coupling to immunogenic carriers for an optimal response. This has been observed to be the .... Popham, A. M., and D. W. Dresser. 1980. Rheumatoid factors in.
Vol. 53, No. 2
INFECTION AND IMMUNITY, Aug. 1986, p. 317-323
0019-9567/86/080317-07$02.00/0 Copyright © 1986, American Society for Microbiology
Detection and Specificity of Antibodies Secreted by Spleen Cells in Mice Immunized with Streptococcus mutans MICHAEL W. RUSSELL,* CECIL CZERKINSKY,t AND ZINA MOLDOVEANU Department of Microbiology and Institute of Dental Research, University of Alabama at Birmingham, Birmingham, Alabama 35294 Received 13 January 1986/Accepted 29 April 1986
Immune responses of mice to Streptococcus mutans serotype c were analyzed by means of the enzyme-linked immunospot assay to determine the predominant specificities of the antibodies developed. In general, the numbers of splenic antibody-secreting cells correlated with serum antibody levels. A low dose (108 CFU) of killed whole cells injected twice intraperitoneally induced antibodies mainly against sur-face protein antigen I/IT. A higher dose (109 CFU) given two to six times also resulted in a predominance of antigen I/lI antibody-secreting cells and, in addition, antibody responses to surface protein antigen III and lipoteichoic acid occurred. Cells producing antibodies to serotype c polysaccharide were elicited only on repeated immunization. These results agreed with the development of antibodies in rabbits repeatedly immunized intravenously with killed whole ceils of S. mutans, S. rattus, and S. sobrinus, which induced specific antibodies in accordance with the surface antigens that they express. Mice immunized twice with the same dose of purified antigens I/It and III developed greater numbers of antigen I/IT spienic antibody-forming cells than antigen III splenic antibody-forming cells and higher serum antibody levels to antigen I/TI than to antigen III. Furthermore, a single injection of antigen I/TI but not of antigen III was sufficient to induce a strong specific-antibody response. Some evidence was also obtained for weak polyclonal stimulation of spleen cells by S. mutans cells and by antigen I/I, a result which could be relevant to the induction by S. mutans of antibodies reactive with mammalian tissues. It was concluded that for the antigens examined, S. mutans elicited the strongest antibody response against antigen I/TI, which was also highly immunogenic in purified form.
antigens in the spleens of mice immunized with S. mutans whole cells. The levels obtained were compared with serum antibody levels in the same animals and also with levels of antibodies in rabbits.
Streptococcus mutans presents an array of antigens located in or on the cell wall, including polysaccharides, glycerol teichoic acid, and proteins. Serotype specificity, which has been the most convenient basis for the subclassification of S. mutans, is based on differences in the carbohydrate moieties of the cell wall (16). Glycerol teichoic acid is widely distributed in gram-positive bacteria and constitutes a heterophile antigen (4, 36). Cell-wall-associated proteins have been described by several groups (9, 11, 22, 23, 26-28, 30), and of these, antigens I/II and III appear to be prominently displayed on the cell surface (20). Although originally described in serotype c (26), antigen I/IT is also found in serotypes e and f, in serotypes d and g (S. sobrinus), and in serotype a (S. cricetus) (23), while antigen III occurs in serotypes c, e, and f and in serotype b (S. rattus) (22). Although the biological functions of these antigens are unknown, antigen I/TI (but not antigen III) has been implicated in the induction of immunity to dental caries and has been successfully used as a vaccine to protect monkeys against dental caries (15). No correlations between protection against dental caries and serum antibodies to antigen III, serotype polysaccharide, or lipoteichoic acid (LTA) have been observed in monkeys immunized with whole cells (24). Others have reported that immunization with antigen A, a protein which appears to be identical to antigen III, does confer protection (29). To obtain further information on the immunogenic potential of these antigens, we used the recently developed enzyme-linked immunospot (ELISPOT) assay (5) to enumerate cells secreting antibodies specific for various purified
MATERIALS AND METHODS Bacteria for immunization. S. mutans Guy (serotype c) was grown in 500 ml of semidefined medium (2) containing 3 g of Casamino Acids (Difco Laboratories, Detroit, Mich.) per liter in place of acidicase peptone and 10 g of glucose per liter. After overnight incubation at 37°C, the culture was treated overnight with 0.5 ml of 40% formaldehyde, and the cells were collected by centrifugation and washed three times in filtered saline. The density of the suspension was adjusted to 2 x 109 CFU/ml by turbidimetry at 660 nm. S. mutans antigens. Protein antigens I/IT and III were
prepared essentially as described previously (22, 23) from the culture supernatant of S. mutans IB-162 (supplied by K. W. Knox) grown for 16 to 18 h in semidefined medium (pH 6.0) (2) containing 1/10 the amount of phosphates and 1% fructose in a 28-liter fermentor (New I3runswick Scientific Co., Inc., Edison, N.J.). Serotype c polysaccharide (cPS) was extracted from the cells with nitrous acid (31). Washed cells (approximately 70 g [wet weight]) were suspended in 400 ml of 1 M NaNO2, and 100 ml of acetic acid was added. Foaming was controlled by the addition of 3 drops of Antifoam B (Sigma Chemical Co., St. Louis, Mo.). After being stirred for h at room temperature, the extract was collected by centrifugation, and polysaccharide was precipitated with 3 volumes of ethanol. The neutral polysaccharide fraction was obtained by chromatography on DEAE-cellulose in 0.01 M Tris hydrochloride (pH 7.4), dialyzed exhaustively against water, and lyophilized. Gas-chromatographic analysis of the product after meth-
* Corresponding author. t Present address: Department of Medical Microbiology, University of Goteborg, Guldhedsgatan 10, Goteborg, Sweden.
