Attenuation of virulence by P and V plasmids in Vibrio cholerae: strains suitable for oral immunization*

Bulletin of the World Health Organization, 57 (4): 643-647 (1979) Attenuation of virulence by P and V plasmids in Vibrio cholerae: strains suitable for oral immunization * V. B. SINHA' & BRAHM S. SRIVASTAVA 2 In a virulent strain of Vibrio cholerae, KB9, P and V plasmids were introduced by bacterial conjugation. Characterization ofPV isolates and systematic screening of them in animal models of cholera revealed that a large number ofPV isolates were non-pathogenic, owing to the loss of ability to synthesize toxin. Results obtained with two such strains, designated as KB9:PV and CD24, are described. The strains with plasmids were stable during in vitro cultivation or during two successive passages in rabbit intestine. Protection conferred by PV strains was determined in mouse protection tests and in the rabbit ileal loop model. The plasmid strains were immunogenic. In view of the results, it is proposed that PV-bearing attenuated strains should be tried in oral immunization. Several attempts have been made to isolate attenuated strains of Vibrio cholerae suitable for use as live oral vaccines (for reviews see 7, 13). Unfor- tunately, some of these strains have proved to be unsatisfactory (3). In an attempt to isolate new promising strains, we discovered that the pathogenicity of V. cholerae was significantly suppressed by introducing the P and V plasmids (12). The present paper demonstrates a new procedure for producing attenuated PV strains of V. cholerae. Plasmids did not alter the immuno- genicity, and they were not lost in vitro or in vivo. MATERIALS AND METHODS Vibrio strains The V. cholerae strains used, with their relevant markers and plasmid status, are shown in Table 1. Details of the maintenance and isolation of strains harbouring plasmids have been described elsewhere (12). Culture media Nutrient broth (Difco) and brain heart infusion (BHI) (Difco) were prepared according to the * From the Cholera Immunology Laboratory, Central Drug Research Institute, Lucknow-226001, India. ' Scientist. 2 Scientist. manufacturer's directions. " Syncase " sucrose minimal medium was prepared as described by Finkelstein & Lospalluto (8). Phosphate-buffered saline (PBS) and minimal medium supplemented with the required amino acids were prepared as described earlier (12, 2). All the media, when required, were solidified with 1 % agar powder (Oxoid) and were autoclaved at 103 kPa (15 lbf/in2) for 10 min. Table 1. Vibrio cholerae strains used in these experi- ments Strain Relevant markers Plasmid Reference KB9 ilv arq his, strl, 0-Og P-V- 2 K89: P iW arg his, strl, 0-Og P+V- 2 KB9:V lv arg his, strr, 0-Og P-V+ 2 KB9: PV ilv arg his, str, 0-Og P+V+ 2 KB92 Prototroph, 0-Og P-V- K. Bhaskaran KB92: P Prototroph, 0-Og P+V- 12 KB93 Prototroph, 0-In P-V- K. Bhaskaran CD24 i/v, arg his, str', 0-Og P-V+ This work 569B Prototroph, 0-In P-V- K. Bhaskaran 569B: P Prototroph, 0-In P+V- 12 569B: V Prototroph, 0-In P-V+ 12 CD23 Prototroph, 0-In P+V+ This work 3837 643 V. B. SINHA & B. S. SRIVASTAVA BHI culture Broth culture and growth conditions have been reported earlier (12). Bacterial dilutions were made in PBS. Animal experiments Mouse virulence tests and ileal loop experiments in adult rabbits (4) were performed as described elsewhere (12). The 10-day-old infant rabbit model (5) was em- ployed for the enteropathogenicity test. The animals were infected intraintestinally with 1 ml of culture suspension and the occurrence of diarrhoea or death was noted up to 48 h. Later, an autopsy was done and the amount of fluid accumulated in the caecum was recorded. The mouse protection test was based on the method of Pittman & Feeley (11). Swiss mice, 10-12 g in weight and reared in this Institute, were immunized by injecting intraperitoneally 