Attenuated recombinant strains of Vibrio cholerae for oral immunization*

Bulletin of the World Health Organizaton, 57 (4): 649-654 (1979) Attenuated recombinant strains of Vibrio cholerae for oral immunization* BRAHM S. SRIVASTAVA,1 V. B. SINHA,2 & RANJANA SRIVASTAVA3 Two attenuated strains of Vibrio cholerae, CDI and CD3, have been isolated that have remained stable since 1976. These strains are motile, adhere to and multiply in rabbit intestine, and colonize the gut of infant mice for 6-7 days. Both strains are antigenic and provide protection to challenge in the mouse protection test and in the rabbit ileal loop model. Because of their ability to adhere to and colonize the gut, and since they are antigenic, strains CD1 and CD3 have the potentiality of oral vaccines. Success of the oral vaccine against Shigella (15), and the limited protection offered by parenteral administration of cholera vaccine (6, 16), prompted a search for an oral vaccine against cholera. Indeed, various attempts resulted in isolation of attenuated strains (25); some of these strains have been tested in man but the results were not satisfactory (5). Logically, an ideal oral vaccine strain of Vibrio cholerae should have all the qualities of a virulent strain except for the production of toxin (22). Besides being attenuated, genetically stable, and antigenic, the vaccine strain must be highly motile, adhere to the intestine, and be able to multiply in and colonize the intestine (3, 12, 14). Therefore, a search was made for attenuated strains with these characteristics. In this report we describe the iso- lation and properties of the strains of V. cholerae that offer promise of a live oral vaccine. A brief report on the isolation and initial charac- terization of these strains was presented earlier (22). MATERIALS AND METHODS Bacteria Strains of V. cholerae with their relevant genetic markers are listed in Table 1. * From the Cholera Immunology Laboratory, Central Drug Research Institute, Lucknow-226001, India. I Scientist. 2 Scientist. I Senior Research Fellow. Table 1. Strains of V. cholerae Strains Relevant genetic markers Source/ KB207 wild-type e/ tor, Ogawa K. Bhaskaran KB92 wild-type, classical, Ogawa K. Bhaskaran KB93 wild-type, classical, Inaba K. Bhaskaran KB14 classical, Inaba, P+, pur, lu str' (1) KB1 1 classical, Ogawa, P+, ilv, arg, his, str' (1) KB599 classical, Ogawa, erg, met, str', attenuated (19) CD1 classical, Inaba, pur, str', attenuated this work CD3 classical, Ogawa, ilv, arg, strr, attenuated this work Media and buffer Bacteria were grown overnight on nutrient agar slopes. The medium contained, per litre, 8 g of Difco nutrient broth, 5 g of NaCl, and 10 g of Oxoid agar powder. Viable counts were made on nutrient agar plates. Exponentially-grown broth cultures were prepared in Difco brain heart infusion (BHI; 3.7 g in 100 ml). The composition of the minimal glucose-salt medium has been reported previously (2) and amino acids were added according to the design of particular experiments. All dilutions were made in phosphate-buffered saline (PBS ;20). Bacterial mating Two crosses were performed: (a) KB11 x KB599 and (b) KB14 x KB599. KB599 was the recipient in both crosses. From an overnight culture on nutrient agar slopes, a small inoculum was taken and diluted 3838 649 B. S. SRIVASTAVA ET AL. in Difco BHI broth and gently shaken to produce a suspension of up to 109 cells/ml. A mixture of 0.1 ml of the suspension of the donor strain and 1 ml of suspension of the recipient strain was placed in a prewarmed (37°C) 25-ml flask and very gently shaken for 2 h at 37 'C. Samples were withdrawn from this mating mixture and after rough agitation bacteria were plated on selective media. Serology The antisera against V. cholerae Ogawa and Inaba were prepared in our laboratory by immunizing rabbits. Slide agglutination tests were performed for serotyping. Virulence in animals For the ileal loop of adult rabbit test (7), 1 ml of culture containing 107 vibrios was injected in each loop, and an examination was made after 18 h. Albino rabbits weighing 1.5 kg and reared in this Institute were used. They were starved for 48 h prior to challenge. For the infant rabbit test (8), 10-day-old albino rabbits, weighing between 100 and 120 g, were infected intraintestinally with 107 organisms. Ani- mals were examined during 48 h for diarrhoea and, death, and those surviving were opened to examine the gut. For the infant mouse