317
318
RUSSELL ET AL.
anolysis and trifluoracetate derivatization revealed a composition of rhamnose and glucose in a 2:1 ratio. The cPS preparation was esterified with palmitoyl chloride as described previously (10). LTA was extracted from cells by the hot phenol method of Wicken et al. (35), purified by treatment with RNase, DNase, and pronase, and chromatographed on Ultrogel AcA22 (LKB Instruments, Inc., Gaithersburg, Md.). The high-molecular-weight component identified by immunodiffusion analysis against an anti-poly(glycerophosphate) serum (supplied by B. Rosan) was used. Immunization of mice. Barrier-bred BALB/c mice of either sex and 2 to 5 months old were immunized with S. mutans cells by intraperitoneal injection of 108 or 109 CFU/ml in 0.5 ml of sterile phosphate-buffered saline (PBS). Secondary immunizations were given similarly after at least 7 days and subsequently at intervals of 3 or 4 days. Purified S. mutans antigen (50 jig) or keyhole limpet hemocyanin (KLH; Sigma) was injected intraperitoneally in Freund complete adjuvant (FCA). Secondary injections were given 7 days later with the same antigens dissolved in sterile PBS. For control purposes, animals received FCA containing PBS initially and PBS alone subsequently. Animals were bled and spleens were removed 4 or 5 days after the last injection. Immunization of rabbits. New Zealand White female rabbits weighing 2 to 2.5 kg were injected intravenously with suspensions of formolized S. mutans Guy (serotype c), BHT (serotype b, S. rattus), or 6715DP (serotype g, S. sobrinus) cells. Cell suspensions were adjusted to 5 x 109 CFU/ml by turbidimetry, and injections of 0.1 ml were given three times on alternate days in week 1, followed by 0.2-ml and 0.3-ml injections on three alternate days in weeks 2 and 3, respectively. Rabbits were bled at the end of week 4, and the cycle of injections was repeated twice more but with 0.5 ml of bacterial suspension. Enumeration of antibody-secreting cells. Enumeration was performed with the ELISPOT assay (5). Petri dishes (Falcon 1006; Becton Dickinson Labware, Oxnard, Calif.) were coated by overnight incubation with previously determined optimal concentrations of the following in PBS: antigen I/MI, 10 ,ug/ml; antigen III, 25 jig/ml; LTA, 50 ,ug/ml; KLH, 50 jig/ml; and cPS ester, 50 jig/ml on dishes precoated with methylated bovine serum albumin (MeBSA) at 200 ,ug/ml. Control dishes were coated with 10% fetal calf serum in PBS or MeBSA as appropriate. All dishes were blocked by treatment with 5% fetal calf serum in PBS for 1 to 2 h. Single-cell suspensions were prepared from mouse spleens, and 0.5-ml samples containing 1 x 105 to 2 x 106 lymphoid cells were incubated in the coated dishes for 3 h at 37°C. After thorough washing with PBS containing 0.05% Tween 20, antibodies secreted by the cells and bound locally to antigen coats were detected by developing the plates with biotinylated goat anti-mouse immunoglobulin (Cooper Biomedical, Inc., Malvern, Pa.) diluted 1:750 overnight at room temperature, followed by avidin-peroxidase (Sigma) at 1 ,g/ml for 2 h at room temperature. Spots indicating the positions where antibody-secreting cells had settled were revealed by the addition of p-phenylenediamine (0.5 mg/ml)H202 (0.01%) substrate in molten 1% agar in PBS. The spots were examined under a low-power stereomicroscope and counted. Whenever possible, counts were made in plates having between 10 and 100 spots. Spots in experiment 1 (see below) were developed by an alternative procedure with peroxidase-conjugated rabbit anti-mouse immunoglobulin (Dako Corp., Santa Barbara, Calif.) diluted 1:200.
INFECT. IMMUN.
