OUP user menu

Immunogenicity of synthetic saccharide fragments of Vibrio cholerae O1 (Ogawa and Inaba) bound to Exotoxin A

Terri K. Wade, Rina Saksena, Joseph Shiloach, Pavol Kováč, William F. Wade
DOI: http://dx.doi.org/10.1111/j.1574-695X.2006.00143.x 237-251 First published online: 1 November 2006

Abstract

Recombinant exotoxin A (rEPA) from Pseudomonas aeruginosa conjugated to Vibrio cholerae O1 serotype-specific polysaccharides (mono-, di- and hexasaccharide) were immunogenic in mice. Monosaccharide conjugates boosted the humoral responses to the hexasaccharide conjugates. Prior exposure to purified Ogawa lipopolysaccharide (LPS) enabled contra-serotype hexasaccharide conjugates to boost the vibriocidal response, but Inaba LPS did not prime for an enhanced vibriocidal response by a contra-serotype conjugate. Prior exposure to the carrier, and priming B cells with the LPS of either serotype, resulted in enhanced vibriocidal titers if the Ogawa hexasaccharides were used, but a diminished response to the Inaba LPS. These studies demonstrate that the ‘functional’ B cell epitopes on the LPS differ from those of the neoglycoconjugates and that the order of immunization and the serotype of the boosting conjugate can influence the epitope specificity and function of the antisera.

Keywords
  • Vibrio cholerae
  • Cholera vaccine
  • Ogawa O1 LPS
  • neoglycoconjugates
  • synthetic oligosaccharide antigens

Introduction

Vibrio cholerae is a gram-negative bacterium that causes the diarrheal disease cholera. A surface structure, the lipopolysaccharide (LPS) of V. cholerae (O1 serogroup), induces protective humoral immune responses in humans and experimental animals (Mosley, 1969; Qadri et al. 1999; Chernyak et al. 2002; Meeks et al. 2004). Vibrio cholerae-specific, anti-LPS antibodies are linked to protection against cholera (Mosley, 1969). Development of LPS-based epitopes for a cholera subunit vaccine is the focus of several groups (Ariosa-Alvarez et al. 1998; Gupta et al. 1998). Inaba and Ogawa are the two major serotypes of V. cholerae O1 that cause either endemic or epidemic cholera (Manning et al. 1994). Vibrio cholerae O-specific polysaccharide (O-SP) consists of a polymer of (1→2)-α-linked 4-amino-4, 6-dideoxy-d-mannose (perosamine), the amino group of which is acylated with 3-deoxy-l-glycero-tetronic acid (Kenne et al. 1982; Hisatsune et al. 1993). The upstream terminal sugar of the O-SP of V. cholerae LPS differentiates the Ogawa and Inaba serotypes. A 2-O-methyl group (Pykett & Preston, 1975; Ito et al. 1994; Liao et al. 2002; Villeneuve et al. 2002) defines Ogawa LPS (sero-epitope B) while the Inaba terminal sugar with a hydroxyl group at that position is thought to define sero-epitope C (Manning et al. 1994; Liao et al. 2002; Villeneuve et al. 2002). Protective Ogawa-specific mAbs bind the O-SP upstream terminal sugar (Wang et al. 1998; Chernyak et al. 2002). Other LPS structures (core-O-SP junction) that are the same for Ogawa and Inaba LPS can also induce protective antibody (Villeneuve et al. 1999).

We previously reported that neoglycoconjugates (hexasaccharide, carbohydrate-bovine serum albumin conjugate — hereafter CHO-BSA) with carbohydrate components that mimic the upstream elements of Ogawa LPS, induced vibriocidal antisera (Chernyak et al. 2002). To approximate a clinically relevant vaccine, we prepared conjugates composed of mono-, di- or hexasaccharide fragments of the O-SP of V. cholerae O1, serotype Inaba and Ogawa as the B cell epitopes and recombinant Pseudomonas aeruginosa exotoxin A (rEPA). Compared to EPA (Pseudomonas aeruginosa exotoxin A), rEPA has one of the glutamic acids in EPA replaced with aspartic acid (Hickey & Mohn, 1996). rEPA has been used previously as a carrier in conjugate vaccines (Fattom et al. 1990, 2004; Schneerson et al. 2004). These neoglycoconjugates were used in experiments to determine if monosaccharides could boost a hexasaccharide-primed host. Other studies addressed the efficacy of the hexasaccharide conjugates if the host was primed with LPS (Inaba or Ogawa) prior to boosting with the neoglycoconjugates. Finally, the efficacy of the conjugates was tested in a host that had prior immunity to the carrier and B cells primed to V. cholerae LPS.