test (23), 5-day-old Swiss mice were infected with 0.1 ml from a cell suspen- sion of vibrios (108/ml). Care was taken to introduce the inoculum into the stomach. Deaths occurring be- tween 5 and 24 h were scored. The intestines of the dead and surviving mice were examined. For the mouse virulence test, Swiss mice, weighing 14-16 g and reared in this Institute, were inoculated intraperitoneally with several dilutions of mucinized vibrios. The procedure described elsewhere was followed exactly (20). Adherence to and multiplication in rabbit intestine Tests for adherence of vibrios to isolated rabbit intestinal discs were done as published earlier (3). In the ileal loop model, vibrios were injected and after 18 h loops were opened, washed, and homogenized and counts were made of the viable bacteria adher- ing firmly to the intestine. Total counts of the vibrios, that is, those adhering to the intestine plus those in the lumen, were also made to determine the extent of multiplication. Survival of vibrios in the intestine of infant mice Five-day old infant mice were divided into groups and each group was fed orally with one strain (106-1o7 vibrios in 0.1 ml). Mice were put back in groups to their respective mothers. Every 24 h the intestines of 4 mice were examined, then removed and homogenized in PBS. Viable counts of vibrios were made on nutrient agar plates containing 500 Fg of streptomycin per ml. All the challenge strains were streptomycin-resistant; other intestinal organ- isms did not grow on these plates. Motility and haemagglutination Recombinant strains were examined for motility and ability to agglutinate human red blood cells. The method described in our earlier paper was followed (4). Mouse protection test This test was performed according to the method of Pittman & Feeley (18). Swiss mice weighing 12-14 g were immunized by injecting intraperiton- eally 1-4 x 101 organisms of CD1 and CD3, and they were challenged by injecting intraperitoneally on the 15th day 500 LD50 of mucinized KB92 and KB93. Death was recorded up to 72 h. Immunization and evaluation of immunity in rabbits Immunity to challenge was evaluated in the rabbit ileal loop model (7). Strains CD1 and CD3 were grown in BHI up to 109 cells per ml. Equal volumes of the two cultures were mixed and treated as live vaccine. A volume of 1 ml of the vaccine was administered twice subcutaneously, the interval be- tween two doses being 21 days. For comparison, rabbits were immunized with cholera vaccine pre- pared by the Central Research Institute, Kasauli (batch No. 076049F-5); two injections each of 1 ml were given. Immunized rabbits were challenged with 107 organisms of KB92 and KB93. Hence, in each rabbit two loops were made. Equal numbers of control animals were also challenged. Results were expressed as volume of fluid per cm of loop. Serum antibody titration Rabbits were bled by cardiac puncture one day before immunization as well as at the time of challenge. Sera were separated by centrifugation and were inactived at 56°C for 30 min. Vibriocidal antibodies were estimated according to the method of Finkelstein (9). 650 STRAINS OF V. CHOLERAE FOR ORAL IMMUNIZATION Table 2. Comparison of pathogenicity of the attenuated and virulent strains in experimental animals Adult rabbit Infant rabbit Infant Swiss mice Swiss mice Strains No. of No. of No. of Amount No. of No. of Condition No. of No. of Condition rabbits 1°°Ps loops (ml/cm) animals deaths of gut animals deaths of gut LD50 CD1 9 9 nil nil 6 nil normal 20 nil normal 8 x 105 CD3 9 9 nil nil 9 nil normal 20 nil normal 7 x 105 KB207 8 8 8 2-3 2 2 fluid 20 19 fluid ND KB92 9 9 9 1.5-2.5 2 2 fluid ND ND ND <13 ND = not done. RESULTS Isolation of CDI and CD3 Bhaskaran (1) mapped the sequence of certain loci on the chromosome of V. cholerae; and the gene tox, which controls enterotoxin production, has been found linked to the his locus (24). Thus, in the cross between KB14 and KB599, str' pur 1eu+ recombi- nants were selected. In the other cross between KBll and KB599, strr ilv arg his+ recombinants were selected. There were 16 recombinants of the former type and 30 of the latter. After a ri,,ourous test of pathogenicity of these recombinants, two representing each type were selected and designated as CD1 and CD3. The genetic markers of CD1 and CD3 are strr, pur, 0-1, and strr, ilv, arg, 0-1, respectively. The studies reported below have been carried out with CD1 and CD3. Attenuation