Evaluation of Surface Components of Vibrio cholerae as Protective ...

Vol. 15, No. 2

INFECTION AND IMMUNITY, Feb. 1977, p. 533-538 Copyright ©) 1977 American Society for Microbiology

Printed in U.S.A.

Evaluation of Surface Components of Vibrio cholerae as Protective Immunogens ELIZABETH R. EUBANKS,' M. NEAL GUENTZEL,2 AND L. J. BERRY* Department of Microbiology, The University of Texas, Austin, Texas 78712

Received for publication 27 July 1976

Surface components of a motile Inaba strain (CA401) were removed from washed cells by low-speed shearing. Flagella contaminated with a vesicular material (designated as crude flagella [CF]) were obtained by differential centrifugation of the shear fluid. Vesicles were obtained from a nonflagellated mutant by the same procedure. Homogeneous small vesicles were obtained in diminished yield from CsCl gradients of CF preparations. Treatment of CF with sodium deoxycholate removed the vesicular material and flagellar sheaths and yielded naked flagella (NF). The ability of these preparations to passively protect infant mice suckled by CFW mothers that had been immunized at the time of mating was compared, on a dry-weight basis, with commercial vaccine (CV). Eight-day-old mice were challenged orally with more than 1,000 50% lethal doses of either the homologous or a heterologous (Ogawa CA411) strain. The most effective immunogen was CF, which provided complete protection at 1 ,gg against both challenges. CF and vesicles provided 50- to 100-fold greater protection than CV against homologous challenge. With heterologous challenge, vesicles were 10-fold more protective than CV, but markedly less protective than CF. The NF offered only slightly greater protection than CV against both challenges. Immunoelectrophoresis revealed an antigen in CF distinct from vesicles, cell wall lipopolysaccharide or NF. This antigen is not present in the nonflagellated mutant and is apparently associated with motility. Passive protection of infant mice via colostrum and milk of appropriately immunized mothers can be demonstrated against oral (20) and intraperitoneal (17) challenge with virulent Vibrio cholerae. Guentzel and Berry (9) used passively immunized infant mice to compare the efficacy of commercial vaccine, cholera enterotoxin, and an Ogawa-derived subcellular vaccine. Both antibacterial and antitoxic dosedependent immunity was demonstrable against oral challenge with highly virulent Inaba, Ogawa, and El Tor Ogawa strains. In each case, the subcellular vaccine was markedly more protective on a weight basis than the other test vaccines. A number of reports (6-8, 10, 14-16) have suggested that adherence of vibrios to the mucosa of the infected host is an essential prerequisite to initiation of disease, even though cholera is a disease caused by the release of a single protein enterotoxin (4) that results in the copious loss of water and electrolytes from the small I Present address: Department of Botany and Microbiology, Arizona State University, Tempe, AZ 85281. 2 Present address: Division of Allied Health and Life Sciences, The University of Texas at San Antonio, San Antonio, TX 78285.

intestine. This intimate association of vibrios with the mucosa may be necessary for the delivery of toxin to its site of action or for its release in vivo. Guentzel and Berry (10) have established that nonmotile mutants of highly virulent strains of V. cholerae are markedly reduced in virulence and in their capacity to adhere to the surface of mouse intestine. Loss of motility did not alter enterotoxin production or phage susceptibility, and the strains remained prototrophic. The above considerations made it reasonable to assume that vibrio flagella, when used as a vaccine, would protect against the onset of disease by immobilizing the pathogens. Even though the results presented in this report fail to confirm this premise insofar as showing that the vibrios lose motility in vivo, they do show that, as a vaccine, preparations of flagella, contaminated with what may be vesicular particles derived from the flagellar sheath or the cell surface, provide the highest degree of protection yet seen in the infant mouse model. The vaccine also has the advantage of offering protection against heterologous challenge that differs only slightly from that seen after homologous challenge. 533

