Download phoPlphoQ-Deleted Salmonella typhi (Ty800) Is a Safe and

1408
phoPlphoQ-Deleted Salmonella typhi (Ty800) Is a Safe and Immunogenic SingleDose Typhoid Fever Vaccine in Volunteers
Elizabeth L. Hohmann, Carmen A. Oletta,
Kevin P. Killeen, and Samuel I. Miller*
Infectious Disease Unit, Massachusetts General Hospital, Boston. and
Virus Research Institute, Cambridge, Massachusetts
The phoPlphoQ virulenceregulatory genesof Salmonella typhi Ty2 were deleted, and the resultant
strain (Ty800) was tested as a live attenuated typhoid fever vaccine in human volunteers. Groups
of 2 or 3 subjects receivedsingle oral doses of 107, 108, 109, or 1010 cfu. Two volunteers who received
the largest dose had self-limited side effects; no bacteremias were detected. Ten of 11 subjects
had evidence of intestinal immune responses to the vaccine as measured by increases in S. typhi
lipopolysaccharide-specific IgA-secreting cells in peripheral blood samples. Humoral immune responses were measured and compared with those of control vaccinees who received 4 oral doses of
S. typhi Ty21a. In the most sensitive assays, 9 of 11 volunteers and 5 of 8 Ty21a control vaccinees
had evidence of IgG directed against S. typhi antigens. Ty800 is safe, and single oral doses are
highly immunogenic in humans.
Typhoid fever remains a worldwide health problem [1]. Parenteral vaccines are available; Ty21a, a live attenuated oral
typhoid fever vaccine (approved by the Food and Drug Administration [FDA]), confers ~60%-70% protective efficacy in
areas in which the disease is endemic (reviewed in [2]). At
least 3 doses of Ty21 a are required to achieve such rates of
efficacy [2], and protection may be lower in naive populations
[3, 4]. Development of a live attenuated, single-dose, oral typhoid fever vaccine could significantly improve public health.
Such a strain could be engineered as a multivalent vector for
oral delivery of heterologous antigens.
New candidate Salmonella typhi vaccines have been attenuated by the introduction of auxotrophic mutations [5- 7]. Purine
auxotrophs have proved poorly immunogenic [6], and aromatic
amino acid auxotrophs have caused bacteremias in some studies
[7, 8]. The phoPlphoQ virulence regulon of Salmonella typhimurium [9, 10] is a bacterial two-component regulatory system
consisting of a membrane-associated sensor kinase (PhoQ) and
a cytoplasmic transcriptional regulator (PhoP). These regulators control the transcription of multiple unlinked phoP-acti-
Received 20 November 1995; revised 2 February 1996.
Presented in part: 31st Joint Conference: United States-Japan Cooperative
Medical Science Program on Cholera and Related Diarrheal Diseases, Kiawah
Island, North Carolina, December 1995.
The studies were reviewed and approved by the Subcommittee on Human
Studies, Massachusetts General Hospital, and adhere to the committee's guidelines. Volunteers gave informed consent before study participation.
Financial support: National Institutes of Health (AI-30497 to S.I.M.; DK01410 to E.L.H.; RR-01066 to Mallinckrodt General Clinical Research Center);
Virus Research Institute.
Reprints or correspondence: Dr. Elizabeth L. Hohmann, Infectious Disease
Unit, Gray 5, Massachusetts General Hospital, Boston, MA 02114.
* Present affiliation: Departments of Medicine and Microbiology, University
of Washington, Seattle.
The Journal ofInfectious Diseases 1996;173:1408-14
© 1996 by The Universityof Chicago. All rights reserved.
0022-1899/96/7306-0014$01.00
vated and phoP-repressed genes [11-13]. Originally characterized in the S. typhimurium- BALB/c mouse model of typhoid
fever, this operon can modulate important virulence functions,
including survival within macrophages [9, 14] and resistance
to endogenous antimicrobial peptides of the defensin-cryptdin
family [14, 15]. PhoP/PhoQ null mutants are markedly attenuated in BALB/c mice and are effective vaccines in these animals [9, 16, 17]. It is logical to hypothesize that deletion of
these sensory-regulatory genes would attenuate the human
pathogen S. typhi and that such deletions could be exploited
for rational development of live attenuated vaccine strains.
Our previous work suggested that deletion of phoPlphoQ
attenuates S. typhi in humans [18]. Creation of a defined genetic
deletion of the phoPlphoQ genes in the aromatic amino acid
auxotrophic S. typhi 514Ty (Ty2 ,6.aroA hisG46) resulted in
strain Ty445. This candidate vaccine strain was nonreactogenic
in human volunteers given 2 oral doses of 5 X 1010 colonyforming units (cfu) but poorly immunogenic: Only 2 of 14
subjects who received 2 maximal doses developed significant
serologic responses to S. typhi antigens. Given that very large
doses of Ty445 were well tolerated, an S. typhi Ty2 strain
deleted only for the phoPlphoQ genes (Ty800) was constructed
in an attempt to provide a better compromise between virulence
and immunogenicity. Here we describe a dose-escalation study
of Ty800 in human volunteers.