of CDI and CD3 in experimental animals Attenuation was tested in 4 different experimental models of cholera and the results have been summarized in Table 2. Infection with CD1 and CD3 in the ileal loop of the adult rabbit and in the intestine of the infant rabbit, and oral feeding of infant mice, caused no symptoms of cholera, where- as the animals infected with pathogenic strains succumbed to cholera. The LD50 in the mouse virulence test was less than 13 for KB92 and about 106 for CD1 and CD3. Survival of CDI and CD3 in the gut The survival and persistence of CD1 and CD3 have been studied in the gut of infant mice. Table 3 records the viable counts recovered from the gut after up to six days. It shows that CD1 and CD3 are capable of multiplying and persisting for a few days before vanishing completely from the gut. We have Table 3. Survival of attenuated strains in the gut of infant mice No. of Viable counts days CD1 CD3 0 7.5x106 1.6x106 1 2.3 x 107 7.3 x 106 3 2.1 X 107 7.0 x 106 4 6.3 x 106 3.6 x 106 5 2.5 x10 8.2 x 105 6 5.0X104 8.0x104 characterized a few strains that disappeared from the gut of infant mice after 48 h of oral challenge (unpublished). Other properties of CDI and CD3 The two strains were examined microscopically and were found to be motile. They adhered like KB92 and KB207 to freshly isolated intestinal discs in vitro (3) and also in vivo in ileal loops of adult rabbits. As many as 90% of the total vibrios were found firmly adhering to the intestine. These strains multiplied in vitro as fast as KB92 and KB207 (3). Similar results were obtained in the ileal loop model, where the number of viable counts rose to 109 in 10 h, compared with 105 in the inoculum. Results on adherence and multiplication of these strains in the ileal loop model have been submitted elsewhere for publication in detail. CD1 and CD3 agglutinate human red blood cells. CDJ and CD3 in the mouse protection test The results obtained in the mouse protection test are given in Table 4. Controls with PBS and KB92 (an immunogenic strain from our laboratory) have 651 B. S. SRIVASTAVA ET AL. also been included. It shows that CD1 and CD3 are both immunogenic and give rise to good protection, particularly against identical serotypes. Table 4. Protection protection testa by CD1 and CD3 in the mouse No. of deaths after challenge with Vaccine KB92 KB93 CD1 0 1 CD3 0 3 KB92 1 2 PBS 14 16 Each group consisted of 16 mice. Protection conferred by CDI and CD3 in the rabbit Protection to challenge against pathogenic strains KB92 and KB93 in rabbits immunized subcutane- ously with live bivalent vaccine (CD1 and CD3) was evaluated and data are given in Table 5. Immunized rabbits were challenged at regular intervals after the second dose of vaccine. Protection of 80-90% was obtained up to 60 days after the second dose of vaccine, and the serum vibriocidal antibody titre was high. Challenge data on the 60th day have been included in Table 5. Immunized rabbits were then challenged on the 98th day after the second dose of vaccine. Protection was less and varied between 65 and 70% without any significant fall in the level of serum vibriocidal titre. Control rabbits showed no protection and the vibriocidal titre was below 50. A group of rabbits were also immunized with cholera vaccine prepared by the Central Research Institute, Kasauli. Protec- tion up to 60 days was as high as with CD1 and CD3 but on the 98th day protection declined significantly to about 30% and the vibriocidal titre was an order of magnitude lower (Table 5). Table 5. Immunity conferred by CD1 and CD3 in rabbit ileal loop challenge experiments Amountof fluid Immunity Mean serum Vaccine Challenge No. of Challenge in ileal loop (ml/cm) to challenge vibriocidalday animals strain tirRange Mean (%tir CD1 and 59 3 KB92 0.00-0.94 0.31 83 105 CD3 KB93 0.00-0.86 0.29 83 105 98 9 KB92 0.00-1.65 0.48 68 5.5 x 105 KB93 0.00-1.88 0.51 67 9 X 104 Control 59 4 KB92 1.43-2.00 1.77 - < 50 K893 1.08-2.07 1.7 - < 50 98 8 KB92 0.33-2.38 1.52 - < 50 KB93 0.00-2.63 1.53 - < 50 Kasauli 98 5 KB92 0.25-1.67 1.1 28 2.6 x 104 vaccine KB93 0.00-2.05 1.03 33 8.2 x 103 DISCUSSION We have described the isolation of two strains, CD1 and CD3, and have presented data relating to their trial as a live oral vaccine against cholera. The strains were derived by conjugation between two strains of V. cholerae. The male donor strain was pathogenic in experimental models of cholera and the recipient