Materials and methods

Animals and immunization protocols

Six-week-old, female BALB/c mice from the National Cancer Institute (Bethesda, MD) were used as hosts for the immunogenicity studies. Individual adult mice were immunized intraperitoneally (ip) with 10 µg (based on carbohydrate weight) of either the Ogawa or Inaba CHO- rEPA conjugate (~5 moles CHO/mole rEPA) suspended in 150 mM NaCl and mixed 1 : 1 with RIBI® adjuvant (Sigma, St Louis, MO). Three immunization schedules (A C) were used in this study. Schedule A had four doses of immunogen. The first two doses were with the hexasaccharide conjugates and the last two with the monosaccharide conjugate (Fig. 1a). Schedule B was developed to assess if priming with nonvibriocidal doses of purified LPS would enhance the efficacy of a booster immunization (5 days later) with the hexasaccharide of the contra serotype (Fig. 1b). Finally, schedule C was an attempt to replicate the steady-state immune status of an individual with prior exposure to the carrier and to V. cholerae LPS. Mice were immunized twice with rEPA in RIBI® 2 weeks before priming with a low dose of either Inaba or Ogawa LPS, followed 5 days later with a booster containing the Ogawa hexasaccharide conjugate (Fig. 1c). The Ogawa hexasaccharide conjugate was used to boost mice immunized with Inaba LPS because we wanted to see if it could expand B cells activated by the contra-serotype of LPS.

Figure 1

Time lines for mouse immunizations and serum collection. (a) For schedule A, the first two inoculations were with either Ogawa or Inaba hexasaccharide conjugates and the last two with the monosaccharide conjugate of the same serotype as the hexasaccharide. (b). Mice were primed with 2.5 µg purified LPS from either Ogawa or Inaba and then boosted (5 and 14 days later) with the hexasaccharide of the contra-serotype. (c) Mice were immunized twice with rEPA in RIBI® 2 weeks apart before priming with 2.5 µg of either Inaba or Ogawa LPS, followed 5 days later with a booster of 10 µg of the Ogawa hexasaccharide conjugate. The day post priming for the immunizations and the blood collection are shown above and below the time lines, respectively.

Blood collection via retro-orbital sinus/plexus was done on days indicated in the individual time lines (Fig. 1). Sera were designated either: prebleed (pb), primary (p), secondary (s), tertiary (t) or quaternary (q) on the graphs. Resulting sera from individual mice was stored separately at 4°C or −20°C until used. ELISA and the vibriocidal assay assessed individual serum samples.

Ogawa CHO-rEPA constructs

rEPA was prepared as described (Johansson et al. 1996). The 5-(methoxycarbonylpentyl) linker-equipped Ogawa hexasaccharide (19) was assembled (Saksena et al. 2005b) by condensation of the disaccharide glycosyl donor (1) with tetrasaccharide glycosyl acceptor (2) and further conversions, as shown in Fig. 2a and b. The analogous Inaba hexasaccharide (20), mono- and the disaccharides that mimic the O-PS of Vibrio cholerae O:1, serotypes Inaba and Ogawa used here (21 24), analogous to hexasaccharides 19 and 20, were from a stock of materials that were made and characterized as reported previously (Gotoh & Ková, 1994; Gotoh et al. 1994; Lei et al. 1995, 1996; Ogawa et al. 1996; Saksena et al. 2005b). All linker-equipped saccharides were converted to the squaric acid derivatives and conjugated to rEPA as shown for the Ogawa hexasaccharide (Fig. 2b). Briefly, an N-iodosuccinimide/silver trifluorosulfonate-mediated reaction of glycosyl donor 1 (Johansson et al. 1996) and glycosyl acceptor 2 (Peters & Bundle, 1989; Zhang & Kovac, 1997) gave tetrasaccharide 3, which was sequentially deacetylated (→4) and methylated, to give glycosyl donor 5. It was treated with alcohol 11 to obtain 5-(methoxycarbonylpentyl) linker equipped hexasaccharide 12. The required compound 11 was obtained as follows. Crystalline (Poirot et al. 2001) diacetate 6 was treated (Saksena et al. 2005b) with methyl 6-hydroxyhexanoate in the presence of boron trifluoride etherate and the 5-(methoxycarbonylpentyl) glycoside 7 formed was deacetylated (→8). Reaction of 8 with thioglycoside 9 followed by deacetylation of the formed 10 gave 11. To introduce the 3-deoxy-l-glycerotetronic acid side chain, the hexaazide 12 was selectively reduced, and the corresponding hexamine 13 was treated with 2-O-acetyl-4-O-benzyl-3-deoxy-l-glycerotetronic acid, to give the fully protected hexasaccharide 14. The foregoing substance was subjected to two-step deprotection, deacetylation (→15) and hydrogenolytic debenzylation (→16), and subsequent amidation with ethylenediamine gave amine 17, whose reaction with squaric acid diethyl ester at pH 7 gave the aquaric acid monoester 18. Treatment of the latter with rEPA at pH 9 was monitored by surface- enhanced laser desorption-ionization time-of-flight mass spectrometry (SELDI-TOF MS) (Chernyak et al. 2001). The reaction was terminated when a molar carbohydrate protein ratio ~5 was reached, as indicated by SELDI-TOF MS, to afford conjugate 19.

