Correspondence: William F. Wade, Department of Microbiology and Immunology, Geisel School of Medicine at Dartmouth, Lebanon, NH 03756, USA. Tel.: +1 603 650 6896; fax: +1 603 650 6223; e-mail: firstname.lastname@example.org
The number of cholera vaccine doses required for immunity is a constraint during epidemic cholera. Protective immunity following one dose of multiple Vibrio cholerae (Vc) colonization factors (Inaba LPS El Tor, TcpA, TcpF, and CBP-A) has not been directly tested even though individual Vc colonization factors are the protective antigens. Inaba LPS consistently induced vibriocidal and protective antibodies at low doses. A LPS booster, regardless of dose, induced highly protective secondary sera. Vc protein immunogens emulsified in adjuvant were variably immunogenic. CBP-A was proficient at inducing high IgG serum titers compared with TcpA or TcpF. After one immunization, TcpA or TcpF antisera protected only when the toxin co-regulated pilus operon of the challenge Vc was induced by AKI culture conditions. CBP-A was not consistently able to induce protection independent of the challenge Vc culture conditions. These results reveal the need to understand how best to leverage the ‘right’ Vc immunogens to obtain durable immunity after one dose of a cholera subunit vaccine. The dominance of the protective anti-LPS antibody response over other Vc antigen antibody response needs to be controlled to find other protective antigens that can add to anti-LPS antibody-based immunity.
Vibrio cholerae (Vc) causes cholera, a diarrheal disease that is endemic in some parts of the world. Perhaps, more concerning is the large-scale cholera epidemics that occur in fragile populations comprised of Vc-antigen naïve individuals. Epidemic cholera recently caused the deaths of over 5000 Haitians between October 2010 and July 2011 (WHO, 2011). Endemic cholera as in the delta areas of south Asia or parts of Africa is an annual crisis (Deen et al., 2008). The cholera case load in Africa has become so significant, some consider Africa the new home for endemic cholera rather than the Indian subcontinent and Bangladesh, the historic home of endemic cholera (Deen et al., 2008). Cholera in both forms results in the reported death of many thousands of people a year, which most health care agencies conclude does not represent the true number of cases (Zuckerman et al., 2007). Whatever the statistic, children between 2 and 5 years of age sustain the major burden of cholera (Deen et al., 2008).
The cholera outbreak in Haiti refreshed the scientific debate as to what constitutes an effective cholera vaccine for those individuals naïve to Vc antigens (Ivers et al., 2010; Chao et al., 2011; Cyranoski, 2011). The licensed, WHO-approved, oral cholera vaccine Dukoral (and now Shanchol) is provided in a two dose regimen that features killed, classical, and El Tor biotypes of Vc that express either the Inaba or Ogawa LPS serotype (Anonymous, 2010; Lopez-Gigosos et al., 2011). A major target of the immune response to oral cholera vaccines is LPS (reviewed in Provenzano et al., 2006). Antibodies that bind LPS are vibriocidal and also protective as evidenced by small field trials conducted some 40 years ago (Wasserman et al., 1994). The protective potential of vibriocidal anti-LPS antibodies in human cholera is contested by some, but vibriocidal antibodies are considered the best correlate of cholera immunity (Eubanks et al., 1977; Glass et al., 1985; Saha et al., 2004). Anti-LPS antibodies are protective for infant mice and rabbits where they are thought to function in reducing access to preferred replication niches (Svennerholm, 1975; Mukhopadhyay et al., 2000; Vijayashree et al., 2003; Bishop et al., 2010).
Many experimental cholera vaccines use multiple (three or more) immunizations to demonstrate immunogenicity and protective efficacy of Vc antigens (Wu et al., 2001; Meeks et al., 2004; Taylor et al., 2004; Acevedo et al., 2009; Perez et al., 2009). Intranasal immunization of mice with two doses of either whole-cell Vc or Vc outer membrane vesicles (OMV) can induce protection against challenge (Nygren et al., 2008; Bishop et al., 2010). Bishop et al. hypothesized protection to be mainly due to anti-LPS IgG1 antibodies, the predominant isotype transferred to infant mice in utero and during suckling. If more than two doses are required for robust cholera immunity, then these formulations will not be practical in the field for contesting epidemic cholera that really demands a one dose vaccine.
The utility of cholera toxin (CT) antibodies induced by infection or by some oral cholera vaccines has been debated for years. Killed oral cholera vaccines with CT do not perform better (long-term protective efficacy) than cholera vaccines without CT B (CTB) subunit, but greater short-term (8 months) protection was seen in the group that received CTB as part of the cholera vaccine (Clemens et al., 1988). The licensed Vietnamese killed cholera vaccine, mORCVAC, and another vaccine, Shanchol, the new entry for killed cholera vaccines do not contain CTB, yet both are proven to provide immunity against cholera (Trach et al., 2002; Mahalanabis et al., 2008). B cell memory for CT is present in those acutely infected with Vc suggesting that anti-CT antibodies are either not protective or need to be at high concentrations for efficacy (Sack et al., 1991; Harris et al., 2009). The role of CT-specific antibody at protecting against or moderating cholera in the short term may be considered for epidemic cholera vaccines as CT antibodies prevent acute disease in a closed system (ligated ileal loops) in rabbits that are a very good model of diarrhea that causes cholera in humans (Kundu et al., 2009). The role of CT antibodies in preventing Vc colonization in humans is not known, and thus, we did not examine CT as one of the immunogens in this study as we wanted to focus on antibodies that target Vc colonization, not downstream events.
