A Vibrio cholerae serogroup O1 vaccine candidate against CTXETΦ infection

Vaccine 25 (2007) 4046–4055

A Vibrio cholerae serogroup O1 vaccine candidate against CTXET infection Meiying Yan, Guangwen Liu, Baowei Diao, Haiyan Qiu, Lijuan Zhang, Weili Liang, Shouyi Gao, Biao Kan ∗ State Key Laboratory for Infectious Disease Prevention and Control, National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, P.O. Box 5, Changping, Beijing 102206, PR China Received 11 June 2006; received in revised form 22 November 2006; accepted 10 February 2007 Available online 5 March 2007

Abstract Cholera is a severe diarrheal disease that may spread rapidly. Vaccination is considered a valid measure against it. We developed a new vaccine candidate, IEM109, against Vibrio cholerae. To generate this candidate, a chromosomal fragment containing the TLC element, attB of the CTX integration site, and RTX cluster responsible for the cytotoxic activity for mammalian cells was deleted through homologous recombination from the previously described El Tor biotype, IEM101. The protective genes ctxB and rstR, which establish resistance to CTX infections, were inserted into that same location on the chromosome of IEM109 to enhance the safety and genetic stability of the vaccine candidate and to prevent horizontal gene transfer. In in vivo tests, cell cultures showed that the cytotoxic effect of IEM109 on Hep-2 was negative. Furthermore, the infection rate of El Tor biotype CTX to that of IEM109 in the rabbit intestine is 3000-fold lower than that of IEM101. Intraintestinal vaccination of rabbits with a single dose of IEM109 elicits high titers of anti-CTB IgG and vibriocidal antibodies. When challenged with 0.5–2 ␮g CT and 105 to 108 CFU of four wild toxigenic strains of different biotypes and serogroups, IEM109 conferred full protection. Thus, IEM109 is a stable vaccine candidate that evokes not only antitoxic and vibriocidal immunities, but also resistance to the El Tor biotype CTX infection. © 2007 Elsevier Ltd. All rights reserved. Keywords: Vibrio cholerae; Vaccine; Immunity

1. Introduction Cholera is a severe diarrheal disease that may spread rapidly and remains a threat to peoples who cannot obtain safe water and proper sanitation, particularly so in developing countries. Seven cholera pandemics have been recorded in history. Most outbreaks of cholera, including the seventh, still ongoing pandemic, are caused by the O1 El Tor biotype. Aside from Vibrio cholerae O1, the only other known serotype responsible for epidemic and endemic outbreaks is V. cholerae O139 [1]. Vaccination is considered a valid measure against the cholera threat. Ideally, whole cell vaccines against enteric ∗

Corresponding author. Tel.: +86 10 61739458; fax: +86 10 61739156. E-mail address: [email protected] (B. Kan).

0264-410X/$ – see front matter © 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2007.02.043

pathogens should be administered by the mucosal route to elicit mucosal protection [2,3]. Two different strategies, inactivated whole cell vaccines (plus B subunit of cholera toxin) and live attenuated vaccines, are commonly used in cholera vaccine development [2]. Attenuated live V. cholerae O1 vaccine candidates [3–10], including Peru-15 [4,10], CVD103-HgR [5,7,9], CVD111 [6], and some others are under development or licensed. These vaccines evoke an immune response similar to that of natural cholera infection, but exhibit some side reactions or reactogenicity in humans. A major and common toxic element of V. cholerae next to the cholera toxin is the RTX toxin. The RTX toxin can cause cell-rouding pathogenic effects by cross-linking actin in mammalian cells. The toxin is different from other poreforming RTX toxins present in enteric pathogens [11,12].

M. Yan et al. / Vaccine 25 (2007) 4046–4055

It can be a potential virulent factor that causes side reactions during vaccination. In mouse models, for example, this toxin can induce acute proinflammatory reactions [13]. Less reactogenicity was observed for attenuated vaccines where RTX toxin is not secreted, such as for some vaccine strains developed from classical O395 [8]. Attenuated live oral vaccines are not without problems, however, due to the potential transfer of hazardous genes [14]. An attenuated vaccine strain could reacquire cholera toxin genes and become virulent after a subsequent phage infection. One such toxin gene is ctxAB, which is carried by a filamentous phage CTX and can be horizontally transferred from toxigenic strains to nontoxigenic ones [15]. Efforts have been taken to circumvent this problem by removing the attB site for the CTX integration, for example, or by introducing mutations into the recA gene [10,16]. Another example in overcoming horizontal gene transfer is our recently constructed vaccine strain IEM108. IEM108 is based on a plasmid-chromosome lethal balanced system that introduces a repression gene, rstR, into the prototype strain IEM101, which is naturally CTX negative and proved be safe in human volunteers [17,18]. It is showed that IEM108 strain is more efficient in resisting CTXET  infections than isogenic strains [17,19]. The vaccine candidate not only represses the same type of CTX infection [20], but also elicits high vibriocidal and antitoxic responses in rabbit models and protects animals against the challenge of the virulent wild type strains [17]. However, it is possible that the recombinant plasmid containing the thyA, rstR, and ctxB genes can be lost due to the presence of thymidine in the intestines similar with other auxotroph of V. cholerae [21–23].