Serum antibodies. Mouse serum antibodies were assayed by an enzyme-linked immunosorbent assay (ELISA) on polyvinyl microtiter plates (Dynatech Laboratories, Inc., Alexandria, Va.). Wells were coated with the following antigens: antigen I/II, 5 jig/ml; antigen III, 10 pLg/ml; LTA, 5 jig/ml; cPS ester, 25 pLg/ml in wells precoated with MeBSA at 200 jig/ml; and KLH, 10,ug/ml. Control wells were coated with bovine serum albumin (BSA) or MeBSA as appropriate. All wells were blocked with 1% BSA in PBS, which was also used as the diluent for the samples and reagents. Serum samples diluted 1:10,000 were incubated in quadruplicate wells overnight at room temperature. Bound antibodies were developed with biotinylated goat anti-mouse immunoglobulin (Cooper) diluted 1:2,500, followed by avidin-peroxidase (Sigma) at 0.5 jig/ml. The substrate was 60 mg of 2,2'-azinobis(3-ethylbenzthiazolinesulfonic acid) (Sigma) in 100 ml of 0.1 M citrate buffer (pH 4.2) plus 6,il of 30% H202. The A414 was measured after 30 min with a Titertek Multiskan photometer (Flow Laboratories, Inc., McLean, Va.). Rabbit serum antibodies to antigens I/IT and III and LTA were assayed by a radioimmunoassay on Immulon IT Removawell strips (Dynatech) coated with antigens as described above for the ELISA. Serum samples diluted 1:1,000 were incubated in duplicate wells overnight at room temperature. Bound antibodies were measured by adding 50,000 cpm of 125I-labeled (17) affinity-purified goat anti-rabbit immunoglobulin (Southern Biotechnology Associates, Inc., Birmingham, Ala.) and counting the bound radioactivity in a Gamma 4000 spectrometer (Beckman Instruments, Inc., Irvine, Calif.). Calibrations were made by incubating serial threefold dilutions of normal rabbit seurm, starting from a dilution of 1:100,000, on plates coated with affinity-purified anti-rabbit immunoglobulin and developing the plates as described above. Calibration curves were constructed with a computer by using the four-parameter logistic model (6). Antibody levels were expressed by comparison with these curves in terms of arbitrary units, taking the 10-5 dilution of the standard as 1,000 units. Because the radioimmunoassay failed to give reproducible results for the estimation of antibodies to cPS, an ELISA similar to that for mouse serum antibodies was used. Plates were coated with cPS ester in MeBSA as described above (controls consisted of MeBSA-coated plates), and serum samples were diluted 1:1,000. Biotinylated anti-rabbit immunoglobulin (Cooper) was used at a dilution of 1:1,000, followed by avidin-peroxidase and 2,2'-azino-bis(3-ethylbenzthiazolinesulfonic acid)-H202 substrate as described above. Calibration curves were constructed in a manner analogous to that described for the radioimmunoassay. RESULTS Antibody-secreting cell responses in mice immunized with S. mutans cells. Mice immunized with two doses of 108 cells or with two, three, four, or six doses of 109 cells were assayed for numbers of splenic antibody-secreting cells detected as spot-forming cells (SFC) in plates coated with antigen I/TI, antigen III, LTA, cPS, or KLH and in control plates. The last always showed zero or negligible SFC (-1 SFC per 106 cells). In mice immunized with 108 cells, substantial numbers of SFC were detected only against antigen I/IT (Fig. 1, experiments 1 and 2). In experiment 1, some mice immunized with S. mutans also displayed a significant number of cells secreting antibodies to KLH, but this result was not seen in experiment 2, in which responses to antigen I/IT were also lower. As similar observations were made in mice
VOL. 53, 1986
ANTIBODY RESPONSE TO S. MUTANS
319
1000 Ir
I
500[ 2001100
-
i
501
1-
N.
0
I
10F 5
2 1L I Experiment
I
Sb 5a Con I O',x6 Con 10 ,x3 10@,x4 10O',x2 6* 6* 2* 1 3 2 5# 3 N 2 2* 5 FIG. 1. Splenic SFC responses of mice immunized with 108 or 109 S. mutans cells two to six times. Also shown are results for control mice immunized with KLH or unimmunized mice (Con). Bars represent the mean + standard deviation of the splenic SFC response determined
Immunization
la
10,x2
lb KLH
2a 100,x2
2b Con
3a
3b Con
against the following: *, antigen I/II; 02, antigen III; 1, LTA; 0, cPS; and a, KLH. N, Number of animals used; *, spleens pooled in pairs; #, pool of three spleens plus two individuals. immunized with larger numbers of S. mutans or purified antigens (see below), this result was believed to indicate polyclonal stimulation of B cells. For comparison, mice immunized with KLH developed very large numbers of SFC to KLH and none to S. mutans antigens (Fig. 1). Spleen cells from mice immunized with 109 cells two to six times displayed the highest SFC responses (-50 to 350 SFC per 106 cells) against antigen I/IT (Fig. 1). In addition, high responses were also seen against LTA, whereas three or more doses were required to elicit modest responses (-4 to 50 SFC per 106 cells) to antigen III (Fig. 1). Small numbers of cells secreting antibodies to cPS (-10 SFC per 106 cells) were detected only after repeated immunization with 109 S. mutans.