3-7.4 x 106 organisms. These immunized mice were challenged intraperitoneally on day 15 with 500 LD50 of homologous (KB92, Ogawa) and heterologous (KB93, Inaba) virulent strains. The challenge dose was prepared in 5 % mucin (Type 1701-W, Wilson Laboratories, Ill.). Mortality was recorded up to 72 h. The ligated ileal loop model in the adult rabbit (4) was employed for determining the immunogenicity of the attenuated strains. Two doses of live vaccine (BHI culture), each 9f 1 ml, were administered subcutaneously with an interval of 21 days. The vaccine was prepared on the day of immunization. The bacterial counts varied from 1.24 to 1.86 X 109 per ml. All'the rabbits were bled before immuniz- ation and before the challenge. Two ileat loops were made in each rabbit for challenge with homologous (KB92, Ogawa) and heterologous (KB93, Inaba)' strains. Challenge inoculum consisted of approxi- mately 106 organisms contained in 1 ml. A post-- mortem was performed after 18 h. Loop contents and length were measured in order to determine the immunity to challenge. Serum vibriocidal antibodies were measured by Finkelstein's method (6). RESULTS Attenuation of virulence Mouse virulence test. The LD50s of the virulent parent strain, KB9, and its attenuated derivatives, KB9: PV and CD24, were determined. The LD50 of KB9 was approximately 9 organisms, whereas the LD50s of KB9:PV and CD24 were found to be 2.77 x 106 and 4.2 X 107 organisms, respectively. Enteropathogenicity in rabbits. In the adult rabbit model, 1.2-1.6 X 109 organisms of KB9:PV and CD24 were administered to each ligated ileal loop but they failed to cause any outpouring of fluid. In similar tests, 1-1.3 x 104 organisms of KB9 pro- duced a significant amount of fluid (12). The nonpathogenic nature of plasmid-carrying strains was further confirmed by experiments with the infant rabbit model. None of the 16 infant rabbits infected with KB9: PV or CD24 showed any evidence of diarrhoea or accumulation of fluid in the caecum. Postmortem findings revealed a normal appearance of the gut. On the other hand, rabbits infected with KB9, with a dose as low as 1.4 x 105 organisms, developed diarrhoea, and the post- mortem examination done after 48 h showed a characteristic distension of the caecum. Assay of toxin in guinea-pigs. Earlier we reported that a strain of 569B harbouring both P and V plasmids produced less toxin in vitro (12). Since previously pathogenic strains lost their pathogenicity completely when carrying P and V, several strains of 569B containing P and V were prepared and their capacity to synthesize toxin in vitro was examined. Most of the strains harbouring P and' V did not produce any detectable amount of toxin, whereas under identical conditions 569B gave a significant quantity of toxin. A bacteria-free culture filtrate of CD24 did not produce induration in the skin of the guinea-pigs. Stability ofplasmids Our present findings show that the inclusion of P and V plasmids in V. cholerae strains causes attenuation of virulence in various experimental models. Therefore, it becomes necessary to examine Preparation and assay of cholra toxin, the stability of these plasmids.- preparation o. t tx an its.......assa have In conjugation. A conjugation experiment be-ThTbe prprtO Itetxl n t sa ae tween KB9: Pand KB92 showed that while the P been described earlier (12).been describedearlier(12).plasmid was transferred to KB92 with normal fre- quency, KB9:P did not become -P. This was re- Preparation of strains with P and V plasmids vealed in the following experiment. The two strains This, together with detection of P and V plasmids, were plated on a streptomycin nutrient agar, which have been described in earlier papers (2, 12). allowed only the donor to grow. Simultaneously, 644 VIBRIO CHOLERAE STRAINS FOR ORAL IMMUNIZATION samples were plated on minimal medium on