Materials and Methods
Construction of Ty800. S. typhi Ty2 (gift of C. Hardegree,
FDA, Rockville, MD) was used as the parent strain for construction
ofTy800. A defined chromosomal deletion of956 bp in the contiguous phoP and phoQ genes was created as described [18] (nucleotides 376-1332 of the published sequence [9]). Briefly, a 2.2-kb
fragment of S. typhimurium DNA internally deleted for 956 bp
(phoPlphoQ ~956) was cloned into the suicide vector pCVD442
[19, 20] and maintained in the permissive host Escherichia coli
JID 1996; 173 (June)
Ty800 Typhoid Fever Vaccine in Volunteers
SM10 A pir. The resultant suicide vector was mobilized via conjugal transfer to the chromosome of Ty2, using plasmid-encoded
ampicillin resistance and the leucine, threonine, and thiamine auxotrophies of E. coli SM lOA pir for selection. Subsequently, selection
against the vector-encoded sacB gene on sucrose agar plates was
used as previously described [20] to identify S. typhi clones that
eliminated plasmid sequences via a second homologous recombination event, reconstituting either a wild type or deleted phoP/
phoQ locus. Clones containing the phoP/phoQ ~956 locus were
easily recognized as white colonies on indicator plates containing
the phosphatase substrate 5-bromo-4-choloro-3-indolyl phosphate
(BCIP). The nonspecific acid phosphatase of salmonellae, encoded
by the phoN gene [21], is regulated by the PhoP and PhoQ proteins
and is responsible for most of the blue coloration of S. typhi Ty2
on this indicator plate. Several white clones were characterized
by Southern blotting of chromosomal DNA to demonstrate the
anticipated chromosomal deletion in the phoP/phoQ genes, and
one was designated Ty800.
Preparation ofvaccine inocula. Stock cultures of Ty800 were
stored at -70°C in 20% glycerol and thawed for growth on Luria
broth agar plates for identification and confirmation of species (by
agglutination with antisera) and phenotype on BCIP plates before
growth of inocula. Multiple colonies (identified as correct) were
subsequently inoculated into Luria broth and grown for 16 h at
37°C on a rotary shaker. This culture was then suspended in 0.9%
saline, standardized spectrophotometrically, and diluted as needed
in saline to attain the appropriate number of viable cfu in 35 mL.
A sample of each inoculum was diluted and plated for determination of the exact cfu delivered at each log dose.
Human volunteers and study protocol. Human studies were
done as described [18]. In brief, healthy adult men and women
ages 18-49 years with no prior typhoid fever vaccination were
screened for health status with a complete medical history and
physical examination. Laboratory screening tests included complete blood counts; a chemistry panel; serologies for human immunodeficiency virus, hepatitis, and syphilis; urinalysis; chest radiography; and stool culture and ova and parasite examinations.
Volunteers were admitted to the General Clinical Research Center,
Massachusetts General Hospital, between April and September
1995. All received a single oral inoculum on the day of admission
(study day 0) and were followed in hospital for 14 subsequent
days. Volunteers were examined daily by a physician; vital signs
were taken every 6 h. Blood was drawn for serology at the screening visit and on study days 0, 7, 14, 21, and 28. Blood was drawn
routinely for peripheral blood mononuclear cell (PBMC) isolation
on study days 0 and 7; some subjects also had blood drawn on
days 4 and 10.
Twelve volunteers enrolled and received vaccine. A fourth subject in the highest dose group was excluded when Salmonella
brandenburg was isolated from 2 of 12 inpatient stool samples
shortly after receipt of the experimental vaccine, despite a stool
culture that grew only normal flora at a screening visit. This subject, who completed only the inpatient observation period and
was completely well, denied recent gastrointestinal illness but was
presumably asymptomatically infected with S. brandenburg some
time previously and was excreting this organism at very low levels.
Ty21a vaccinees (microbiology laboratory workers) have been
described [18]. Each received 4 Ty21a capsules (Swiss Serum
1409
and Vaccine Institute, Berne, Switzerland): 1-6 X 109 cfu viable
organisms and 5-50 X 109 nonviable bacterial cells/capsule on
alternate days as directed by the manufacturer.
Bacteriology. Bacteriology specimens were processed in the
Clinical Microbiology Laboratory, Massachusetts General Hospital. A single daily stool culture was obtained from each volunteer.