Small numbers of SFC to KLH (-2 to 20 SFC per 106 cells) were observed in most animals immunized with 109 S. mutans but also in some unimmunized control animals (Fig. 1, experiment 4). None of the unimmunized animals displayed cells secreting antibodies to antigens of S. mutans.
Antibody-secreting cell responses in mice immunized with purified protein antigens. Mice immunized with S. mutans cells displayed the highest SFC responses against antigen I/IT, compared with other antigens, particularly antigen III. These results may indicate that antigen I/IT is expressed at a relatively higher density on the cell surface or that it is inherently more immunogenic. Experiments were therefore performed by immunizing mice with purified antigens I/IT and III in similar doses and under the same conditions. Mice were primed with 50-,ug doses of these antigens in FCA and
boosted with a similar dose of the antigens in PBS. Under these conditions, antigen I/II again elicited very strong responses (-60 to 500 SFC per 106 cells) (Fig. 2, experiments 6 and 7), whereas antigen III induced less marked responses (16 to 90 SFC per 106 cells). The specificity of the response and of the assay was shown in these experiments, insofar as immunization with antigen I/TI induced little or no response to antigen III and vice versa. However, there appeared to be a weak polyclonal response, which was also shown in the control animals (sham immunized with FCA), as indicated by responses to KLH (Fig. 2). Because of the strong secondary response, especially to antigen I/IT, antibody-secreting cells were also examined in response to primary immunization with antigens I/TI and III. Mice given a single dose of antigen I/IT in FCA or in Freund incomplete adjuvant (FIA) showed very large numbers of spleen cells secreting specific antibodies (>200 SFC per 106 cells) by day 7 (Fig. 2, experiments 8 and 9), whereas mice primed with antigen III showed much weaker splenic SFC responses to antigen III (-8 SFC per 106 cells). Putative polyclonal responses (on KLH-coated plates) were seen in some animals, including those immunized with antigen I/II in FCA or FIA or with FCA alone (Fig. 2, experiment 9). Serum antibody responses in mice. Serum samples from mice immunized with S. mutans cells or purified antigens were collected at the same time that the spleen cells were assayed by the ELISPOT assay and were examined for levels of S. mutans-specific antibodies by a conventional ELISA. Mice immunized with 108 S. mutans cells exhibited high
INFECT. IMMUN.
RUSSELL ET AL.
320
1000
500
200
100 S
50
0
201.
h.
C.) LL
co)
101_
2
Experiment Immunization
N=
Ksa i/II,x2 2
6b IlI,x2 2
6c con
7a
7b
7c
Ba
8b
9b
con
IIiI,x1
II/1,x1
9c con
9d con
2*
l/ll,X1 2*
III,XI
2
IIi,X2 2*
8c con
98
I/lI,X2
2-
2
4*
2*
2-
2*
2*
FIG. 2. Splenic SFC responses of mice immunized with antigens I/II and III twice or only once. All primary immunizations were given in FCA, except for experiment 9b, in which FIA was used. con, Corresponding control mice sham immunized with FCA (experiments 6c, 7c, 8c, and 9c) or unimmunized (experiment 9d). N, Number of animals; *, spleens pooled in pairs. Bars represent the mean and range of the splenic SFC response determined against the following: *, antigen I/II; 02, antigen III; and OL, KLH. serum antibody levels against antigen I/II but only weak to moderate levels against antigen III (Table 1). Immunization with 109 cells induced still higher levels of antibodies to antigen I/HI and also elicited increased levels of antibodies to antigen III, but only after three or four doses. Antibodies to LTA seemed variable, being highest in those animals given
two doses of 109 cells, but antibodies to cPS remained
insignificant. Modest levels of antibodies to KLH appeared in most immunized animals, but similar amounts were seen in some unimmunized control animals. However, most immunized animals showed weakly elevated antibodies to MeBSA in
TABLE 1. Serum antibody responses to S. mutans antigens in mice immunized with S. mutans cells Antibody level (mean + SD A414) to antigen:
No. of
Immu animals
10' cells, two doses 109 cells, two doses 109 cells, three to four doses KLH, two doses None an
=
4.