which only KB92 would grow. It was found that 100% of the donor cells were PI whereas 44-64 % of KB92 were PI. Hence, the donor cells always retained a copy of the P plasmid. This is suggestive of plasmid replication during sex transfer. In vivo. Strains KB9:P and KB9:V were given two serial passages in the ileal loop of adult rabbits. About 100 colonies, isolated after animal passage, were tested for their plasmid status. In no instance was there any loss of the plasmids. In vitro. Incubation with a nonlethal concentration of acridine orange and sodium dodecyl sulfate did not eliminate the plasmids from cells. Two successive subcultures of plasmid-bearing strains in BHI did not eliminate plasmids. Strains KB9:P, KB9:V, and KlB9:PV were kept for 3-4 weeks at 37°C without agitation. Discrete colonies were obtained on nutrient agar plates. In all, 144 colonies were tested and all of them were found to have retained their respective plasmids. When KB9: PV was incubated for 4 weeks, 7 colonies did not show the P plasmid. CD24 was one of these isolates. On the assumption that CD24 harbourecr only V, this strain was crossed with a PI strain to transfer P into CD24. In two independent matings, P could not be transferred to CD24. A control was also run with KB9 V in which P transfer occurred with normal frequency. This suggests that CD24 is not indeed devoid of P, which may be present in a repressed state (2). Immunogenicity of K69:PV and CD24 Before using them as vaccines, it is necessary to determine whether the attenuated strains are able to confer immunity. Mouse protection test. Table 2 demonstrates the protection conferred by these strains in the mouse protection test. It appears that all the strains are quite immunogenic and give rise to good protection, particularly against homologous (KB92, Ogawa) challenge. Table 2. Protection conferred by attenuated strains in the mouse protection testa N.of ognss No. of deaths/no, challengedImmunizing No. org nisng after challenge with: strain zigose_dose ~~~KB92 KB93 KB9 4.0x 106 0/16 3/16 KB9:P 3.0x105 0/16 7/16 KB9:V 6.1 x105 0/16 5/16 KB9:PV 7.4X106 1/16 10/16 CD24 5.5x106 1/16 6/16 Challenge dose was 500 LD50. Ileal loop challenge experiments in adult rabbits. The degree of protection conferred by live bacterial vaccines prepared from attenuated strains KB9: PV and CD24 was investigatend in adult rabbits. The immunogenic efficacy of the two strains was evalu- ated against homologous (KB92, Ogawa) and heterologous (KB93, Inaba) challenge strains. The immunized rabbits were challenged 35 days after the second dose of vaccine. Immunity to challenge and assay of vibriocidal antibodies was detergiined. Table 3 shows that both the vaccines induced significant protection against tie two challenge strains, within the range from 73 to 100%. High vibriocidal titres were also seen towards the chal- Table 3. hleat loop chailenge experiments in adult rabbits at 35 days after the second dose of vaccine Vaccine No. of Challenge Fluid in ileal loop Immunity Vibriocidal strain rabbits strain to challenge titrebRange Mean %)(mean) KB9:PV 7 KB92 0.00-1.58 0.37 73 2.23 x 105 K693 0.00-0.00 0.00 100 8.71 x 104 CD24 - 6 KB92 0.00-1.65 0.27 80 f.00 x 106 KB93 0.00-1.33 . 0.39 76 7.78 x 104 Control 7 K892 0.00-2.13 1.39 - < 50 K893 1.23-1.88 1.6S - <50 ml/cm of loop. bTitre is the reciprocal of the highest dilution causing a positive reaction. The vibriocidal titre of the pre- immunized sera was less than 50. 