4047

To improve the stability of rstR and ctxB genes cloned in the balanced plasmid, we removed a fragment containing TLC element, attB site for CTX integration, and RTX cluster from chromosomal DNA of IEM101. At the same time, we inserted rstR and ctxB genes into the same chromosomal location. This newly developed vaccine strain is called IEM109. In this paper, we described its construction and sensitivity to CTXET  infection in vivo as well as the immune responses elicited by it in rabbit models.

2. Materials and methods 2.1. Strains and culture media Bacterial strains and plasmids used in this study are described in Table 1. V. cholerae and E. coli strains were maintained in LB broth with 20% glycerol at −70 ◦ C and streaked on Luria-Bertani (LB) agar for experiments. Transformants containing recombinant plasmids were selected on LB agar containing ampicillin (100 ␮g/ml) or chloramphenicol (34 ␮g/ml for E. coli, 15 ␮g/ml for V. cholerae). The E. coli strain JM109 was used for general transformations during cloning, and SM10␭pir was used as the recipient strain of the recombinant suicide plasmids. 2.2. Construction of IEM109 The strategy for constructing IEM109 is showed in Fig. 1. To knockout the target genes, two homologous fragments rtxD and Y were amplified from V. cholerae prototype strain

Table 1 Bacterial strains and plasmids used in this study Strains and plasmids

Description

Source and reference

V. cholerae IEM101 N-c 1119 O395 Wujiang-2 Bin-43 IEM101-T IEM108 IEM109

El Tor, Ogawa, ctxAB− , CTXET − , CTXclass − , tcpA+ N16961 ctxAB::cat, CAFr Classical, Inaba, CTXET − , CTXclass + , tcpA+ Classical, Ogawa, CTXET − , CTXclass + , tcpA+ El Tor, Inaba, CTXET + , CTXclass − , tcpA+ El Tor, Ogawa, CTXET + , CTXclass − , tcpA+ El Tor, Ogawa, IEM101thyA El Tor, Ogawa, pUTBL3 cloned into IEM101-T From IEM101, TLC-attB-RTX::ctxB-rstR

[18] [19] Patient Patient Patient Patient [17] [17] This study

E. coli JM109 SM10␭pir

recA1supE44endA1hsdR17thi(Lac-proAB)F [traD36proAB+ LacIq LacZM15] supE, recA::RP4-2-Tc::Mu, Kmr ␭pir

Our lab Our lab

Plasmids pCOS5 pUC18 pUX pUXY pUXYCm pUXYBR pCVD442 pCXYCm pCXYBR

Cosmid vector, oriT, Apr , CAFr Clone vector, oriMB1, lacZ+ , Apr pUC18::rtxD, Apr pUX::Y, Apr pUXY::cat, Apr , CAFr pUXY::ctxB-rstR, Apr Suicide plasmid, mob, ori, bla, sacB, Apr pCVD442::rtxD-cat-Y, Apr , CAFr pCVD442::rtxD-ctxB-rstR-Y, Apr

[24] Our lab This study This study This study This study [25] This study This study

4048

M. Yan et al. / Vaccine 25 (2007) 4046–4055

Fig. 1. Construction pattern of IEM109 by homogenous recombination. (A) Target genes on chromosome of V. cholerae IEM101. RTX cluster is composed of the rtxA/C/B and rtxD genes (not to scale). (B) Fragments Y and rtxD of IEM101, the homogenous fragments were cloned into the suicide plasmid pCVD442 to produce pCXYCm. TLC, attB and rtxA/C/B of V. cholerae IEM101 were replaced by the cat gene by homogenous exchange using recombinant plasmid pCXYCm to construct IEM109-Cm. (C) pCXYBR, a recombinant plasmid with Y, rtxD (the homogenous fragments) and rstR, ctxB genes cloned into suicide plasmid pCVD442. The cat gene was deleted and the rstR and ctxB genes were introduced into IEM109-Cm by homogenous recombination to generate IEM109.

IEM101. The rtxD fragment was amplified using the primers rtxD-up (5 GCT CTA GAT GAC TCA TGA CCC AAT G3 ) and rtxD-down (5 ACG CGT CGA CAT CAC ACG TCG TTT ATC3 ). The Y fragment was amplified using the primer set Y-up (5 CAG GAG CTC ATC CGC AAC GTA TTC CCA CAC C3 ) and Y-down (5 TGG GGT ACC TGC TCC GAG TTA TTT CGA AAC C3 ). Endonuclease sites (italic alphabets) were added to the N-terminal of each primer. Then the two purified PCR products were double digested with XbaI-SalI and SacI-KpnI, respectively, and subsequently cloned into pUC18 to create a recombinant plasmid pUXY. An 800 bp fragment encoding chloramphenicol resistance was amplified from a cosmid pCOS5 using the following primers: cat-up (5 CGT AGC ACC AGG CGT TTA AG3 ) and cat-down (5 GAT CGG CAC GTA AGA GGT TC3 ). This cat (chloramphenicol) gene was blunt-ligated at the SmaI site of plasmid pUXY to create the recombinant plasmid pUXYCm. Transformants were selected by growth on LB-chloramphenicol and confirmed by colony PCR using the cat-up/down primers. The 2.25 kb fragment containing rtxD-cat-Y fusion was excised from pUXYCm using SalI and SacI and the excised fragment was cloned into the location between the SalI and SacI site of the suicide plasmid pCVD442 to create the recombinant suicide plasmid pCXYCm. pCXYCm was conjugatively transferred to V. cholerae IEM101 by filter mating as described previously [26]. The conjugants, as a result of the first crossover event, were selected on LB agar plates containing ampicillin, chloramphenicol, and polymyxin B. The resultant mutants were grown in a modified LB medium containing 15% sucrose but no NaCl. The clones sensitive to ampicillin, but resistant to chloramphenicol and polymyxin B were designated as IEM109-Cm. To confirm the presence of the cat gene and