10
1/11
0.609 >2
6
III
LTA
cPS
KLH
BSA
MeBSA
0.682 0.105
0.063 0.118
0.157 0.007 + 0.011 0.162
0.040
0.043
0.013
0.012
0.005
0.088
0.020 0.551
0.317 0.004
0.007 0.244
0.085
0.033
0.012
0.021
0.009
0.007 0.051
0.033 0.105
0.014
0.050 + 0.017
0.010
0.007
0.006
5
1.609
±
0.400 0.411
0.426 0.033
4
0.152
±
0.121 0.115
0.016 0.030
±
0.042 0.012
6
0.013
±
0.012 0.018
0.005 0.015
±
0.007 0.002 + 0.003 0.102
>2 ±
0.021
0.054
±
0.013
0.012
0.002
0.012a
±
0.004
0.028a
0.025
VOL. 53, 1986
ANTIBODY RESPONSE TO S. MUTANS
321
TABLE 2. Serum antibody responses to S. mutans antigens in mice immunized with purified protein antigens No. of
Immunizatinanimals Antigen I/II, two doses Antigen III, two doses Antigen I/II, one dose Antigen I/II, one dose (FIA) Antigen III, one dose FCA control None an = 2. b NT, Not tested. Cn
=
Antibody level (mean
I/II
III
>2 0.156 ± 0.022 >2 >2 0.157 ± 0.016 0.030 ± 0.007 0.016 ± 0.001
4 4 6 2 2 6 2
0.202 0.626 0.134 0.100 0.188
± ± ± ± ± 0.047 ± 0.026 ±
0.092 0.187 0.075 0.012 0.084 0.017 0.004
t
SD A414) to antigen:
KLH
MeBSA
0.243 ± 0.135 0.227 ± 0.114 0.104 ± 0.046 0.091 ± 0.016 0.101 ± 0.025 0.091 ± 0.032 0.037 + 0.005
0.053 ± 0.034 0.049 t 0.010 0.034a ± 0.006 NTb 0.032 ± 0.015 0.029 ± 0.006 NT
BSA 0.024 0.017 0.023 0.043 0.009 0.025 0.021
± 0.002 ± 0.017 + 0.015 ± 0.012 ± 0.009 + 0.004 ± 0.003
4.
comparison with both antibodies in unimmunized control animals and antibodies measured in control wells coated with unconjugated BSA. These data may indicate a weak polyclonal activation of the immune system in these animals. Mice immunized with two doses of purified antigens I/II and III showed serum antibody responses specific for the homologous antigens, although responses to antigen III were weaker than those to antigen I/IT (Table 2). A single dose of antigen I/II in either FCA or FIA was also highly effective in eliciting specific antibodies, whereas antigen III was much less effective. Putative polyclonal antibody responses were also apparent, although weak, as measured in KLH- or MeBSA-coated wells. In general, the serum antibody levels agreed with the splenic SFC responses. Significant positive correlations were obtained by the Spearman rank procedure for all ELISA and ELISPOT assay results obtained for antigen I/IT (p = 0.889; P < 0.001) antigen III (p = 0.415; P = 0.02), and LTA (p = 0.585; P = 0.004). The data obtained for cPS and KLH were inadequately distributed to make correlations valid. Serum antibody responses in rabbits. In rabbits given repeated intravenous injections of S. mutans Guy or S. sobrinus 6715DP cells, serum antibodies to antigen I/II were predominant after 1, 2, and 3 months (Table 3). Antibodies to antigen I/TI were not elicited by S. rattus BHT cells, which lack this antigen (23). Rabbits immunized with strains Guy and BHT developed antibodies to antigen III generally at a lower level than antibodies to antigen I/II (Table 3), whereas strain 6715DP, which lacks antigen III (22), failed to induce
corresponding antibodies. Antibodies to LTA developed in all rabbits (Table 3), reaching very high levels in those immunized with strain BHT, which produces abundant amounts of glycerol teichoic acid (34). Only the serotype c strain induced antibodies to cPS, but at modest levels, even after repeated immunization (Table 3). A survey of 13 rabbit antisera raised against a variety of S. mutans serotype c preparations, including crude and purified cell walls and crude and purified soluble antigens, revealed that antibodies to cPS were difficult to induce, and then only at modest levels as compared with those seen in antisera to whole cells (data not shown). These observations are consistent with the failure to find precipitating antibodies to cPS in any of the sera examined and stand in marked contrast to the ease with which antibodies to protein antigens I/II and III or to LTA were induced. DISCUSSION Splenic antibody responses of mice to S. mutans were analyzed at the single-cell level with the ELISPOT assay (5), which enumerates cells secreting antibodies specific for any antigen that can be adsorbed onto plastic surfaces. This method has particular advantages over the more familiar hemolytic plaque assay, for which antigens must be attached to erythrocytes, and is not biased towards the detection of certain isotypes. It has been shown that the spots are due to the synthesis of specific antibodies by metabolically active cells, because pretreatment of cells with cycloheximide inhibits spot formation (4a). Antigenic specificity was also shown in the present experiments with mice immunized with
TABLE 3. Serum antibody responses to S. mutans antigens in rabbits immunized with S. mutans cells
Immunization
Sampling Samp
time (mo)
Antibody levela to antigen: ____ III LTA 1/11
cPS
S. mutans (Guy) (serotype c)
0 1 2 3
0, 0 860, 360 780, 380 NTb, 670
40, 0 250, 260 460, 270 NT, 320
30, 0 380, 160 730, 730 NT, 480
0, 0 0, 26 13, 22 17, 0
S. sobrinus (6715DP) (serotype g)
0 1 2 3
0, 20 610, 720 NT, 800 NT, 710
4, 40 0, 2 NT, 20 NT, 40
8, 0 410, 30 NT, 310 NT, 190
0, 0 0, 0 NT, 0 NT, 0
S. rattus (BHT)
0 1 2 3
70, 10 10, 0 2, 30 10, 20
0, 3 240, 30 330, 520 300, 250
90, 60 440, 700 990, 1150 600, 800
0, 0 0, 0 0, 0 0, 0
(serotype b)
Antibody units (see text) for individual animals. bNT, Not tested.