Stools were planted both as primary cultures and after overnight
enrichment in selenite- F broth on Hektoen enteric agar and MacConkey agar plates. Quantitative stool cultures for some volunteers
were done by dissolution and serial dilution of 0.5-1.0 g of stool
in PBS and plating on selective media. On study days 2, 4, 6, 8,
and 10,2 sets of blood cultures were collected from each volunteer
and inoculated into Bactec (Becton Dickinson, Sparks, MD) aerobic and anaerobic blood culture bottles (5 mL ofinoculum/bottle).
Blood cultures were held 7 days.
Immunologic assays. Sterile ELISA plates with nitrocellulose
membrane bottoms (Millititer HA; Millipore, Bedford MA) were
used in ELISPOT assays designed to enumerate the number of
peripheral blood IgA antibody-secreting cells (ASC) [22, 23] directed against S. typhi lipopolysaccharide (LPS). Membranes were
incubated overnight with either 50 mM sodium carbonate buffer
(pH 10) alone or with buffer with LPS from S. typhi or Vibrio
cholerae: The latter was used as a control LPS antigen (both at
20 mg/mL, Sigma, St. Louis MO). Plates were then blocked overnight with PBS containing 2% bovine serum albumin (BSA) and
subsequently rinsed 3 times with PBS.
PBMC were isolated from 16 mL of citrated blood using cell
preparation tubes (Vacutainer; Becton Dickinson, Franklin Lakes,
NJ) as directed by the manufacturer; 1-2 X 107 PBMC were
consistently obtained per draw. PBMC were washed three times
with PBS, counted, and suspended at known concentrations in
RPMI tissue culture medium (Mediatech, Herndon, VA) containing 10% fetal calf serum and 1% penicillin-streptomycin solution (Sigma). Duplicate 106 PBMC aliquots and serial 5-fold dilutions of the aliquots were applied to membrane-bottom wells
treated with the various antigens and incubated for 4 h in a tissue
culture incubator. Nonadherent cells were removed by washing
wells 3 times with PBS and 3 times with PBS containing 0.05%
Tween 20 (Sigma). Goat anti-human IgA conjugated to peroxidase
(Kirkegaard & Perry Laboratories, Gaithersburg, MD) was applied
at a dilution of 1:2500 in PBS~2% BSA and incubated overnight
at 4°C. Plates were then washed 5 times with PBS-0.05% Tween
20 and 3 times with PBS. Membranes were developed for 15 min
with 3-amino-9-ethyl-carbazole substrate (Sigma) in 0.1 M acetate
buffer (pH 5) as directed by the manufacturer. Dry membranes
were punched from the ELISA wells using a cork borer, and
brown-red spots representing IgA-bearing cells specifically bound
to antigen-coated membranes were counted at the most appropriate
dilution under X20 magnification. Numbers reported represent the
mean value for duplicate wells at an appropriate dilution for counting.
Serologic analyses. ELISAs were done as described [18] to
measure increases in serum IgG directed against either whole
Ty800 vaccine organisms or S. typhi 0 antigen (Difco, Detroit).
The same protocol was used to measure increases in serum IgA
directed against S. typhi LPS (coating solution, 5 j.lg/mL LPS) and
serum IgG directed against purified Vi antigen (gift of T. Barnett,
CDC, Atlanta; coating solution, 2 j.lg/mL). In each of these assays,
1410
Hohmann et al.
we used a standard serum dilution of 1:80, and each diluted serum
sample was applied in parallel to wells containing buffer alone
or buffer containing antigen. We used affinity-purified, alkaline
phosphatase-labeled goat antihuman IgG or antihuman IgA secondary antibodies (Kirkegaard & Perry).
Specific optical density (Ofr) was determined for each serum
sample by subtracting the OD value of the buffer well from the
antigen well. Increases in specific OD from preimmune to peak
values were obtained by subtracting the preimmune value from
values on subsequent days. Significant OD increases exceeded a
statistically derived significant Of) threshold (0.22 Ol) units
[ODU] for the whole cell IgG assay, 0.13 ODU for the a antigen
IgG assay, and 0.07 ODU for the LPS IgA assay). These threshold
values were determined by testing pre- and postimmune paired
sera from 15 volunteers who received a live attenuated cholera
vaccine and equal the mean net increase in Ol) + 3 SD of these
paired sera.
To provide another measurement of serologic response, endpoint dilution studies were done using 5 ILg/mL S. typhi LPS to
coat antigen wells. Sera were serially diluted 2-fold in PBS-2%
BSA in a microtiter plate beginning at 1:40. The end-point titer
was defined as the highest serum dilution at which specific OD
was >0.15 ODU. An increase in titer >4-fold from preirnmune
to peak was deemed significant. The Widal tube test using H
antigen was done as directed by the manufacturer (Difco). We
used Fisher's exact test to generate P values for comparison of
the proportion of Ty800 and Ty21a vaccinees with positive serologic tests.