the absent of the target fragment TLC-attB-RTX (including two copies of TLC, the attB site, and the RTX cluster) on the chromosome of IEM109-Cm, PCR and colony hybridizations were performed using cat-up/down primers, RTX specific primers (rtxA-up 5 TTG GGT ATG GGT GTT GTA GC3 rtxA-down 5 TTG GGT ATG GGT GTT GTA GC3 ), TLC specific primers (cri-up 5 TTG TTG GAT GTC GGC TTA G3 and cri-down 5 GGT TAT GGA TTG GGT CTT C3 ) and their corresponding probes (cat, rtxA and cri). A similar protocol as described above was used to construct IEM109. To introduce the functional rstR repression gene into V. cholerae, an 874 bp gene cassette was extracted from EcoRI digested plasmid pUTBL2 and blunt-ended using Klenow I large fragment. Then the blunt-ended rstR fragment and the pUXY-SmaI digestion product were ligated and transformed into E. coli JM109. The recombinant plasmid containing a rtxD-rstR-Y fusion fragment was annotated as pUXYR. Furthermore, a functional 1.15 kb ctxB fragment obtained from XbaI digested-pUTBL2 was subcloned into plasmid pUXYR to generate a recombinant plasmid pUXYBR which contains a rtxD-rstR-ctxB-Y fusion fragment. To ensure that the rstR and ctxB genes were introduced into the chromosome of V. cholerae, the corresponding recombinant plasmids pUXYR and pUXYBR were sequenced and examined for CTB expression by GM1ELISA. The plasmid pUXYBR expressing CTB was digested with SalI and SacI and a 3.42 kb fragment containing the rtxD-rstR-ctxB-Y fusion genes was cloned into the suicide plasmid pCVD442 to create the recombinant suicide plasmid pCXYBR. pCXYBR was conjugatively transferred to V. cholerae IEM109-Cm by filter mating. The clone that had the cat gene replaced with the rstR and ctxB genes was named IEM109. PCR analysis and Southern blot analysis were used to verify the exact deletion and replacement of target genes.

M. Yan et al. / Vaccine 25 (2007) 4046–4055

4049

2.3. Genetic stability assay

2.6. Colonization and immunization assays

The vaccine strain IEM109 was cultured overnight on LB agar plates as “one generation” and continually streaked on plates for 100 generations. To detect the presence of the rstR and ctxB genes on the chromosome, PCR was performed on the randomly selected 50 colonies of each generation.

One day before immunization, the blood of the adult New Zealand White rabbits was collected as blank ground control. Eight rabbits were randomly divided into two groups: three rabbits were immunized with prototype IEM101 as the control group, the other five rabbits were inoculated with IEM109. For each group, 1 ml of 107 cfu/ml bacteria was inoculated into the intestines via ilea loop ligation [17]. Fresh fecal materials collected every day were suspended in alkaline peptone water and cultured on LB agar containing gentamycin and polymyxin B. Peripheral blood was also collected once a week to detect the anti-CTB and vibriocidal antibodies.

2.4. Cytotoxic effect assay Bacteria were grown for 16 h in LB broth and washed twice with phosphate buffer (pH 7.4). Then the bacteria of 107 cfu/ml was added to 105 cfu/ml mammalian Hep-2 cells covered with fresh RPMI 1640 containing 10% fetal bovine serum without antibiotics. Cells were cultured at 37 ◦ C in the presence of 5% CO2 . Ninety minutes later morphological changes of the mammalian cells were viewed with an Olympus inverted microscope (model CKX41SF). For visualization of actin, cells growing on eight-well CultureSlide (BD) were stained with Rodamin phalloidin (Invitrogen) according the manufacture’s instruction. The cells were then mounted onto glass slides with mounting medium H-1200 (Vector, England). Glass slide were viewed with a fluorescence microscope Axiovert 200 M (Carl Zeiss). 2.5. In vivo CTX infection assays CTXc, a chloramphenicol marked CTXET  constructed in our previous study [19], was used to estimate whether IEM109 was resistant to CTX infection. CTXc particles were induced by mitomycin-C as previously described [16,23]. The infection activity of CTXc particles was verified with an in vitro infection assay. Briefly, V. cholerae strains IEM109, IEM108, 1119, Wujiang-2, and IEM101 were cultured at 30 ◦ C for 18–24 h for maximum TCP expression [15]. Then the strains were individually mixed with CTXc, and cultured at 30 ◦ C for 3 h. To assess the amount of each strain infected with CTXc, the resulting infected clones were selected on LB plates containing chloramphenicol. The infection rate was calculated as the ratio of the recovered chloramphenicol resistant bacteria to the mixtures. Adult New Zealand White rabbits (2–2.5 kg) were used for in vivo CTXc infection assays. The different mixtures of the strains IEM109, IEM108, 1119, Wujiang-2, and IEM101 with the CTXc particles were injected into ileal loops. After 10 h, the rabbits were sacrificed and the mucous membrane of each loop was washed with 1 ml of 0.9% saline. For the loops with accumulated fluid, 1 ml liquid was collected as the ileal loop fluid. Infection rates were calculated as the ratio of the bacteria strains grown on the LB agar containing chloramphenicol to the total number of recovered strains from the intestine. To verify CTXc infection, some colonies were randomly selected to detect the cat gene by amplification using cat-up/down primer set.

