- Email: [email protected]
Monoclonal antibodies to Brucella rough lipopolysaccharide: characterization and evaluation of their protective effect against B. abortus
Q INSTITUT PASTEUR/EI.sEVIER Paris 1993
Res. Microbiol.
1993, 144, 475-484
Monoclonal antibodies to Brucella rough lipopolysaccharide: characterization and evaluation of their protective effect against B. abortus A. Cloeckaert (=) (*), I. Jacques 0), R.A. Bowden (0, G. Dubray (I) and J.N. Limet (2) ¢o Institut National de la Recherche Agronomique, Centre de Recherches de Tours, Laboratoire de Pathologic infectieuse et hnmunologie, 37380 Nouzilty (France), and (2) Facult~s Universitaires Notre-Dame de la Paix, 61 rue de Bruxelles, 5009 Namur (Belgium)
SUMMARY We characterized 4 monoclonal antibodies (mAb) specific for rough lipopolysaccharide (R-LPS) of Bruce#a. mAb were selected by enzyme-linked immunosorbent assay IELISA) on whole R. abortus 45/20 rough cells and R-LPS from B. melitensis B115 rough cells. Specificity ~as confirmed by immunoblot analysis using R-LPS and smooth LPS (S-LPS) preparations. Anti-R-LPS revealed the low molecular mass R-LPS molecules below 20.1 kDa in the R-LPS and S-LPS preparations as well as the typical A and M patterns in high molecular mass S-LPS molecules (between 21.5 and 66 kDa) in the S-LPS preparations. An O-polysaccharide-specific mAb revealed only high molecular mass S-LPS molecules in the S-LPS preparation. In ELISA the anti-R-LPS mAb bound better on rough than on smooth B. abortus 544 whole cells, and this was confirmed by immunoelectron microscopy. Protective activity of anti-R-LPS mAb of different isotypes was tested on mice and compared with an S-LPS-specific mAb. Only the IgG3 mAb reduced significantly the splenic infection but did not reach the level of protection conferred by the S-LPS-specific mAb.
Key-words: LPS, Brucella abortus, Brucella melitensis, IgG, mAb; Mouse, Protection, Epitopes.
INTRODUCTION The smooth lipopolysaccharides (S-LPS) of Brucella spp. and other Gram-negative bacteria
are complex molecules comprising three structurally and serologicaUy distinct regions, i.e. a glycolipid moiety called lipid A, a core oligosaccharide and the O-polysaccharide lOPS) chain.
Submitted March II, 1993, accepted July 28, 1993. {*) Corresponding author.
The Brucella 0 chain contains two distinct epitopes, designated A and M (Bundle et aL, 1989). Their relative amounts vary among the smooth Brucella strains, and these epitopes are absent on the rough strains, which lack OPS (Alton et aL, 1988). mAb to Brucella S-LPS have been reported with specificity for the A, the M and the common epitopes, some of which cross-react
476
A. C L O E C K A E R T E T A L .
with Yersinia enterocolitica 0:9 S-LPS (Bundle et al., 1989; Cloeckaert et al., 1992; GarinBastuji et al., 1990; Limet et al., 1989; Palmer and Douglas, 1989; Vizcaino et al., 1991). These mAb confer protection on mice by reducing the number of Brucella in the spleen and liver (Cloeckaert et al., 1992; Limet et al., 1989, 1987; Montaraz et al., 1986; Phillips et al., 1989). Brucella rough LPS (R-LPS) lacking OPS and expressed predominantly in rough strains of Brucella has been immunochemically characterized (Moreno et al., 1979), and antibody response against it has been observed in B. ovis infections (Riezu-Boj et al., 1986). We obtained mAb specific for R-LPS by fusion with the NSO myeloma of spleen cells of mice infected by the rough B. a b o r t u s 45/20 and B. melitensis B115 strains (Cloeckaert et al., 1990). Our purpose here was to characterize the anti-R-LPS mAb in terms of molecular and cellular binding and to compare their protective activity in mice against B. a b o r t u s with an anti-S-LPS mAb already shown to be protective (CIoeckaert et al., 1992; Limet et al., 1989; Jacques et al., 1992; Limet et al., 1987; CIoeckaert et al., 1991).
MATERIALS
AND METIIODS
(Difco Laboratories, Detroit, MI). Cells were harvested by gentle agitation in sterile distilled water. Purity and phase (smooth, S, or rough, R) were checked using standard procedures (Alton et aL, 1988). Dilutions were made in phosphate-buffered saline (PBS) containing 0.5 % iv/v) Tween-80 to avoid autoagglutination, and the number of cells was determined by optical density measurements at 600 nm in a spectrophotometer (OD=0.165 for 1 0 9 cells/ml, with I-cm light path). Cells were killed by addition of 20 ptl/ml to 10 ~°cells/ml of a 5 °70 (v/v) peracetic acid solution and incubated overnight at room temperature. After three washes in sterile distilled water, sterility was checked by plating 0.2 ml of each cell suspension on petri dishes containing the culture medium, LPS fractions
S-LPS fractions of B. melitensis 16M and B. abortus 99 were prepared by the phenol-water method (Leong et al., 1970). The R-LPS fraction of B. mefitensis B115 was obtained by the phenol-waterchloroform-petroleum-ether method (Galanos et al., 1969). Proteinase-K-digested S-LPS fractions of B. rnelitensis 16M and B. abortus 544 were prepared as described previously (Garin-Bastuji et al., 1990).
OPS of B. abortus 99 and B. melitensis 16M were extracted by autoclaving S cells in 2 070 iv/v) acetic acid/10 070(w/v) NaCI (Jacques et al., 1991) and purified by high-performance liquid chromatography (Zygmunt et al., 1988).
Brucella strains Monoclonal antibodies
B. abortus 544 (S) and B. abortus 544 (R), B. abortus 99 (S), B. abortus 45/20 (R), B. mefitensis 16M (S) and B. melitensis B! 15 (R) were obtained
from the Institut National de la Recherche Agronomique, Nouziily, France. Whole Brucella S and R cells used for binding studies
Cultures were grown at 37°C for 48 h on trypticase soy agar (BioM6rieux, Marcy-I'Etoile, France) slants supplemented with 0.1% (w/v) yeast extract
CFU L.PS
mAb OMP OPS PBS PBS-T
= = = = = = =
colony-forming u:fit. lipopolysaccharide. monoclonal antibody. outer m e m b r a n e protein. O polysaccharide. phosphate-buffered saline. PBS+Twccn-20.
Anti-R-LPS mAb used were mAb A68/25B03/ D08 (lgGI), A68/03F03/D05 (lgG2b), A68/29C05/ B06 (IgG3) and A68/10A06/BI 1 (lgM). These mAb were derived fi'om fusion of spleen cells from B. abortus-45/20-infected BALB/c mice with the NSO myeloma (CIoeckaert el al., 1990). Hybridomas were cultured in 200-mi flasks (Falcon) at 37°C with 6.5 070 CO 2 in Dulbecco modified Eagle's medium supplemented with 10 °70 foetal calf serum, 2 °70 hypoxanthine-thymidine (HT supplement), 2 070nonessential amino acid supplement, 2 mM sodium pyru-
R-LPS S-LPS SDS-PAGE
= = =
TBS Tw-TBS
= =
rough LPS. smooth LPS. sodium dodecyl sulphate/polyacrylamide gel el¢ctrophoresis. Tris-buffered saline. TBS+Tween-20.
M O N O C L O N A L A N T I B O D I E S TO BRUCELLA R - L P S
vate, 4 mM glutamine, 200 IU/ml penicillin and 200 tzg/ml streptomycin (all from Gibco). Hybridoma supernates were taken after confluent growth and iysis of the cells and were centrifuged (i,500 g) to remove cell debris. Screening for specificity was done by ELISA using rough B. abortus 45/20 cells and purified B. meliwnsis Bll5 R-LPS as antigens (Cloeckaert et al., 1990). The anti-S-LPS mAb O4F9 (IgG2a) and 12Gl2 (IgGI), obtained as described previously (Limet et al., 1987; Cloeckaert et ai., 1990), were shown to be specific for OPS. ELISA Anti-R-LPS and anti-S-LPS (12G12) hybridoma supernates or purified anti-S-LPS (04F9) immunoglobulin (Ig) fraction were assayed for antibody activity by solid-phase ELISA against antigens coated by overnight incubation at 37°C. Antigens were S-LPS (phenol phase or proteinase-K-digested) (4 izg/ml), R-LPS (I0 ~tg/ml), OPS (4 v~g/ml) and killed S and R B. abortus 544 cells. Brucella cells (OD60o = 1.5) were immobilized on microtitre plates (Greiner, Stuttgart, Germany) by capture with rabbit anti-Brucella immunoglobulins (Cloeckaert et al., 1990) at 10 v.g/ml. For binding studies on whole bacteria, 20-fold concentrated hybridoma supernates or Ig fraction were serially diluted (1/10 to 1/590,490) in PBS 0.02 M pH 7.2 containing 0.05 070Tween-20 (PBS-T). Binding of mAb to the cells was detected with peroxidase-conjugated goat anti-mouse immunoglobulins (Jackson, West Grove, PA) diluted in PBS-T. ABTS (2,2'-azino-di-(3-ethylbenzthiazoline sulphonate) was used as a coloured reagent in the substrate solution as described (Alton et a/., 1988). Light absorbance w,~s measured at 414 nm after l-h incubation with the substrate solution. For binding studies on LPS and OPS, hybridoma supernates were serially diluted as above, starting from !/3 dilution.
