In vivo evidence of immunological masking of the Vibrio cholerae 0 antigen of a hybrid Salmonella typhi Ty21a-Vibrio cholerae oral vaccine in humans Bruce D. Forrest* and Justin T. LaBrooy The immunogenicity o f the live oral hybrid vaccine organism Salmonella typhi Ty21a/ V. cholerae Inaba (EX210) following its growth in media containing variable concentrations o f supplemental galactose was examined in human volunteer subjects. The local intestinal IgA-specific antibody responses to both typhoid and cholera lipopolysaccharide (LPS) preparations were determined. It was observed that the immunogenicity o f the galactose-independent Vibrio cholerae O antigen in vivo was dependent upon the variation in galactose-dependent long chain S. typhi O antigen production which was directly proportional to the media galactose concentration. It is likely that this observation was a result o f steric hindrance o f the presentation o f the V. cholerae O antigen by S. typhi Ty21a in the presence o f the longer, immunodominant S. typhi Ty21a 0 antigen. This observation may have relevance to the use o f S. typhi vectors in vaccine development involving the presentation o f LPS-associated heterologous antigens. Keywords:Typhoidvaccine;lipopolysaccharide;IgA; 0 antigen
INTRODUCTION
Salmonella typhi Ty21a, a safe, effective live, orally administrable typhoid vaccine ~-5, has been proposed as one possible vector for the carriage of foreign putative protective antigens 6-14. One of these hybrid strains, designated EX645, producing the O antigen of a Vibrio cholerae Inaba strain, was shown recently to provide a low level of protection against clinical cholera in human volunteers challenged with V. cholerae ~4. Their immune response to the V. cholerae 0 antigen was poor, despite the stimulation of a strong local and systemic humoral anti-typhoid lipopolysaccharide (LPS) antibody response following oral vaccination with EX645. The low level of protection observed against the V. cholerae challenge was attributable to this inadequate local antibody response 14. It has been reported recently that the in vitro identification of V. cholerae 0 antigen on a similar S. typhi Ty21a/V. cholerae hybrid strain, designated EX210, was highly dependent upon the amount of O antigen synthesized by the S. typhi Ty21a vector ~s. This synthesis, unlike that of the foreign V. cholerae 0 antigen, was confirmed to be highly dependent upon the University of Adelaide Department of Medicine, Royal Adelaide Hospital, Adelaide, South Australia. *To whom correspondence should be addressed at: Communicable Disease Surveillance Centre, Public Health Laboratory Service, 61 Colindale Avenue, London NW9 5EQ, UK. (Received 29 October 1990; revised 29 January 1991; accepted 8 February 1991) 0264-410X/91/070515-06 © 1991Butterworth-HeinemannLtd
concentration of exogenous galactose supplementing the growth media, production being directly proportional to supplemental galactose concentration. Attridge et al. 15 demonstrated in vitro that the immunological evidence of V. cholerae 0 antigen surface expression decreased as the concentration of supplemental galactose increased. However, despite this apparent reduction in V. cholerae O antigen production, they demonstrated the presence of significant quantities of the V. cholerae 0 antigen by silver staining SDS-PAGE gels of LPS preparations 15. In this subsequent study we have examined the clinical significance of this reported phenomenon on the immunogenicity of EX210 in vivo in humans, following its in vitro culture in variably galactose-supplemented nutrient broth. MATERIALS AND METHODS Vaccine strain
The development of EX210 has been described previously 12'~4'15. Briefly, EX210 is based on a rifampicinresistant S. typhi Ty21a in which the chromosomal rfa region was replaced with the homologous region from Escherichia coli K12 due to the better polymerization of the V. cholerae 0 antigen onto this core oligosaccharide than onto the Salmonella one. Into this strain, a plasmid was conjugated which encoded tetracycline resistance and carriea the genes encoding the V. cholerae Inaba O antigen. The resulting strain was shown in vitro to produce V. cholerae 0 antigen when grown in the absence
Vaccine, Vol. 9, July 1991 515
Lipopolysaccharide antigenic masking: B.D. Forrest and J.T. LaBrooy Table 1
Strain preparation and level of 0 antigen expression"
Vaccination group A B C D
Supplemental galactose concentration c (%) 0.001 ~
0.0001e
0~ 0~
Level of O antigen surface expression b (%)
S. typhi
V. cholerae Inaba
100 36 1 1
2 71 71 71
aTable adapted from Attridge et al. is bLevels of surface expression of O antigen represented as the percentage inhibition of the strain compared to homologous control LPS samples in haemagglutination inhibition assay TM CAll doses were grown in nutrient broth supplemented with the above detailed concentrations of exogenous galactose and freshly harvested for administration ~ ~Vaccine doses administered in 40 ml skimmed milk ~Vaccine doses administered in 40 ml 0.9% w/v saline
of exogenous galactose, and both V. cholerae and S. typhi O antigens when grown in the presence of exogenous galactosel 5.