2.7. GM1-ELISA A GM1 ganglioside dependent enzyme linked immunosorbent assay (GM1-ELISA) was used to determine the expression of CTB in pUXYBR and in IEM109 and to detect anti-CTB titers in sera of immunized rabbits as described [17]. IEM109 was inoculated in AKI media and statically cultured at 30 ◦ C followed by shaking at 37 ◦ C for CTB production. A 96 microwell plate coated with GM1(2 ␮g/ml, Sigma) was incubated with 100 ␮l supernatant of JM109 cell lysate containing pUXYBR at 37 ◦ C for 1 h. The plate was then washed with PBS containing 0.05% Tween 20 and incubated with 100 ␮l polyclonal rabbit anti-CT antiserum for 1 h at 37 ◦ C. The microwell plate was then incubated with 100 ␮l of the HRP-conjugated secondary antibody and subsequently o-phenylenediamine (Sigma) was added as the substrate. For the plate, absorption at 492 nm was determined by a microplate reader (Bio-Rad model 550). Positive results were determined if the ratio of the sample to the control was ≥2. JM109 cell lysate was added as a control. The titer of the serum anti-CTB IgG antibody in immunized rabbits was also measured with GM1-ELISA as described above [17]. 2.8. Serum vibriocidal antibody assay To measure the serum vibriocidal antibody titers, microassays were performed using 96-well plates. Briefly sera were inactivated at 56 ◦ C for 30 min and diluted 1:5 in PBS as the initial samples. The pre-diluted rabbit sera were then added into the first well and then serially diluted in subsequent wells. PBS alone was used as a negative control. After incubation with 100 CFU of indicator V. cholerae Bin-43 and guinea pig serum as a complement source in PBS, 0.01% 2,3,5-trihenyltetrazolium chloride in LB broth were added to each well. The plates were further incubated at 37 ◦ C until negative controls showed a color change. The vibriocidal titer is defined as the highest dilution of serum that completely inhibits growth of Bin-43.

4050

M. Yan et al. / Vaccine 25 (2007) 4046–4055

2.9. Immune protection assessment in rabbit models To evaluate the protection efficacy of IEM109 in vivo, immunized rabbits were challenged 28 days after the singledose immunization with pure cholera toxin (Sigma) and four virulent V. cholerae strains (O395, 1119, Wujiang-2 and Bin-43) with different serotypes and biotypes. Rabbits were anesthetized and the intestines were tied into 4–5 cm long loops. Then 105 to 108 CFU of challenge strains and 0.5, 1, 1.5, 2, and 3 ␮g of cholera toxin were injected into each loop. Normal saline was used as a negative control. Sixteen to 18 h later, the rabbits were sacrificed and the accumulated fluid from each loop was collected. The ratio of the volume of accumulated fluid to the length of each loop was calculated for the challenged rabbits.

3. Results 3.1. Construction and phenotypic characterization of the recombinant vaccine IEM109

Cm. When this cat gene was replaced with the ctxB and rstR genes, we obtained IEM109. The exact deletion and replacement of target genes was verified with PCR and Southern blot. We observed no difference between IEM109 and the initial strain IEM101 in terms of production of hemolysis, phage typing, and growth (data not shown). Moreover, the amplifications of ctxB and rstR genes are positive for 100 generations, demonstrating the two genes could stably exist on the chromosome of IEM109. To ensure the expression of ctxB in IEM109, GM1-ELISA was used to detect CTB extracts from JM109 containing pUXYBR and from the supernatant of IEM109. We determined that the absorption ratio of JM109 containing pUXYBR to JM109 was higher than the threshold, indicating that the ctxB gene cloned in plasmid pUXYBR can express toxin subunit B. The absorption ratio of IEM109 to control sample is 10, indicating that IEM109 can secret CTB to the supernatant. We detected no CTB signal in the IEM101 supernatant, consistent with the fact that IEM101 is naturally defective in ctxAB. 3.2. Cytotoxic effect assay

The prototype strain for developing IEM109 was IEM101, an El Tor biotype strain naturally defective in the CTX element. The DNA section between the TLC and the RTX cluster, including the attB site, was deleted by the insertion of the cat gene to produce the intermediate strain IEM109-

Cell rounding was observed after Hep-2 cells were co-incubated with strain IEM101. Cell rounding was not observed after adding IEM109 (Fig. 2), however, indicating that the cytotoxic activity disappeared by removing the

Fig. 2. Cytotoxic effect of IEM109 and IEM101 on Hep-2 cells (40×) and depolymerization of actin caused by RTX toxin. The V. cholerae strains added were mock (PBS) (C1 and C2), RTX− strain IEM109 (A1 and A2) and RTX+ strain IEM101 (B1 and B2). The actin depolymerized in rounding cells (B1) showed dense palloidin-staining aggregates (B2).