A major focus of research on Vc pathogenesis has highlighted the colonization factors that Vc uses to secure a replication niche in humans and experimental animals. The outer membrane component, LPS, is well known for its role in the pathogenesis of Gram negative bacteria. In addition to its barrier function, Vc LPS plays a role in adhesion to and subsequent colonization of the gut (Chitnis et al., 1982; Mukhopadhyay et al., 2000; Vijayashree et al., 2003). The major colonization factors identified for mice are components of the toxin co-regulated pilus (TCP) operon (TcpA, TcpF) and the chitin response genes (e.g. CBP-A) required for Vc association with N-acetyl glucosamine-marked structures of aquatic and human hosts (Kirn & Taylor, 2005; Kirn et al., 2005). (Note, we use CBP-A to designate the chitin binding protein product VCA0118; GBP-A is used by some, but since chitin is a polymer of different forms of glucosamine, we elected to use chitin as the modifier as it represents the product that Vc binds in nature.) TcpF and CBP-A are both secreted proteins that enhance Vc's colonization of infant mice (Kirn & Taylor, 2005; Kirn et al., 2005). The antibody response of humans to Vc LPS, and to a lesser extent TcpA, is described (Attridge et al., 2004; Asaduzzaman et al., 2004; reviewed in Provenzano et al., 2006). The major component of classical and El Tor TCP, TcpA, is required for effective colonization and is immunogenic for humans during natural infection (Hall et al., 1991; Asaduzzaman et al., 2004; Attridge et al., 2004). The protective potential of rabbit anti-TcpA (classical or El Tor) was reported (Sun et al., 1990). The protective capacity of human El Tor TcpA-specific antibodies in an experimental or natural Vc challenge of humans is not reported, nor is the immunogenicity of TcpF and CBP-A for humans apparent in the biomedical literature.
The initial studies of the protective capacity of Vc TcpF and CBP-A used rabbits as host for antibody generation (Kirn & Taylor, 2005; Kirn et al., 2005). Rabbits are substantially different from humans in the manner they achieve diversification of their mature antibody repertoire. It is not clear whether the antibody response of rabbits to Vc antigens is predictive of how humans will respond especially to fewer immunizations that define the need of epidemic cholera vaccination. The issue of a one dose cholera vaccine in mind, we investigated the adaptive antibody response to four Vc colonization factors: LPS, TcpA, TcpF, and CBP-A. We used mice as host for antibody production as they diversify their antibody repertoire using the same genetics mechanisms (random exon recombination and somatic mutation of assembled variable gene segments) as do humans (Mage et al., 1999; Lanning et al., 2000). We tested antisera obtained 21 days after the initial immunization and at a similar interval after a booster. The rationale for the booster was if initial immunization was promising (some efficacy at protecting infant mice), we wanted to know whether another immunization with the same antigens would significantly improve the protective efficacy of the antisera. While the Vc protein colonization factors were immunogenic, the protective efficacy of induced antibodies was dependent on the culture conditions of the Vc challenge bacteria. The partial protection provided by CBP-A antisera was not reflected in sera of mice immunized with all three Vc protein colonization factors or animals boosted with CBP-A. There was an additive effect of LPS-specific antibodies and Vc protein colonization factor-specific antibodies after one dose of the three Vc protein immunogens if the challenge Vc for the challenge inoculum was grown in Luria–Bertani (LB) at 30 °C but not if the challenge bacteria were grown under AKI culture conditions, which results in very high CT expression (Iwanaga & Yamamoto, 1985; Iwanaga et al., 1986). Our data advise more studies in mice to rank the protective efficacy of other Vc antigens such as Vc's flagella antigens, but more critical is the need to demonstrate that humans can use these Vc colonization factor to generate antibodies that will be effective in the context of a natural cholera infection.
Material and methods
Bacterial strains used for these studies are listed in Table 1. Glycerol stocks were maintained at −80 °C.
Table 1. Vc strains and plasmids used in this study
Vibrio cholerae, El Tor, Inaba
Vibrio cholerae, classical, Inaba
R. Taylor, Dartmouth Medical School
Super Competent E. coli cells
Vector that carries an N-terminal His-Tag® sequence
Competent E. coli cells
His-tagged ET TcpA. clone with pET15b/tcpA1
R. Taylor, Dartmouth Medical School (Rollenhagen et al., 2006)
Vc immunogen cloning and purification
El Tor protein colonization factors, TcpF and CBP-A used in this study were cloned from N16961 DNA using the following oligonucleotides: TcpF 5′-GCGCTCGAGGATATGAGATATAAAAAAACCTTAATG-3′ and 5′-GCCGGATCCTCTTATTTAAAGTTCTCTGAATATGC-3′; CBP-A 5′-CGCTCGAGGATATGAAAAAACAACCTAAAATGACCGC-3′ and 5′-GCCGGATCCTCTTAACGTTTATCCCACGCCATTTC-3′. Vc genomic DNA encoding the Vc colonization factors was amplified, digested with the appropriate restriction enzymes, and ligated into the pET-15b (Novagen) vector containing an N-terminal His tag®. The resulting constructs were transformed into XL-1 Blue competent cells (Agilent), and the fidelity of the Vc colonization factor sequences was verified by DNA sequencing. Competent Origami2 (DE3) cells (Novagen) were transformed with plasmid DNA for the individual Vc protein colonization factors. Escherichia coli expressing El Tor TcpA cDNA were provided by R. Taylor, Dartmouth Medical School (Rollenhagen et al., 2006). His-tagged immunogens were purified with a Ni-column following the manufacturer's instructions (GE Healthcare). Inaba LPS was purified from Vc strain 569B using the hot phenol–chloroform method as described (Chernyak et al., 2002). Extracted LPS was treated with DNase, Rnase, and proteinase K to reduce contaminating molecular species.
Female BALB/c mice were purchased from the National Cancer Institute (Bethesda, MD) for the generation of antisera. Untimed, pregnant CD-1 mice were purchased from Charles River (Raleigh, NC) for the infant mouse assay. Mice were housed at the Animal Resource Center at Dartmouth-Hitchcock Medical Center that is accredited by the American Association for Accreditation of Laboratory Animal Care.
For all experimental and control formulations, mice were immunized twice intraperitoneally (i.p.) on days 0 and 21 using the MPL® + TDM Adjuvant system (formally known as RIBI®) from Sigma (Table 2). Blood was collected via retro-orbital puncture on days 0 (prebleed) and 21 (primary bleed). Mice were ensanguined on day 42 (secondary bleed). Sera were pooled for the protection assay, but individual sera were analyzed for endpoint titers (ELISA and vibriocidal).
Mice were inoculated i.p. with the Vc immunogens emulsified in RIBI adjuvant.