Immunoblot techniques The S-LPS and R-LPS antigens were separated by sodium dodecyl sulphate/polyacrylamide gel electrophoresis (SDS-PAGE) as previously described (Garin-Bastuji et ai., 1990). After electrophoresis, the antigens were transferred to nitrocellulose at 0.8 mA/cm 2 for 1 h. Unoccupied sites on the nitrocellulose membranes were blocked by a 30-min incubation in Tris-buffered saline (TBS) (0.15 M NaCI, 10 mM Tris hydrochloride pH 7.5) containing 3 070(w/v) skim milk, at room temperature with agitation. The membranes were then successively incubated overnight at room temperature with hybridoma supernates diluted 1/2 in TBS containing 0.05 070 (v/v) Tween-20 (Tw-TBS) and ! 070 (w/v)
477
skim milk, for l h with rabbit anti-mouse immunoglobulin serum (Nordic, The Netherlands) and l h with peroxidase-conjugated protein A (Sigma) diluted 1/250 and l/l,000, respectively, in the same buffer. Washings between incubations were performed with Tw-TBS. Finally, after 3 washings in Tw-TBS and 1 in TBS, the blots were developed by incubation at room temperature in a solution of TBS containing 0.06 070 (w/v) 4-chloro-l-naphthol and 5 mM H202. The reaction was stopped by washing in distilled water. Membranes were air-dried and photographed.
Immunoelectron microscopy lmmunoelectron microscopy was performed as described previously (Cloeckaert et al., 1990). Binding of the mAb (3-fold diluted hybridoma supernates) was detected with sheep anti-mouse biotinylated immunoglobulins (Amersham, UK) and gold-labelled streptavidin (15 nm) (Amersham, UK).
Mice protection test mAb were screened for protection of mice as described previously (Limet et al., 1987). mAb preparations were controlled for endotoxin contamination by the chromogenie Limulus amoebocyte lysate test (Toxicolor, Seikagaku, Japan), following the recommendations of the manufacturer. Endotoxin contamination was in the order of l0 pg/ml as compared with a standard of Escherichia coil Ol I l LPS. AntiR-LPS mAb (20-fold concentrated hybridoma superhates) and anti-S-LPS mAb 04F9 (4 t~g) were injected subcutaneously (0.1 ml) to groups of six CD-I mice l day before the standard intravenous challenge with B. abortus 544 smooth strain (2 x 105 CFU). Spleen counts were done 7 and 21 days after challenge. The experiment included one negative control group which received saline, and as a positive control we included the anti-S-LPS mAb. The statistical significance was determined after variance analysis by one-side comparison of means with the negative control mean (F-test) (Lison, 1968).
RESULTS mAb specificity by E L I S A The anti-Brucella R-LPS mAb showed weak or absence o f reactivity against S-LPS (phenol phase or proteinase-K-treated S-LPS preparat i o n s ) o r O P S o f B. m e l i t e n s i s 16M (M dominant) (fig. l) and B. a b o r t u s 99
A. CL OECKA ER T E T A L.
478
(A dominant) (not shown), but bound at high dilution on B. melitensis Bi 15 R-LPS (fig. IA). They thus recognize epitopes present in the core lipid A moiety. The mAb 12G!2, specific for OPS, did not react with the R-LPS preparation (fig. 1).
A
I11 o z¢[ m nO (n ¢B <
2"01 1.5 1.0 0.5!
mAb specificity by immunoblot analysis
0.0100" .----..-.-~101" ~ ' " ~ ; 2 " " " - ' ; ; 3
.......104
DILUTION(l/X) B
2.52.0" ml O
z< m nO (n m <
1.5. 1.00.50.0
1 0
. . . . . . . .
,
101
.':
......
,
102
.-- ....
~.
103
~...~-
104
DILUTION(l/X) C 2.0 ILl 0
1.5 1.0 0.5
0.
0 ......... 101 •
.........
102
- .....
3 "~ ....
--",,o
Reactivity of OPS-specific mAb 12G12 and anti-R-LPS mAb in immunoblotting with R-LPS of B. melitensis Bll5 (R) and proteinase-Ktreated S-LPS preparations of B. melitensis 16M (M dominant) and B. abortus 544 (A dominant) is illustrated in figure 2. mAb 12Gl2, which was previously shown to react equally well with S-LPS of A and M dominant strains (Cloeckaert et al., 1992), did not react with the R-LPS preparation and revealed only the high molecular mass S-LPS molecules (between 29 and 66 kDa) (protein molecular mass markers) of both S-LPS preparations. The anti-R-LPS mAb revealed the low molecular mass R-LPS molecules (below 20 kDa) in the R-LPS and S-LPS preparations. The differences in intensity of reactivity to the low molecular mass R-LPS are probably due to different amounts of the latter in each strain. The anti-R-LPS mAb also reacted with the higher molecular mass S-LPS molecules of both S-LPS preparations. The banding patterns of the S-LPS portion detected by the anti-R-LPS mAb were a close succession of regularly spaced narrow bands for the A dominant strain and regularly spaced triplets of bands for the M dominant strain. The mAb A68/10A06/BI 1 detected more short bands of S-LPS than the other mAb, especially on B. abortus. The anti-R-LPS mAb also reacted with the high molecular mass S-LPS molecules of phenol-extracted LPS (data not shown).
DILUTION(l/X) Fig. I. ELISA binding of the anti-K~LPS mAb (hybridoma supernates) A68/25B03/D08 (IgGI) ((), A68/03F03/D05 (IgG2b) (O), A68/29C05/B06 (lgG3) ([]), A68/10A06/BI I (IgM) (A) and anti-OPS mAb (hybridoma supernate) 12G12 (lgGI) ( e ) on R-LPS of B. melitensis B115 (R) (A) and on S-LPS (phenol-phase or proteinase-K.digested) (B) and OPS (C) of B. melitensis 16M (M dominant).
ELISA on whole Brucella cells Binding of the anti-R-LPS mAb was compared to binding of the OPS-specific mAb 04F9 by ELISA using R or S B. abortus 544 whole cells. The anti-R-LPS mAb bound on both R
M O N O C L O N A L A N T I B O D I E S TO BRUCELLA R-LPS
479
kDa
9
-.
•
11 I! I1 Ii l i
Fig. 2. lmmunoblot of B. melitensis BI 15 R-LPS (a) and proleina~e-K-treated S-LPS of B. abortus 544 (A dominant) (b) and B. melitensis 16M (M domhiant) (c) with anti-R-LPS mAb (hybridoma supernates) A68/03F03/D05 (IgG2b) (!), A68/10A06/BI 1 (lgM) (2), A68/29C05/B06 (IgG3) (3), A68/25B03/D08 (IgGl) (4) and anti-OPS mAb 12G12 (lgGI) (5).
and S cells, but binding was superior on R cells (fig. 3). In contrast, S-LPS-specific mAb bound only on S cells.
lmmunoelectron microscopy Binding of the anti-R-LPS mAb on whole R and S B. abortus 544 cells was studied by using immunogold labelling for electron microscopy. lmmunogold labelling using the anti-R-LPS mAb (3-fold diluted hybridoma suparnates) was only observed on the R ceils (fig. 4, C and D). At higher antibody concentration (1/50 diluted
ascitic fluid) immunogold labelling was observed on both S and R cells, but labelling was always more intense on the R cells (data not shown). With the O-specific mAb 12G12, labelling was only seen on S cells (fig. 4, A and B). No immunolabelling was observed with a negative control mAb (unrelated to Brucella) (not shown).
Protection conferred on mice Among the anti-R-LPS mAb used in this study, only one (mAb A68/29C05/B06 of lgG3 isotype) significantly reduced (p < 0.05) spleen
480
A. CLOECKAERT ET AL.
A 2.5 2.0I&l
o
z < m nO r/J
1.5" 1.0"
m ,< 0.5 0.0
~
10 1
......
,
• -.
....
10 2
.--.,
10 3
.-- .......