Vaccine preparation EX210 was grown, and the vaccine doses prepared as described previously 15. Three batches of vaccine doses were prepared by growing EX210 overnight in nutrient broth, containing one of three concentrations of supplemental galactose as detailed in Table 1. These concentrations were known to result in specific S. typhi 0 antigen: V. cholerae 0 antigen ratios 15.
Subjects Twenty two healthy adults (13 men and 9 women, 18-55 years of age) agreed to participate in this study. Written and informed consent was obtained from all volunteers prior to their entry into the study. The use of human subjects was in accordance with the ethical standards of the Human Ethics Committee of the Royal Adelaide Hospital, the Committee on the Ethics of Human Experimentation of the University of Adelaide, and with the Helsinki Declaration of 1975. None of the subjects had any previous exposure to typhoid fever or cholera, and only two subjects (members of group C) had been vaccinated against typhoid and cholera (with the parenterally administered heat-inactivated vaccines) in the past 7 years. Only one of the subjects (subject X, a member of group D) had any history or current symptoms of gastrointestinal tract disease having undergone a Bilroth type 1 gastrectomy for recurrent duodenal ulcer disease, 3 years previously. The subjects were allocated randomly for four study groups, designated A to D. The vaccine doses and regimens used in the respective groups are detailed in Table 2. The oral vaccination schedule required that freshly harvested vaccine doses were administered to the subjects after an 8 h fast and 5 min after the oral administration of 50 ml 2% sodium hydrogen carbonate solution. The vaccine dose was ingested after suspension in either 40 ml skimmed milk or 40 ml 0.9% saline and washed down with 100 ml distilled water. Skimmed milk was initially used in group A as an additional buffer against gastric acid and intestinal proteolytic enzymes; its potential
516
V a c c i n e , Vol. 9, July 1991
effects on S. typhi 0 antigen production were evaluated in group C.
Collection of samples Intestinal fluid samples were obtained from the upper jejunum as previously described 5 where possible on day 14 or 15 after the commencement of the vaccination course. Briefly, after an 8 h fast, each subject gargled a local anaesthetic solution comprising 5 ml 2% amethocaine solution (supplied by the Royal Adelaide Hospital Pharmacy). Subsequently, a polyvinyl ANPRO AN 20 Andersen tungsten-weighted intestinal sump tube (H.W. Andersen Products, Oyster Bay, New York, USA) was passed orally and positioned in the upper jejunum. Correct positioning of the tube was confirmed by fluoroscopy. Intestinal fluid samples with pH > 6.4 were collected and kept on ice until 25-30ml had been collected from each subject. The samples were then centrifuged at 40009 at 4°C and stored at -70°C until required for assaying. Antigen-specific intestinal IgA levels have been previously well documented to be the most useful differential indicator of a local immune response in the gut to an orally administered antigen5'16"1
Assays for quantifying specific antibody Class-specific anti-typhoid and anti-cholera LPS antibodies in intestinal fluid were quantified by ELISA 5. Briefly, 96-well polyvinyl microtitre ELISA plates (Costar cat. no. 2595, Data Packaging Corp., Cambridge, MA, USA) were coated with either S. typhi Ty2 LPS (Sigma Chemical Co., St Louis, MO, USA, cat. no. L 6386) or V. cholerae Inaba 569B LPS (Sigma Chemical Co., St Louis, MO, USA, cat. no. L 0385) and blocked with 0.05% (w/v) bovine serum albumin in phosphatebuffered saline. Subsequently, intestinal fluid obtained from individual vaccinated subjects was added to duplicate wells on the plates commencing with a 1:2 dilution and with subsequent twofold dilutions down the plate, and incubated at 37°C for 16 h. After washing, alkaline-phosphatase conjugated goat anti-human IgA antiserum ('KPL', Kirkegaard and Perry Laboratories, Gaithersburg, MD, USA) was added and the plates were incubated at 37°C for 4 h. After washing, 0.1 ml of a 1 mgm1-1 solution of the substrate p-nitrophenyiphosphate (Sigma Phosphatase Substrate 104-105, Sigma
Table 2
Vaccine doses and schedules Vaccination
Vaccination group
Day
Dose" ( x 1011)
A
0 2 4 0 2 4 0 2 4 0 2 4
1.2 0.9 1.1 2.1 1.3 1.2 1.1 1.2 1.1 1.6 1.4 1.2