Results
In vitro characterization of Ty800. Restriction endonuclease digestion and Southern blotting showed that chromosomal
Ty800 DNA contained the anticipated chromosomal deletion
in the phoP and phoQ genes. Ty800 had a growth rate in Luria
broth identical to that of the Ty2 parent strain, as measured by
serial determination of A 600 values for cultures in logarithmic
growth phase. Ty800 was more sensitive to rabbit defensin
NP-l than was Ty2, a phenotype of other PhoP/PhoQ null
Salmonella strains [15, 18] (data not shown).
Clinical response and side effects. After vaccination, 9 of
11 volunteers remained completely well and had no temperature elevations or other adverse effects. Two subjects who
received the highest dose (4 X 1010) had transient side effects.
Volunteer 9 had an elevated temperature (37.9°C) ~20 h after
receiving the vaccine; the temperature was associated with 1
loose stool and myalgias. He was completely well within 12
h and had no other adverse effects for the duration of the study.
Volunteer 10 had an elevated temperature (38°C) 20 h after
vaccination accompanied by acute gastroenteritis-like symptoms (10 loose-to-liquid stools, grade 3-4 [24], and cramping)
12-36 h after vaccination. The stools were culture-positive for
the vaccine organism; there was no vomiting or bloody stools.
The syndrome resolved without therapy within 48 h after vaccination, and the volunteer felt well for the rest of the study.
JID 1996; 173 (June)
Bacteriology. No subject had blood cultures positive for
the vaccine organism during the study (l0 sets/volunteer). Limiting-dilution studies using strain Ty800 and sterile human
blood demonstrated that the blood culture system used was
very sensitive. Blood culture bottles consistently turned positive within 48 h after introduction of experimental inocula
calculated to deliver 1 organism (measured by parallel plating
on Luria broth agar plates).
The vaccine organism was detected in stool cultures in 10
of 11 subjects. Volunteer 2 (l of 2 given the lowest dose) had
no positive stool culture. The duration of positive stool cultures
was 1-3 days after vaccination in 9 volunteers. Subject 3, who
received 6 X 108 cfu, had positive cultures on days 1,2, and 3
and then intermittently positive stool cultures only after selenite
enrichment until day 15 of the study. This volunteer was completely asymptomatic during this time and received a 7-day
course of oral ciprofloxacin to eradicate the vaccine organism.
The ciprofloxacin therapy was successful (shown by multiple
negative follow-up stool cultures over the ensuing month).
Quantitative stool cultures, done only for subjects who received
the highest 2 doses, showed that vaccine organism burdens
ranged from 102_105 cfu/g of stool (8 X 109 cfu/cohort) to
102-107 cfu/g (4 X 1010 cfu/cohort). The most vaccine organisms were detected on days 1 and 2 after vaccination and
decreased rapidly thereafter. Vaccine organisms recovered
from volunteers retained their PhoP null phenotype on BCIP
plates (see above).
Immunologic analyses. All volunteers except number 1,
who received the lowest dose, had increases in IgA-bearing
cells directed against S. typhi LPS detected in peripheral blood
7 days after vaccination (table 1). Control studies on day 0 and
control wells containing buffer alone or V cholerae LPS on
days 0 and 7 were negative in all cases «2 spots/well with
the highest cell inoculum). Additional ELISPOT studies on
days 4 and 10 for the 2 lower-dose groups showed that IgAsecreting cells were maximal on day 7 (data not shown), as
previously demonstrated for Ty21a [23]. The excluded volunteer with nontyphoidal salmonellosis had 9 IgA-secreting cells
detected on day 0 and 1325 on day 7 after vaccination.
Because a variety of measures for seroconversion were reported in other studies of oral typhoid fever vaccines [5- 7,
25], we used several serologic methods to examine humoral
immune responses to different S. typhi antigen preparations.
ELISAs using standard 1:80 dilutions of sera to measure increases in serum IgG against 0 antigen or whole vaccine organisms were positive in 9 of 11 subjects over the range of doses
in each assay (figure 1). ELISAs for a control group given 4
doses of Ty21a were positive in 3 of 8 or 5 of 8 vaccinees. Two
control vaccinees (numbers 13 and 14) and 1 Ty800 volunteer
(number 6) had high background OD values in these assays.
Control 13 had been vaccinated with a parenteral typhoid fever
vaccine >30 years ago; the other 2 subjects (to their knowledge) had never received typhoid fever vaccines. Peak IgG
JID 1996; 173 (June)
Ty800 Typhoid Fever Vaccine in Volunteers
1411
Table 1. Immunologic responses to S. typhi Ty800 and Ty21 a vaccines.
IgG end-point titer vs. LPS~
Vaccine/dose
Ty800/l dose
7 X 107
6 X 108
8 X 109
4 X 10 10
Volunteer
IgA ASC*
Peak widal tube test vs.
H antigen!