,
10 4
DILUTION ( l / X )
B 2.5 2.0
o z
1.5
m m 0
1.0 °
m 0.5 0.0 10 1
10 2
10 3
10 4
DILUTION ( l / X )
Fig. 3. ELISA binding of the anti-R-LPS mAb (20-fold concentrated hybridoma supernates) A68/25B03/D08 (IgGl) ("~,). A68/03F03/D05 (IgG2b) (C)), A68/29C05/ B06 (lgG3) ([:]), A 6 8 / 1 0 A 0 6 / B l l (IgM) (A) and antiS-LPS (A determinant) m A b 04F9 (IgG2a) (immunoglobulin fraction at 40 ~tg/ml) (O) on rough (A) and smooth (B) B. aborlus 544 bacterial cells.
colonization at 7 and 21 days after the B. abortus 544 (S) challenge (table I). Reduction of spleen colonization using the S-LPS-specific mAb 04F9 was, however, greater.
DISCUSSION
mAb against Brucella R-LPS were produced by fusion of the NSO myeioma cells with spleen cells from mice infected with B. abortus 45/20
rough cells (Cloeckaert et al., 1990). They were selected by ELISA using whole B. abortus 45/20 rough cells and R-LPS from B. melitensis B115 as antigens. Thus, selected mAb recognized cell surface-exposed epitopes and probably the most exposed core oligosaccharide region of R-LPS. Specificity was confirmed by immunoblotting using R-LPS and S-LPS preparations. The banding patterns of ~he high molecular mass S-LPS portion detected by anti-R-LPS were similar to those detected by S-LPS-specific mAb described elsewhere (Garin-Bastuji et al., 1990). In ELISA, these mAb bound weakly or not at all to S-LPS. However, in immunoblotting of either proteinase-K-treated or phenol-extracted S-LPS preparations, they reacted with both the low molecular mass R-LPS molecules and with the higher molecular mass S-LPS molecules containing O chains of variable length. This indicated that in S-LPS molecules, when separated in SDSoPAGE, epitopes of the core lipid A region are accessible to the anti-R-LPS mAb. Absence of reactivity or weak reactivity to S-LPS have been reported for anti-core LPS mAb derived from mice immunized with the rough E. coil J5 strain (Bogard et al., 1987). Others (Appelmelk et al., 1988) have also produced mAb against the LPS core ofE. coli strain J5 and also reacting with other Gram-negative bacilli. They concluded that OPS does not necessarily render the core inaccessible to antibody because the mAb produced still reacted with intact smooth E. coil cells. In contrast, Gigliotti and Shenep (1985) reported, using the same E. coli strain, the production of mAb against the core LPS and these mAb did not bind to whole cells of smooth E. coil strains. However, in both studies no immunoblot analysis of LPS preparations was perf o r m e d . As m e a s u r e d by E L I S A , our anti-R-LPS mAb bound less to S than to R intact B. abortus. This was also observed on R and S B. melitensis cell~ (data not shown). These results were confirmed by immunoelectron microscopy. Therefore, different binding patterns of anti-R-LPS mAb on S bacteria probably results from reactivity with R-form LPS non-substituted with O chains and present in various amounts on bacterial surfaces.
MONOCLONAL
A N T I B O D I E S TO B R U C E L L A R - L P S
481
Cm
AO
i ¸
~"
-
C
-
.;t
i¸
:i< !! i Fig. 4. lmmunogold labelling of 8. abortus 544 smooth (A, C) and rough (B, D) cells with antiOPS mAb 12G12 (lgGl) (A, B) and anti-R-LPS mAb A68/03F03/D05 (lgG2b) (C, D); x 30,000.
Despite the accessibility of R-LPS epitopes to mAb on the S B. a b o r t u s 544 strain used to infect CD-I mice, 3 out of the 4 anti-R-LPS mAb did not confer significant protection. Only I antiR-LPS mAb, of IgG3 isotype, reduced splenic
infection significantly, but protection was lower than with the anti-OPS mAb used in this study. This can be connected with the lower binding of the anti-R-LPS mAb observed on the S B. a b o r t u s 544 as compared with that of the
482
A. C L O E C K A E R T E T AL.
Table !. Protective activity of anti-R-LPS mAb to B. abortus 544 (S).
mAb ~a~ A68/O3F03/D05 A68/IOAO6/Bl I A68/25B03/D08 A68/29C05/B06 04F9 Control (saline)
Specificity
Isotype
R-LPS R-LPS R-LPS R-LPS S-LPS
lgG2b lgM lgGI lgG3 igG2a
Mean (SD) Ioglo Brucella CFU at days post-challenge the. 7 21 5.86 (0.20) 5.75 (0.50) 5.70 (0.52) 5.59 (0.3 ! ) ~'~ 5.13 (0.18) ~''~ 6.07 (0.21)
4.75 (0.20) 4.80 (0.50) 4.81 (0.34) 4.13 (0.47) ~'~ 3.90 (0.41) ~'*~ 4.60 (0.29)
,Jr Injection of mAb was performed subcutaneously I day prior to challenge. t~, CD-I mice were challenged with 200,000 CFU of strain 544 (S), killed after 7 and 21 days, and Brucella from spleen were counted. (") p < 0.05. {**) p < 0.01. in comparison with negative cor,!rol group according to variance analysis.
OPS-specific mAb (except for the IgM mAb). This could be due not only to a lower number of R-LPS molecules on the surface of smooth Brucella cells but also to steric hindrance caused by the long LPS O chains, as also described for accessibility of outer membrane proteins (OMP) to mAb (Cloeckaert et al., 1990). It was also previously shown that anti-OMP m?,b conferred no protection or much lower protection than anti-S-LPS mAb (Jacques et al., 1992; Montaraz et aL, 1986; Cloeckaert et al., 1991). In vivo, antibodies against the most exposed surface antigens, such as the O chains of S-LPS, are probably the most effective in fully achieving opsonization of Brucella organisms. The isotype of mAb probably also plays an important role in protection (Oishi et al., 1992). Non-agglutinating antibodies which bind to rough Brucella cells have been found in the sera of immunized cattle (Santisteban et al., 1986). Interestingly, they lacked protective capacities (viz opsonisation, complement fixation, antibody-dependent cytotoxicity) (Parma et al., 1987). As they were obtained by immunoadsorption with rough Brucella cells from sera previously adsorbed with smooth cells, they were most probably directed to R-LPS and also, as claimed, to outer membrane proteins. Lack of protection by anti-R-LPS mAb has been described for other Gram-negative infec-
tions (Appelmelk et al., 1988; Miner et al., 1986), i.e., mAb to the LPS-core region were unable to protect mice from lethal infection or endotoxaemia. Comparison is however difficult since this mouse model was based on lethality. Other authors (Dunn et al., 1986; Johns et aL, 1983 ; Teng et ai., 1985 ; Terashima et aL, 1991) described antibodies reacting with the LPS core, thus preventing death from Gram-negative sepsis. In most cases, antibodies were directed against the lipid A, the toxic moiety of LPS molecules. Work is in progress to define which epitopes of the LPS-core region, i.e. lipid A or core oligosaccharide, our mAb botmd and to explore their role in infections by natura!ly rough Brucella species.
Acknowledgements
We thank J.M. Verger and M. Grayon for providing Brucella strains. A. Cloeckaert was supported by the lnstitut National pour l'Encouragementde la Recherche Scientifiquedans I'lndustrie et l'Agriculture. R.A. Bowden is indebted to AUPELF (France), Fundaci6n Antorchas, Facuitad de Ciencias Veterinarias-U.N.C. and SECYT(Argentina) for support. Part of this paper corresponds to the thesis work of the third author.
MONOCLONAL
A N T I B O D I E S TO B R U C E L L A R - L P S
Anlicorps monoclonaux anti-LPS-R de Brucella: caract~risation el ~valuation de leur effet protecteur contre B. abortus
Nous avons caract6ris6 4 anticorps monoclonaux (AcM) sp6cifiques du lipopolyoside rugueux (LPS-R) de Brucella. Les AcM ont 6t~ s~lectionn~s par une 6preuve immunoenzymatique (ELISA) sur des cellules enti~res de B. abortus 45/20, souche rugueuse, et sur du LPS-R de B. melitensis B115. La sp~cificit6 a ~t6 confirm6e par analyse des immunoempreintes, utilisant le LPS-R et le LPS lisse (LPS-S). Les AcM anti-LPS-R ont r~v~l~ les bandes de basse masse mol6culaire, en dessous de 20,1 kDa, dans les pr6parations de LPS-R et LPS-S, ainsi que les profils typiques A et M dans les mol6cules de LPS~S de haute masse mol6culaire (entre 21,5 et 66 kDa) dans les ~chantillons de LPS-S. Un AcM sp~cifique du polyoside O a r6v~16 uniquement les bandes de haute masse mol~culaire dans les preparations de LPS-S. En E L I S A , les AcM anti-LPS-R se sont fix~ mieux sur les variantes rugueuses de B. abortus 544 que sur les lisses, et cela a ~t~ confirm6 par immunomarquage /l l ' o r coiloi'dal en microscopie ~lectronique. L'activit6 protectrice des AcM anti-LPS-R de differents isotypes a ~t~ ~tudi6e chez la souris et compar~e/~ celle d ' u n AcM sp6cifique du LPS-S. Seul I ' A c M d ' i s o t y p e lgG3 r6duit significativement l'infection spl~nique, sans pourtant atteindre l'activit6 pr~sent i e par i ' A c M sp6cifique du LPS-S (utilis6 comme t~moin positif). Mots-cl~s : LPS, Brucella abortus, Brucella melitensis, igG, Anticorps m o n o c l o n a l ; Souris, Protection, Epitopes.