Figure 2

Schema for generation of neoglycoconjugate immunogens from the Ogawa hexasaccharide and rEPA carrier.

Serology

The presence of anti-O-SP Ogawa or Inaba antibody was measured by ELISA as described in detail previously (Chernyak et al. 2002; Meeks et al. 2004; Saksena et al. 2005a). The ELISA test antigen was purified Ogawa LPS P1418 or Inaba LPS (Sigma, St Louis, MO). Endpoint titers for ELISA were defined as the reciprocal of the antibody dilution for the last well in a column with a positive optical density for each sample after subtracting twice the average optical density background. The disaccharide-BSA was used as an ELISA test antigen at 0.5 µg well−1(Saksena et al. 2005a).

Vibriocidal microtiter assay

The microtiter vibriocidal test developed by Fournier's group (Boutonnier et al. 2003; Meeks et al. 2004; Saksena et al. 2005a) was used. This assay is based on metabolically active bacteria metabolizing a substrate to produce a violet color in the well, which indicates the presence of live vibrios. Inhibition of bacterial growth (end-point titer) is reported as the reciprocal of the antibody dilution for the negative well containing the lowest concentration of antibody. A titer of <1 : 40 is considered not vibriocidal. The negative control (−cont.) included bacteria and complement only, while the positive control (+cont.) was quaternary sera from mice immunized four times with either purified Ogawa or Inaba LPS.

Results

Rationale for studies

We reported that synthetic B cell epitopes, which mimic the terminal sugars of Ogawa or Inaba LPS, are immunogenic (Chernyak et al. 2002). Ogawa immunogens, especially the terminal hexasaccharide, efficiently induce vibriocidal and protective antibody, whereas the Inaba-based immunogens did not induce protective responses (Chernyak et al. 2002; Meeks et al. 2004; Saksena et al. 2005a, 2006). Bovine serum albumin (BSA) was the carrier component of the neoglycoconjugates for our early studies. Because BSA is not an appropriate carrier for human vaccines, we tested the nontoxic rEPA for its suitability as a carrier.

Anti-LPS (IgM) titers induced by hexasaccharide conjugates can be boosted by monosaccharide conjugates

We assessed the immunogenicity and efficacy of the rEPA-based conjugates for mice given two doses of hexassacharide followed by two doses of the monosaccharide conjugates (Fig. 1a). We chose this approach because two doses of Ogawa hexasaccharide were effective at inducing vibriocidal serum antibody, and Ogawa monosaccharides were found to be able to enhance the humoral response of mice primed with low doses of LPS (Saksena et al. 2006). This experimental design will provide information that has implications for the vaccination protocol and also the production of the vaccine. If monosaccharides can boost the anti-LPS antibody response as efficiently as hexasaccharide conjugates, then six times more doses of vaccine will be available from the same production level of carbohydrate.

Four immunizations (schedule A) induced comparable if not higher anti-LPS (IgM) end-point titers (Fig. 3a) in the tertiary and quaternary sera than three immunizations with hexasaccharide immunogens (data not shown). The four-dose immunization protocol resulted in anti-LPS IgM end-point titers in quaternary sera that were similar in magnitude to those induced by four immunizations with purified LPS (Fig. 3aand b, +cont.). The kinetics of the anti-Inaba and anti-Ogawa responses to the neoglycoconjugates was similar, but the anti-LPS IgM titers were higher by a log for mice immunized with Ogawa conjugates. The antibodies induced, regardless of serotype, were cross-reactive to the LPS that was heterologous to the immunizing LPS. An anti-LPS IgG1 response was apparent in tertiary and quaternary sera of mice immunized with Ogawa-based conjugates and in the quaternary sera of mice immunized the Inaba-based conjugates (Fig. 3c).