645 V. B. SINHA & B. S. SRIVASTAVA Table 4. Multiplication and adherence of vibrios in the ileal loop of adult rabbits No. of vibrios No. of vibrios Total no. Fluid in Strains No. of vibrios adherin to in lumen of vibrios ileal loopinjected wall of intestine(per ml) (per ml) (per ml) (ml/cm) KB9 1.4 x 105 6.0 x 108 7.5 x 109 8.1 x 109 1.63 KB9: PV 3.6 x 104 2.3 x 100 2.1 x 109 2.33 x 109 0.00 CD24 1.2 x 105 7.5 x 108 7.5 x 109 8.3 x 109 0.00 lenge strains to which the animals were immune. Findings are suggestive of good protection associ- ated with a high titre of vibriocidal antibodies. Multiplication and adherence ofKB9:PVand CD24 The growth of the plasmid-carrying strains was compared with that of the parent KB9 and no difference was found in vivo or in vitro. Adherence of CD24 and KB9: PV to the intestine was similar to that of KB9. The data recorded in Table 4 suggest that bacterial adherence to the intestinal mucosa and multiplication in the lumen were not affected. DISCUSSION Two strains of V. cholerae harbouring P and V plasmids demonstrated loss of pathogenicity in vari- ous experimental models. In an earlier report we described the suppression of pathogenicity by these plasmids (12). The present study was more exten- sive, using a large number of recombinants harbour- ing both these plasmids, and we investigated whether infection of V. cholerae with these plasmids caused loss of pathogenicity due to loss of ability to produce biologically active toxin. It was also ob- served that these attenuated strains do have the ability to multiply and adhere in the rabbit gut, which is a vital point for a strain that may be used for the production of a live oral vaccine (7, 9). The plasmid status of these strains appears to be quite stable both in vitro and in vivo. Some of these strains have been maintained and passaged on nutrient agar for several years and loss of plasmid(s) has not been observed so far. Successive passages in rabbit intestine do not have any effect on the plasmids. It has been reported that plasmid P can be eliminated by lengthy incubation on solid medium (1) and by sodium dodecyl sulfate (10). However, none of these treatments, including exposure to acridine orange, eliminated these plasmids in our studies. Lengthy incubation in BHI had no effect, except in one instance where P appears to have been lost, resulting in the isolation of CD24. Information on bacterial mating, however, suggests that P may be present in a repressed state in this strain. These attenuated strains appear to be immunogenic as they give good protection in the mouse protection test (11) and ileal loop challenge experiments in adult rabbits. This protection was also associated with the high vibriocidal antibody in the latter model. A number of attenuated strains of cholera vibrio have previously been reported for the possible development of live oral vaccine (13). Some of these either failed to provide effective immunity (3) or reverted to toxicity (strain M13, R. A. Finkelstein, personal communication, 1976). The attenuated strains described here appear to be well suited for use in the production of live oral vaccines since they are quite stable under various conditions, able to multiply and adhere to the gut of the rabbit, provide good protection, and carry genetic markers for their identification in the environment. Such attenuated strains have the advantage of not having been exposed to mutagenic treatment. In principle, there- fore, these strains should possess all the features of the wild type strain except toxin production. This is confirmed by our data. ACKNOWLEDGEMENTS We are grateful to Dr Nitya Nand for his interest and support and to Mr A. K. Sarkar and Mr M. U. Khan for their laboratory help. This is communication No. 2436 of this Institute. 