was an attenuated, slow growing strain of slightly low adherence. By appropriate selection after mating, attenuated recombinant strains were isolated that grew as fast as the donor parent and adhered well to intestinal mucosa. The recombinant strains inherited the attenuated feature of the re- cipient. Since the location of the tox gene on the chromosome of V. cholerae is near the his locus (24), selection of leu+ or his' recombinants permitted the isolation of a few recombinants that behaved in an attenuated fashion like KB599. These strains have remained genetically stable since the first announcement of their isolation in 1976 (22). They bear auxotrophic and streptomycin- resistant markers and have been shown to be attenu- ated in experimental animals. They protected mice 652 STRAINS OF V. CHOLERAE FOR ORAL IMMUNIZATION in Pittman & Feeley mouse protection tests (18) and provided an immunogenic stimulus in rabbits after subcutaneous administration. It remains to be seen, however, whether oral feeding would provoke an immunogenic stimulus. CD1 and CD3 are motile, adhere to rabbit intestine, and are capable of multiplying and sur- viving in the gut of orally fed infant mice for a week. The vibrios disappeared from the intestine completely after a week. Thus CD1 and CD3 colonized the intestine temporarily. To avoid a hazard of any kind, a live vaccine strain that colonizes the intestine for a short period but long enough to induce immunity, should be preferred to a strain that gives permanent colonization. As in induced cholera, or with attenuated poliovirus vaccines, a short period of colonization may be sufficient to induce immunity. From our findings, it is clear that CD1 and CD3 meet the requirements expected of a live oral vaccine (14). Recent findings on the pathogenesis of cholera in experimental models suggest clearly that motility, adherence to intestinal mucosa, and multiplication and colonization in the intestine are some essential factors contributing to virulence of vibrios (3, 10, 11, 12, 17). Therefore, an ideal strain for an oral cholera vaccine must undergo all the pathogenic events and possess all the features of a virulent strain except that of releasing the toxin (22). If this rationale is accepted, CD1 and CD3 appear to be ideal oral vaccine strains. In this light it would be interesting to examine those attenuated strains that have been tried in the past in man. One of those strains, C14-S5, isolated in this laboratory several years ago, was feebly motile, adhered poorly to the intestinal mucosa, and was slow-growing (3). Its lack of the desirable characteristics might be associated with its failure in human trials (5). We have found complete loss of pathogenicity of V. cholerae if P and V plasmids were introduced in a pathogenic strain, and this effect was due to the inability of vabrios to produce toxin (20, 21). CD1 and CD3 could be made even safer by introducing both plasmids so that any possible reversion to toxicity would be suppressed by the plasmids. Ju- dicious administration of CD1 and CD3 and CD24 (21) together with an immunogenic non-toxic subunit of cholera toxin (13) might offer durable immunity to cholera in man. ACKNOWLEDGEMENTS We are grateful to Dr Nitya Nand, Director of the Central Drug Research Institute, for encouragement and facilities. Ranjana Srivastava is a Senior Research Fellow of the Indian Council of Medical Research. We thank Mr A. K. Sarkar and Mr M. U. Khan for excellent technical and laboratory assistance. R]ESUMJt SOUCHES RECOMBINANTES A1TTNUEES DE VIBRIO CHOLERAE POUR L'IMMUNISATION ORALE CONTRE LE CHOLERA Pour l'immunisation orale, la souche ideale de Vibrio cholerae devrait logiquement avoir toutes les proprietes d'une souche virulente- sauf celle d'elaborer une toxine. Cette souche - attenuee, genetiquement stable et anti- genique - devrait donc posseder une grande motilite, la capacite d'adherer a la muqueuse intestinale ainsi que de se multiplier et de former des colonies dans l'intestin. L'article decrit la methode d'isolement et les caracteristiques des deux souches de V. cholerae, CD1 et CD3, qui semblent pr6senter toutes ces caracteristiques. Les souches CD1 et CD3 sont issues d'un croisement KB14 X KB599 et KB11 x KB599 respectivement. KB14 et KB11 ont servi de donneurs virulents, KB599 etant une souche attenuee a croissance lente dont la motilite, comme la capacite d'adherence et de multiplication, est faible. Parmi les