References
Alton, G.G., Jones, L.M., Angus, R.D. & Verger, J.M. (1988), Techniques for the brucellosis laboratory. lnstitut National de la Recherche Agronomique, Paris. Appelmelk, B.J., Verweij-Van Vught, A.M.J., Maaskant, J.J., Schouten, W.F., De Jonge, A.J.R., Thijs, L.G. & Maclaren, D.M. (1988), Production and characterization of mouse monoclonal antibodies reacting with the lipopolysaccharide core region of Gram-negative bacilli. J. Med. MicrobioL, 26, 107-114. Bogard, W.C., Dunn, D.L., Abernethy, K., Kilgarriff, C. & Kung, P.C. (1987), Isolation and characterization of murine monoclonal antibodies specific for Gramnegative bacterial lipopolysaccharide: association of cross-genus reactivity with lipid A specificity. Infect. hnmun., 55, 899-908. Bundle, D.R., Cherwonogrodzky, J.W., Gidney, M.A.J., Meikle, P.J., Perry, M.B. & Peters, T. (1989), Deft-
483
nition of Brucella A and M epitopes by monoclonal typing reagents and synthetic oligosaccharides. Infect. immun., 57, 2829-2836. CIoeckaert, A., de Wergifosse, P., Dubray, G. & Limet, J.N. (1990), Identification of seven surface-exposed Brucella outer membrane proteins by use of monoclonal antibodies: immunogold labeling for electron microscopy and enzyme-linked immunosorbent assay. Infect. immun., 58, 3980-3987. Cloeckaert, A., Jacques, l., Bosseray, N., Limet, J.N., Bowden, R., Dubray, G. & Plommet, M. (1991), Protection conferred on mice by monoclonai antibodies directed against outer-membrane-protein antigens of Brucella. J. Med. Microbiol., 34, 175-180. Cloeckaert, A., Jacques, !., de Wergifosse, P., Dubray, G. & Limet, J.N. (1992), Protection against Brucella melitensis or Brucella abortus in mice with immunoglobulin G (lgG), lgA and lgM monoclonal antibodies specific for a common epitope shared by the Brucella A and M smooth lipopolysaccharide types. Infect. Immun., 60, 312-315. Dunn, D.L., Ewald, D.C., Chandan, N. & Cerra, F.B. (1986), Immunotherapy of Gram-negative bacterial sepsis. Arch. Surg., 121, 58-62. Galanos, C., Luderitz, O. & Westphal, O. (1969), A new method for the extraction of R-lipopolysaccharides. Eur. J. Biochem., 9, 245-249. Garin-Bastuji, B., Bowden, R.A., Dubray, G. & Limet, J.N. (1990), Sodium dodecyl sulfate polyacrylamide gel electrophoresis and immunoblotting analysis of smooth lipopolysaccharide heterogeneity among Brucella biovars related to A and M specificities. J. Clin. Microbiol., 28, 2169-2174. Gigliotti, F. & Shencp, J.L. (1985), Failure of monocional antibodies to core glycolipid to bind intact smooth strains of Escherichia coil J. Infect. Dis., 151, 1005-1011. Jacques, l., Olivier-Bernardin, V. & Dubray, G. (1991), Induction of antibody and protective responses in mice by Brucella O-polysaccharide-BSA conjugate. Vaccine, 9, 896-900. Jacques, l., Cloeckaert, A., Limet, J.N. & Dubray, G. 0992), Protection conferred on mice by combinations of monoclonal antibodies directed against outermembrane proteins or smooth lipopolysaccharide of Brucella. J. Med. MicrobioL, 37, 100-103. Johns, M., SkehiU, A. & McCabe, W.R. (1983), Immunization with rough mutants of Salmonella minnesota. - - IV. Protection by antisera to O and rough antigens against endotoxin. J. Infect. Dis., 147, 57-67. Leong, D., Diaz, R., Milner, K., Rudbach, J. & Wilson, J.B. (1970), Some structural and biological properties of Brucella endotoxin. Infect. hnmun., I, 174-182. Limet, J.N., Bosseray, N., Garin-Bastuji, B., Dubray, G. & Plommet, M. (1989), Humoral immunity in mice mediated by monoclonal antibodies against the A and M antigens of Brucella. J. Med. MicrobioL, 30, 37-43. Limet, J.N., Plommet, A.M., Dubray, G. & Plommet, M. (1987), Immunity conferred to mice by anti-LPS monoclonal antibodies in murine brucellosis. Ann. Inst. Pasteur/lmmunol., 138, 417-424. Lison, L. (1968), Statistique appliqu~e ~ la biologie exp6rimentale. Collection Science et Techniques d'Aujourd'hui. Gauthier-Villars, Paris.
484
A. CLOIzCKAERT
Miner, K.M., Manyak, C.L.. Williams, E., Jacksoll, J., Jewell, M., Gammon, M.T., Ehrenfreund, C., Hayes, E., Callahan, L.T., III, Zweerink, H. & Sigal, N.H. (1986), Characterization of routine monoclonal antibodies to Escherichia coli J5. Infect. hnmun., 52, 56-62. Montaraz, J.A., Winter, A.J., Hunter, D.M., Sowa, B.A., Wu, A.M. & Adams, L.G. (1986), Protection against Brucella abortus in mice with O-polysaccharidespecific monoclonal antibodies. Infect. lmmun., 5 I, 961-963. Moreno, E., Pitt, M.W., Jones, L.M., Schurig, G.G. & Berman, D.T. (1979), Purification and characterization of smooth and rough lipopolysaccharides from Brucella abortus. J. Bacteriol., 138, 361-369. Oishi, K., Koles, N.L., Guelde, G. & Pollack, M. (1992), Antibacterial and protective properties of monoclonal antibodies reactive with Escherichia coli OI II:B4 lipopolysaccharide: relation to antibody isotype and complement-fixingactivity. J. Infect. Dis., 165, 34-45. Palmer, D.A. & Douglas, J.T. (1989), Analysis of Brucella lipopolysaccharide with specific and cross-reacting monoclonal antibodies. J. Clin. Microbiol., 27, 2331-2337. Parma, A.E., Bowden, R.A., Santisteban, C.G., Cerone, S. & Fernfindez, A. (1987), Effect of bovine nonagglutinating antibodies on the blood clearance of 1-131 Brucella abortus 45/20. Vet. MicrobioL, 15, 121-128. Phillips, M., Deyoe, B.L. & Canning, P.C. (1989), Protection of mice against Brucella abortus infection by
ET AL.
inoculation with monoclonal antibodies recognizing Brucella O-antigen. Am. J. Vet. Res., 50, 2158-2161. Riezu-Boj, J.l., Moriy6n, I., Blasco, J.M., Matin, C.M. & Diaz, R. (1986), Comparison of lipopolysaccharide and outer membrane protein-lipopolysaccharide extracts in an enzyme-linked immunosorbent assay for the diagnosis of Brucella ovis infection. J. Clin. Microbiol., 23, 938-942. Santisteban, C.G., Parma, A.E., Cerone, S.I. & Bowden, R.A. (1986), Detection of antigenic fractions from Brttcella abortus s45/20 which bind to non-agglutinating antibodies using electroblotting and enzymelinked antibody probes. Vet. Microbiol., 12, 353-358. Teng, N.N.H., Kaplan, H.S., Hebert, J.M., Moore, C., Douglas, H., Wunderlich, A. & Braude, A.I. (1985), Protection against Gram-negative bacteremia and endotoxemia with human monoclonal IgM antibodies. Proc. nat. Acad. Sci. (Wash.), 82, 1790-1794. Terashima, M., Uezumi, I., Tomio, T., Kato, M., lrie, K., Okuda, T., Yokota, S.I. & Noguehi, H.A. (1991), Protective human monoclonal antibody directed to the outer core region of Pseudomonas aeruginosa lipopolysaccharide. Infect. Immun., 59, I-6. Vizcaino, N., Chordi, A. & Fernfindez-Lago, L. (1991), Characterization of smooth Brucella lipopolysaccharides and polysaecharides by monoclonal antibodies. Res. Microbiol., 142, 971-978. Zygmunt, M.S., Dubray, G., Bundle, D.R. & Perry, M.B. (1988), Purified native haptens of Brucella abortus BI9 and Bruce#a melitensis 16M reveal the lipopolysaccharide origin of the antigens. Ann. Inst. Pasteur/Microbiol., 139, 421-433.