M. Yan et al. / Vaccine 25 (2007) 4046–4055

4051

Table 2 The infection rates of CTXET c for five strains of V. cholerae Strains

Infection rates (95% confidence interval) In vivo

1119 IEM101 IEM108 IEM109 Wujiang-2

1.79 × 10−3

In vitro (7.56 × 10−4

4.23 × 10−3 )

to 3.88 × 10−7 (1.54 × 10−7 to 9.8 × 10−7 ) 6.78 × 10−10 (7.19 × 10−11 to 6.39 × 10−9 ) 9.85 × 10−11 (1.57 × 10−11 to 6.17 × 10−10 ) 5.36 × 10−9 (2.1 × 10−10 to 1.37 × 10−7 )

RTX cluster from IEM101. To verify that cell rounding was not attributed to pore formation or necrosis, the leakage of lactate dehydrogenase was assayed. Our results show that Hep-2 cells co-incubated with IEM101 did not release lactate dehydrogenase into the culture medium (data not shown). At 1.5 h, the actin within the rounded cells was found in dense palloidin-staining aggregates, which indicates that the actin stress fibers in Hep-2 cells exposed to IEM101 are completely depolymerized (Fig. 2). Depolymerization was not observed in cells exposed to IEM109, demonstrating that RTX toxin is required for the disassembly of actin fibers the same to the previous report [12]. 3.3. Resistance to the infection of CTXc To evaluate the resistance of IEM109 and other reference strains to infection with CTXc, extracted CTXc particles were used. After five strains (1119, Wujiang-2, IEM101, IEM108 and IEM109) was mixed with the induced CTX, we found that only the classical strain 1119 could be infected by CTXc in vitro, though with extremely low frequency. The average infection rate for this strain was 9.1 × 10−7 (95% confidence interval 1.3 × 10−7 to 6.4 × 10−6 ). Rabbit ileal loops were used as in vivo CTXc infection models because the small intestine provides an appropriate environment for the growth of the V. cholerae strain and CTXc infection [13]. Colonies grown on the selective agars were verified by antisera and cat specific PCR, and were all positive. Subsequently, colonies growing on the selective media were counted as infected with CTXc. The average rate of CTXc infection to 1119 was 1.79 × 10−3 , while that for the four El Tor strains was about 10−7 to 10−11 (Table 2). Thus, it appears that the classical V. choleae strain 1119 is much more susceptible than the El Tor strains to infection with the El Tor derived CTXc in vivo. The average infection frequency for IEM109 in the rabbit intestine was 9.85 × 10−11 , about 3000-fold lower than that of its prototype IEM101. It was also the lowest rates among all the tested strains. Although there was no significant difference in infection rates between IEM109 and IEM108, about seven-fold lower rate of infection in IEM109 was observed comparing with that of IEM108. The rates of these two strains are statistically lower than that of the other two strains 1119 and Wujiang-2 that carried classical and El Tor biotype CTX,

9.1 × 10−7 (1.3 × 10−7 to 6.4 × 10−6 ) 0 0 0 0

respectively. These results indicate that RstR exert specific immunity to the same type of CTX and that IEM109 has the characteristics to effectively prevent CTXET  infection. 3.4. Intestinal colonization of IEM109 in rabbit models A prolonged shedding of vibrios in coproculture is indicative of successful colonization and subsequent multiplication in the small intestine of rabbits as well as in humans. The shedding of IEM109 and IEM101 was assayed by coproculture (feces was cultured in alkaline peptone water) every day after initial vaccination of the rabbits with one dose of a strain. The coproculture of the rabbits vaccinated with IEM109 was positive for on average 7 days (6–11 days) and shedding time in the rabbits vaccinated with IEM101 was 6.5 days (5–8 days), consistent with previous studies [18]. There was no significant difference in bacterial shedding or clearance between these two strains and the vaccine candidate strain IEM109 efficiently colonizes rabbit intestines. 3.5. Detection of anti-CTB IgG and vibriocidal antibodies None of the preimmune sera from the rabbits showed any detectable anti-CTB antibodies when analyzed with GM1ELISA. Specific high anti-CTB IgG antibodies were detected

Fig. 3. Serum anti-CTB IgG response after immunization with IEM109 and IEM101.

4052

M. Yan et al. / Vaccine 25 (2007) 4046–4055

Fig. 6. Average fluid accumulations in the rabbit ileal loops after challenge with pure CT. Different animal groups are indicated as follows: naive, white bars; IEM101-immunized, gray bars; IEM109-immunized, black bars. Fig. 4. Serum vibriocidal response after immunization with IEM109 and IEM101.

in rabbits immunized with IEM109 (Fig. 3). The antibodies became detectable on day 7 and reached peak titers (mean titer 1:1351) on day 28 after immunization (Fig. 3). These results indicate that IEM109 express CTB subunits eliciting high titers of antibodies. There was no anti-CTB antibody detected in rabbits vaccinated with IEM101 due to the lack of its ctxB gene. The preimmune sera did not show any detectable level of vibriocidal antibodies in rabbits. A similar antibody profile was viewed in animals immunized with IEM109 and IEM101. The vibriocidal antibodies became detectable on day 7 post-immunization, significantly increased on day 21 and the peak titers were reached on day 28 after immunization (Fig. 4). The mean antibody titer for the IEM109 group was higher than that of IEM101 group.