534

EUBANKS, GUENTZEL, AND BERRY

MATERIALS AND METHODS Mice. CFW mice, purchased from Carworth Division, Charles Rivers Farms, were bred and their offspring raised in departmental animal facilities. The animals were housed and fed as described in a previous communication by Guentzel and Berry (9). Bacterial strains. The highly virulent motile strains of V. cholerae used in this study, CA401 (Inaba) and CA411 (Ogawa), and a weakly virulent nonmotile mutant, CA401 M-5, have been described previously (10). All cultures were maintained in the lyophilized state and restored as needed. Vaccines. Commercial bivalent cholera vaccines were obtained from the Health Center, The University of Texas at Austin. Equal mixtures of Lederle lot no. 329-383 and Lilly lot no. 6WE36A were administered in doses equivalent to cell dry weights (9) of 10, 1.0, and 0.1 Ag unless sp-ecified otherwise. An Ogawa-derived subcellular vaccine (HS222) was kindly provided by P. Actor of Smith, Kline and French Laboratories. The purification and properties of this material have been reported (12). Crude flagella (CF) were prepared from CA401 (Inaba) by first selecting highly motile cells from 'race tubes" to use as inocula. Cultures were grown to the early stationary phase (12 h) with aeration in carboys. The medium contained, in grams per liter: tryptone, 10; yeast extract, 1; glucose, 1; NaCl, 5; K2HPO4, 2; MgSO4, 0.1; pH 7.5. Cells were harvested by low-speed centrifugation, and 50 g of cells was suspended to 200 ml in buffer (0.02 M borate, 3% NaCl, pH 8.4). Surface components were removed by low-speed shearing for 5 min with a Sorvall Omnimix. The shearing procedure did not affect cellular morphology in Gram-stained smears and did not appreciably reduce the viability of the vibrios. The sheared preparations were then subjected to two centrifugations (15,900 x g) to remove cells and large debris. The flagella were pelleted from the supernatant fluid by high-speed centrifugation (85,000 x g, 1 h). The differential centrifugations were repeated three times. The washed pellet suspended in buffer [0.1 M tris(hydroxymethyl)aminomethane, 0.005 M ethylenediaminetetraacetate, pH 7.8], called CF, consisted of flagella contaminated with small and large "vesicles" and phage parts. Vesicles free of flagella were obtained by subjecting the nonflagellated mutant of CA401, M-5, to the same procedure described above. These preparations, free of flagella, were called M-5 vesicles (V). Small vesicles also were obtained in lower yield from CsCl gradients of CF preparations. For this separation, cesium chloride was added to CF [3 mg of protein per ml in 0.1 M tris(hydroxymethyl)aminomethane, 0.005 M ethylenediaminetetraacetate, pH 7.8] at a concentration of 0.45 g/ml. Separation of components was achieved by centrifugation at 40,000 rpm (SW50L rotor) for 20 h. Collection of 10-drop fractions yielded homogeneous small vesicles and purified flagella containing small and large vesicles. The two fractions were dialyzed twice against 100 volumes of phosphatebuffered saline (pH 7.2) for 8 h. Treatment of CF with sodium deoxycholate according to the procedure outlined in the flow diagram (Fig. 1) removed

INFECT. IMMUN. Crude flagella (8 mg of plrotein/ml) 1% Sodium deoxycholate 0.1 M Tris, pH 8.0

Centrifuge - 85,000 x g

Pellet (flagella)

Supernatant (LPS)

Wash 3 x in 0.5% sodium deoxycholate

Add 6 volumes of ethanol Centrifuge - 16,000 x g

Purified flagella LPS Wash 3 x, 0.1 M Tris, pH 8.0 FIG. 1. Flow diagram ofprocedure used to obtain purified flagella (NF) and vibrio LPS from CF preparations.