Res. Microbiol.
1993, 144, 475-484
Monoclonal antibodies to Brucella rough lipopolysaccharide: characterization and evaluation of their protective effect against B. abortus A. Cloeckaert (=) (*), I. Jacques 0), R.A. Bowden (0, G. Dubray (I) and J.N. Limet (2) ¢o Institut National de la Recherche Agronomique, Centre de Recherches de Tours, Laboratoire de Pathologic infectieuse et hnmunologie, 37380 Nouzilty (France), and (2) Facult~s Universitaires Notre-Dame de la Paix, 61 rue de Bruxelles, 5009 Namur (Belgium)
SUMMARY We characterized 4 monoclonal antibodies (mAb) specific for rough lipopolysaccharide (R-LPS) of Bruce#a. mAb were selected by enzyme-linked immunosorbent assay IELISA) on whole R. abortus 45/20 rough cells and R-LPS from B. melitensis B115 rough cells. Specificity ~as confirmed by immunoblot analysis using R-LPS and smooth LPS (S-LPS) preparations. Anti-R-LPS revealed the low molecular mass R-LPS molecules below 20.1 kDa in the R-LPS and S-LPS preparations as well as the typical A and M patterns in high molecular mass S-LPS molecules (between 21.5 and 66 kDa) in the S-LPS preparations. An O-polysaccharide-specific mAb revealed only high molecular mass S-LPS molecules in the S-LPS preparation. In ELISA the anti-R-LPS mAb bound better on rough than on smooth B. abortus 544 whole cells, and this was confirmed by immunoelectron microscopy. Protective activity of anti-R-LPS mAb of different isotypes was tested on mice and compared with an S-LPS-specific mAb. Only the IgG3 mAb reduced significantly the splenic infection but did not reach the level of protection conferred by the S-LPS-specific mAb.
Key-words: LPS, Brucella abortus, Brucella melitensis, IgG, mAb; Mouse, Protection, Epitopes.
INTRODUCTION The smooth lipopolysaccharides (S-LPS) of Brucella spp. and other Gram-negative bacteria
are complex molecules comprising three structurally and serologicaUy distinct regions, i.e. a glycolipid moiety called lipid A, a core oligosaccharide and the O-polysaccharide lOPS) chain.
Submitted March II, 1993, accepted July 28, 1993. {*) Corresponding author.
The Brucella 0 chain contains two distinct epitopes, designated A and M (Bundle et aL, 1989). Their relative amounts vary among the smooth Brucella strains, and these epitopes are absent on the rough strains, which lack OPS (Alton et aL, 1988). mAb to Brucella S-LPS have been reported with specificity for the A, the M and the common epitopes, some of which cross-react
476
A. C L O E C K A E R T E T A L .
with Yersinia enterocolitica 0:9 S-LPS (Bundle et al., 1989; Cloeckaert et al., 1992; GarinBastuji et al., 1990; Limet et al., 1989; Palmer and Douglas, 1989; Vizcaino et al., 1991). These mAb confer protection on mice by reducing the number of Brucella in the spleen and liver (Cloeckaert et al., 1992; Limet et al., 1989, 1987; Montaraz et al., 1986; Phillips et al., 1989). Brucella rough LPS (R-LPS) lacking OPS and expressed predominantly in rough strains of Brucella has been immunochemically characterized (Moreno et al., 1979), and antibody response against it has been observed in B. ovis infections (Riezu-Boj et al., 1986). We obtained mAb specific for R-LPS by fusion with the NSO myeloma of spleen cells of mice infected by the rough B. a b o r t u s 45/20 and B. melitensis B115 strains (Cloeckaert et al., 1990). Our purpose here was to characterize the anti-R-LPS mAb in terms of molecular and cellular binding and to compare their protective activity in mice against B. a b o r t u s with an anti-S-LPS mAb already shown to be protective (CIoeckaert et al., 1992; Limet et al., 1989; Jacques et al., 1992; Limet et al., 1987; CIoeckaert et al., 1991).
MATERIALS
AND METIIODS
(Difco Laboratories, Detroit, MI). Cells were harvested by gentle agitation in sterile distilled water. Purity and phase (smooth, S, or rough, R) were checked using standard procedures (Alton et aL, 1988). Dilutions were made in phosphate-buffered saline (PBS) containing 0.5 % iv/v) Tween-80 to avoid autoagglutination, and the number of cells was determined by optical density measurements at 600 nm in a spectrophotometer (OD=0.165 for 1 0 9 cells/ml, with I-cm light path). Cells were killed by addition of 20 ptl/ml to 10 ~°cells/ml of a 5 °70 (v/v) peracetic acid solution and incubated overnight at room temperature. After three washes in sterile distilled water, sterility was checked by plating 0.2 ml of each cell suspension on petri dishes containing the culture medium, LPS fractions
S-LPS fractions of B. melitensis 16M and B. abortus 99 were prepared by the phenol-water method (Leong et al., 1970). The R-LPS fraction of B. mefitensis B115 was obtained by the phenol-waterchloroform-petroleum-ether method (Galanos et al., 1969). Proteinase-K-digested S-LPS fractions of B. rnelitensis 16M and B. abortus 544 were prepared as described previously (Garin-Bastuji et al., 1990).
OPS of B. abortus 99 and B. melitensis 16M were extracted by autoclaving S cells in 2 070 iv/v) acetic acid/10 070(w/v) NaCI (Jacques et al., 1991) and purified by high-performance liquid chromatography (Zygmunt et al., 1988).
Brucella strains Monoclonal antibodies
B. abortus 544 (S) and B. abortus 544 (R), B. abortus 99 (S), B. abortus 45/20 (R), B. mefitensis 16M (S) and B. melitensis B! 15 (R) were obtained
from the Institut National de la Recherche Agronomique, Nouziily, France. Whole Brucella S and R cells used for binding studies
Cultures were grown at 37°C for 48 h on trypticase soy agar (BioM6rieux, Marcy-I'Etoile, France) slants supplemented with 0.1% (w/v) yeast extract
CFU L.PS
mAb OMP OPS PBS PBS-T
= = = = = = =
colony-forming u:fit. lipopolysaccharide. monoclonal antibody. outer m e m b r a n e protein. O polysaccharide. phosphate-buffered saline. PBS+Twccn-20.
Anti-R-LPS mAb used were mAb A68/25B03/ D08 (lgGI), A68/03F03/D05 (lgG2b), A68/29C05/ B06 (IgG3) and A68/10A06/BI 1 (lgM). These mAb were derived fi'om fusion of spleen cells from B. abortus-45/20-infected BALB/c mice with the NSO myeloma (CIoeckaert el al., 1990). Hybridomas were cultured in 200-mi flasks (Falcon) at 37°C with 6.5 070 CO 2 in Dulbecco modified Eagle's medium supplemented with 10 °70 foetal calf serum, 2 °70 hypoxanthine-thymidine (HT supplement), 2 070nonessential amino acid supplement, 2 mM sodium pyru-
R-LPS S-LPS SDS-PAGE
= = =
TBS Tw-TBS
= =
rough LPS. smooth LPS. sodium dodecyl sulphate/polyacrylamide gel el¢ctrophoresis. Tris-buffered saline. TBS+Tween-20.
M O N O C L O N A L A N T I B O D I E S TO BRUCELLA R - L P S
vate, 4 mM glutamine, 200 IU/ml penicillin and 200 tzg/ml streptomycin (all from Gibco). Hybridoma supernates were taken after confluent growth and iysis of the cells and were centrifuged (i,500 g) to remove cell debris. Screening for specificity was done by ELISA using rough B. abortus 45/20 cells and purified B. meliwnsis Bll5 R-LPS as antigens (Cloeckaert et al., 1990). The anti-S-LPS mAb O4F9 (IgG2a) and 12Gl2 (IgGI), obtained as described previously (Limet et al., 1987; Cloeckaert et ai., 1990), were shown to be specific for OPS. ELISA Anti-R-LPS and anti-S-LPS (12G12) hybridoma supernates or purified anti-S-LPS (04F9) immunoglobulin (Ig) fraction were assayed for antibody activity by solid-phase ELISA against antigens coated by overnight incubation at 37°C. Antigens were S-LPS (phenol phase or proteinase-K-digested) (4 izg/ml), R-LPS (I0 ~tg/ml), OPS (4 v~g/ml) and killed S and R B. abortus 544 cells. Brucella cells (OD60o = 1.5) were immobilized on microtitre plates (Greiner, Stuttgart, Germany) by capture with rabbit anti-Brucella immunoglobulins (Cloeckaert et al., 1990) at 10 v.g/ml. For binding studies on whole bacteria, 20-fold concentrated hybridoma supernates or Ig fraction were serially diluted (1/10 to 1/590,490) in PBS 0.02 M pH 7.2 containing 0.05 070Tween-20 (PBS-T). Binding of mAb to the cells was detected with peroxidase-conjugated goat anti-mouse immunoglobulins (Jackson, West Grove, PA) diluted in PBS-T. ABTS (2,2'-azino-di-(3-ethylbenzthiazoline sulphonate) was used as a coloured reagent in the substrate solution as described (Alton et a/., 1988). Light absorbance w,~s measured at 414 nm after l-h incubation with the substrate solution. For binding studies on LPS and OPS, hybridoma supernates were serially diluted as above, starting from !/3 dilution.