3.6. Protection against the challenge of toxigenic V. cholerae in vaccinated rabbits In the naive-animal group, all the loops had significant amounts of fluid accumulation after being challenged with either CT or toxigenic bacteria (Fig. 5). Furthermore, more accumulated fluid was recorded for higher doses of challenge. In the IEM101-vaccinated group, little fluid accumulation was detected in the loops exposed to a lower dose of 105 to 107 CFU of the bacterial challenge. However, there was significant fluid production in loops of rabbits challenged with strain 1119, Bin-43, and Wujiang-2 of 108 CFU. No fluid accumulation was observed in animals challenged with the O395 strain. It is notable that no fluid accumulation was detected in the loops of the IEM109-vaccinated rabbits except one subjected to 108 CFU of 1119. Moreover, there was no fluid accumulation detected in IEM109-vaccinated animals when their loops

Fig. 5. Average fluid accumulations in the rabbit ileal loops after challenge with the strains O395, 1119, Bin-43, and Wujiang-2. Different animal groups are indicated as follows: naive, white bars; IEM101-immunized, gray bars; IEM109-immunized, black bars.

M. Yan et al. / Vaccine 25 (2007) 4046–4055

were exposed to 2 ␮g cholerae toxin, whereas significant fluid accumulated in loops of IEM101-vaccinated rabbits (Fig. 6). These results suggested that IEM109 elicits more effective protection against wild virulent strains than the initial strain IEM101.

4. Discussion In this paper, we describe the construction of IEM109, a newly developed vaccine strain against V. cholerae, and its less sensitivity to CTXET  infection in vivo. IEM109 was developed by replacing a fragment containing the TLC element, the attB site, and the RTX cluster in the chromosomal DNA of IEM101 with the rstR and ctxB genes formerly carried on a plasmid in the strain IEM108. IEM109 cannot only overcome horizontal gene transfer but can also prevent the loss of any plasmids carrying the thyA, rstR, and ctxB genes, which establish resistance to CTXET  infection in the vaccine candidate strain IEM108. The cholera vaccine prototype IEM101 has been evaluated in animal models and shown to be safe and protective against V. cholerae [18]. Single dose vaccination in volunteers showed that IEM101 could colonize in intestine for 4–9 days and elicited protective serum and mucosal immunity with no apparent side reactions [18]. However, there is no antitoxic response due to the absence of ctxAB genes in IEM101. To enhance its immuogenicity toxin subunit gene ctxB was integrated in the chromosome of IEM101. With the attenuation and improvement of immunogenicity in live bacterial vaccine candidates, safety becomes a concern in avoiding the contamination of environment with vaccine strains or their derivative strains, which may reacquire hazardous genes [3]. Nontoxigenic V. cholerae strains can become toxigenic strains due to the horizontal transfer of the lysogenic bacteriophage CTX in the environment or intestine. To decrease the probability of acquisition of ctxAB and improve safety, the repression gene of CTX infection, rstR, was also inserted into the chromosomal DNA of IEM101. However, the inhibition of the RstR is biotype specific, which means certain types of rstR allele only repress the reproduction of the same type of CTX. Given that El Tor biotype of V. cholerae are predominately observed in current years, the same type of rstR prefer to be introduced into the vaccines. In fact, most of serogroup O139 isolates carry El Tor-derived CTX. An El Tor-derived rstR gene cloned into a balance plasmid was introduced into IEM101T to develop a new vaccine candidate IEM108 previously [17]. CTX infection assays in vivo indicated that the ratio of CTXET  infection to IEM108 is 1000-fold lower than that of the initial strain IEM101, demonstrating the repressive effect of rstR introduced into IEM101. In addition, the deletion of the TLC element and attB site would further decrease the replication, reproduction and integration of CTX. The TLC element on the upstream of CTX element is tightly linked