30 μg TcpA
15 μg TcpA
30 μg TcpF
15 μg TcpF
30 μg CBP-A
15 μg CBP-A
30 μg each (TcpA + TcpF + CBP-A)
15 μg each (TcpA + TcpF + CBP-A)
0.1 μg LPS
0.1 μg LPS
0.5 μg LPS
0.5 μg LPS
5.0 μg LPS
5.0 μg LPS
0.1 μg LPS + 30 μg each (TcpA + TcpF + CBP-A)
0.1 μg LPS + 15 μg each (TcpA + TcpF + CBP-A)
0.5 μg LPS + 30 μg each (TcpA + TcpF + CBP-A)
0.5 μg LPS + 15 μg each (TcpA + TcpF + CBP-A)
5.0 μg LPS + 30 μg each (TcpA + TcpF + CBP-A)
5.0 μg LPS + 15 μg each (TcpA + TcpF + CBP-A)
The protein colonization antigens TcpA, TcpF, CBP-A, or purified Inaba LPS were bound overnight to ELISA plates in carbonate buffer, and the mean antigen-specific antibody endpoint titers were determined as previously described (Chernyak et al., 2002; Meeks et al., 2004; Wade et al., 2006). Endpoint titers are the reciprocal of the antibody dilution for the last positive well (optical density, OD) after subtracting twice the average optical density of the negative controls (prebleed) wells. Optical densities were measured using a Dynatech Laboratories MRX microplate reader. The positive control sera (PCS) for these studies came from two sources. PCS-1 was pooled from mice immunized i.p. (28 days apart) with Inaba LPS. The first two doses were 9 µg of Inaba LPS, followed by two 5 µg doses. Twelve days after the final 2.5 µg LPS booster that was given after 3 months of rest, mice were exsanguinated and sera pooled. The ELISA endpoint titers of PCS-1 were 102400 (IgM) and 2560 (IgG) and 5120 for the vibriocidal assay. PCS-1 was used for ELISA and vibriocidal assays and for protection assays depicted in Figs 1d, e, 2c and 3d. When the PCS-1 was exhausted, PCS-2 was comprised of pooled BALB/c mouse sera similarly to PCS-1 sera but without the final booster and immunization were 21 days apart not 28. The ELISA endpoint titers for PCS-2 were 25600 (IgM) and 2560 (IgG), while the vibriocidal endpoint titer was 320. PCS-2 was used for the protection assay controls described in Figs 2d and 3e. The negative control sera (prebleed, pb) for ELISA were individual serum samples from mice in the various groups before they were immunized with Vc immunogens. The pb sera controlled for nonspecific immunity of mice in the experimental groups that were later immunized with various Vc antigens. The negative control serum titer for ELISA analysis was assessed as the baseline dilution < 200 (reciprocal of the starting serum dilution) of the pb serum. The pb starting dilution never shows any signal (OD) at the baseline dilution. The vibriocidal titer of the pb sera was < 40.
Vibriocidal microtiter assay
To measure vibriocidal activity, we used the microtiter test developed by Fournier's group that is described in detail (Boutonnier et al., 2003; Meeks et al., 2004). The microtiter vibriocidal assay is based on the effect of complement-mediated lysis of metabolically active bacteria. Metabolically active bacterial cells render the colorimetric substrate to a violet color in wells with growing bacteria. Inhibition of bacterial growth (vibriocidal capacity) by mouse antisera is reported as the reciprocal of the antibody dilution of the negative well containing the lowest concentration of antibody. A titer > 40 was considered vibriocidal. The negative control included Vc N16961 bacteria and complement only; the PCS were described previously and generated to purified Inaba LPS. The positive control titer for vibriocidal anti-LPS antibody was either 5210 or 320 depending on the source of the control antisera that were used. The negative control (pb) titer was the baseline dilution < 40 (reciprocal of the serum dilution). The pb titers never registered OD units when tested at the baseline dilution.
Two different methods were used to grow Vc for use in the vibriocidal assay. For vibriocidal assays that did not use AKI media to grow Vc, an L-agar plate was streaked with N16961 (ATCC # 39315) and grown overnight at 37 °C. One colony was inoculated into 2 mL alkaline peptone water media (16 × 125 mm tube) and incubated overnight at 37 °C with shaking. Two ml of the overnight culture suspension was placed onto a prewarmed (37 °C) L-agar plate for 1 min to adsorb the liquid. After 90 min at 37 °C the plate was rinsed with 5 mL cold, sterile PBS, the plate swirled to resuspend the cells for transfer to a 15-mL centrifuge tube to adjust the bacterial cell concentration to an OD600 nm = 0.9 with PBS to approximate 109 cells mL−1. The AKI method of growing Vc (Iwanaga et al., 1986) was used for some experiments and involves media formulated as follows: 1.5% Bacto peptone, 0.4% yeast extract, 0.5% NaCl, 0.3% NaHCO3. Three 13 × 100 mm tubes containing 5 mL freshly made AKI media were inoculated with one colony of N16961 (ATCC 39315) from a streaked L-agar plate grown overnight at 37 °C. The tubes were placed in a 30 °C incubator, and the bacteria were grown without shaking until the OD600 nm was 0.1 and then pooled together and 12.5 mL was placed into a 125-mL Erlenmeyer flask and grown at 30 °C with shaking until the OD600 nm = 0.9. The bacterial cells were collected by centrifugation and resuspended in ice cold PBS. This preparation was then added to the guinea pig complement and then to the heat-inactivated antisera. From this point on, the procedure was identical and followed the established protocol.
Infant mouse challenge assay
Three- to 5-day-old CD-1 neonates were used in the cholera challenge model to assess the protective quality of anti-Vc colonization factor antibodies (Chernyak et al., 2002; Taylor et al., 2004). For some experiments, cultures of Vc N16961 were grown for 17 h in 2 mL of LB broth, pH 6.5, at 30 °C on a rotating wheel. Twenty-five microlitres of bacteria representing 2–12 LD50 were combined with 25 μL of pooled prebleed (pb) serum, pooled anti-Inaba LPS hyperimmune serum, or pooled primary or secondary antiserum diluted 1 : 10 with normal heat-inactivated mouse serum (AbSerotec) immediately before intragastric administration to the infant mice. Challenged mice were kept at 30 °C and monitored every 4 h, beginning 24 h after challenge for clinical signs of cholera: dehydration, loss of vitality, and diarrhea (Taylor et al., 2004).