Immunoblot techniques The S-LPS and R-LPS antigens were separated by sodium dodecyl sulphate/polyacrylamide gel electrophoresis (SDS-PAGE) as previously described (Garin-Bastuji et ai., 1990). After electrophoresis, the antigens were transferred to nitrocellulose at 0.8 mA/cm 2 for 1 h. Unoccupied sites on the nitrocellulose membranes were blocked by a 30-min incubation in Tris-buffered saline (TBS) (0.15 M NaCI, 10 mM Tris hydrochloride pH 7.5) containing 3 070(w/v) skim milk, at room temperature with agitation. The membranes were then successively incubated overnight at room temperature with hybridoma supernates diluted 1/2 in TBS containing 0.05 070 (v/v) Tween-20 (Tw-TBS) and ! 070 (w/v)
477
skim milk, for l h with rabbit anti-mouse immunoglobulin serum (Nordic, The Netherlands) and l h with peroxidase-conjugated protein A (Sigma) diluted 1/250 and l/l,000, respectively, in the same buffer. Washings between incubations were performed with Tw-TBS. Finally, after 3 washings in Tw-TBS and 1 in TBS, the blots were developed by incubation at room temperature in a solution of TBS containing 0.06 070 (w/v) 4-chloro-l-naphthol and 5 mM H202. The reaction was stopped by washing in distilled water. Membranes were air-dried and photographed.
Immunoelectron microscopy lmmunoelectron microscopy was performed as described previously (Cloeckaert et al., 1990). Binding of the mAb (3-fold diluted hybridoma supernates) was detected with sheep anti-mouse biotinylated immunoglobulins (Amersham, UK) and gold-labelled streptavidin (15 nm) (Amersham, UK).
Mice protection test mAb were screened for protection of mice as described previously (Limet et al., 1987). mAb preparations were controlled for endotoxin contamination by the chromogenie Limulus amoebocyte lysate test (Toxicolor, Seikagaku, Japan), following the recommendations of the manufacturer. Endotoxin contamination was in the order of l0 pg/ml as compared with a standard of Escherichia coil Ol I l LPS. AntiR-LPS mAb (20-fold concentrated hybridoma superhates) and anti-S-LPS mAb 04F9 (4 t~g) were injected subcutaneously (0.1 ml) to groups of six CD-I mice l day before the standard intravenous challenge with B. abortus 544 smooth strain (2 x 105 CFU). Spleen counts were done 7 and 21 days after challenge. The experiment included one negative control group which received saline, and as a positive control we included the anti-S-LPS mAb. The statistical significance was determined after variance analysis by one-side comparison of means with the negative control mean (F-test) (Lison, 1968).
RESULTS mAb specificity by E L I S A The anti-Brucella R-LPS mAb showed weak or absence o f reactivity against S-LPS (phenol phase or proteinase-K-treated S-LPS preparat i o n s ) o r O P S o f B. m e l i t e n s i s 16M (M dominant) (fig. l) and B. a b o r t u s 99
A. CL OECKA ER T E T A L.
478
(A dominant) (not shown), but bound at high dilution on B. melitensis Bi 15 R-LPS (fig. IA). They thus recognize epitopes present in the core lipid A moiety. The mAb 12G!2, specific for OPS, did not react with the R-LPS preparation (fig. 1).
A
I11 o z¢[ m nO (n ¢B <
2"01 1.5 1.0 0.5!
mAb specificity by immunoblot analysis
0.0100" .----..-.-~101" ~ ' " ~ ; 2 " " " - ' ; ; 3
.......104
DILUTION(l/X) B
2.52.0" ml O
z< m nO (n m <
1.5. 1.00.50.0
1 0
. . . . . . . .
,
101
.':
......
,
102
.-- ....
~.
103
~...~-
104
DILUTION(l/X) C 2.0 ILl 0
1.5 1.0 0.5
0.
0 ......... 101 •
.........
102
- .....
3 "~ ....
--",,o
Reactivity of OPS-specific mAb 12G12 and anti-R-LPS mAb in immunoblotting with R-LPS of B. melitensis Bll5 (R) and proteinase-Ktreated S-LPS preparations of B. melitensis 16M (M dominant) and B. abortus 544 (A dominant) is illustrated in figure 2. mAb 12Gl2, which was previously shown to react equally well with S-LPS of A and M dominant strains (Cloeckaert et al., 1992), did not react with the R-LPS preparation and revealed only the high molecular mass S-LPS molecules (between 29 and 66 kDa) (protein molecular mass markers) of both S-LPS preparations. The anti-R-LPS mAb revealed the low molecular mass R-LPS molecules (below 20 kDa) in the R-LPS and S-LPS preparations. The differences in intensity of reactivity to the low molecular mass R-LPS are probably due to different amounts of the latter in each strain. The anti-R-LPS mAb also reacted with the higher molecular mass S-LPS molecules of both S-LPS preparations. The banding patterns of the S-LPS portion detected by the anti-R-LPS mAb were a close succession of regularly spaced narrow bands for the A dominant strain and regularly spaced triplets of bands for the M dominant strain. The mAb A68/10A06/BI 1 detected more short bands of S-LPS than the other mAb, especially on B. abortus. The anti-R-LPS mAb also reacted with the high molecular mass S-LPS molecules of phenol-extracted LPS (data not shown).
DILUTION(l/X) Fig. I. ELISA binding of the anti-K~LPS mAb (hybridoma supernates) A68/25B03/D08 (IgGI) ((), A68/03F03/D05 (IgG2b) (O), A68/29C05/B06 (lgG3) ([]), A68/10A06/BI I (IgM) (A) and anti-OPS mAb (hybridoma supernate) 12G12 (lgGI) ( e ) on R-LPS of B. melitensis B115 (R) (A) and on S-LPS (phenol-phase or proteinase-K.digested) (B) and OPS (C) of B. melitensis 16M (M dominant).
ELISA on whole Brucella cells Binding of the anti-R-LPS mAb was compared to binding of the OPS-specific mAb 04F9 by ELISA using R or S B. abortus 544 whole cells. The anti-R-LPS mAb bound on both R
M O N O C L O N A L A N T I B O D I E S TO BRUCELLA R-LPS
479
kDa
9
-.
•
11 I! I1 Ii l i
Fig. 2. lmmunoblot of B. melitensis BI 15 R-LPS (a) and proleina~e-K-treated S-LPS of B. abortus 544 (A dominant) (b) and B. melitensis 16M (M domhiant) (c) with anti-R-LPS mAb (hybridoma supernates) A68/03F03/D05 (IgG2b) (!), A68/10A06/BI 1 (lgM) (2), A68/29C05/B06 (IgG3) (3), A68/25B03/D08 (IgGl) (4) and anti-OPS mAb 12G12 (lgGI) (5).
and S cells, but binding was superior on R cells (fig. 3). In contrast, S-LPS-specific mAb bound only on S cells.
lmmunoelectron microscopy Binding of the anti-R-LPS mAb on whole R and S B. abortus 544 cells was studied by using immunogold labelling for electron microscopy. lmmunogold labelling using the anti-R-LPS mAb (3-fold diluted hybridoma suparnates) was only observed on the R ceils (fig. 4, C and D). At higher antibody concentration (1/50 diluted
ascitic fluid) immunogold labelling was observed on both S and R cells, but labelling was always more intense on the R cells (data not shown). With the O-specific mAb 12G12, labelling was only seen on S cells (fig. 4, A and B). No immunolabelling was observed with a negative control mAb (unrelated to Brucella) (not shown).
Protection conferred on mice Among the anti-R-LPS mAb used in this study, only one (mAb A68/29C05/B06 of lgG3 isotype) significantly reduced (p < 0.05) spleen
480
A. CLOECKAERT ET AL.
A 2.5 2.0I&l
o
z < m nO r/J
1.5" 1.0"
m ,< 0.5 0.0
~
10 1
......
,
• -.
....
10 2
.--.,
10 3
.-- .......
