mammary tumour virus. In the case of Tacaribe virus, at least two sites have been defined in the present study although more may become apparent with more monoclonal antibodies. Confirmation of the importance of the epitope recognized by antibody 2.25.4 was the ready selection of stable variants resistant to neutralization (Table 4 and Fig. 7).

Reaction of these variants with the heterologous monoclonal antibodies manifesting incomplete neutralization suggests that point mutations occur in the gene for the glycoprotein resulting in changes in or around the epitope recognized by antibody 2.25.4. This selection may also alter the conformation of non-overlapping distinct epitopes recognized by the second subset of antibodies. For example, some variants lost their reactivity with antibody 2.14. D4 (Fig. 7 b); this may be explained on the basis that under the selection pressure of the monoclonal antibody such mutations may not necessarily occur at the same site in the gene coding for the envelope glycoprotein. As a result, substitutions are not restricted to any one particular amino acid of the protein sequence and some of these changes may affect the ability of the glycoprotein to bind other antibodies.

Alternatively, antibody 2.14. D4 may recognize an epitope which partially overlaps both sites defined by the remaining antibodies, but owing to a high affinity is not clearly differentiated by competitive radioimmunoassay from low-affinity antibodies recognizing a site distinct from the epitope reactive for antibody 2.25.4. Further analysis of these monoclonal antibodies will help to clarify the role of antibody in the establishment and maintenance of arenavirus persistence in vivo, particularly by distinguishing the relative roles of neutralizing and non-neutralizing antibody. Further studies are also in progress to define at the molecular level the antigenic sites involved in these neutralization reactions.

We are indebted to Dr K. Johnson for the supply of Machupo virus antigens and Dr G. Lloyd and Miss S. A1 Mufti for immunofluorescent screening of antibodies to Lassa and Amapari viruses respectively and P. Young for helpful discussions. We should also like to thank Mrs S. Giles and Miss T. Gidman for typing the manuscript. This work is supported by N IH grants AI 16102 and NS 12428, NATO grant 82-237, and the Medical Research Council of Great Britain, and is publication No. 3448/Imm. from the Department of Immunology, Scripps Clinic and Research Foundation. M.J.B. is an Established Investigator of the American Heart Association.

REFERENCES BUCHMEIER, M. J. & OLDSTONE, M. B. A. (1978). Virus-induced immune complex disease: identification of specific

viral antigens and antibodies deposited in complexes during chronic lymphocytic choriomeningitis virus infection. Journal of Immunology 120, 1297-1304.

Tacaribe virus monoclonal antibodies 1395

BUCHMEIER, M. J. & OLDSTONE, M. B. A. (1981). Molecular studies of LCM virus-induced immunopathology: development and characterization of monoclonal antibodies to LCM virus. In The Replication of Negative Strand Viruses, pp. 71 77. Edited by D. H. L. Bishop & R. W. Compans. New York: Elsevier.

BUCHMEIER, M. J., GEE, S. R. & RAWLS, W. E. (1977). Antigens of Pichinde virus. I. Relationship of soluble antigens derived from infected BHK-21 cells to the structural components of the virion. Journal of Virology 22, 175- 186.

BUCItMEIER, M. J., ELDER, J. H. & OLDSTONE, M. B. A. (1978). Protein structure of lymphocytic choriomeningitis virus: identification of the virus structural and cell associated polypeptides. Virology 89, 133-145.

BUCHMEIER, M. J., LEWICKI, H. A., TOMORI, O. & JOHNSON, K. (1980). Monoclonal antibodies to lymphocytic choriomeningitis virus react with pathogenic arenaviruses. Nature, London 288, 486-487.

BUCHMEIER, M. J., LEWICKI, H. A., TOMOR1, O. & OLDSTONE, M. B. A. (1981). Monoclonal antibodies to lymphocytic choriomeningitis virus and Pichinde viruses: generation, characterization and cross-reactivity with other arenaviruses. Virology 113, 73-85.

CASALS, J., BUCKLEY, S. M. & CEDENO, R. (1975). Antigenic properties of the arenaviruses. Bulletin of the World Health Organization 52, 421-427.

CItANAS, A. C., MANN, G., YOUNG, P., ELLIS, D. S., STAMFORD, S. & HOWARD, C. R. (1980). Evaluation of plaque size reduction as a method for the detection of Pichinde virus antibody. Archives of Virology 65, 157-167.

COLLINS, A. R., KNOBLER, R. L., POWELL, H. & BUCHMEIER, M. J. (1982). Monoclonal antibodies to murine hepatitis virus-4 (strain JHM) define the viral glycoprotein responsible for attachment and cell-cell fusion. Virology 119, 358-371.