4053

to the CTX prophage among pathogenic V. cholerae strains especially in toxigenic strains. The TLC element may be related to the replication and reproduction of CTX[27]. Our results showed that the infection ratio of CTXc to IEM109 is significantly (3000-fold) lower than that of the prototype IEM101. The infection rate is even lower than that of IEM108 and the native strain carrying El Tor biotype CTX. These observations suggest a further decrease in the possibility of reacquiring toxic genes to become toxigenic strains in vivo, and thus improve the safety of vaccine strains. It is documented that nontoxigenic V. cholerae strains may become toxigenic strains by reacquirement of CTX. The probability of CTX conversion of live vaccine strain would not be negligible in that new emerging toxigenic strains may originated from the source of nontoxigenic strains. In this study, CTX particles were artificially mixed at high concentration with vaccine strain IEM109 to evaluate the infection rate of the CTX phage to it. Only extremely few infected colonies were detected (9.85 × 10−11 ), indicating a low probability for IEM109 to reacquire cholera toxin genes, even more so. It goes without saying the actual low level of CTX in vivo and in vitro environments, in their natural environment where concentrations of CTX particles are much lower. CTX particles present in IEM109 occurred in a plasmid form (data not shown). This plasmid form is prone to exclusion in animal intestines [28]. The strains containing CTX as a plasmid form may be diluted during reproduction either in vivo or in vitro and may reduce the infection rate for IEM109 even further. All these data indicate that IEM109 is efficient in resisting the infection of CTXET  and safe in acquiring hazardous genes either in the intestinal environment or in vitro as a vaccine candidate. Both antibacterial and antitoxic immunities are important for the exclusion of V. cholerae. In addition, mucosal colonization of V. cholerae is essential for eliciting a long and effective immune response. Our experiments show that IEM109 can effectively colonize intestines after one dose of vaccination (107 CFU). The introduction of the ctxB gene into the IEM109 chromosome results in the strain eliciting antitoxic response as well as antibacterial immunity. High titers of anti-CTB IgG were detected in peripheral blood, indicating that the ctxB gene integrated on the chromosome of IEM109 elicits antitoxic immunity in rabbit models. Rabbits immunized with IEM109 produced stronger protection than those immunized with IEM101 against the challenge of pure cholera toxin and live toxigenic strains. Furthermore, the titer of vibriocidal antibody elicited by IEM109 is markedly higher than that of IEM101, suggesting that the antibacterial and antitoxic immunities act synergistically by interfering with two separate events in cholera pathogenesis. Similar observations were made in previous studies [17,29,30]. Levine and Kaper [6] reported that the vibriocidal response of serum was correlated with the protective reaction of intestinal mucus. The stronger the vibriocidal response of serum elicited by live oral vaccine, the more effective the

4054

M. Yan et al. / Vaccine 25 (2007) 4046–4055

protective reaction of intestinal mucous. The active protection against wild virulent V. cholerae in the rabbit model suggested the synergistic action. In the RTX cluster genes of V. cholerae, rtxA encodes cytotoxin and rtxC encodes cytotoxin activator [11,12]. The intact RTX cluster is essential for the cytotoxic response in cell culture models for the El Tor and O139 strains. The RTX toxin of V. cholerae is different from other enteric bacterial pore-forming toxins in the mechanism of cell rounding. The toxin causes depolymerization of actin stress fibers and covelant cross-linking of cellular actin into dimmers, trimers and multimers [12]. Aside from HAP and hemolysin the RTX toxin can be one factor inducing reactogenicity. The toxin also contributes to acute inflammation response in mouse models. The serum level of proinflammatory factors such as IL-6 and MIP decreased in mice inoculated with strains containing a knocked out rtxA gene and significantly improved clinical symptoms [13]. Less reactogenicity was also observed in some RtxA inactivated vaccine strain such as CVD103, O395-N1 [16]. In the construction of IEM109, the RTX genes were deleted. Thereby, we obtained a candidate that evokes antitoxic and vibriocidal immunities and resistance to CTX infections. In summary, we developed a new vaccine candidate, IEM109, against V. cholerae by replacing a fragment containing the TLC element, the attB site, and the RTX cluster in the chromosomal DNA of IEM101 with the rstR and ctxB genes. Compared with IEM108 we developed before, attB and RTX gene cluster were removed and ctxB and rstR genes were integrated into chromosome DNA. IEM109 cannot only overcome horizontal gene transfer but also enhance the stability of the newly introduced rstR and ctxB genes. Thus, it evokes antitoxic and vibriocidal immunities as well as resistance to CTX infections. The disappearance of the cytotoxic activity RTX in IEM109, compared to IEM101 and IEM108, may improve the safety of this vaccine. The actual protection and reactogenicity of IEM109 now should to be assessed in a field trial.

Acknowledgement This work was supported by grant of the High Tech Research and Development Program (2001AA215191) from the Ministry of Scientific Technology, PR China.

References [1] Albert MJ, Siddique AK, Islam MS, Faruque AS, Ansaruzzaman M, Faruque SM, et al. Large outbreak of clinical cholera due to Vibrio cholerae non-O1 in Bangladesh. Lancet 1993;341:704. [2] Marc PG, Duncan S, Claire LC, Marie PK. A review of vaccine research and development: human enteric infections. Vaccine 2006;24:2732– 50. [3] Richard IW. Considerations for development of whole cell bacterial vaccines to prevent diarrheal diseases in children in developing countries. Vaccine 2005;23:3369–85.