,
10 4
DILUTION ( l / X )
B 2.5 2.0
o z
1.5
m m 0
1.0 °
m 0.5 0.0 10 1
10 2
10 3
10 4
DILUTION ( l / X )
Fig. 3. ELISA binding of the anti-R-LPS mAb (20-fold concentrated hybridoma supernates) A68/25B03/D08 (IgGl) ("~,). A68/03F03/D05 (IgG2b) (C)), A68/29C05/ B06 (lgG3) ([:]), A 6 8 / 1 0 A 0 6 / B l l (IgM) (A) and antiS-LPS (A determinant) m A b 04F9 (IgG2a) (immunoglobulin fraction at 40 ~tg/ml) (O) on rough (A) and smooth (B) B. aborlus 544 bacterial cells.
colonization at 7 and 21 days after the B. abortus 544 (S) challenge (table I). Reduction of spleen colonization using the S-LPS-specific mAb 04F9 was, however, greater.
DISCUSSION
mAb against Brucella R-LPS were produced by fusion of the NSO myeioma cells with spleen cells from mice infected with B. abortus 45/20
rough cells (Cloeckaert et al., 1990). They were selected by ELISA using whole B. abortus 45/20 rough cells and R-LPS from B. melitensis B115 as antigens. Thus, selected mAb recognized cell surface-exposed epitopes and probably the most exposed core oligosaccharide region of R-LPS. Specificity was confirmed by immunoblotting using R-LPS and S-LPS preparations. The banding patterns of ~he high molecular mass S-LPS portion detected by anti-R-LPS were similar to those detected by S-LPS-specific mAb described elsewhere (Garin-Bastuji et al., 1990). In ELISA, these mAb bound weakly or not at all to S-LPS. However, in immunoblotting of either proteinase-K-treated or phenol-extracted S-LPS preparations, they reacted with both the low molecular mass R-LPS molecules and with the higher molecular mass S-LPS molecules containing O chains of variable length. This indicated that in S-LPS molecules, when separated in SDSoPAGE, epitopes of the core lipid A region are accessible to the anti-R-LPS mAb. Absence of reactivity or weak reactivity to S-LPS have been reported for anti-core LPS mAb derived from mice immunized with the rough E. coil J5 strain (Bogard et al., 1987). Others (Appelmelk et al., 1988) have also produced mAb against the LPS core ofE. coli strain J5 and also reacting with other Gram-negative bacilli. They concluded that OPS does not necessarily render the core inaccessible to antibody because the mAb produced still reacted with intact smooth E. coil cells. In contrast, Gigliotti and Shenep (1985) reported, using the same E. coli strain, the production of mAb against the core LPS and these mAb did not bind to whole cells of smooth E. coil strains. However, in both studies no immunoblot analysis of LPS preparations was perf o r m e d . As m e a s u r e d by E L I S A , our anti-R-LPS mAb bound less to S than to R intact B. abortus. This was also observed on R and S B. melitensis cell~ (data not shown). These results were confirmed by immunoelectron microscopy. Therefore, different binding patterns of anti-R-LPS mAb on S bacteria probably results from reactivity with R-form LPS non-substituted with O chains and present in various amounts on bacterial surfaces.
MONOCLONAL
A N T I B O D I E S TO B R U C E L L A R - L P S
481
Cm
AO
i ¸
~"
-
C
-
.;t
i¸
:i< !! i Fig. 4. lmmunogold labelling of 8. abortus 544 smooth (A, C) and rough (B, D) cells with antiOPS mAb 12G12 (lgGl) (A, B) and anti-R-LPS mAb A68/03F03/D05 (lgG2b) (C, D); x 30,000.
Despite the accessibility of R-LPS epitopes to mAb on the S B. a b o r t u s 544 strain used to infect CD-I mice, 3 out of the 4 anti-R-LPS mAb did not confer significant protection. Only I antiR-LPS mAb, of IgG3 isotype, reduced splenic
infection significantly, but protection was lower than with the anti-OPS mAb used in this study. This can be connected with the lower binding of the anti-R-LPS mAb observed on the S B. a b o r t u s 544 as compared with that of the
482
A. C L O E C K A E R T E T AL.
Table !. Protective activity of anti-R-LPS mAb to B. abortus 544 (S).
mAb ~a~ A68/O3F03/D05 A68/IOAO6/Bl I A68/25B03/D08 A68/29C05/B06 04F9 Control (saline)
Specificity
Isotype
R-LPS R-LPS R-LPS R-LPS S-LPS
lgG2b lgM lgGI lgG3 igG2a
Mean (SD) Ioglo Brucella CFU at days post-challenge the. 7 21 5.86 (0.20) 5.75 (0.50) 5.70 (0.52) 5.59 (0.3 ! ) ~'~ 5.13 (0.18) ~''~ 6.07 (0.21)
4.75 (0.20) 4.80 (0.50) 4.81 (0.34) 4.13 (0.47) ~'~ 3.90 (0.41) ~'*~ 4.60 (0.29)
,Jr Injection of mAb was performed subcutaneously I day prior to challenge. t~, CD-I mice were challenged with 200,000 CFU of strain 544 (S), killed after 7 and 21 days, and Brucella from spleen were counted. (") p < 0.05. {**) p < 0.01. in comparison with negative cor,!rol group according to variance analysis.
OPS-specific mAb (except for the IgM mAb). This could be due not only to a lower number of R-LPS molecules on the surface of smooth Brucella cells but also to steric hindrance caused by the long LPS O chains, as also described for accessibility of outer membrane proteins (OMP) to mAb (Cloeckaert et al., 1990). It was also previously shown that anti-OMP m?,b conferred no protection or much lower protection than anti-S-LPS mAb (Jacques et al., 1992; Montaraz et aL, 1986; Cloeckaert et al., 1991). In vivo, antibodies against the most exposed surface antigens, such as the O chains of S-LPS, are probably the most effective in fully achieving opsonization of Brucella organisms. The isotype of mAb probably also plays an important role in protection (Oishi et al., 1992). Non-agglutinating antibodies which bind to rough Brucella cells have been found in the sera of immunized cattle (Santisteban et al., 1986). Interestingly, they lacked protective capacities (viz opsonisation, complement fixation, antibody-dependent cytotoxicity) (Parma et al., 1987). As they were obtained by immunoadsorption with rough Brucella cells from sera previously adsorbed with smooth cells, they were most probably directed to R-LPS and also, as claimed, to outer membrane proteins. Lack of protection by anti-R-LPS mAb has been described for other Gram-negative infec-
tions (Appelmelk et al., 1988; Miner et al., 1986), i.e., mAb to the LPS-core region were unable to protect mice from lethal infection or endotoxaemia. Comparison is however difficult since this mouse model was based on lethality. Other authors (Dunn et al., 1986; Johns et aL, 1983 ; Teng et ai., 1985 ; Terashima et aL, 1991) described antibodies reacting with the LPS core, thus preventing death from Gram-negative sepsis. In most cases, antibodies were directed against the lipid A, the toxic moiety of LPS molecules. Work is in progress to define which epitopes of the LPS-core region, i.e. lipid A or core oligosaccharide, our mAb botmd and to explore their role in infections by natura!ly rough Brucella species.
Acknowledgements
We thank J.M. Verger and M. Grayon for providing Brucella strains. A. Cloeckaert was supported by the lnstitut National pour l'Encouragementde la Recherche Scientifiquedans I'lndustrie et l'Agriculture. R.A. Bowden is indebted to AUPELF (France), Fundaci6n Antorchas, Facuitad de Ciencias Veterinarias-U.N.C. and SECYT(Argentina) for support. Part of this paper corresponds to the thesis work of the third author.
MONOCLONAL
A N T I B O D I E S TO B R U C E L L A R - L P S
Anlicorps monoclonaux anti-LPS-R de Brucella: caract~risation el ~valuation de leur effet protecteur contre B. abortus
Nous avons caract6ris6 4 anticorps monoclonaux (AcM) sp6cifiques du lipopolyoside rugueux (LPS-R) de Brucella. Les AcM ont 6t~ s~lectionn~s par une 6preuve immunoenzymatique (ELISA) sur des cellules enti~res de B. abortus 45/20, souche rugueuse, et sur du LPS-R de B. melitensis B115. La sp~cificit6 a ~t6 confirm6e par analyse des immunoempreintes, utilisant le LPS-R et le LPS lisse (LPS-S). Les AcM anti-LPS-R ont r~v~l~ les bandes de basse masse mol6culaire, en dessous de 20,1 kDa, dans les pr6parations de LPS-R et LPS-S, ainsi que les profils typiques A et M dans les mol6cules de LPS~S de haute masse mol6culaire (entre 21,5 et 66 kDa) dans les ~chantillons de LPS-S. Un AcM sp~cifique du polyoside O a r6v~16 uniquement les bandes de haute masse mol~culaire dans les preparations de LPS-S. En E L I S A , les AcM anti-LPS-R se sont fix~ mieux sur les variantes rugueuses de B. abortus 544 que sur les lisses, et cela a ~t~ confirm6 par immunomarquage /l l ' o r coiloi'dal en microscopie ~lectronique. L'activit6 protectrice des AcM anti-LPS-R de differents isotypes a ~t~ ~tudi6e chez la souris et compar~e/~ celle d ' u n AcM sp6cifique du LPS-S. Seul I ' A c M d ' i s o t y p e lgG3 r6duit significativement l'infection spl~nique, sans pourtant atteindre l'activit6 pr~sent i e par i ' A c M sp6cifique du LPS-S (utilis6 comme t~moin positif). Mots-cl~s : LPS, Brucella abortus, Brucella melitensis, igG, Anticorps m o n o c l o n a l ; Souris, Protection, Epitopes.