1
2
3
4
5
6
7
8
9
10
11
0
400 11
300 11
1725 11
475"
1075"
90 11
1250 11
2400 11
2388 11
2950 11
<1:20
1:640 11
1:160 11
1:320 11
1:40
<1:20
<1:20
<1:20
1:80 11
1:1280 11
1:160 11
1:40
1:80
<1:40
1:40
1:80
1:640
1:40
1:160
1:40
1:40
1:40
1:40
1:640 11
1:80
1:320 11
1:640 11
1:640
1:40
1:320
1:160 11
1:80
1:640 11
0.41 11
0.47 11
0.18 11
0.60 11
12
ND
ND
ND
ND
ND
ND
ND
ND
<1:20
<1:20
<1:20
<1:20
<1:20
<1:20
<1:20
<1:20
1:80
1:640
1:640
<1:40
1:160
1:40
1:80
<1:40
1:640 11
1:640
1:640
1:40
1:640 11
1:80
1:1280 11
<1:40
0.17 11
0.26 11
0.04
0.02
0.06
0.22 11
0.10 11
0.06
Pre immune
Peak
Serum IgA vs. LPS§
O.oI
0.20 11
0.04
0.35 11
0.17 11
0.14 11
O.oI
Ty21 a/4 doses
13~
14
15
16
17
18
19
NOTE. ND, not done.
* Antibody-secretingcells (ASC) directed against S. typhi lipopolysaccharide (LPS) in ELISPOT assay reported as spots/l O" mononuclear cells; >6, significant (II).
t All preimmune titers were < 1:20. Results are dilution of serum giving 2+ positive reaction (preimmune vs. peak); >4-fold increase, significant (II).
~ Serum dilution at which specific optical density (OD) exceeded 0.15 OD units (ODU) in ELISA against S. typhi LPS; >4-fold increase, significant (II).
§ ODU for 1:80 serum dilutions in ELISA that detected IgA directed against S. typhi LPS; >0.07 ODU, significant rise (II).
~ Subject received parenteral typhoid vaccine > 30 years earlier.
responses in ELISAs occurred on study days 14 or 21 for both
Ty800 and Ty21 vaccinees.
Five of 11 volunteers had evidence of seroconversion by
~4-fold increases in end-point titers using S. typhi LPS as an
antigen, and 6 of 11 seroconverted with ~4-fold increases by
the Widal tube agglutination test using H antigen. We considered seroconversion to either antigen to be evidence of a systemic immune response. Thus, 7 of 11 subjects over the wide
dose range seroconverted after single-dose vaccinations. In
comparison, only 3 of 8 volunteers given 4 oral doses of Ty21a
seroconverted by IgG end-point titer versus LPS (vs. Ty800,
P = .65), and none had positive Widal tube agglutination tests
(vs. Ty800, P = .018).
Serum IgA directed against S. typhi LPS increases significantly in some persons after vaccination with live attenuated
S. typhi vaccines [8, 25]. Because IgA ELISPOT studies were
not done for the Ty21a group, we measured serum IgA by
ELISA to provide another measure of comparison between
these 2 groups. Eight of 11 experimental vaccinees and 4 of 8
Ty21a vaccinees had significant Of) increases, reflecting serum
IgA directed against S. typhi LPS (table 1; P = .38).
Three of 11 volunteers (numbers 6, 10, and 11) also had
measurable increases in serum IgG directed against purified Vi
antigen. In this assay, Ty21a vaccinees were used as a control
population, as Ty21 does not express the Vi antigen. Volunteers
6, 8, and 10 had peak specific Of) increases of 0.56, 0.26, and
0.34 ODD, respectively, at a 1:80 serum dilution, while the
mean net increase (±3 SD) for ty21a vaccinees from days 0
to 21 was 0.02 ± 0.04 ODU.
Discussion
This study demonstrated that Ty800, a genetically defined
phoPlphoQ-deleted S. typhi Ty2, is markedly attenuated in
humans and is a promising candidate vaccine and vector strain.
Of importance, no subject developed serious side effects or
bacteremias-even at very high doses. A dose that induced
some toxicity was reached, but only 1 subject had a self-limited
gastroenteritis-like adverse event, and lower doses were clearly
immunogenic. The pathogenesis of this diarrhea of rapid onset
is unclear. In a previous study, we saw much milder diarrhea
in I of 14 volunteers given 1010 cfu of a more attenuated strain
(Ty445) and have hypothesized that this represents an acute
gastrointestinal inflammatory response to a large bolus of S.
typhi LPS.
Hohmann et al.
1412
0.8
A.
E_
c:::
••
••
G)
Lt)
c:::
o :s 0.6
~E
ci
..E 0.4
dQ.
.5
G)
II)
.
C'lS
G)
u
0.2
I
C'lS
e:-
•
•
•••
••
Ty21a
Ty800
----_.!.._---_!!_---
.:.:::
G)
•
•
••
•
'G)
O
.E
-0.2
-
1.2
c:::
Lt)
•
B.