[4] Kenner JR, Coster TS, Taylor DN, Trofa AF, Barrera-Oro M, Hyman F, et al. Peru-l 5, an improved live attenuated oral vaccine candidate for Vibrio cholerae Ol. J Infect Dis 1995;172:1126–9. [5] Kotloff KL, Wasserman SS, O’Dounad, Losonsky GA, Cryz SJ, Levine MM, et al. Safty and immumogenicity in North Americans of a single dose of live oral cholera vaccine CVD103-HgR: results of a randomized, placebo-controlled, double-blind crossover trial. Infect Immun 1992;60:4430–2. [6] Levine MM, Kaper JB. Live oral cholera vaccines: from principle to product. Bull Inst Pasteur 1995;93:243–53. [7] Taylor DN, Killeen KP, Hack DC, Kenner JR, Coster TS, Beattie DT, et al. Development of a live, oral, attenuated vaccine against El Tor cholera. J Infect Dis 1994;170:1518–23. [8] Tacket CO, Kotloff KL, Lonsonsky G, Nataro JP, Michalski J, Kaper JB, et al. Volunteer studies investigating the safety and efficacy of live oral El Tor Vibrio cholerae O1 vaccine strain CVD 111. Am J Trop Med Hyg 1997;56:533–7. [9] Richie EE, Punjabi NH, Sidharta YY, Peetosutan KK, Sukandar MM, Wasserman SS, et al. Efficacy trial of single-dose live oral cholerav accine CVD 103-HgR in North Jakarta, Indonesia, a cholera-endemic area. Vaccine 2000;18:2399–410. [10] Cohen MB, Giannella RA, Bean J, Taylor DN, Parker S, Hoeper A, et al. Randomized, controlled human challenge study of the safety, immunogenicity, and protective efficacy of a single dose of Peru-l5: a live aftenuated oral cholera vaccine. Infect Immun 2002;70:1965–70. [11] Wei L, Karla JF, Rebecca C, Jessica AS, Marc BR, Kerstin EC, et al. Identification of a Vibrio cholerae RTX toxin gene cluster that is tightly linked to the cholera toxin prophage. Proc Natl Acad Sci USA 1999;96:1071–6. [12] Karla JF, John JM. In vivo covalent cross-linking of cellular actin by the Vibrio cholerae RTX toxin. EMBO J 2000;19:5315–23. [13] Fullner KJ, Boucher JC, Hanes MA, Haines III GK, Meehan BM, Walchle C, et al. The contribution of accessory toxins of Vibrio cholerae O1 El Tor to the proinflammatory response in a murine pulmonary cholera model. J Exp Med 2002;195:1455–62. [14] Viret JF, Dietrich G, Favre D. Biosafety aspects of the recombinant live oral Vibrio cholerae vaccine strain CVD 103-HgR. Vaccine 2004;22:2457–69. [15] Waldor MK, Mekalanos JJ. Lysogenic conversion by a filaemtous phage encoding cholera toxin. Science 1996;272:1910–4. [16] Ketley IM, Kaper JB, Herrington DA, Losonsky C, Levine MM. Diminished immunogenicity of a recombination-deficient derivative of Vibrio cholerae vaccine strain CVD103. Infect Immun 1990;58:1481–4. [17] Liang WL, Wang SX, Yu FG, Qi GM, Kan B. Construction and evaluation of a safe, live, oral Vibrio cholerae vaccine candidate IEM108. Infect Immun 2003;71:5498–504. [18] Liu YQ, Qi GM, Wang SX, Yu YM, Duan GC, Zhang LJ, et al. A natural vaccine candidate strain against cholera. Biomed Environ Sci 1995;8:350–8. [19] Liu GW, Yan MY, Liang WL, Qi GM, Liu YQ, Gao SY, et al. Resistance of the cholera vaccine candidate IEM108 against CTX infection. Vaccine 2006;24:1749–55. [20] Kimsey HH, Waldor MK. CTX immunity: application in the development of cholera vaccines. Proc Natl Acad Sci USA 1998;95:7035–9. [21] Attridge SR. Thymine auxotrophy as an attenuating marker in Vibrio cholerae. Microb Pathog 1995;19:11–8. [22] Ryan ET, Crean TI, Kochi SK, John M, Luciano AA, Killeen KP, et al. Development of a Delta glnA balanced lethal plasmid system for expression of heterologous antigens by attenuated vaccine vector strains of Vibrio cholerae. Infect Immun 2000;68:221–6. [23] Valle E, Ledon T, Cedre B, Campos J, Valmaseda T, Rodriguez B, et al. Construction and characterization of a nonproliferative El Tor cholera vaccine candidate derived from strain 638. Infect Immun 2000;68:6411–8. [24] Terry DC, Andrew JM, Randall KH. A new mobilizable cosmid vector for use in Vibrio cholerae and other gram-negative bacteria. Gene 1995;153:85–7.

M. Yan et al. / Vaccine 25 (2007) 4046–4055 [25] Donnenberg MS, Kaper JB. Construction of an eae deletion mutant of enteropathogenic Escherichia coli by using a positive-selection suicide vector. Infect Immun 1991;59:4310–7. [26] Maria RF, Elisabetta M, Liu YQ, Qi GM, Duan GC, Rino R, et al. IEM101, a naturally attenuated Vibrio cholerae strain as carrier for genetically detoxified derivatives of cholera toxin. Vaccine 2001;19:75–85. [27] Rubin EJ, Lin W, Mekalanos JJ, Waldor MK. Replication and integration of a Vibrio cholerae cryptic plasmid linked to the CTX prophage. Mol Microbiol 1998;28:1247–54.

4055

[28] Faruque SM, Rahman MM, Hasan AK, Nair GB, Mekalanos JJ, Sack DA, et al. Diminished diarrheal response to Vibrio cholerae strains carrying the replicative form of the CTX genome instead of CTX lysogens in adult rabbits. Infect Immun 2001;69:6084–90. [29] Svennerholm AM, Holmgren J. Synergistic protective effect in rabbits of immunization with Vibrio cholerae lipopolysaccharide and toxin/toxoid. Infect Immun 1976;13:735–40. [30] Bondre VP, Sinha VB, Srivastava BS. Evaluation of different subcellular fractions of Vibrio cholerae O139 in protection to challenge in experimental cholera. FEMS Immunol Med Microbiol 1998;19:323–9.