vesicular material and flagellar sheaths and yielded naked flagella (NF). The flagellar pellet, obtained by high-speed centrifugation, was washed three times in 0.5% sodium deoxycholate and dialyzed against 100 volumes of phosphate-buffered saline. Lipopolysaccharide was precipitated from the deoxycholate supernatant fluid as shown in Fig. 1. Immunoelectrophoresis. Antibody was raised in rabbits against CF. Immunoelectrophoresis was performed according to the manufacturer's directions (Gelman Instruments Co.), applying a current of 3 mA/frame for 2 h. Immunization. Female CFW mice, 10 weeks of age, were administered a single subcutaneous dose of one of the test vaccines just prior to mating, as described previously (9). All vaccines were diluted in sterile nonpyrogenic saline (Travenol, Inc.) just prior to use. Control animals were injected with saline alone. Oral challenge. Seven-day-old offspring (eight to ten mice per litter) of appropriately vaccinated mothers were fasted overnight and then challenged by the oral route as described previously by Guentzel and Berry (9). The challenge dose, which varied in individual experiments from 5 x 107 to 108 colonyforming units, represented greater than 1,000 50% lethal doses of CA401 (Inaba) or CA411 (Ogawa).

RESULTS Preliminary experiments. The isolation of flagella from highly virulent Inaba CA401 was carried out to determine the possible utility of surface components as protective immunogens. Electron microscopy on samples of CF (Fig. 2) negatively stained with phosphotungstic acid revealed the presence of long flagella contaminated with a heterogeneous population of vesicles and a few phage parts. Passive protection afforded infant mice suckled on mothers immu-

535

SURFACE ANTIGENS OF V. CHOLERAE

VOL. 15, 1977

't

,.

'i

vesicles

of dIfferent sIz

an

p

FIG. 2. Electron photomicrograph of a CF preparation. Flagella, vesicles of different sizes, and phage pieces are visible. x73,800.

nized with CF, homogeneous small vesicles, and purified flagella was compared with that observed with commercial vaccine and the Ogawa-derived subcellular vaccine (HS222) (Table 1). Each value represents the challenge of 20 to 40 mice with greater than 1,000 50% lethal doses of the Inaba strain. CF and purified flagella and vesicles at 1 ,rg offered essentially complete protection against the homologous

challenge and were more protective than HS222. Each of the subcellular vaccines proved to be manyfold more protective than the wholecell commercial vaccine. Since similar results were obtained when survivors were scored at 36 h or 7 days, and most infant mice succumb to oral challenge between 16 and 24 h (9), the data in the following experiments are expressed as survivorship at 36 h.

536


INFECT. IMMUN.

TABLE 1. Preliminary comparison of V. cholerae vaccines: oral challenge with Inaba CA401 of passively immunized infant mice

that of the commercial vaccine. NF were superior to commercial vaccine with homologous challenge. Vesicles offered protection comparable to that of CF against homologous challenge but were demonstrably less protective against heterologous challenge. Vesicles appear to be superior to CF at the smallest test dose used (Fig. 4). Effect of an oral injection of CF on experimental cholera in infant mice. Since CF serve as an excellent vaccine in providing passive immunity in infant mice nursed by immunized mothers, it was reasonable to assume that the antibody blocked attachment of vibrios by combining with the flagella. It was considered

% Survivors at:

Vaccine

Dose (ug) 36 h

Saline controls Comercial killed vaccine HS222 CF Vesicles Purified flagella

38.0 (2 x 108 cells) 3.8 (2 x 107cells) 0.38 (2 x 106 cells) 10 1 10 1 10 1 10 1

7

8

days 6

80 31 5 92 67 90 100 100 93 100 87

75 31 5 83 55 87 94 100 90 86 83

0)

x

INABA

loo

o0

(33)

c

EQ

Comparison of the efficacy of CF, vesicles, and NF as vaccines. Since the flagella band in the CsCl gradient was poorly separated from the vesicle band, two other approaches were used to determine the relative contribution of flagella and contaminating vesicles to the observed passive immunity conferred by crude flagellar preparations. Vesicles (V) were prepared from cultures of nonflagellated mutant of CA401 by the same procedure used to prepare CF. Second, CF preparations were treated with sodium deoxycholate to remove the vesicles. This procedure yielded NF, since electron microscopy revealed that both vesicles and flagellar sheaths were removed by this procedure. Commerical bivalent vaccine and the Ogawaderived subcellular vaccine were used as standard to which the various preparations were compared. Eight-day-old offspring of immunized mothers were challenged orally with greater than 1,000 50% lethal doses of the homologous motile strain (CA401) or a heterologous Ogawa strain (CA411). The results of one of several determinations are illustrated in Fig. 3 and 4. CF provided approximately 100-fold greater protection than commercial vaccine, as judged by vaccine dose, against homologous (Inaba) challenge. Protection conferred by CF against heterologous (Ogawa) challenge was similar to that observed against homologous challenge (Fig. 3). Passive protection conferred on offspring of mice immunized with vesicles and NF is illustrated in Fig. 4 (the data from Fig. 3 for CF and commercial vaccine are plotted as hatched lines for comparison). NF were markedly less protective than vesicles against either homologous or heterologous challenge but offered protection similar to