In other studies to hyperinduce TcpA and TcpF expression, we used AKI to grow Vc for the challenge assays. Vc was grown in AKI broth for 3 h without shaking and then transferred into culture tubes and incubated at 30 °C on a rotating wheel for 17 h (Iwanaga et al., 1986). The growth of Vc at 30 °C in LB and AKI at two temperatures is both compatible with Vc colonization, but different numbers of bacteria in the challenge inoculum give different results (W.F. Wade, pers. obs.). The amount of AKI Vc grown can either be normalized for virulence or cell number to Vc grown in LB. The number of LB-grown bacteria (≈ 107) if compared to the same number of AKI-grown Vc is much less virulent with 50% (LB grown) of the mice dying during the test periods compared with 100% (AKI grown). If mice are normalized for the same percent death (50%) over the observation period, then 1–3 × 106 AKI-grown Vc is equivalent to 107 LB-grown Vc (W.F. Wade, pers. obs.).
Antibody titers were scored as endpoint titers for individual serum sample for all groups. Sera that did not generate a signal in the ELISA at the starting dilutions were considered negative, but designated as 1 for the endpoint titer, so log10 transformation could be performed and means and standard deviations calculated that were analyzed by analysis of variance (anova) and Tukey's posttest comparison both of which are available on graphpad prism software (Version 5; GraphPad Software, San Diego, CA). The survival curves were also generated using graphpad prism software that uses the Kaplan–Meier method to generate curves. To test the null hypothesis that survival curves did not differ, the Mantel–Haenzel test was used. P values > 0.05 are not considered significant.
LPS is a dominant immunogen for mice inducing vibriocidal antibodies after one dose with as little as 0.5 µg of purified Inaba LPS
We used purified Inaba LPS emulsified in RIBI® adjuvant to induce serum (IgG or IgM) and vibriocidal antibodies in BALB/c mice. Serum endpoint titers were measured after one or two i.p. immunizations. We chose i.p. immunization to be consistent with our previous studies of the immunogenicity of Vc LPS (Chernyak et al., 2002; Meeks et al., 2004; Wade et al., 2006), and further, the developing literature indicates both marginal zone and B1 B cells that are activated by i.p. introduced antigen are responsive to LPS (Oliver et al., 1997; Cole et al., 2009). Responses (IgM) measured at 21 days postpriming indicated that LPS is immunogenic in a dose-dependent manner with the higher dose of 5.0 µg inducing a response in 80% of the mice, whereas lower LPS doses were less able to invoke serum anti-LPS antibodies (Fig. 1a). The mean anti-Inaba LPS IgM responses after the booster were variably increased. Regardless of the priming LPS dose, the anti-LPS IgG response (one outlier aside) was not a constant feature until after the booster immunization and the use of the higher doses of LPS (Fig. 1b). The pairwise comparisons that are significant are listed in the legend of Fig. 1a and b. The vibriocidal activity of the various Inaba LPS antisera was assessed in both primary and secondary sera (Fig. 1c). The higher the immunizing dose of LPS, the greater the percentage of mice in the group responding with anti-Inaba LPS vibriocidal antibody was. The LPS booster increased the serum vibriocidal endpoint titers at day 42, which was quantitatively reflected in all the groups. The pairwise comparison of the vibriocidal titers of the primary or secondary sera is reported in the legend (Fig. 1c).
The infant mouse protection assay basically assesses colonization potential and uses virulent Vc admixed with antibodies that are gavaged into the stomach of individual pups (Bellamy et al., 1975; Taylor et al., 2004). N16961 grown in LB at 30 °C can be found in neonatal mouse intestines 1 h postgavage (Angelichio et al., 1999). This represents a small percentage of the input inoculum, and it takes between 10 and 24 h for Vc replication (following colonization) that is TCP dependent for the number of Vc in the small intestines to exceed the input by about 3 logs. Pooled primary or secondary antisera from groups of mice immunized with Inaba LPS were analyzed for their protective capacity (Fig. 1d). None of the different group's primary anti-LPS sera were protective compared to the positive control anti-Inaba LPS. A booster immunization, regardless of dose, produced LPS antisera that protected 90–100% (statistically significant) of the challenged mice (Fig. 1e).
Individual Vc protein colonization factors are variably immunogenic and less effective at producing protective antibodies than purified Inaba LPS
Column-purified, El Tor TcpA, TcpF, or CBP-A were used to immunize mice twice with a decreasing amount of antigen emulsified in RIBI® adjuvant. The primary and secondary IgM response to the individual immunogens showed a hierarchy of serum titers with TcpA titers being lower than TcpF or CBP-A (Fig. 2a). A booster with an individual protein colonization factor increased the anti-Vc protein colonization factor secondary IgM antibody responses with the exception of the CBP-A response. In general, the antigen-specific IgG titers for protein colonization factors paralleled the IgM titers, with the exception that the TcpA IgG response was apparent after a single dose (Fig. 2b, open squares). The IgG response for CBP-A immunogen was significantly higher after the booster with the titer close to 1 : 100 000 000. The TcpA and TcpF IgG titers after the booster were not significantly different from each other (shaded squares vs. shaded triangles) but were different from the corresponding group titers or primary antisera.
Pooled antisera (primary and secondary) from mice immunized with individual protein colonization factors were tested for their protective efficacy against challenge of approximately 10 LD50 of N16961 (Fig. 2c and d). Primary antisera generated against CBP-A (P = 0.0018, triangles) were partially protective for infant mice, but the secondary generated to any of the Vc protein colonization factors used at the same dilution, albeit against a slightly increased lethal challenge dose, was not (Fig. 2d). Antibodies in TcpA primary antisera did not protect well enough to be statistically significant (P = 0.074) at a P value cutoff of 0.05, but there was a trend for mice to survive more readily (Fig. 2c). None of the secondary sera provided protection in the infant mouse model where the challenge dose (number of Vc cells) was slightly higher, and thus, more bacterial cells might have masked the protective effect if the amount of protective antibodies was marginal (Fig. 2d). The reason for the seemingly lower survival rates of some of the experimental groups compared to that seen in the pb sera group is not known and in general was not significant (Fig. 2d). Only the CBP-A immunized group survival was significantly different. The inoculation order of the groups was randomized suggesting that the differences in survival are not associated with a methodological error. We do not know what accounts for the seemingly enhanced virulence of Vc for mice immunized with CBP-A. It is more likely that this group's survival represents chance events and is not linked to any important biology of the challenge system or potential detrimental effects of CBP-A-specific antibodies.