E_
G)
c:::
o :s
~E 0.8
ci 'G)E 0.6
d
..
Q.
0.4
.5
G)
II)
.:.:::
C'lS
G)
C'lS
G) Q.
.-
u
.E
0.2
0
-0.2
••
•
•
•
•
•••
•
Ty21a
Ty800
•
•
•
•••
----------.... - - - -
Figure 1. Serum IgO responses against S. typhi Ty800 whole vaccine organisms (A) or S. typhi 0 antigen (B). Sera were diluted 1:80
and applied in parallel to ELISA plates coated with antigen. Plates
were developed with goat anti-human IgO antibody conjugated to
alkaline phosphatase, and optical density (OD) was read at 405 nm.
Results are expressed as net increases in specific OD from preimmune
to peak values. OD increases above dotted lines are considered significant (see Methods).
One subject was possibly colonized with the vaccine organism. There was no medical history (e.g., recent antibiotic use)
that might have explained the finding. Unfortunately, quantitative stool cultures were not done for this subject. Because
this volunteer received antibiotics to eliminate the recombinant
organism, the true duration of colonization could not be assessed. The subject received only 108 cfu of bacteria, and others
who received much larger doses did not have prolonged shedding, suggesting that colonization patterns may be more dependent upon host factors than on vaccine-specific factors. It has
been previously suggested that the presence or absence of secreted ABH carbohydrate determinants may modulate adherence of S. typhi to epithelial cells [26].
After vaccination with Ty21a, IgA production by PBMC in
vitro is a very sensitive measure of immunologic response that
JID 1996; 173 (June)
correlates more highly with intestinal IgA production than with
serum measurement of IgG or IgA [27]. Ten (91%) of II
volunteers given any single dose of Ty800 had evidence of
intestinal immune responses as measured by marked increases
in S. typhi LPS-specific IgA-bearing PBMC on day 7 after
vaccination. In previous studies, 3 doses of Ty21 a resulted in
detection of IgA-bearing cells by ELISPOT in 56%-80% of
subjects [23, 28]. S. typhi-specific IgA production by PBMC
in vitro was detected in 93%-100% of volunteers after vaccination with Ty21 a [22, 27]. CVD908, a single-dose aromatic
amino auxotrophic S. typhi vaccine strain, induced S. typhispecific ASC or in vitro IgA production by PBMC in 100%
of volunteers [5, 8] at doses that resulted in bacteremias in
some persons. Ty800 induced >2300 ASC/1 06 PBMC in 3
volunteers, larger numbers than reported in studies of attenuated live typhoid vaccines in which ASC numbers were specifically reported [5, 23]. These results suggest that Ty800 may
be a particularly promising vector strain for antigens relevant
to the gastrointestinal tract or in which secretory (s) IgA at
other sites is important. Direct measurement of mucosal sIgA
would be useful in future Ty800 studies.
The relative importance of intestinal and systemic serologic
responses in protection against typhoid fever is debated [I,
29]. Systemic seroconversion has correlated with protection in
endemic areas [I], and parenteral vaccines that do not generate
intestinal immunity have had protective efficacy in endemic
areas [1] and in presumed naive vaccinees [4]. In addition,
systemic humoral immune responses are relevant to other antigens potentially engineered into a multivalent vector strain.
Although not compared head-to-head here, single doses of
Ty800 resulted in comparable serologic results in ELISAs and
apparently greater seroconversion rates by the Widal test than
in controls who received 4 doses of Ty21 a. Lack of seroconversion in the Widal test was reported in a previous Ty21 study
[28], and ~ 50% seroconversion by the Widal test was seen in
a recent study of subjects given 2 oral doses of CVD908 expressing a malarial circumsporozoite antigen [5].
We found that the most sensitive assays for detecting increases in serum IgG were ELISAs against whole vaccine organisms or S. typhi a antigen. We used this technique because
it successfully documented serologic responses to attenuated
S. typhi strains at the Center for Vaccine Development [5, 7,
8]. With these ELISAs, controls and volunteers with high baseline ODs in an end-point dilution study (Ty800 subjects 6 and
8 and Ty21 a controls 13 and 14) mounted increases in specific
ODU that exceeded by 3 SD the mean net increase of the
control cholera-vaccinated population. For example, volunteer
8, who had large increases in IgA-secreting cells but end-point
titers less than needed for formal seroconversion, had relatively
large increases in Ol) ELISAs (0.6-0.8 ODU). This subject
did not meet criteria for seroconversion by the less sensitive
but presumably more specific end-point dilution titer tests
against S. typhi LPS (table I). Nevertheless, it is likely that
JID 1996;173 (June)
Ty800 Typhoid Fever Vaccine in Volunteers
this subject had a true serologic response to S. typhi. Similarly,
controls 13 and 14 had high baseline titers in the LPS endpoint dilution study and no increase after vaccination; these 2
control vaccinees exceeded the statistically significant threshold in the ELISAs. We presume that like control 13, these 3
persons with high baseline titers or ODs may have unknowingly
previously received typhoid fever vaccine or that they had
cross-reacting antibodies generated by exposure to cross-reacting bacterial antigens. Most likely, nontyphoidal salmonellosis, which can impair response to Ty21a [30], was responsible for the high baseline serologic test results and the
relatively poor serologic responses in these subjects.