I-

*0-

(44) | I~tt2 .A

w 9) 40-

0.

(4)(36)

z

VAC

(33

20

(51)

.(421 ,

CONTROL I

4

3

LOG (ng) IMMUNIZING DOSE FIG. 3. Percentage of survival in 8-day-old mice suckled by mothers vaccinated at the time of mating with the indicated amount of one of several vaccines. The infants were challenged orally with 1,000 50% lethal doses of either the homologous CA401 Inaba strain or the CA411 Ogawa strain. The numbers in parentheses show the number of infants used for each point in the graph. CF is crude flagella, VAC is the bivalent commercial killed vaccine, and HS222 is the Ogawa-derived subcellular vaccine. 0

CD

INABA

OGAWA

40

A CF

so-

_

--(4

WA

0 *o o

|

/~~~~~~(22)

{

4Car

so

-I

w

/

I

/

1-Go 0

a

_

f,---

Va

/

I

40

*0

VC

(44)/

z Wlltoe

I (47), I -

(41

6 / (66)Mr

I} 1W){

C.

20

__2

(42)

I CONTROL

i3

CONTROL

i

LOG (ng) IMMUNIZING DOSE FIG. 4. Same as Fig. 3. The hashed curves are the same as those shown in Fig. 3. VES is purified vesicles and NF is naked (purified) flagella.

VOL. 15, 1977

TABLE 2. Challenge of suckling mice with V. choleraea mixed with CF % Survivors at: Concn of CF (/g)