B
C
D
aMean number of viable organisms comprising vaccine dose
Lipopolysaccharide antigenic masking: B.D. Forrest and J.T. LaBrooy Chemical Company, St Louis, MO, USA) in 10% diethanolamine buffer was added to all wells and after further incubation at 37°C for 2 h the plates were read in a Titertek ELISA reader at 405 nm. In each assay, serum obtained from convalescent typhoid and cholera patients with known high antibody titres directed against S. typhi LPS and V. cholerae Inaba were included as positive controls, and serum obtained from unexposed individuals with known low IgA antibody titres against S. typhi and V. cholerae Inaba LPS were included as negative controls. Intestinal specific IgA antibody responses were presented as the reciprocal of the final titration that gave an absorbance of 0.15 ELISA units/0.1 ml (the volume added to each well) and are expressed as units of antibody. This absorbance was chosen as it represented the upper limit of the 95% confidence intervals (CI) above background levels18. The intestinal fluid specific antibody units were adjusted for total class-specific immunoglobulin content and were expressed as units of specific antibody mg- 1 of total class-specific immunoglobulin, A single radial immunodiffusion method x9 was used to determine the total class-specific immunoglobulin content of intestinal fluid, with human colostrum of known secretory IgA content as the standard. Statistics
Comparisons of the fold rises in specific antibody titres between groups were analysed using the two-tailed
Wilcoxon Rank Sum Test (WRST) for independent samples.
RESULTS The individual anti-typhoid intestinal IgA antibody responses are presented in Figure I and summarized in Table 3. It is evident from Figure I that a decline in both the numbers of subjects achieving a fourfold rise in specific anti-typhoid antibody titre and the magnitude of the individual antibody responses was observed as the quantity of S. typhi 0 antigen was reduced through a reduction in the media supplemented galactose concentration. In those subjects ingesting vaccine organisms completely smooth for S. typhi 0 antigen (group A) a strong intestinal antibody response was stimulated with 5/6 having a fourfold or greater response. A comparable Table 3 LPS
Specific jejunallgA antibody against S. typhi and V. cholerae
Group
S. typhi LPS (%)
V. cho/erae LPS (%)
A B C D
5/6 (83) 4/6 (67) 5/10 (50) 1/5 (20)
0/6 (0) 0/6 (0) 2/10 (20) 4/5 (80)
Numbers of subjects having a fourfold or greater rise in specific IgA antibody titres (day 14/day 0) as determined by enzyme-linked immunosorbent assay (ELISA)
100
100
to >. "1o
>. I0
10
t0
+.s ttu
tfo
u
u
'G Q.
1
2
3
q
5
6
1
2
3
q
Subject
5
6
Subject
100
1 O0
== t-
t-
>~
L. >. "O
&
INTRODUCTION
Salmonella typhi Ty21a, a safe, effective live, orally administrable typhoid vaccine ~-5, has been proposed as one possible vector for the carriage of foreign putative protective antigens 6-14. One of these hybrid strains, designated EX645, producing the O antigen of a Vibrio cholerae Inaba strain, was shown recently to provide a low level of protection against clinical cholera in human volunteers challenged with V. cholerae ~4. Their immune response to the V. cholerae 0 antigen was poor, despite the stimulation of a strong local and systemic humoral anti-typhoid lipopolysaccharide (LPS) antibody response following oral vaccination with EX645. The low level of protection observed against the V. cholerae challenge was attributable to this inadequate local antibody response 14. It has been reported recently that the in vitro identification of V. cholerae 0 antigen on a similar S. typhi Ty21a/V. cholerae hybrid strain, designated EX210, was highly dependent upon the amount of O antigen synthesized by the S. typhi Ty21a vector ~s. This synthesis, unlike that of the foreign V. cholerae 0 antigen, was confirmed to be highly dependent upon the University of Adelaide Department of Medicine, Royal Adelaide Hospital, Adelaide, South Australia. *To whom correspondence should be addressed at: Communicable Disease Surveillance Centre, Public Health Laboratory Service, 61 Colindale Avenue, London NW9 5EQ, UK. (Received 29 October 1990; revised 29 January 1991; accepted 8 February 1991) 0264-410X/91/070515-06 © 1991Butterworth-HeinemannLtd