In both groups, some subjects had very low baseline serology
results in all assays (volunteers I and 7 and controls 15 and
19). These were apparent serologic nonresponders, presumably
because of unknown host factors. Volunteer 7 had no evidence
of a serologic response in any assay but still generated increases
in specific IgA-secreting cells, confirming that this is a very
sensitive measure of immunologic response to enteric vaccines.
The results of the excluded volunteer with coexisting S. brandenburg infection are more difficult to interpret. Presumably
the low level positive ELISPOT study on day 0 represented
cells generated by a nontyphoidal infection that cross-reacted
'with S. typhi LPS. It is not possible to say what proportion of
day 7 ELISPOT cells were attributable to the experimental
vaccine.
An unexpected finding in our study was the apparent systemic immune response to orally administered Vi antigen.
There was no seroconversion to Vi antigen in a previous trial
of 9 subjects who received 3 oral doses of 5 X 109 cfu of
Ty21a engineered to express this antigen, despite expected rates
of seroconversion to other antigens [28]. Serologic studies of
Vi antigen responses have not been consistently reported in
studies of other attenuated S. typhi vaccines [5, 7, 8], but occasional unquantified responses have been noted [7], suggesting
that the more immunogenic live oral vaccines may also engender responses to Vi. Two of 3 subjects in our study (numbers
6 and 10) who responded to this antigen were not among those
with the most vigorous serologic responses in other assays,
suggesting that host factors may direct responses to individual
antigens.
In summary, we demonstrated that Ty800 is a promising
single-dose oral typhoid vaccine candidate. Although we tested
a small number of volunteers, several important features of this
vaccine strain were apparent. Large doses were well tolerated
and lower doses were immunogenic. No bacteremias occurred,
making this vaccine strain more widely applicable to large
populations than less attenuated strains. Very large numbers
of vaccine-specific IgA-secreting cells were detected, suggesting that the strain may be particularly useful in generating
mucosal immune responses. phoP-activated genes, which will
be transcriptionally repressed in Ty800, have been implicated
in the inhibition of antigen processing by S. typhimurium in
1413
vitro, and this could contribute to the immunogenicity ofTy800
[31]. We conclude that Ty800 may be a particularly useful
vaccine strain for development of a single oral dose typhoid
fever vaccine and for engineering an S. typhi vector capable of
delivering heterologous antigens to the gastrointestinal immune
system.
Acknowledgments
We gratefully acknowledge the contributions of the volunteers
and the staffs of Mallinckrodt General Clinical Research Center
and Clinical Microbiology Laboratories, Massachusetts General
Hospital.
References
1. Levine MM, Fereccio C, Black RE, Tacket CO, Gennainer R. Progress
in vaccines against typhoid fever. Rev Infect Dis 1989; I I(suppl 3):
S552-67.
2. Levine MM. Typhoid fever vaccines. In: Plotkin SA, Mortimer EA Jr,
eds. Vaccines. Philadelphia: WB Saunders, 1994:597-633.
3. Hirschell B, Wuthrick R, Somaini B, Staffen R. Inefficacy of the commercial live oral Ty21 a vaccine in the prevention of typhoid fever. Eur J
Clin Microbiol 1986;4:925-8.
4. Schwartz E, Shlim RR, Eaton M, Jenks N, Houston R. The effect of
oral and parenteral typhoid vaccination on the rate of infection with
Salmonella typhi and Salmonella paratyphi A among foreigners in Nepal. Arch Intern Med 1990; 150:349-51.
5. Gonzales C, Hone D, Noriega FR, et al. Salmonella typhi vaccine strain
CVD908 expressing the circum sporozoite protein of Plasmodium falciparum: strain construction and safety and immunogenicity in humans.
J Infect Dis 1994; 169:927 - 31.
6. Levine MM, Herrington D, Murphy MR, et al. Safety, infectivity, immunogenicity and in vivo stability of two attenuated auxotrophic mutant
strains of Salmonella typhi 54lTy and 543Ty, as live oral vaccines in
man. J Clin Invest 1987; 79:888-902.
7. Tacket CO, Hone DM, Curtiss R III, et al. Comparison of the safety and
immunogenicity of llaroCllaroD and Ilcya!lcrp S. typhi strains in
adult volunteers. Infect Immun 1992; 60:536-41.
8. Tacket CO, Hone DM, Losonsky GA, Guers L, Edelman R, Levine MM.
Clinical acceptability ofCVD 908 Salmonella typhi vaccine strain. Vaccine 1992; 10:443 -6.