Combined protein colonization factors with Vc Inaba LPS induce protection in the infant mouse assay
The additive potential of anti-LPS antibodies and anti-TcpA (or MSHA) antibodies has been reported (Osek et al., 1994). The suggestion of an additive response of TcpA and TcpF antibodies has also been presented (Megli et al., 2011). We immunized mice with four protective Vc immunogens, one also being a toll receptor agonist (LPS) to generate passive antibody to test their additive protective potential. The study design used variable doses of LPS (Fig. 3a = 0.1 µg, Fig. 3b = 0.5 µg, Fig. 3c = 5.0 µg) combined with 30 µg each of individual protein colonization factor antigens for the primary immunization. The booster had the same dose of LPS used for priming, but 15 µg each of the protein colonization factors. The inclusion of multiple immunogens did not grossly affect the relative immunogenicity of the colonization factor's serum endpoint titers (Fig. 3a–c). In general, LPS antibody titers for the mice immunized with the four Vc colonization factors were similar to those obtained for mice immunized with Inaba LPS alone. The lack of a measurable IgG (Fig. 3a) of mice receiving two doses of the four Vc colonization factors that included 0.1 µg of LPS is notable and will be discussed later.
Pooled antisera from mice immunized with different amounts of LPS but the same amount of the three protein colonization factors as in previous groups of mice (Fig. 2) were tested for their additive protective capacity against LB-grown challenge Vc. Primary antisera generated to the multiple Vc protective antigens showed an additive protective effect (Fig. 3d). The booster immunization, with half the mass of the four protein colonization factors and different amounts of LPS, also induced protective antibodies if the Inaba LPS was 0.5 µg or greater (Fig. 3e). Mice immunized with two doses of either 0.5 or 5.0 µg of LPS alone generated secondary antisera that were protective (90–100%), and thus, it is difficult to determine what components of the secondary antisera for mice immunized with multiple colonization factors contributed to protection.
Vibriocidal antibody titers of mice immunized with four colonization factors
To gain additional insight into the protective potential of the antisera mice generated to the four Vc colonization factors, we measured the vibriocidal endpoint titers in sera of mice immunized once or twice with the cocktail of Vc colonization antigens (Fig. 4). Vc LPS as shown in Fig. 1c readily induces vibriocidal antibody in a dose-dependent manner. The vibriocidal endpoint titers for the same groups of mice that responded were not significantly different (Fig. 4a and b, 0.5 µg and 5.0 µg, respectively). This relative magnitude of vibriocidal serum titers did not change if the Vc was grown in AKI media (data not shown).
Protective capacity against AKI-grown Vc of antibodies induced by four Vc colonization-specific antibodies
The lack of protection by anti-TcpA and TcpF antibodies against challenge bacteria grown in LB broth was perhaps not unexpected. While LB-grown Vc can induce disease, the expression of TcpA and CT is not as high as when Vc is cultured under special conditions. AKI media were formulated principally to increase the CT production from El Tor Vc because the El Tor biotype did not produce as much CT as the classical biotype making it difficult to study CT from El Tor Vc. (Iwanaga & Yamamoto, 1985). Subsequent papers that closely examined Vc culture conditions concluded that temperature 30 °C rather than 37 °C was the important component for upregulating virulence factor operons (Mukhopadhyay et al., 1996). Others felt it was bicarbonate that was controlling (Iwanaga & Yamamoto, 1985). It was suggested that AKI was not as relevant as the exposed surface area of the Vc during culture (Sánchez et al., 2004). The prior induction of virulence factors such as TcpA and TcpF is thought to provide increased initial target expression and acquisition for admixed anti-TcpA and anti-TcpF antibodies before the gavaged inoculum reaches the small intestines where TcpA, TcpF, and CBP-A are normally upregulated as would be the case with LB-grown Vc. We reexamined the protective potential of primary and secondary antisera for Vc grown under AKI conditions (Fig. 5). The three levels of LPS used to immunize were very protective providing antibodies that were approximately 75% protection after one immunization with a serum dilution of 1 : 30. Further diluting the secondary anti-LPS sera (1 : 60) showed that 0.1 µg of LPS did not induce as much protective antibody after two doses compared to higher doses of LPS used to immunize (Fig. 5d).
A primary dose of 30 µg of the protein colonization factors TcpA or TcpF (as well as the combination) again induced around 75% in pooled primary sera (Fig. 5b). Primary anti-CBP-A antibody was not effective in inducing protective antibody against AKI-grown Vc. Dilution of secondary antisera reduced the protective effect of anti-TcpA and TcpF colonization factor antibodies (Fig. 5e). The combination of the TcpA, TcpF, and CBP-A did not induce addition protective antibody. There was no additive effect of TcpA and TcpF antibody compared to the antibodies functioning alone. This suggests that either no additive effects in the AKI system or the dilution of antisera to normalize the bacteria and antibody levels reduced the levels of antibodies to nonadditive levels.
The combination of the four colonization factors in the immunogen inoculum induced around 70% protection in primary sera with the exception of pooled sera from mice immunized with the protein colonization factors and 0.1 µg of Inaba LPS (Fig. 5c). The inclusion of either 0.5 µg or 5.0 µg of Inaba LPS did not change the protection compared to mice immunized with the same amount of the protein colonization factors but no Inaba LPS (Fig. 5c, filled triangles). Secondary sera showed intermediate protection of mice receiving the four colonization factor immunogens (Fig. 5f). The best protection with the most dilute sera to approximately 33% more lethal doses of Vc was the group that received the highest amount of Inaba LPS and the three protein immunogens.
The challenge experiments for the primary and secondary sera were performed on different days with different dilutions of sera. This later parameter was manipulated to try to establish conditions where either anti-LPS antibody or antiprotein colonization factor antibody would not be sufficient for optimal protection. Clearly, this goal was achieved for sera of mice immunized with the protein colonization factors and to a lesser extends with the anti-LPS sera. How close we came to not diluting the protective antibodies in a given sera of mice immunized with the four Vc immunogens to a level where the antibodies could not function independently to protect but still be present in enough concentration and participate in an additive effect is not known but should be considered for future studies.
