Jun 16, 1975 - toxin suggests that antitoxin antibodies must react with the antigen within the gut lumen or at the lining surface of the small bowel to be effective.
INFECTION AND IMMUNITY, Dec. 1975, p. 1331-1340 Copyright C) 1975 American Society for Microbiology
Vol. 12, No. 6 Printed in USA.
Antitoxic Immunity in Experimental Cholera: Protection, and Serum and Local Antibody Responses in Rabbits After Enteral and Parenteral Immunization J. HOLMGREN,* A.-M. SVENNERHOLM, 0. OUCHTERLONY, A. ANDERSSON, G. WALLERSTROM, AND U. WESTERBERG-BERNDTSSON
Institute of Medical Microbiology, University of Goteborg, G6teborg, Sweden Received for publication 16 June 1975
The protective effect of enteral and parenteral immunization with cholera toxin antigen against experimental cholera in rabbits was studied by using the small-bowel loop technique. Subcutaneous injection of crude toxin as well as purified toxin or toxoids gave rise to significant protection against toxin challenge. The enhanced resistance to toxin was found to correspond to a many-fold higher magnitude of protection against challenge with live vibrios. In the primary response the protection increased during the first month. Booster immunization gave rise to a further increased immunity which, however, declined rapidly. Multiple oral or repeated intraintestinal antigen administrations also induced protective antitoxic immunity although of less magnitude than that obtained by parenteral immunization. Enteral and, to a lesser extent, parenteral immunization gave rise to increased antitoxic antibody titers and immunoglobulin levals in intestinal washings and mucosa scrapings. Immunoglobulin G (IgG) and IgG antitoxins predominated, but after enteral immunization total IgA and specific IgA antibodies occasionally reached levels similar to those for IgG. In serum, significantly increased antibody levels (IgG) were only recorded after parenteral immunization. Both the primary binding and the neutralizing antitoxin titers showed a statistically significant correlation with the degree of protection against toxin challenge; however, for the neutralizing antibodies this correlation was not without exceptions. No relation to protection was found for intestinal antibodies. The results of the present study indicate that enteral as well as parenteral immunization with toxin antigen can give rise to effective cholera immunity. After enteral immunization, the protection appears to be mediated by locally synthesized antibodies. After parenteral vaccination both serum-derived and locally produced antibodies seem to be effective.
Cholera is caused by the action on the smallintestinal mucosa of a protein exotoxin of Vibrio cholerae (7, 28). The toxin is antigenic, and the possibility of developing effective antitoxic cholera immunoprophylaxis in man, which is not achieved by presently used cholera vaccines, is a pressing subject for investigation. Recent studies have demonstrated that immunization with toxin or toxoid preparations can induce protection against experimental cholera in dogs (4), rabbits (1, 6, 17), and mice (12). The information about the magnitude and duration of the functional antitoxic cholera immunity achieved by different immunization conditions is, however, limited. The local diarrheagenic action of cholera toxin suggests that antitoxin antibodies must react with the antigen within the gut lumen or at the lining surface of the small bowel to be effective. Such coproantibodies may originate
from intestinal sites of synthesis as suggested by the finding that local antigen application can induce protection (12, 17) even in the absence of significant serum antibody formation (17). However, after parenteral immunization, antibodies derived from the circulation may also contribute significantly to the local protective antitoxic immunity as indicated by perfusion experiments with serum antibodies (4, 17, 29). In the present investigation, we have analyzed how the protective antitoxic cholera immunity in rabbits is influenced by the route of antigen administration, the number of antigen exposures, and the time interval after immunization. In addition we have examined how the immunity induced relates to the levels of antitoxin antibody of different immunoglobulin classes in serum and intestinal washings. The data reported strongly support previous indica-
1331
1332
HOLMGREN ET AL.