9. Miller SI, Kukral AM, Mekalanos JJ. A two component regulatory system
phoP phoQ controls Salmonella typhimurium virulence. Proc Nat! Acad
Sci USA 1989; 86:5054-8.
lo. Groisman E, Chiao E, Lipps C, Heffron F. Salmonella typhimurium phoP
virulence gene is a transcriptional regulator. Proc Nat! Acad Sci USA
1989; 86:7077 -81.
II. Belden WJ, Miller SI. Further characterization of the PhoP regulon: identification of new pag virulence loci. Infect Immun 1994;62:5095-lOl.
12. Behlau I, Miller S. A PhoP-repressed gene promotes S. typhimurium invasion of epithelial cells. J Bacteriol 1993; 175:4475-84.
13. Pulkkinen W, Miller S. AS. typhimurium virulence protein is similar to
a Yersinia enterocolitica invasion protein and a bacteriophage lambda
outer membrane protein. J Bacteriol 1991; 173:86-93.
14. Fields P, Groisman E, Heffron F. A Salmonella locus that controls resistance to microbicidal proteins from phagocytic cells. Science 1986;243:
lO59-62.
15. Miller SI, Pulkkinen WS, Selsted ME, Mekalanos JJ. Characterization of
defensin resistance phenotypes associated with mutations in the phoP
1414
Hohmann et al.
virulence regulon of Salmonella typhimurium. Infect Immun 1990; 58:
3706-10.
16. Galan JE, Curtiss R III. Virulence and vaccine potential of phoP mutants
of Salmonella typhimurium. Microb Pathog 1989;6:433-43.
17. Miller SI, Mekalanos JJ, Pulkkinen WS. Salmonella vaccines with mutations in the phoP virulence regulon. Res Microbioll990; 141:817-21.
18. Hohmann EL, Oletta CA, Miller SI. Evaluation of a phoP/phoQ-deleted
aroA-deleted live oral Salmonella typhi vaccine strain in human volunteers. Vaccine 1995; 14:19-24.
19. Blofield 1, Vaughn V, Rest R, Eisenstein B. Allelic exchange in E. coli
using the Bacillus subtilis sacB gene and a temperature sensitive pSC101
replicon. Mol Microbiol 1991;5:1447-57.
20. Donnenberg M, Kaper 1. Construction of an eae deletion mutant of enteropathogenic E. coli by using a positive-selection suicide vector. Infect
Immun 1991; 59:4310- 7.
21. Kier LD, Weppelman RM, Ames BN. Regulation of nonspecific acid
phosphatase in Salmonella: phoN and phoP genes. J Bacteriol 1979;
138:155-61.
22. Forrest BD. Identification of an intestinal immune response using peripheral blood lymphocytes. Lancet 1988; 1:81-3.
23. Kantele A, Arvilommi H, Jokinen 1. Specific immunoglobulin secreting
human blood cells after peroral vaccination against S. typhi. J Infect
Dis 1986; 153:1126-31.
JID 1996; 173 (June)
24. Levine MM, Kaper J, Herrington D, et al. Safety, immunogenicity and
efficacy of recombinant live oral cholera vaccine CVD 103 and CVD
103-HgR. Lancet 1988;2:467-70.
25. Forrest BD, Labrooy JT, Beyer L, Dearlove CE, Shearson DJe. The human
humoral immune response to Salmonella typhi Ty21a. J Infect Dis 1991;
163:336-45.
26. Hoffmann E, Chianale J, Rollan A, Pereira J, Ferrecio C, Sotomayor V.
Blood group antigen secretion and gallstone disease in the Salmonella
typhi chronic carrier state [letter]. J Infect Dis 1993; 167:993-4.
27. Forrest BD. Indirect measurement of intestinal immune responses to an
orally administered attenuated bacterial vaccine. Infect Immun 1992;
60:2023-9.
28. Tacket CO, Losonsky G, Taylor DN, et al. Lack of immune response to
the Vi component of a Vi-positive variant of the Salmonella typhi live
oral vaccine strain Ty21a in human studies. J Infect Dis 1991; 163:
901-4.
29. Robbins JB, Schneerson R, Szu Se. Perspective: hypothesis: serum IgG
antibody is sufficient to confer protection against infectious diseases by
inactivating the inoculum. J Infect Dis 1995; 171:1387-98.
30. Forrest BD. Impairment of immunogenicity of Salmonella typhi Ty21a
due to preexisting cross-reacting intestinal antibodies [letter]. J Infect
Dis 1992; 166:210-2.
31. Wick M, Harding CV, Twesten NJ, Normark SJ, Pfeifer JD. The phoP
locus influences processing and presentation of S. typhimurium antigens
by activated macrophages. Mol MicrobioI1995; 16:465-70.