646 VIBRIO CHOLERAE STRAINS FOR ORAL IMMUNIZATION 647 RtSUMJt ATTNUATION DE LA VIRULENCE DE VIBRIO CHOLERAE PAR DES PLASMIDES P ET V: SOUCHES CONVENANT POUR L'IMMUNISATION ORALE Dans un precedent article, les auteurs avaient rapporte que les plasmides P et V supprimaient de facon significative la pathogenicite de Vibrio cholerae. Dans le present article, ils exposent une nouvelle methode permettant d'isoler des souches att6nuees de V. cholerae apres introduction de plasmides P et V. Ces plasmides ne modifient pas l'immu- nogenicite des vibrions, et sont stables aussi bien in vitro qu'in vivo. Des plasmides P et V ont ete introduits dans une souche virulente KB9 par conjugaison. Plusieurs cellules porteuses de P et V ont ete isolees. La virulence de ces souches a ete etudiee dans des modeles animaux du cholera, et il a ete demontre qu'un grand nombre d'isolats PV etaient depour- vus de pathogenicite. Par ailleurs, tandis que la souche 569B etait toxinogene in vitro, son isolat PV (CD23) ne produisait aucune quantite decelable de toxines. Deux souches porteuses de plasmides, KB9:PV et CD24, ont ete caracterisees en detail. Dans l'epreuve de protection de la souris, la DL50 6tait respectivement de 9, 2,7 x106 et 4,2 x 107 pour les souches KB9, KB9:PV et CD24. Chez 16 lapereaux nouveau-nes infectes par voie intestinale avec les souches KB9:PV et CD24, on n'a observe ni diarrhee, ni accumulation de liquide dans le caecum. L'administration des souches KB9: PV et CD24 a un lapin adulte dans une anse ileale ne provoquait aucune perte excessive de liquide. En revanche, l'administration de la souche KB9 a des lapins adultes et a des lapereaux nouveau-nes entrainait une forte diarrhee et une perte de liquide notable. La stabilite des plasmides a ete etudiee in vitro et in vivo. L'orange d'acridine et le dodecylsulfate de sodium n'ont pu les eliminer des cellules. Plusieurs passages sur bouillon et une incubation prolongee a 37°C etaient egalement sans effet. Apres deux passages successifs dans une anse ileale de lapin adulte, on a recherche les plasmides dans environ 100 colonies de vibrions: aucun plasmide P et V n'avait disparu. La protection conferee par les souches KB9: PV et CD24 a ete determinee par l'epreuve de protection de la souris et dans le modele constitue par l'anse ileale du lapin. Dans la premiere epreuve, on a observe une bonne protection, notamment contre la reinfection par une souche homologue. Dans le modele de l'anse ileale du lapin, des lapins ont ete immunises par des souches KB9:PV et CD24 vivantes, et eprouves 35 jours plus tard. Le degre de protection se situait entre 73 et 100 %, avec un titre eleve d'anticorps vibriocides. Ces deux souches sont donc immu- nogenes. Les souches porteuses de plasmides etant stables et immunogenes, elles peuvent etre utilisees pour la prepa- ration d'un vaccin oral vivant, et pourraient convenir pour l'immunisation orale. REFERENCES 1. BHASKARAN, K. J. Journal of general microbiology, 23: 47-54 (1960). 2. BHASKARAN, K. & SINHA, V. B. J. Journal of general microbiology, 40: 281-284 (1971). 3. CASH, R. A. ET AL. Infection and immunity, 10: 762- 764 (1974). 4. DE, S. N. & CHATTERJEE, D. N. Journal ofpathology and bacteriology, 66: 559-562 (1953). 5. DurrA, N. K. & HABBU, M. K. British journal of pharmacology, 10: 153-159 (1955). 6. FINKELSTEIN, R. A. Journal of immunology, 59: 264- 271 (1962). 7. FINKELSTEIN, R. A. Current topics in microbiology and immunology, 69: 137-195 (1975). 8. FINKELSTEIN, R. A. & LOSPALLUTFO, J. J. Journal of experimental medicine, 130: 185-202 (1969). 9. JoO, I. Cholera vaccines. In: Barua, D. & Burrows, W. ed. Cholera, Philadelphia, W. B. Saunders Co., 1974, p. 333. 10. PARKER, C. ET AL. Recombination in Vibrio cholerae. In: Fukumi, H. & Ohashi, M., ed. Proceedings of a Symposium on Cholera, Woods Hole, Massachusetts, 9-11 August 1971. 11. PrTMAN, M. & FEELEY, J. C. Laboratory assay of cholera vaccine potency. In: Proceedings of the Chol- era Research Symposium, 24-29 January 1965, Hon- olulu, Hawaii, Washington, DC, US Government Printing Office, 1965, p. 163 (PHS Publication No. 1328). 12. SINHA, V. B. & SRIVASTAVA, B. S. Journal of general microbiology, 104: 251-255 (1978). 13. WHO Technical Report Series, No. 500, 1972 (Oral enteric bacterial vaccines. Report of a WHO Scientific Group).