concentration of exogenous galactose supplementing the growth media, production being directly proportional to supplemental galactose concentration. Attridge et al. 15 demonstrated in vitro that the immunological evidence of V. cholerae 0 antigen surface expression decreased as the concentration of supplemental galactose increased. However, despite this apparent reduction in V. cholerae O antigen production, they demonstrated the presence of significant quantities of the V. cholerae 0 antigen by silver staining SDS-PAGE gels of LPS preparations 15. In this subsequent study we have examined the clinical significance of this reported phenomenon on the immunogenicity of EX210 in vivo in humans, following its in vitro culture in variably galactose-supplemented nutrient broth. MATERIALS AND METHODS Vaccine strain
The development of EX210 has been described previously 12'~4'15. Briefly, EX210 is based on a rifampicinresistant S. typhi Ty21a in which the chromosomal rfa region was replaced with the homologous region from Escherichia coli K12 due to the better polymerization of the V. cholerae 0 antigen onto this core oligosaccharide than onto the Salmonella one. Into this strain, a plasmid was conjugated which encoded tetracycline resistance and carriea the genes encoding the V. cholerae Inaba O antigen. The resulting strain was shown in vitro to produce V. cholerae 0 antigen when grown in the absence
Vaccine, Vol. 9, July 1991 515
Lipopolysaccharide antigenic masking: B.D. Forrest and J.T. LaBrooy Table 1
Strain preparation and level of 0 antigen expression"
Vaccination group A B C D
Supplemental galactose concentration c (%) 0.001 ~
0.0001e
0~ 0~
Level of O antigen surface expression b (%)
S. typhi
V. cholerae Inaba
100 36 1 1
2 71 71 71
aTable adapted from Attridge et al. is bLevels of surface expression of O antigen represented as the percentage inhibition of the strain compared to homologous control LPS samples in haemagglutination inhibition assay TM CAll doses were grown in nutrient broth supplemented with the above detailed concentrations of exogenous galactose and freshly harvested for administration ~ ~Vaccine doses administered in 40 ml skimmed milk ~Vaccine doses administered in 40 ml 0.9% w/v saline
of exogenous galactose, and both V. cholerae and S. typhi O antigens when grown in the presence of exogenous galactosel 5.
Vaccine preparation EX210 was grown, and the vaccine doses prepared as described previously 15. Three batches of vaccine doses were prepared by growing EX210 overnight in nutrient broth, containing one of three concentrations of supplemental galactose as detailed in Table 1. These concentrations were known to result in specific S. typhi 0 antigen: V. cholerae 0 antigen ratios 15.
Subjects Twenty two healthy adults (13 men and 9 women, 18-55 years of age) agreed to participate in this study. Written and informed consent was obtained from all volunteers prior to their entry into the study. The use of human subjects was in accordance with the ethical standards of the Human Ethics Committee of the Royal Adelaide Hospital, the Committee on the Ethics of Human Experimentation of the University of Adelaide, and with the Helsinki Declaration of 1975. None of the subjects had any previous exposure to typhoid fever or cholera, and only two subjects (members of group C) had been vaccinated against typhoid and cholera (with the parenterally administered heat-inactivated vaccines) in the past 7 years. Only one of the subjects (subject X, a member of group D) had any history or current symptoms of gastrointestinal tract disease having undergone a Bilroth type 1 gastrectomy for recurrent duodenal ulcer disease, 3 years previously. The subjects were allocated randomly for four study groups, designated A to D. The vaccine doses and regimens used in the respective groups are detailed in Table 2. The oral vaccination schedule required that freshly harvested vaccine doses were administered to the subjects after an 8 h fast and 5 min after the oral administration of 50 ml 2% sodium hydrogen carbonate solution. The vaccine dose was ingested after suspension in either 40 ml skimmed milk or 40 ml 0.9% saline and washed down with 100 ml distilled water. Skimmed milk was initially used in group A as an additional buffer against gastric acid and intestinal proteolytic enzymes; its potential
516
V a c c i n e , Vol. 9, July 1991
effects on S. typhi 0 antigen production were evaluated in group C.