a
322
RUSSELL ET AL.
antigens I/II and III or KLH. Although the method can be used to detect isotype-specific responses by means of appropriate developing reagents, we assessed the total antibodysecreting spleen cell responses of mice to S. mutans cells and protein antigens irrespective of isotype. The results obtained in this way for the enumeration of cells secreting antibodies to antigens I/MI and III and LTA correlated with those obtained in the ELISA for the measurement of serum antibodies to these antigens. While there was appreciable variation in responses between individual mice and especially between batches of mice used on different occasions, the overall pattern of results was essentially the same. Mice immunized with whole cells of S. mutans serotype c showed the strongest response (SFC and serum antibodies) to antigen I/TI, irrespective of the dose or duration of immunization. The response to antigen III was lower and required a higher dose for induction. LTA induced variable responses, generally at moderate levels. Antibody responses to cPS were weak and could be detected only after repeated immunization. All of these antigens are located at or on the cell surface, as shown by immunoelectron microscopy (20), immunofluorescence (22, 23), and inferentially by agglutination with specific antisera (M. W. Russell, unpublished observations), but their density and precise configuration may be different in ways that could affect the immune response. Of the protein antigens, I/II does appear to be inherently more immunogenic than III when injected in purified form at the same dose. Whether this is because antigen I/IT is a much larger molecule (molecular weight, 185,000) than antigen III (molecular weight, 39,000) or because of structural differences is not known. Purified polysaccharide and teichoic acid antigens are known to be often poorly immunogenic and to require coupling to immunogenic carriers for an optimal response. This has been observed to be the case previously with cPS [R. Linzer and R. J. Genco, J. Dent. Res. 56(Special Issue A):131, abstr. no. 355, 1977] and LTA (1, 8). LTA is a ubiquitous antigen of gram-positive bacteria (36), so that the variability of the response observed against it could be due to previous exposure and presensitization of some of the mice to indigenous organisms, even though elevated serum antibodies were not evident in the control animals. cPS is an abundant material in the cell wall and is probably important as a structural component, and repeated immunization of mice did elicit antibody-secreting cells against it. However, cPS, like many polysaccharide antigens with repetitive epitopes, is believed to be thymus independent (32), so that it would tend to induce a primary antibody response confined to immunoglobulin M, with secondary response limited to the minor immunoglobulin G3 isotype. These would have been detected by our assays but would not be expected to result in high levels of antibodies. The results obtained from rabbits repeatedly immunized with whole cells of three different organisms broadly agreed with those obtained with mice. That is, antigen I/IT appeared to be more immunogenic in situ than the other antigens examined. The strains were chosen because they were known to contain both antigens I/IT and III (Guy, serotype c), antigen I/IT alone (6715DP, S. sobrinus), or antigen III alone (BHT, S. rattus) (22, 23). The serum antibody responses obtained reflected these differences. In addition, antibodies to LTA were induced, especially by immunization with strain BHT, which contains cell wall glycerol teichoic acid (34). Low levels of antibody to cPS were induced only by the serotype c strain. It is interesting to note that rhesus monkeys appeared to
INFECT. IMMUN.
respond in a similar manner to immunization with S. mutans cells, cell walls, or purified antigens. That is, serum antibodies were readily developed against protein antigens I/TI and III and also against LTA, while anti-cPS responses were weak (24). Furthermore, immunization with purified antigen III appeared to induce less marked specific-antibody responses than did that with purified antigen I/II (15). Whether this was responsible for the lack of observed protection against caries by antigen III (15), while others (29) found that a similar antigen A was protective, remains a matter of speculation. S. mutans and purified antigens displayed weak polyclonal-antibody-activating properties, as has been shown for other bacterial cell wall components, such as peptidoglycan, muramyl dipeptide, and lipopolysaccharide (7, 12, 21). An explanation for this will require more detailed knowledge of the structure of antigens I/IT and III and the immunological properties of their epitopes. However, the immunogenic properties of these antigens, including their apparent ability to stimulate polyclonal immune responses, may be relevant to the induction of antibodies reactive with mammalian tissue components when animals are hyperimmunized with S. mutans (33). Further studies are in progress to investigate this possibility. Antibodies to S. mutans have frequently been measured against whole cells in ELISA or other assays to assess the totality of the response. Comparison with different strains representing different serotypes has been used to show alleged serotype specificity. However, cells express a complex mosaic of different antigens, including serotyperestricted antigens, common antigens, and even widespread heterophile antigens like LTA. The major part of the antibody response to S. mutans appears to be directed against the common protein antigen I/IT, which is found in all serotypes except b (S. rattus) (23). However, the density of this antigen on the cell surface may vary between strains and with culture conditions (13, 18), so that these factors could influence antibody assays based on whole cells. The presence of antigen I/IT on most strains of S. mutans found in humans and its high immunogenicity may explain why serum and salivary antibodies against it are found frequently and at high levels in normal individuals (3, 25). Oral immunization of volunteers with capsules containing S. mutans induces the appearance in the peripheral blood of a population of cells having the capacity to secrete immunoglobulin A antibody predominantly to antigen I/TI, as shown by the ELISPOT assay (19). These cells are believed to be components of the common mucosal immune system on the way to populating secretory sites, including the salivary glands, where they will produce secretory immunoglobulin A antibodies (14). Since parenteral immunization of monkeys with antigen I/IT has been shown to induce protection against dental caries (15), it will be of interest to determine if oral immunization with this antigen can be of benefit for human caries control. Analysis of antibody responses to S. mutans and to its component antigens with the ELISPOT assay will be valuable for this purpose. ACKNOWLEDGMENTS We thank K. W. Knox and B. Rosan for providing S. mutans 1B-162 and antiserum to LTA, respectively, and D. Pritchard for carbohydrate analysis. We are grateful to Darrah K. Hammond and Yvonne T. Pfanenstiel for technical assistance and Maria E. Paulson for typing. C. Czerkinsky was a visitor in Jiri Mestecky's laboratory, and we thank him for his support and encouragement. This work was supported by Public Health Service grants
VOL.