COTO, C. E., DAMONTE, E. B., CALELLO, M. A. & WEISSENBACHER, M. C. (1980). Protection of guinea pigs inoculated with Tacaribe virus against lethal doses of Junin virus. Journal oflnfectious Diseases 141, 389-393.

COULON, P., ROLLIN, P. E. & FLAMAND, A. (1983). Molecular basis of rabies virus virulence. II. Identification of a site on the CVS glycoprotein associated with virulence. Journal of General Virology 64, 693-696.

DELLA-PORTA, A. J. & WESTAWAY, E. G. (1978). A multi-hit model for the neutralization of animal viruses. Journalof General Virology 38, 1-19.

EY, P. L., PROWSE, S. J. & JENKIN, C. R. (1978). Isolation of pure IgGl, IgG2a and IgG2b immunoglobulins from mouse serum using protein A-Sepharose. lmmunochemistry 15, 429-436.

GARD, G. P., VEZZA, A. C., BISHOP, D. H. L. & COMPANS, R. W. (1977). Structural proteins of Tacaribe and Tamiami virions. Virology 83, 84-95.

GRAU, O., FERNANDEZ, M. T. F., ROMANOWSKI, V., RUSTIA, S. M. & ROSAS, M. F. (1981). Junin virus structure. In The Replication of Negative Strand Viruses, pp. 11-14. Edited by D. H. L. Bishop & R. W. Compans. New York: Elsevier.

HENDERSON, J. R. & DOWNS, W. G. (1965). Junin and Tacaribe plaque production in rhesus monkey kidney cell monolayers. American Journal of Tropical Medicine and Hygiene 14, 796-799.

JOHNSON, K. M., WEBB, P~ A. & JUSTINES, G. (1973). Biology of Tacaribe-complex viruses. In Lymphocytic Choriomeningitis Virus and Other Arenaviruses, pp. 241-248. Edited by F. Lehmann-Grube. Berlin: Springer- Verlag.

KILEY, M. P., TOMORI, O., REGNERY, R. L. & JOHNSON, K. M. (1981). Characterization of the arenaviruses Lassa and Mozambique. In The Replication of Negative Strand Viruses, pp. 1-9. Edited by D. H. L. Bishop & R. W. Compans. New York: Elsevier.

K(JHLER, G. & MILSTEIN, C. (1975). Continuous cultures of fused cells secreting antibody of predefined specificity. Nature, London 256, 495-497.

LAFFERTY, K. J. (1963). The interaction between virus and antibody. 1. Kinetic studies. Virology 21, 61-75. LAFON, M., WIKTOR, T. J. & MACFARLAN, R. I. (1983). Antigenic sites on the CVS rabies virus glycoprotein: analysis

with monoclonal antibodies. Journal of General Virology 64, 843-851. MANN, G. F., ALLISON, L. M., COPELAND, J. A., AGOSTINI, C. F. & ZUCKERMAN, A. J. (1980). A simplified plaque assay

system for measles virus. Journal of Biological Standardization 8, 219-225. rCO.SSEY, R. S. & SCHOCHETMAN, G. (1981). Viral epitopes and monoclonal antibodies: isolation of blocking

antibodies that inhibit virus neutralization. Science 213, 447M49. BAMOS, B. A., COURTNEY, R. J. & RAWLS, W. E. (1972). Structural proteins of Pichinde virus. Journal of Virology 10,

661-667. VEZZA, A. C., CLEWLEY, J. P., GARD, G. P., ABRAHAM, N. Z., COMPANS, R. W. & BISHOP, D. H. L. (I978). Virion RNA

species of the arenaviruses Pichinde, Tacaribe and Tamiami. Journal of Virology 26, 485-497. VOLK, W. A., SNYDER, R. M., BENJAMIN, D. C. & WAGNER, R. R. (1982). Monoclonal antibodies to the glycoprotein of

vesicular stomatitis virus: comparative neutralizing activity. Journal of Virology 42, 220-227. WEISSENBACHER, M. C., COTO, C. E. & CALELLO, M. A. (1975/1976). Cross-protection between Tacaribe complex

viruses. Presence of neutralizing antibodies against Junin virus (Argentine haemorrhagic fever) in guinea pigs infected with Tacaribe virus. Intervirology 6, 42-49.

WEISSENBACHER, M. C., COTO, C. E., CALELLO, M. A., RONDINONE, S. N., DAMONTE, E. B. & FRIGERIO, M. J. (1982). Cross- protection in non-human primates against Argentine haemorrhagic fever. Infection and Immunity 35, 425- 430.

WULFF, H., LANGE, J. V. & WEBB, P. A. (1978). Interrelationships among arenaviruses measured by indirect immunoftuorescence. Intervirology 9, 344-350.

(Received 13 February 1985)