SURFACE ANTIGENS OF V. CHOLERAE

537

associated with flagellated vibrios was not present. This observation and the reduced protection of NF compared with that of CF is discussed in relation to the complex architecture of 36 h 7 days the V. cholera flagellum. 0 0 0 DISCUSSION 10 8 8 100 12 12 The relative merit of antibacterial and antia Oral challenge with 108 colony-forming units of toxic immunity in providing effective prophylaxis against cholera has not been determined Inaba CA401. in spite of the growing number of reports on cholera immunity, recently reviewed by Finkelstein (4). We have used oral infection of CRUDE FLAGELLA passively immunized infant mice to compare ANTI-CRUDE FLAGELLA the efficacy of various types of vaccines. In addition to the advantages of low cost and M-5 VESICLES availability in large numbers, infant mice are FIG. 5. Tracing of the immunoelectrophoresis susceptible to orally administered toxin (Guentpattern obtained when rabbit antiserum against CF zel, unpublished observations), are highly susderived from the Inaba CA401 strain was reacted ceptible to oral infection with small doses of against the CF vaccine and the vesicles derived from virulent strains of V. cholerae (10), and suca nonmotile mutant of the same strain. An antigen cumb to infection with symptoms resembling present in CF is missing in the vesicles. those seen in the human disease. In a previous report (9), we observed that the levels of protecworthwhile to determine whether vibrios would tion conferred against challenge with the have reduced ability to attach and hence cause highly virulent Ogawa and Inaba strains were disease if they were mixed with the CF vaccine similar when mice were suckled on mothers prior to infectious challenge. Two dose levels immunized with commercial bivalent vaccine were used, 10 and 100 jig of crude flagella per or cholera enterotoxin. However, an Ogawa108 challenge organisms of the homologous In- derived subcellular vaccine was demonstrably aba CA401 strain. The results are given in more protective at all dose levels tested. In the Table 2 and indicate that a statistical level of present report, the level of passive immunity protection was not obtained with either amount conferred on the offspring of female mice vacciof flagella vaccine. The larger dose represents a nated with CF at the time of mating is superior dry-weight ratio of 1:7 of challenge organisms to that obtained with any preparation that has to CF. Larger amounts could not be tested be- been tested in this model. Significant survival cause of technical difficulties. The fact that against homologous and heterologous challenge there was 12% survival leaves unresolved is afforded by antigenic doses of less than 100 ng whether more flagella would have reduced the (Fig. 3 and 4), a remarkably small amount outcome of the infection. since resistance to infection is passively conComparison of vesicles and CF by immuno- ferred. The yield of NF is less than 5% of the Lowry electrophoresis. Antisera were prepared against CF to compare the immunoelectropho- protein in the CF. Therefore, if the NF were retic pattern of antigens present in V with that responsible for the superior protection of the of CF (Fig. 5). Every preparation of vesicles and CF, the purified flagella should be highly proCF gave the same pattern. Only one antigen tective, but they were not. Since the flagellum was detected in CF that was not present in of V. cholerae is sheathed (2, 5), it is not survesicles. The other antigens removed from the prising that antibody against NF is not very vibrios by shearing appear to be identical. Of protective. The CF should not be more protecimportance is the fact that NF failed to produce tive against the heterologous challenge than precipitin lines with the anti-CF. Evidently, vesicles (essentially all Inaba LPS) if the LPS of the antigen band associated solely with flagel- the cell envelope and the sheath are the same lated vibrios was not flagella. To determine antigen. However, CF were markedly more whether this antigen was lipopolysaccharide protective against the heterologous challenge (LPS), the LPS precipitated from the superna- than vesicles. An antigen, present in CF and tant resulting from deoxycholate treatment of not present in vesicles from the nonflagellated CF and vesicles were subjected to immunoelec- mutant, may be responsible for the excellent trophoresis. The patterns produced by both LPS protection conferred by CF. This antigen is not preparations were identical, and the antigen classical LPS or NF, but it may be flagellar

538


sheath material, although there is no direct evidence for this. The basis for the immunity is not easily established. As shown in the paper that follows (11), the number of vibrios and the depth of their penetration into intervillous spaces and crypts of Lieberkuhn within the ileum are dramatically diminished in passively immunized infant mice. The distribution seen under these conditions is similar to that found in infected infants of unvaccinated mothers challenged with nonmotile organisms. By implication, antibody transmitted via the milk would, therefore, render virulent motile vibrios nonmotile. Benenson et al. (2) observed in 1964 that even somatic antibody (presumably anti-LPS) inhibited vibrio motility in wet mounts of stools from cholera patients. There is no experimental evidence to show that this occurs in the intestine as a result of passive immunity, but the inference is unavoidable. Nonmotile mutants of highly virulent motile strains of V. cholerae are markedly reduced in virulence (3, 10) and in the capacity to adhere to intestinal mucosa (10, 13). In infected ileal loops of normal rabbits, the onset of fluid accumulation was concomitant with the establishment of large masses of organisms in the intervillous spaces and crypts after successful penetration of a mucous zone separating the intestinal contents from the tips ofthe villi (18). This process occurred less readily with nonmotile cells (21). Recent evidence suggests that crosslinking of vibrios by antibody (1, 19) or inhibition of "mobility" by clumping of vibrios in the lumen by antibody plays a significant role in immunity to cholera. Significantly, specific vibrio antiflagella antiserum is at least as protective as specific antisomatic antiserum when administered to orally challenged infant mice (1). ACKNOWLEDGMENTS This work was supported in part by Public Health Service grant AI-10466 from the National Institute of Allergy and Infectious Diseases. We are indebted to Leodocia M. Pope of this Department for the electron photomicrograph.

LITERATURE CITED 1. Bellamy, J. E. C., J. Knop, E. J. Steele, W. Chaicumpa, and D. Rowley. 1975. Antibody cross-linking as a factor in immunity to cholera in infant mice. J. Infect. Dis. 132:181-188. 2. Benenson, A. S., M. R. Islam, and W. B. Greenough III. 1964. Rapid identification of Vibrio cholerae by dark-field microscopy. Bull. W.H.O. 30:827-831. 3. Eubanks, E. R., M. N. Guentzel, and L. J. Berry. 1976. Virulence factors involved in the intraperitoneal infection of adult mice with Vibrio cholerae. Infect.