53, 1986
DE-06746, DE-02670, and DE-06669 from the National Institutes of Health.
LITERATURE CITED 1. Beining, P. R., G. M. Flannery, B. Prescott, and P. J. Baker. 1980. Influence of carrier-specific, thymus-derived cells on the immunoglobulin M response to staphylococcal lipoteichoic acid. Infect. Immun. 29:132-139. 2. Bowden, G. H., J. M. Hardie, and E. D. Fillery. 1976. Antigens from Actinomyces species and their value in identification. J. Dent. Res. 55(Special Issue A):192-204. 3. Challacombe, S. J., L. A. Bergmeier, C. C. Czerkinsky, and A. A. Rees. 1984. Natural antibodies in man to Streptococcus mutans: specificity and quantification. Immunology 52:143-150. 4. Chorpenning, F. W., H. R. Cooper, and S. Rosen. 1975. Crossreactions of Streptococcus mutans due to cell wall teichoic acid. Infect. Immun. 12:586-591. 4a.Czerkinsky, C. 1986. Antibody-secreting cells, p. 23-44. In H. U. Bergmeyer, J. Bergmeyer, and M. Grassl (ed.), Methods of enzymatic analysis, 3rd ed., vol. X. VCH Weinheim. 5. Czerkinsky, C. C., L.-A. Nilsson, H. Nygren, 0. Ouchterlony, and A. Tarkowski. 1983. A solid-phase enzyme-linked immunospot (ELISPOT) assay for enumeration of specific antibody-secreting cells. J. Immunol. Methods 65:109-121. 6. Davis, S. E., P. J. Munson, M. L. Jaffe, and D. Rodbard. 1980. Radioimmunoassay data processing with a small programmable calculator. J. Immunoassay 1:15-25. 7. Dziarski, R. 1982. Preferential induction of autoantibody secretion in polyclonal activation by peptidoglycan and lipopolysaccharide. 1. In vitro studies. J. Immunol. 128:1018-1025. 8. Fiedel, B. A., and R. W. Jackson. 1976. Immunogenicity of a purified and carrier-complexed streptococcal lipoteichoic acid. Infect. Immun. 13:1585-1590. 9. Forester, H., N. Hunter, and K. W. Knox. 1983. Characterization of a high molecular weight extracellular protein of Streptococcus mutans. J. Gen. Microbiol. 129:2779-2788. 10. HAmmerling, U., and 0. Westphal. 1967. Synthesis and use of 0-stearoyl polysaccharides in passive haemagglutination and haemolysis. Eur. J. Biochem. 1:46-50. 11. Holt, R. G., Y. Abiko, S. Saito, M. Smorawinska, J. B. Hansen, and R. Curdss. 1982. Streptococcus mutans genes that code for extracellular proteins in Escherichia coli K-12. Infect. Immun. 38:147-156. 12. Izui, S., R. A. Eisenberg, and F. J. Dixon. 1979. IgM rheumatoid factors in mice injected with bacterial lipopolysaccharides. J. Immunol. 122:2096-2102. 13. Knox, K. W., and A. J. Wicken. 1985. Environmentally induced changes in the surfaces of oral streptococci and lactobacilli, p. 212-219. In S. E. Mergenhagen and B. Rosan (ed.), Molecular basis of oral microbial adhesion. American Society for Microbiology, Washington, D.C. 14. Kutteh, W. H., W. J. Koopman, M. E. Conley, M. L. Egan, and J. Mestecky. 1980. Production of predominantly polymeric IgA by human peripheral blood lymphocytes stimulated in vitro with mitogens. J. Exp. Med. 152:1424-1429. 15. Lehner, T., M. W. Russell, J. Caldwell, and R. Smith. 1981. Immunization with purified protein antigens from Streptococcus mutans against dental caries in rhesus monkeys. Infect. Immun. 34:407-415. 16. Linzer, R. 1976. Serotype polysaccharide antigens of Streptococcus mutans: composition and serological cross-reactions, p. 91-99. In W. H. Bowen, R. J. Genco, and T. C. O'Brien (ed.), Immunologic aspects of dental caries. Information Retrieval, Inc., Washington, D.C. 17. Marchalonis, J. J. 1969. An enzymic method for the trace iodination of immunoglobulins and other proteins. Biochem. J. 113:299-305.