recombinants, et apres un test rigoureux de pathogenicite, CD1 et CD3 ont ete etudiees en detail en tant que souches attenuees. Differents modeles de cholera experimental ont ete utilises a cette fin. Dans l'ileon de lapin adulte et chez le lapereau, CD1 et CD3 n'ont pas 653 654 B. S. SRIVASTAVA ET AL. provoque I'apparition des symptomes du cholera alors que ces animaux infect6s par des souches pathogenes ont succomb6 a la maladie. Dans une epreuve de virulence chez la souris, DL50 6tait <13 organismes pour une souche virulente et d'environ 106 pour CD1 et CD3. La survie de CD1 et CD3 et leur etablissement en colonies dans l'intestin ont ete observ6s chez plusieurs groupes de souriceaux de 5 jours a qui CD1 et CD3 ont ete administr6es oralement et qui sont demeures aupres de leurs meres respectives pendant plusieurs jours. Un comp- tage des vibrions viables contenus dans l'intestin a ete oper6 quotidiennement. CD1 et CD3 ont surv6cu et se sont multipli6es pendant 6 a 7 jours. Elles ont manifeste la meme motilite que des souches virulentes etudiees prece- demment, possedaient la meme capacite d'adh6rer a l'intestin et s'y sont multipli6es de la meme maniere. La protection conf6ree par CD1 et CD3 a ete 6tudiee au moyen d'une 6preuve chez la souris, ainsi que sur un modele constitue par l'ileon de lapin. La premiere a montre que CD 1 et CD3 sont l'une et I'autre immunogenes et determinent I'apparition d'une protection - notamment contre les souches de memes serotypes. Les lapins consti- tuant le modele ont fait l'objet de deux inoculations sous- cutan6es d'un vaccin bivalent CD1-CD3 a 21 jours d'inter- valle et ont e ensuite 6prouves a intervalles r6guliers. Jusqu'au 60C jour, le degr6 de protection a vari6 entre 80 et 90 %, mais il a ensuite diminue et variait entre 65 et 70% le 98e jour. Le titre d'anticorps vibriocides dans le serum 6tait elev6. Ces r6sultats permettent de conclure que CD1 et CD3 possedent les caracteristiques necessaires pour qu'on puisse envisager leur utilisation aux fins d'immunisa- tion orale. REFERENCES 1. BHASKARAN, K. Bulletin of the World Health Organi- zation, 30: 845-853 (1964). 2. BHASKARAN, K. & SINHA, V. B. Journal of general microbiology, 69: 89-97 (1971). 3. BHATTACHARJEE, J. W. & SRIVASTAVA, B. S. Bulletin of the World Health Organization, 57: 123-128 (1979). 4. BHATTACHARJEE, J. W. & SRIVASTAVA, B. S. Journal of general microbiology, 107: 407-410 (1978). 5. CASH, R. A. ET AL. Infection and immunity, 19: 763- 764 (1974). 6. DAs GUPTA, A. ET AL. Bulletin of the World Health Organization, 37: 371-385 (1967). 7. DE, S. N. & CHATTER1EE, D. N. Journal ofpathology and bacteriology, 66: 559-567 (1953). 8. DuTTA, N. K. & HABBU, M. K. British journal of pharmacology, 10: 153-159 (1953). 9. FwNKELSTEIN, R. A. Journal of immunology, 89: 264- 271 (1962). 10. FRETER, R. Texas reports on biology and medicine, 27 (Suppl. 1): 299-316 (1969). 11. GUENTZEL, M. N. & BERRY, L. J. Infection and. immunity, 11: 890-897 (1975). 12. HOLMGREN, J. & SVENNER HOLM, A. M. Journal of infectious diseases, 136 (SuppI.): S105-S112 (1977). 13. HOLMGREN, J. ET AL. Nature (London), 269: 602-604 (1977). 14. JOO, I. Cholera vaccines. In: Barua, D. & Burrows, W. ed. Cholera, Philadelphia, Saunders, 1974, p. 333. 15. MEL, D. M. ET AL. Bulletin of the World Health Organization, 32: 647-655 (1965). 16. MOSELEY, W. H. ET AL. Bulletin of the World Health Organization, 40: 177-185 (1969). 17. NELSON, E. T. ET AL. Infection and immunity, 14: 527- 547 (1976). 18. PrrrMAN, M. & FEELEY, J. C. Laboratory assay of cholera vaccine potency. In: Proceedings of the Chol- era Research Symposium, 24-29 January, 1965, Honolulu, Hawaii, Washington, DC, US Government Printing Office, 1965, p. 163 (PHS Publication No. 1328). 19. SINHA, V. B. & BHATTACHARJEE, J. W. Bulletin of the World Health Organization, 53: 482-484 (1976). 20. SrNHA, V. B. & SRIVASTAVA, B. S. Journal of general microbiology, 105: 251-255 (1978). 21. SINHA, V. B. & SRIVASTAVA, B. S. Bulletin of the World Health Organization, 57: 643-647 (1979). 22. SRIVASTAVA, B. S. In: Abstracts ofpapers presented at the Symposium on Drug Design, Lucknow, India, 1976 Lucknow, Central Drug Research Institute. 23. UJIIYE, A. & KOBARI, K. J. Journal of infectious diseases, 121 (Suppi.): S50-S55 (1970). 24. VASIL, M. L. ETAL. Science, 187: 849-850 (1975). 25. WHO Technical Report Series, No. 500, 1972 (Oral enteric bacterial vaccines. Report of a WHO Scientific Group).