The largely unseen images of epidemic cholera in Africa were brought into stark relief by the kinetic images of Haiti's cholera epidemic. The Haitian's plight may be fresh, but it is not unique with respect to the crises and misery that follows cholera. Haiti's epidemic recharged the debate on how to respond to epidemic cholera. Whether to, and how to deploy the existing killed, oral cholera vaccine for control of the Haitian outbreak was discussed. The Pan American Health Organization did not recommend vaccination initially (http://new.paho.org/hq/dmdocuments/2010/PAHO_position_cholera_vaccination.pdf), but with an increase in Haiti's cholera cases, reactive vaccination was reevaluated and planned for the spring of 2011 but was pushed back to the spring of 2012 after Shanchol was approved by the WHO (Cyranoski, 2011; Hawkes, 2012). Some of the parameters weighed as to vaccinate Haitians or not were the logistics of delivering two doses of a World Health Organization approved cholera vaccine to displaced populations in the needed time frame, the pervasive cost calculus, and the utility of the vaccine distribution design (Enserink, 2010). The decision tree based on these issues would change if a one dose cholera vaccine quickly induced immunity in a broad age range of individuals.
Cholera vaccines today
Dukoral and Shanchol are killed, orally delivered cholera vaccines approved by the WHO (Wade, 2011). They are multicomponent vaccines comprised of intact Vc of different serotypes and biotypes that are killed by either heat or formalin. Dukoral includes CTB, while other oral, killed cholera vaccines, Shanchol and mOCRVAC (Vietnamese version of Shanchol), do not (reviewed in Saha et al., 2004; Wade, 2011). Modified-live, one dose cholera vaccines have been developed and tested (Tacket et al., 1999). Several including Peru-15 are still in development with extensive field trials winding down. The efficacy of Peru-15 and the like will be compared to the field trial results of CVD103-HgR that was developed to have more of the ‘right’ Vc antigens (expressed because of replication) available to the small intestine-associated immune system. The Vc antigens of the modified-live cholera vaccines would produce TcpA, TcpF, and CBP-A and thus may explain why CVD103HgR was able to induce protection in westerners with one dose (reviewed in Chowdhury et al., 2009). Anti-LPS antibodies are a major feature of the human immune response to cholera infection or vaccination, but the immunogenic rank of the available Vc antigens (protective or not) of the different formulations of the oral cholera vaccines is not well known nor a current topic of research. The isolated Vc protein antigens available from killed Vc or modified-live Vc have not been functionally vetted in clinical trials. Which Vc protein antigens might provide the most effective induction of protective immunity may be based on viability or growth conditions. This area of cholera vaccine research is underexplored in human trials.
Oral cholera vaccines were developed in response to the inadequate duration and protective response of some cohorts of humans to parenteral cholera vaccines. Oral immunization presents a set of intrinsic challenges that may not allow complete coverage of all susceptible cohorts until they reach a certain age (Wade, 2011). Therefore, in addition to oral cholera vaccine immunization, other means of mucosal delivery of Vc antigens such as intranasal immunization (Nygren et al., 2008; Bishop et al., 2010) and perhaps parenteral vaccination for those living in endemic areas should be tested in humans (Graves et al., 2010). This suggestion is based on our understanding of the mouse's common mucosal immune system and the parallels of the common mucosal immune system of humans, wherein Vc antigens introduced into the nasal mucosa would be predicted to provide immunity in trans (in the gut mucosa) with respect to effector function (B cell memory and plasma cells). The complementary and simultaneous induction of the nasal-associated immune system with the gut mucosal immune system will increase the effectors (memory B cells or plasma cell precursors) that will track to the gut mucosa (Shahiwala & Amiji, 2008). Further, the anamnestic response of Vc antigen-exposed people can easily be reinvigorated by parenteral immunization that recalls both IgG and mucosal IgA memory B cell responses (Svennerholm et al., 1980). Intranasal delivery of kW-C Vc with additional known protective antigens as subunits (TcpA and TcpF) in a standalone subunit vaccine or attached via chitin to the kW-C may provide auxiliary support for the oral induction of cholera immunity in the young after one dose.
Vc LPS a dominant antigen for mice
The murine response to Vc Inaba LPS we report here again demonstrates the robust immunogenicity of low levels of LPS for mice, and the attending vibriocidal and protective antibodies LPS readily induces (Chernyak et al., 2002; Wade et al., 2006; Wade & Wade, 2008). The reason(s) for the fascicle anti-LPS response following intraperitoneal inoculation of Vc LPS does not occur in mice immunized intranasal (two doses, W-C or OMV) or orally (three doses, OMV) with whole-cell-associated LPS has not been directly evaluated (Nygren et al., 2008; Bishop et al., 2010). A possible explanation is that the different routes of immunization target different subsets of B cells. Intraperitoneal immunization primes all three murine splenic B cell subsets (follicular, marginal zone, and B1) as well as peritoneal B1 B cells. This may be relevant as marginal zone B cells and B1 B cells are evolutionarily adapted for making anti-LPS antibodies (Oliver et al., 1997; Cole et al., 2009). In addition to their more well-known location, the spleen, human marginal zone B cells are present in mucosal tissues such as the subepithelial domes covering Peyer's patches in gut lamina propria (Spencer et al., 1985; Weill et al., 2009). Neither the protective capacity of gut-associated marginal zone B cells nor their developmental limitations are as well studied as splenic marginal zone B cells. The relevance of splenic marginal zone B cells or their gut equivalent for the anti-Vc LPS antibody response in young children is because it takes several years for splenic marginal zone B cells to become functional in humans (Weill et al., 2009). Another potentially relevant B cell subset, circulating CD27+ B cells, takes even longer (5–10 years) than marginal zone B cells (2 years) to mature into the adult repertoire with its attending somatically mutated B cell receptors (Weller et al., 2003). Leung and colleagues assessed several immune response parameters of three groups with clinical cholera (El Tor, Ogawa). Younger Bangladeshi children (3–5 years old) had lower baseline vibriocidal antibody titers than older children and adults (Leung et al., 2011). The protective capacity and the variable domain sequences of the anti-LPS antibodies recovered from in vitro culture of cholera patient peripheral B cells need to be addressed, as well as the B cell phenotype to determine whether marginal zone-like B cells or B1 B cells contribute to the anti-LPS antibody response in humans (Griffin et al., 2011).