tions that effective antitoxic cholera immunity can be achieved by parenteral as well as enteral immunization. MATERIALS AND METHODS Antigens. Lyophilized culture filtrate of Vibrio cholerae strain 569B Inaba and culture supernatant of strain B1307 Ogawa were supplied by the National Institutes of Health, Bethesda, Md. (lot no. 4493G and 001). They were dissolved in phosphatebuffered saline (PBS; 0.05 M phosphate, 0.14 M sodium chloride, pH 7.2) for immediate use as antigen in the oral immunizations to be described. As crude toxin for other immunizations we used the same 569B filtrate after removal of low-molecular-weight material by filtration through a pellicon membrane followed by lyophilization (20). Three different preparations of this material were used: "fresh" nonaggregated toxin (crude toxin A); "aged" aggregated toxin (crude toxin B) (15); and a toxin batch prepared immediately before repeated use for a 2-month period, during which it was stored in a desiccator with silica gel drying powder to avoid aggregate formation (crude toxin C). Highly purified cholera toxin, choleragen, and its spontaneously formed nontoxic derivative, choleragenoid, were prepared by R. A. Finkelstein, Dallas, Tex. (8). The toxin was supplied by C. Miller at the National Institutes of Health, and the toxoid was a gift from R. A. Finkelstein. Both proteins were kept frozen at -70 C at concentrations of 1 or 5 mg/ml until used. A purified formalin toxoid was prepared by treating the choleragen, 4 mg/ml in tris(hydroxymethyl)aminomethane - ethylenediaminetetraacetic acid buffer [0.05 M tris(hydroxymethyl)aminomethane, 0.001 M disodium ethylenediaminetetraacetic acid, 0.003 M NaNi, and 0.2 M NaCl, pH 7.5], with 0.2% formalin at 35 C for 4 days and then diluting the detoxified protein with PBS to a concentration of 10 ,tg/ml. Purified lipopolysaccharides of V. cholerae 35A3 Inaba and of Escherichia coli 06:K2:H1 were prepared by the hot phenol-water extraction method followed by repeated ultracentrifugation (27). Immunizations. Each immunization group consisted of four to five rabbits of similar age and weight. Immunization was performed via the subcutaneous (s.c.), intravenous (i.v.), oral, or intraintestinal route, using the immunogens and vaccination schedules specified under Results. All antigens were given without addition of adjuvants. PBS-treated animals accompanied the immunized groups to serve as nonimmunized controls at the time for challenge. The s.c. immunizations were given as two 0.5-ml injections above each of the posterior legs, and the i.v. immunizations were given as injections of 1.0 ml of the antigen into a marginal ear vein. For the oral immunizations, a plastic catheter was introduced through the mouth into the ventricle after sedation of the animal. A saturated solution of sodium bicarbonate was instilled through the catheter, and when aspirates gave a neutral pH indicator reaction, usually after about 5 ml of the bicarbonate
INFECT. IMMUN. solution, the antigen was instilled in a 5-ml volume. A few minutes after administration of the antigen, another 5 to 10 ml of bicarbonate was given through the catheter. The intraintestinal immunizations were performed as previously described (15) except that 5 mg of morphine chloride was injected into the smallintestinal lumen together with the antigen to inhibit intestinal motility. Sampling of serum and intestinal specimens. The animals were bled before immunization and then at the day of challenge. Sera prepared from the bleedings were kept frozen at -20 C until use. Intestinal washings were collected from most of the rabbits as follows. Two bowel loops (20 to 30 cm long) were tied in different positions. The loops were injected with 10 ml of PBS, and 5 min later the fluid was aspirated and centrifuged at 2,000 x g for 15 min to eliminate fecal contents. The two washings of each animal were pooled, heat inactivated at 56 C for 30 min, and stored at -20 C until use. In one experiment (see Table 4), intestinal washings and mucosa scrapings were collected from the excised entire small bowel. The washings were obtained by rinsing the lumen with 20 ml of cold PBS and handling the fluid as above. The intestine was then everted over a glass rod and the mucosa layer was scraped off. These scrapings were homogenized in an Omnimixer (Sorvall, Norwalk, Conn.; 10,000 rpm, 10 min) and then in an Ultrasonic disintegrator (MSE Ltd., London; 8 A, 20 kc/min, 10 min), centrifuged (2,000 x g, 20 min), heat inactivated (56 C, 30 min), and frozen (-20 C) until use. Challenge with V. cholerae bacteria or toxin. At the time of challenge the rabbits were 12 to 16 weeks old. Operations were performed as earlier reported (17). In the early phase of the study the ileal loop technique with five 10-cm-long test loops and a "negative-control" saline loop in each animal was used. It was later replaced by the small-bowel multiloop technique described by Pierce and Wallace (30), using 20 to 25 approximately 5-cm-long loops, including two saline control loops, in each rabbit. For challenge, each loop was injected with either culture filtrate (toxin) or live vibrios of the Inaba strain 569B. Serially threefold varied amounts of toxin, 0.1 to 10 mg, or tenfold varied numbers of bacteria, 104 to 109, were injected in 1-ml (smallbowel multiloop technique) or 2-ml (ileal loop technique) volumes. The bacteria had been cultivated in Syncase medium at 37 C for 4 to 5 h and diluted in this medium; the toxin was diluted in PBS. EC50 (the concentration of toxin or the density of bacteria which in the standardized volumes mentioned above caused half-maximal fluid accumulation in a loop) was calculated for each group of animals (Group EC5O,) according to Burrows and Musteikis (2). The reproducibility of such determinations is within 10% (15). The protective effect of immunization was determined as the ratio between the Group EC50 values for vaccinated and PBS-injected animals. This ratio is referred to as the protection factor. The introduction of the small-bowel multiloop