References
Alton, G.G., Jones, L.M., Angus, R.D. & Verger, J.M. (1988), Techniques for the brucellosis laboratory. lnstitut National de la Recherche Agronomique, Paris. Appelmelk, B.J., Verweij-Van Vught, A.M.J., Maaskant, J.J., Schouten, W.F., De Jonge, A.J.R., Thijs, L.G. & Maclaren, D.M. (1988), Production and characterization of mouse monoclonal antibodies reacting with the lipopolysaccharide core region of Gram-negative bacilli. J. Med. MicrobioL, 26, 107-114. Bogard, W.C., Dunn, D.L., Abernethy, K., Kilgarriff, C. & Kung, P.C. (1987), Isolation and characterization of murine monoclonal antibodies specific for Gramnegative bacterial lipopolysaccharide: association of cross-genus reactivity with lipid A specificity. Infect. hnmun., 55, 899-908. Bundle, D.R., Cherwonogrodzky, J.W., Gidney, M.A.J., Meikle, P.J., Perry, M.B. & Peters, T. (1989), Deft-
483
nition of Brucella A and M epitopes by monoclonal typing reagents and synthetic oligosaccharides. Infect. immun., 57, 2829-2836. CIoeckaert, A., de Wergifosse, P., Dubray, G. & Limet, J.N. (1990), Identification of seven surface-exposed Brucella outer membrane proteins by use of monoclonal antibodies: immunogold labeling for electron microscopy and enzyme-linked immunosorbent assay. Infect. immun., 58, 3980-3987. Cloeckaert, A., Jacques, l., Bosseray, N., Limet, J.N., Bowden, R., Dubray, G. & Plommet, M. (1991), Protection conferred on mice by monoclonai antibodies directed against outer-membrane-protein antigens of Brucella. J. Med. Microbiol., 34, 175-180. Cloeckaert, A., Jacques, !., de Wergifosse, P., Dubray, G. & Limet, J.N. (1992), Protection against Brucella melitensis or Brucella abortus in mice with immunoglobulin G (lgG), lgA and lgM monoclonal antibodies specific for a common epitope shared by the Brucella A and M smooth lipopolysaccharide types. Infect. Immun., 60, 312-315. Dunn, D.L., Ewald, D.C., Chandan, N. & Cerra, F.B. (1986), Immunotherapy of Gram-negative bacterial sepsis. Arch. Surg., 121, 58-62. Galanos, C., Luderitz, O. & Westphal, O. (1969), A new method for the extraction of R-lipopolysaccharides. Eur. J. Biochem., 9, 245-249. Garin-Bastuji, B., Bowden, R.A., Dubray, G. & Limet, J.N. (1990), Sodium dodecyl sulfate polyacrylamide gel electrophoresis and immunoblotting analysis of smooth lipopolysaccharide heterogeneity among Brucella biovars related to A and M specificities. J. Clin. Microbiol., 28, 2169-2174. Gigliotti, F. & Shencp, J.L. (1985), Failure of monocional antibodies to core glycolipid to bind intact smooth strains of Escherichia coil J. Infect. Dis., 151, 1005-1011. Jacques, l., Olivier-Bernardin, V. & Dubray, G. (1991), Induction of antibody and protective responses in mice by Brucella O-polysaccharide-BSA conjugate. Vaccine, 9, 896-900. Jacques, l., Cloeckaert, A., Limet, J.N. & Dubray, G. 0992), Protection conferred on mice by combinations of monoclonal antibodies directed against outermembrane proteins or smooth lipopolysaccharide of Brucella. J. Med. MicrobioL, 37, 100-103. Johns, M., SkehiU, A. & McCabe, W.R. (1983), Immunization with rough mutants of Salmonella minnesota. - - IV. Protection by antisera to O and rough antigens against endotoxin. J. Infect. Dis., 147, 57-67. Leong, D., Diaz, R., Milner, K., Rudbach, J. & Wilson, J.B. (1970), Some structural and biological properties of Brucella endotoxin. Infect. hnmun., I, 174-182. Limet, J.N., Bosseray, N., Garin-Bastuji, B., Dubray, G. & Plommet, M. (1989), Humoral immunity in mice mediated by monoclonal antibodies against the A and M antigens of Brucella. J. Med. MicrobioL, 30, 37-43. Limet, J.N., Plommet, A.M., Dubray, G. & Plommet, M. (1987), Immunity conferred to mice by anti-LPS monoclonal antibodies in murine brucellosis. Ann. Inst. Pasteur/lmmunol., 138, 417-424. Lison, L. (1968), Statistique appliqu~e ~ la biologie exp6rimentale. Collection Science et Techniques d'Aujourd'hui. Gauthier-Villars, Paris.
484
A. CLOIzCKAERT
Miner, K.M., Manyak, C.L.. Williams, E., Jacksoll, J., Jewell, M., Gammon, M.T., Ehrenfreund, C., Hayes, E., Callahan, L.T., III, Zweerink, H. & Sigal, N.H. (1986), Characterization of routine monoclonal antibodies to Escherichia coli J5. Infect. hnmun., 52, 56-62. Montaraz, J.A., Winter, A.J., Hunter, D.M., Sowa, B.A., Wu, A.M. & Adams, L.G. (1986), Protection against Brucella abortus in mice with O-polysaccharidespecific monoclonal antibodies. Infect. lmmun., 5 I, 961-963. Moreno, E., Pitt, M.W., Jones, L.M., Schurig, G.G. & Berman, D.T. (1979), Purification and characterization of smooth and rough lipopolysaccharides from Brucella abortus. J. Bacteriol., 138, 361-369. Oishi, K., Koles, N.L., Guelde, G. & Pollack, M. (1992), Antibacterial and protective properties of monoclonal antibodies reactive with Escherichia coli OI II:B4 lipopolysaccharide: relation to antibody isotype and complement-fixingactivity. J. Infect. Dis., 165, 34-45. Palmer, D.A. & Douglas, J.T. (1989), Analysis of Brucella lipopolysaccharide with specific and cross-reacting monoclonal antibodies. J. Clin. Microbiol., 27, 2331-2337. Parma, A.E., Bowden, R.A., Santisteban, C.G., Cerone, S. & Fernfindez, A. (1987), Effect of bovine nonagglutinating antibodies on the blood clearance of 1-131 Brucella abortus 45/20. Vet. MicrobioL, 15, 121-128. Phillips, M., Deyoe, B.L. & Canning, P.C. (1989), Protection of mice against Brucella abortus infection by
ET AL.
inoculation with monoclonal antibodies recognizing Brucella O-antigen. Am. J. Vet. Res., 50, 2158-2161. Riezu-Boj, J.l., Moriy6n, I., Blasco, J.M., Matin, C.M. & Diaz, R. (1986), Comparison of lipopolysaccharide and outer membrane protein-lipopolysaccharide extracts in an enzyme-linked immunosorbent assay for the diagnosis of Brucella ovis infection. J. Clin. Microbiol., 23, 938-942. Santisteban, C.G., Parma, A.E., Cerone, S.I. & Bowden, R.A. (1986), Detection of antigenic fractions from Brttcella abortus s45/20 which bind to non-agglutinating antibodies using electroblotting and enzymelinked antibody probes. Vet. Microbiol., 12, 353-358. Teng, N.N.H., Kaplan, H.S., Hebert, J.M., Moore, C., Douglas, H., Wunderlich, A. & Braude, A.I. (1985), Protection against Gram-negative bacteremia and endotoxemia with human monoclonal IgM antibodies. Proc. nat. Acad. Sci. (Wash.), 82, 1790-1794. Terashima, M., Uezumi, I., Tomio, T., Kato, M., lrie, K., Okuda, T., Yokota, S.I. & Noguehi, H.A. (1991), Protective human monoclonal antibody directed to the outer core region of Pseudomonas aeruginosa lipopolysaccharide. Infect. Immun., 59, I-6. Vizcaino, N., Chordi, A. & Fernfindez-Lago, L. (1991), Characterization of smooth Brucella lipopolysaccharides and polysaecharides by monoclonal antibodies. Res. Microbiol., 142, 971-978. Zygmunt, M.S., Dubray, G., Bundle, D.R. & Perry, M.B. (1988), Purified native haptens of Brucella abortus BI9 and Bruce#a melitensis 16M reveal the lipopolysaccharide origin of the antigens. Ann. Inst. Pasteur/Microbiol., 139, 421-433.