Protective antibodies to Vc protein colonization factors – species complications
The original reports of the protective nature of the Vc extracellular protein antigens, TcpA, TcpF, or CBP-A as individual immunogens featured mainly rabbit antisera, which were assessed in the infant mouse protection model (Sun et al., 1990; Kirn & Taylor, 2005; Kirn et al., 2005). The early test immunogens for TcpA (classical or El Tor) were crude pili preparations or synthesized peptides (Sun et al., 1990, 1991; Osek et al., 1994; Taylor et al., 2004). The TcpF and CBP-A antigens were recombinant proteins from classical Vc (Kirn & Taylor, 2005; Kirn et al., 2005).
The goal when using rabbits to generate specific antibodies is to generate a large volume of high affinity antisera that are usually accomplished by multiple (3+) immunizations. The use of rabbits as the host to produce anti-Vc colonization factor antibodies merits consideration. The choice of host to test cholera vaccine candidate antigens should be couched in terms of whether the host and the choice of immunization protocol are consistent with the development of a one dose cholera vaccine. This first round of the rabbit antibody repertoire diversification occurs early in life in the appendix and is independent of cognate antigen. B cell receptor diversity is initiated by bacterial antigens and involves recombination of a few V heavy chain exons that utilize a gene conversion-like mechanism to increase the breadth of the repertoire (Lanning et al., 2000). Somatic point mutations can also occur. The second round of rabbit antibody diversification occurs in germinal centers and uses somatic mutation to further refine the binding attributes (affinity) of antigen-selected B cells (Mage et al., 1999; Lanning et al., 2000).
Mice and humans do not rely on precognate antigen expansion nor gene conversion to diversify their mature B cell repertoire. Their diversification occurs in the sterile bone marrow and uses random recombination of gene segments from diverse families of variable (V), diversity (D), and joining (J) genes in the heavy chain (VDJ) and light chain (VJ only) immunoglobulin loci (Lanning et al., 2000). The two stages of diversification of rabbit B cells ensconces their mature B cell repertoire more rapidly compared to mice which only start to reconfigure the binding sites of immunoglobulin in the germinal centers in a process that usually requires T cell help after B cells have bound and internalized their cognate antigen (Mehr et al., 2004). The different mechanism that rabbits and mice or humans use to shape the mature B cell repertoire may more quickly advance the ‘changes’ in rabbit anti-Vc colonization factor antibodies, and thus, humans might require more immunizations to achieve the selection of protective clones and the functional levels of protective antibodies that rabbits achieve with fewer immunizations. Therefore, we used mice to investigate a one or two dose immunization protocol for the Vc protein colonization factors. The choice of one or two doses represents either the ultimate epidemic cholera vaccine protocol or the current cholera vaccination protocol, respectively.
Evidence for additive affects of antibodies able to bind Vc colonization factors
Reports of the protective efficacy of combinations of Vc antigens (subunit vaccine) tested in animal models are limited. We are unaware of published data that examined the immunogenicity and protective efficacy of the combination of the four Vc colonization factors we used. Osek and colleagues showed that LPS- (either Ogawa or Inaba) and TcpA (either classical or El Tor)-specific antibodies provided additive protection in the infant mouse model (Osek et al., 1994). This was reported as increased percent survival of challenged infant mice over the protection separately provided by different specificity antibodies. Megli et al. (2011) concluded the combination of a suboptimal concentration of TcpA (0.04 mg mL−1) polyclonal antibody with a TcpF (1 mg mL−1) monoclonal antibody enhanced protection against challenge with O395, classical, Ogawa Vc from 60% protection by anti-TcpA antibody alone to 80% for the combination. The species of anti-TcpA antibody in Megli's report was not noted, but the 1991 reference provided described polyclonal rabbit antibody made to different TcpA peptides (three inoculations) that comprise the globular domain of TcpA. Transcutaneous immunization of mice with El Tor TcpA-His (100 µg) did not induce protective anti-TcpA antibodies after two immunizations unless CT (50 µg) was present (Rollenhagen et al., 2006). The reported geometric mean titers measured by ELISA used a kinetic scale (mOD min−1) that does not readily translate to endpoint titers, but sera from mice immunized with TcpA-His did not register on this scale, which may reflect the need for an adjuvant to induce TcpA-specific antibodies. Protection against El Tor Vc challenge of infant mice by anti-CT and anti-TcpA antibodies was statistically significant compared to anti-TcpA antibodies alone. Mother's milk after at least five immunizations was the source of the passive antibody for the challenged neonatal mice.
The ‘functional’ capacity of anti-TcpF monoclonal antibody (1–2 mg mL−1 depending on experiment) used by Megli et al. (2011) is related to affinity and concentration. The protocol used to generate B cells for the anti-TcpF hybridomas, mAb13, used two i.p. doses of 50 µg of TcpF (His tag removed) in adjuvant. Our data support the protective anti-TcpF antibody response reported by Megli et al. as we were able to generate mouse polyclonal anti-TcpF that protected infant mice after one dose of a similar amount of TcpF. Further we showed that TcpA was similarly able to induce protective antisera after only one dose. However, our results did not support an additive response between TcpA and TcpF antibodies. This could be due to the experimental differences involved in diluting the passive antibodies.
Vc biotypes – influences on protection in the gavaged antibody infant mouse
The immunogenicity and protective capacity of the secreted, Vc colonization-aids, TcpF and CBP-A, are not as well studied as TcpA. The limited reports used rabbit immunization regimens to generate antisera against classical biotype, TcpF and CBP-A (Kirn & Taylor, 2005; Kirn et al., 2005). The immunogen forms for the early studies of TcpF and CBP-A were recombinant proteins: GST fusion protein for TcpF and a His-tagged CBP-A suggested that the additional structures for purification do not compromise the potential to induced antibodies to protective epitopes and, therefore, were left in place for our studies (Kirn & Taylor, 2005; Kirn et al., 2005).