VOL. 12, 1975
ANTITOXIC IMMUNITY IN CHOLERA PROTECTION
1333
TABLE 1. Protection against intestinal toxin technique enabled determination of EC50 values also for the individual rabbits (Individual EC5,) since by challenge by subcutaneous immunization with crude toxin this technique each challenge dose could be tested in four different positions in the same animal. No. of imGroup Protection Animal Antibody and immunoglobulin assays. The imEC_' muniza factord gr°UP munoglobulin contents of intestinal washings and tions° (mg/ml) scrapings were analyzed by the single radial diffu1 0 0.16 sion method of Mancini et al. (25). Goat antisera 2 1 0.23 1.41 specific for rabbit immunoglobulin G (IgG), IgM, or 2 0.75 4.69 3 IgA were purchased from Nordic, Tilburg, The Netherlands. Relative immunoglobulin values for 4 0 0.15 test specimens were calculated from standard curves 1 1.40 9.33 5 established with reference preparations consisting of 6 2 13.0 1.95 a serum (for IgG) and an intestinal washing of an unimmunized rabbit (for IgA). 0.18 0 7 The intradermal toxicity test of Craig was used to 1 0.18 1.0 8 determine neutralizing antibodies (3). The titers 4.00 22.2 2 9 were corrected in relation to a reference antiserum included in each test animal. On later comparison a Five animals in each group. Groups 2 and 3 were with the Swiss standard antitoxic cholera serum immunized with crude toxin A, groups 5 and 6 with (4,470 antitoxic units/ml), our reference serum had crude toxin B, and groups 8 and 9 with crude toxin to 115 antia neutralizing capacity corresponding 6 were challenged by the ileal 1 C. through Groups toxic units/ml, which in our assay gave a mean titer loop technique, and groups 7 through 9 were chalof 1:30,000. by the small-bowel multiloop assay. The enzyme-linked immunosorbent assay lenged b Number of s.c. immunizations with 12.5 mg of (ELISA) (5) was used for titration of antiexotoxin crude cholera toxin in each injection; for groups 3 antibodies, using purified cholera toxin as solid- and 6 the were given 2 weeks apart and for phase antigen as previously described (21). Titer group 9, 3injections weeks apart. correction was done in relation to the same reference c Concentration of culture filtrate resulting in serum as that used in the neutralization test. fluid accumulation in small-bowel half-maximal Statistical methods. The statistical analyses loops on challenge 4 to 5 days after the immunizawere done by Student's t test or by the Spearman tions. rank correlation test (13). d Ratio between EC50 values of immunized and PBS-injected (control) animal groups. RESULTS
groupal
Protection by single and booster parenteral immunizations. The protective immunogenicity of one or two s.c. injections of crude cholera toxin was evaluated in three sets of experiments using three different preparations of this antigen (Table 1). Only one preparation (crude toxin B), which probably contained a high proportion of aggregated antigen (15), gave in a single injection rise to pronounced protection against toxin challenge. In a booster immunization all antigen preparations were effective giving up to 22-fold higher resistance than was found after a single injection. A single i.v. injection of crude toxin A also induced significant protection. As soon as 3 days after the immunization a protection factor of 2.4 was registered against toxin challenge, i.e., higher than early in the primary response to antigen given s.c. (group 2 in Table 1). Previous experiments had indicated that the protection induced by booster immunization with crude toxin is highest shortly after the antigen injection (17). In the present investigation the time course of the protective immunity to toxin challenge was studied more thoroughly both in response to a single and booster s.c. immunization with crude toxin C (Fig. la). A
moderate progressive increase in resistance was seen during the 1-month observation period of the primary response. By contrast, after reimmunization the markedly enhanced immunity noted after 4 days decreased by almost onehalf within 10 days and was approximately onethird of the maximal value after 1 month. For the animal groups presented in Fig. 1 the Individual EC50 values were also determined (Table 2). The means of these values were consistently higher than for the Group EC.;, values, but the differences were minor (ratio range, 1.05 to 1.44). Statistical evaluation with Student's t test showed that the protection 30 days after a single immunization was significantly increased compared with the controls, as was the resistance registered in all the secondaryresponse groups. The protection observed during the secondary response also differed significantly from that seen during the primary reponse (Table 2). Antibodies in serum and intestinal washings. The antibody response to the cholera exotoxin was studied in serum specimens and intestinal washings, using the intradermal neutralization test (3) and the ELISA measuring the primary binding of antibodies to solid-phase
1334
HOLMGREN ET AL.
INFECT. IMMUN.
PRIMARY RESPONSE
SECONDARY RESPONSE
a
protection
22.2 4-
3-
12.1
E
E
6
(5
7.7 1-
0-
3.2 1.0 10
16
a m
I
I serum antibody
62 CP 0
2-