ANTIBODY RESPONSE TO S. MUTANS
323
18. McBride, B. C., M. Song, B. Krasse, and J. Olsson. 1984. Biochemical and immunological differences between hydrophobic and hydrophilic strains of Streptococcus mutans. Infect. Immun. 44:68-75. 19. Mestecky, J., C. Czerkinsky, T. A. Brown, S. J. Prince, S. M. Michalek, M. W. Russell, S. Jackson, M. SchOller, and J. R. McGhee. 1986. Human immune responses to Streptococcus mutans, p. 297-306. In S. Hamada, S. M. Michalek, H. Kiyono, L. Menaker, and J. R. McGhee (ed.), Molecular microbiology and immunobiology of Streptococcus mutans. Elsevier Biomedical 2ress, Amsterdam. 20. Moro, I., and M. W. Russell. 1983. Ultrastructural localization of protein antigens I/II and III in Streptococcus mutans. Infect. Immun. 41:410-413. 21. Popham, A. M., and D. W. Dresser. 1980. Rheumatoid factors in mice: nonspecific activators of heterophile rheumatoid factor production. Immunology 41:579-585. 22. Russell, M. W. 1979. Purification and properties of a protein surface antigen of Streptococcus mutans. Microbios 25:7-18. 23. Russell, M. W., L. A. Bergmeier, E. D. Zanders, and T. Lehnier. 1980. Protein antigens of Streptococcus mutans: purification and properties of a double antigen and its protease-resistant component. Infect. Immun. 28:486-493. 24. Russell, M. W., S. J. Challacombe, and T. Lehner. 1980. Specificity of antibodies induced by Streptococcus mutans during immunization against dental caries. Immunology 40: 97-106. 25. Russell, M. W., D. Hammond, J. R6dl, J. J. Haanman, and J. Mestecky. 1984. Secretory IgAl and IgA2 responses to environmental antigens. Protides Biol. Fluids Proc. Colloq. 32:77-80. 26. Russell, M. W., and T. Lehner. 1978. Characterization of antigens extracted from cells and culture fluids of Streptococcus mutans serotype c. Arch. Oral Biol. 23:7-15. 27. Russell, M. W., E. D. Zanders, L. A. Bergmeier, and T. Lehner. 1980. Affinity purification and characterization of proteasesusceptible antigen I of Streptococcus mutans. Infect. Immun. 29:999-1006. 28. Russell, R. R. B. 1979. Wall-associated protein antigens of Streptococcus mutans. J. Gen. Microbiol. 114:109-115. 29. Russell, R. R. B., D. Beighton, and B. Cohen. 1982. Immunization of monkeys (Macaca fascicularis) with antigens purified from Streptococcus mutans. Br. Dent. J. 152:81-84. 30. Scholler, M., J.-P. Klein, P. Sommer, and R. Frank. 1983. Common antigens of streptococcal and nonstreptococcal oral bacteria: characterization of wall-associated protein and comparison with extracellular protein antigen. Infect. Immun. 40:1186-1191. 31. Swanson, J., K. C. Hsu, and E. C. Gottschlich. 1969. Electron microscopic studies on streptococci. I. M antigen. J. Exp. Med. 130:1063-1075. 32. Torii, M., J. R. McGhee, W. J. Koopman, S. Hamada, and S. M. Michalek. 1981. Lymphoid cell responses to bacterial cell wall components: polyclonal and immune responses of murine B cells to Streptococcus mutans carbohydrate antigens. J. Immunol. 127:2106-2112. 33. Van de Rijn, I., A. S. Bleiweis, and J. B. Zabriskie. 1976. Antigens in Streptococcus mutans cross-reactive with human heart muscle. J. Dent. Res. 55(Special Issue C):59-64. 34. Wicken, A. J., J. D. Evans, L. K. Campbell, and K. W. Knox. 1982. Teichoic acids from chemostat-grown cultures of Streptococcus mutans and Lactobacillus plantarum. Infect. Immun. 38:1-7. 35. Wicken, A. J., J. W. Gibbens, and K. W. Knox. 1973. Comparative studies on the isolation of membrane lipoteichoic acid from Lactobacillus fermenti. J. Bacteriol. 113:365-372. 36. Wicken, A. J., and K. W. Knox. 1975. Lipoteichoic acids: a new class of bacterial antigen. Science 187:1161-1167.