The immunization protocols of the early protection studies using TcpF and CBP-A were not revealed making it difficult to estimate affinity or concentration of TcpF- or CBP-A-specific antibodies compared to what we report for one or two immunizations of mice with the El Tor biotype equivalent of the TcpF and CBP-A. We observed protection by primary antisera specific for individual El Tor CBP-A (LB-grown challenge inoculum) and TcpA or TcpF (AKI-grown challenge inoculum), but protection was dependent on the means to grow the challenge bacteria. The ability of CBP-A-specific antibody to provide protection was not consistent for LB-grown Vc and was not seen in mice challenged with Vc grown in AKI media. We do not know the reason for the loss of protection in secondary sera, or the reason primary or secondary sera is not effective against AKI-grown challenge bacteria. In the case of LB-grown Vc, it may be related to the competition between the B cell epitopes on CBP-A (extrinsic and intrinsic B cell epitopes) wherein other antibody targets that do not bind protective epitopes dominate the secondary response. Recent data indicate CBP-A exists as an extended monomer of four domains (Wong et al., 2012). The domain structures are linked to different function such as interaction with Vc or the host. How the different domains are ranked for the induction of protective antibodies is likely important to the diversity of the protective antibodies. In this case, directing antibody responses to domains 1 and 4 (interacts with chitin and mucus and thus prevents colonization) might be expected to be more effective than antibodies to domains 2 and 3 that interact with some structure on Vc (reduce aggregation of bacteria which mediates against serum sensitivity). A consideration for why CBP-A-specific antibodies in primary antibodies do not protect against AKI-grown Vc may be related to the increased virulence of the challenge inoculum and the likely reduced expression of CBP-A relative to TcpA and TcpF.
Availability of Vc targets for protective antibodies
A reasonable explanation for the lack of consistent protective efficacy of antibodies that bind the three Vc protein colonization factors we studied relates to the time frame of target acquisition by antibody that relates to Vc biotype and the media used to grow Vc for the challenge. The original reports of protection afforded by TcpA, TcpF, and CBP-A used passive antibody admixed with virulent classical Vc that was gavaged into infant mice. This system, while well established, may have limitations for specificities of antibodies that are protective against Vc targets that are not well expressed initially because of either the biotype of Vc or the conditions used to grow the Vc inoculum. The bolus of gavaged antibody and the bacteria (grown at 30 °C in LB and pH 6.5) pass into the terminal small intestine within 1 h but the major replication phase occurs later between 4 and 10 h (Angelichio et al., 1999). Gavaged bacterial cells unbound by antibody or unaffected by the antibody initiate infection by traversing the mucus layer that is not easily penetrated by gavaged antibody. If the targets of the gavaged antibodies are expressed at higher levels or by a greater percentage of the Vc population at the epithelial interface, the passive antibodies will not be well positioned for a protective effect. This limitation is overcome by growth of Vc in AKI. The nonphysiologic expression of TcpA and TcpF may provide targets for agglutination of the inoculum that could influence movement to the epithelia and thus reduces the effective inoculum. The increase in CT makes protection perhaps a more difficult protective challenge for passive antibodies especially if the somatic Vc antigen antibody titers are limiting as the CT/ cell can be 60-fold higher in AKI-induced Vc. The infant mouse challenge system is a major issue that confounds cholera vaccine research. The infant mouse challenge model does not inform us as to whether anti-TcpA or TcpF sIgA antibodies are effective at protecting the challenge host if they are displayed in the gut mucosa as compared to admixed with ‘pre-expressed’ antigen targets. Further, the effect on disease (after colonization) of anti-TcpA or TcpF antibodies from serum transudate or newly secreted sIgA is not known. This is an issue that has to be addressed in humans because the timing of TcpA and TcpF expression is controlled by Vc's access to the mucosal epithelia and the expression of the target antigens is also temporally regulated.
Expression of TcpA, TcpF, and CBP-A
The expression pattern of a number of Vc genes during infection of rabbits (ileal ligated-loop-associated Vc), mice (ileal resident Vc), or humans (stool-associated Vc) has been reported (Osorio et al., 2005; Lombardo et al., 2007; Nielsen et al., 2010). In the rabbit ileal loop, TcpA and TcpF gene expression were highest at the mucus and epithelial interface at 4–12 h postinoculation, with TcpA expression being higher and lasting longer than TcpF (Nielsen et al., 2010). The expression of TcpA was lower in the lumen 12 h after inoculation, while TcpF was expressed at baseline levels. The timing of CBP-A expression in vivo is not as well studied as that of TcpA or TcpF. N16961, an El Tor Inaba strain expressing CBP-A, but not a CBP-A deleted N16961 strain, induced thickening of the small intestines of adult mice between 6 and 10 h after inoculation (Bhowmick et al., 2008). Increased mucus production was reported to be induced by isolated CBP-A. CBP-A was not noted to be upregulated in Vc recovered from human stool and was not differentially expressed in Vc recovered from mouse intestines (Lombardo et al., 2007). A comparison of gene expression by classical and El Tor strains of Vc grown in M9 media with additives essential for growth revealed the VieA regulon was dominant in classical Vc, significantly increasing virulence genes compared to the El Tor biotype. CBP-A (VCA0118) was expressed at about 10-fold higher in classical (strain O395) compared to El Tor (strain A1552).
The development of a one dose cholera vaccine must include preclinical testing in animals. The choice of animals for the generation of antisera and the protocol for its generation should be based on the realities of the use of a cholera vaccine in the field that must induce immunity in Vc-antigen naïve humans with a minimal number of immunizations. Mice should be the primary hosts for antibody development and should be immunized with a panel of Vc antigens to first see which are protective by themselves after one immunization and then that panel tested with graded doses of LPS to determine additive affects. Eventually, the utility of Vc vaccine antigens such as TcpA, TcpF, and CBP-A needs to be tested in humans who are challenged with virulent Vc as has been performed for testing the current oral cholera vaccines. This is the only metric to ensure the utility of TcpA or TcpF as candidate cholera vaccine antigens. This approach could be very important as three of the four Vc colonization factors we studied are proteins and, therefore, should not be limited in their immunogenicity by the current state of the immune repertoire of young children who currently do not respond to Vc LPS as do older individuals. We envision partial anti-LPS antibody response combined with other partial responses to protective Vc immunogens such as TcpA and TcpF will provide one dose protection for the cohort that is still the most at risk for cholera.