Collection of samples Intestinal fluid samples were obtained from the upper jejunum as previously described 5 where possible on day 14 or 15 after the commencement of the vaccination course. Briefly, after an 8 h fast, each subject gargled a local anaesthetic solution comprising 5 ml 2% amethocaine solution (supplied by the Royal Adelaide Hospital Pharmacy). Subsequently, a polyvinyl ANPRO AN 20 Andersen tungsten-weighted intestinal sump tube (H.W. Andersen Products, Oyster Bay, New York, USA) was passed orally and positioned in the upper jejunum. Correct positioning of the tube was confirmed by fluoroscopy. Intestinal fluid samples with pH > 6.4 were collected and kept on ice until 25-30ml had been collected from each subject. The samples were then centrifuged at 40009 at 4°C and stored at -70°C until required for assaying. Antigen-specific intestinal IgA levels have been previously well documented to be the most useful differential indicator of a local immune response in the gut to an orally administered antigen5'16"1
Assays for quantifying specific antibody Class-specific anti-typhoid and anti-cholera LPS antibodies in intestinal fluid were quantified by ELISA 5. Briefly, 96-well polyvinyl microtitre ELISA plates (Costar cat. no. 2595, Data Packaging Corp., Cambridge, MA, USA) were coated with either S. typhi Ty2 LPS (Sigma Chemical Co., St Louis, MO, USA, cat. no. L 6386) or V. cholerae Inaba 569B LPS (Sigma Chemical Co., St Louis, MO, USA, cat. no. L 0385) and blocked with 0.05% (w/v) bovine serum albumin in phosphatebuffered saline. Subsequently, intestinal fluid obtained from individual vaccinated subjects was added to duplicate wells on the plates commencing with a 1:2 dilution and with subsequent twofold dilutions down the plate, and incubated at 37°C for 16 h. After washing, alkaline-phosphatase conjugated goat anti-human IgA antiserum ('KPL', Kirkegaard and Perry Laboratories, Gaithersburg, MD, USA) was added and the plates were incubated at 37°C for 4 h. After washing, 0.1 ml of a 1 mgm1-1 solution of the substrate p-nitrophenyiphosphate (Sigma Phosphatase Substrate 104-105, Sigma
Table 2
Vaccine doses and schedules Vaccination
Vaccination group
Day
Dose" ( x 1011)
A
0 2 4 0 2 4 0 2 4 0 2 4
1.2 0.9 1.1 2.1 1.3 1.2 1.1 1.2 1.1 1.6 1.4 1.2
B
C
D
aMean number of viable organisms comprising vaccine dose
Lipopolysaccharide antigenic masking: B.D. Forrest and J.T. LaBrooy Chemical Company, St Louis, MO, USA) in 10% diethanolamine buffer was added to all wells and after further incubation at 37°C for 2 h the plates were read in a Titertek ELISA reader at 405 nm. In each assay, serum obtained from convalescent typhoid and cholera patients with known high antibody titres directed against S. typhi LPS and V. cholerae Inaba were included as positive controls, and serum obtained from unexposed individuals with known low IgA antibody titres against S. typhi and V. cholerae Inaba LPS were included as negative controls. Intestinal specific IgA antibody responses were presented as the reciprocal of the final titration that gave an absorbance of 0.15 ELISA units/0.1 ml (the volume added to each well) and are expressed as units of antibody. This absorbance was chosen as it represented the upper limit of the 95% confidence intervals (CI) above background levels18. The intestinal fluid specific antibody units were adjusted for total class-specific immunoglobulin content and were expressed as units of specific antibody mg- 1 of total class-specific immunoglobulin, A single radial immunodiffusion method x9 was used to determine the total class-specific immunoglobulin content of intestinal fluid, with human colostrum of known secretory IgA content as the standard. Statistics
Comparisons of the fold rises in specific antibody titres between groups were analysed using the two-tailed
Wilcoxon Rank Sum Test (WRST) for independent samples.
RESULTS The individual anti-typhoid intestinal IgA antibody responses are presented in Figure I and summarized in Table 3. It is evident from Figure I that a decline in both the numbers of subjects achieving a fourfold rise in specific anti-typhoid antibody titre and the magnitude of the individual antibody responses was observed as the quantity of S. typhi 0 antigen was reduced through a reduction in the media supplemented galactose concentration. In those subjects ingesting vaccine organisms completely smooth for S. typhi 0 antigen (group A) a strong intestinal antibody response was stimulated with 5/6 having a fourfold or greater response. A comparable Table 3 LPS
Specific jejunallgA antibody against S. typhi and V. cholerae
Group
S. typhi LPS (%)
V. cho/erae LPS (%)
A B C D
5/6 (83) 4/6 (67) 5/10 (50) 1/5 (20)
0/6 (0) 0/6 (0) 2/10 (20) 4/5 (80)
Numbers of subjects having a fourfold or greater rise in specific IgA antibody titres (day 14/day 0) as determined by enzyme-linked immunosorbent assay (ELISA)
100
100
to >. "1o
>. I0
10
t0
+.s ttu
tfo
u
u
'G Q.
1
2
3
q
5
6
1
2
3
q
Subject
5
6
Subject
100
1 O0
== t-
t-
>~
L. >. "O
&