INFECT. IMMUN. Immun. 13:457-463. 4. Finkelstein, R. A. 1975. Immunology of cholera. Curr. Top. Microbiol. Immunol. 69:137-196. 5. Follett, E. A. C., and J. Gordon. 1963. An electron microscope study of vibrio flagella. J. Gen. Microbiol. 32:235-239. 6. Freter, R. 1969. Studies of the mechanism of action of intestinal antibody in experimental cholera. Tex. Rep. Biol. Med. 27(Suppl. 1):299-316. 7. Freter, R. 1970. Mechanism of action of intestinal antibody in experimental cholera. II. Antibody-mediated antibacterial reaction at the mucosal surface. Infect. Immun. 2:556-562. 8. Freter, R. 1972. Parameters affecting the association of vibrios with the intestinal surface in experimental cholera. Infect. Immun. 6:134-141. 9. Guentzel, M. N., and L. J. Berry. 1974. Protection of suckling mice from experimental cholera by maternal immunization: comparison of the efficacy of wholecell, ribosomal-derived, and enterotoxin immunogens. Infect. Immun. 10:167-172. 10. Guentzel, M. N., and L. J. Berry. 1975. Motility as a virulence factor for Vibrio cholerae. Infect. Immun. 11:890-897. 11. Guentzel, M. N., L. H. Field, E. R. Eubanks, and L. J. Berry. 1977. Use of fluorescent antibody in studies of immunity to cholera in infant mice. Infect. Immun. 15:539-548. 12. Jensen, R., B. Gregory, J. Naylor, and P. Actor. 1972. Isolation of protective somatic antigen from Vibrio cholerae (Ogawa) ribosomal preparations. Infect. Immun. 6:156-161. 13. Jones, G. W. 1975. Adhesive properties of enteropathogenic bacteria, p. 137-142. In D. Schlessinger (ed.), Microbiology- 1975. American Society for Microbiology, Washington, D.C. 14. LaBrec, E. H., H. Sprinz, H. Schneider, and H. B. Formal. 1965. Localization of vibrios in experimental cholera: a fluorescent antibody study in guinea pigs, p. 272-276. In Proceedings of the cholera research symposium. U. S. Public Health Service Publ. no. 1328. Government Printing Office, Washington, D.C. 15. Lankford, C. E. 1960. Factors of virulence of Vibrio cholerae. Ann. N.Y. Acad. Sci. 88:1203-1212. 16. Lankford, C. E., and U. Legsomburana. 1965. Virulence factors of choleragenic vibrios, p. 109-121. In Proceedings of the cholera research symposium. U.S. Public Health Service Publ. no. 1328. Government Printing Office, Washington, D.C. 17. Pitkin, D., and P. Actor. 1972. Immunity to Vibrio cholerae in the mouse. I. Passive protection of newborn mice. Infect. Immun. 5:428-432. 18. Schrank, G. D., and W. F. Verwey. 1976. Distribution of cholera organisms in experimental Vibrio cholerae infections: proposed mechanisms of pathogenesis and antibacterial immunity. Infect. Immun. 13:195-203. 19. Steele, E. J., W. Chaicumpa, and D. Rowley. 1975. Further evidence for cross-linking as a protective factor in experimental cholera: properties of antibody fragments. 132:175-180. 20. Ujiiye, A., and K. Kobari. 1970. Protective effect on infections with Vibrio cholerae in suckling mice caused by the passive immunization with milk of immune mothers. J. Infect. Dis. 121(Suppl.):S50S55. 21. Williams, H. R., Jr., W. F. Verwey, G. D. Schrank, and E. K. Hurry. 1973. An in vitro antigen-antibody reaction in relation to an hypothesis of intestinal immunity to cholera. In Symposium on cholera. U.S.Japan Cooperative Medical Science Program.