Figure 3

Humoral and vibriocidal response of mice immunized with Ogawa or Inaba hexasaccharides twice before boosting with two doses of the monosaccharides of the same serotype. (a) The individual end-point titers of anti-Ogawa LPS IgM are shown for prebleed (PB, open squares), primary (P, open triangles), secondary (S, closed inverted triangles), tertiary (T, open diamonds) and quaternary (Q, closed triangles) sera. The bar represents the arithmetic mean of the data. The+cont. is quaternary sera (Q, closed circles) from mice immunized with purified Ogawa LPS, four times every 14 days. (b) The individual end-point titers of anti-Inaba LPS IgM are shown. The symbols are as in a. The +cont. is hyperimmune anti-Inaba LPS sera. (c) The individual end-point titers of anti-Ogawa LPS IgG1 are shown. The symbols and +cont. are as a. ELISA endpoint titers of 100 are considered negative (−cont.). (d) The vibriocidal titers of PB (open circles), secondary (S, closed circles), and quaternary (Q, inverted closed triangles) antisera against Ogawa bacteria. The −cont. is bacteria and complement without sera; the +cont. is quaternary hyperimmune Ogawa LPS antisera. (e) The vibriocidal titers and symbols as in d. but against Inaba bacteria and the +cont. is anti-Inaba LPS hyperimmune sera (open circle).

Vibriocidal titers in response to immunization with hexasaccharide and monosaccharide conjugates

Immunization schedule A resulted in vibriocidal titers in the secondary sera for the mice immunized with the Ogawa conjugates (Fig. 3d). It is noteworthy that the quaternary sera of mice immunized with Ogawa conjugates using rEPA were highly vibriocidal, with the majority of the end-point titers in the range of 1 : 10 000. The serum from mice immunized with the Ogawa conjugate was specific for Ogawa LPS and did not kill Inaba bacteria (Fig. 3e). Neither Inaba conjugate induced vibriocidal antibody (Fig. 3e), consistent with a previous report (Meeks et al. 2004). We performed bacterial adsorption experiments using quaternary antisera from mice immunized with either Inaba LPS or Inaba conjugates. Under incubation conditions used in the vibriocidal assay (3 h at 37°C), antibodies from anti-Inaba LPS and anti-Inaba conjugate sera both bound V. cholerae LPS in situ equally well (data not shown), yet only the anti-Inaba LPS sera was vibriocidal (Fig. 3e). Antibodies in Ogawa LPS antisera bound as well as antibodies in antisera made to Ogawa conjugates; both antisera were vibriocidal. These results suggest that the concentration and affinity of the antibodies (IgM or IgG1) in the Inaba antisera are similar to those in the Ogawa antisera and certainly at a level that could be functional. The reason for nonvibriocidal Inaba response is not known but likely involves antibody epitopes.

Priming with native LPS and enhancement of vibriocidal response by contra-serotype conjugate

This experiment (immunization schedule B) was designed to determine whether priming with purified LPS would enhance the immunogenicity and efficacy of a booster with contra-serotype neoglycoconjugates. A low dose (2.5 µg) of either Inaba or Ogawa LPS was delivered intraperitoneally (ip) on day zero, followed 5 and 14 days later by an ip inoculation of the contra-serotype hexasaccharide neoglycoconjugate (Fig. 1b). This immunization protocol was motivated by the fact that people living in areas of endemic cholera can be exposed to either serotype of V. cholerae and it is, therefore, important to know if that preexisting immunity to either serotype can influence the response to the neoglycoconjugates.

None of the doses of Ogawa LPS used to prime induced measurable anti-LPS IgM titers against Ogawa or Inaba LPS (Fig. 4a). Purified Inaba LPS induced low titers of anti-IgM antibody at all doses tested (Fig. 4b). Surprisingly, a vibriocidal response was present for mice that had no or low anti-LPS IgM titers (Fig. 4c). The Inaba vibriocidal response was higher than that of the anti-Ogawa vibriocidal response, which was less than 1 : 100 (compared to ~1/1000 for mice immunized with Inaba LPS).

Figure 4

Humoral and vibriocidal response of mice primed with purified LPS and boosted with a hexasaccharide of the contra-serotype. (a) ELISA titers of anti-Ogawa LPS (IgM) of mice immunized intraperitoneally with different doses of Inaba or Ogawa LPS. Prebleed titers (open squares) and titers for the primary (P, open triangles) sera are shown. The horizontal bar represents the mean of the particular data set. (b) ELISA titers of anti-Inaba LPS (IgM) for mice immunized with varying doses of either Inaba or Ogawa LPS. ELISA end-point titers of 100 are considered negative (−cont.). (c) The PB sera of mice immunized against Inaba (left-hand panel) or Ogawa (right-hand panel) LPS were not vibriocidal (not shown). The vibriocidal titers against Inaba (open circles) and Ogawa (closed circles) for individual P serum samples are shown. Vibriocidal titers of less than 1 : 40 are considered negative. (d) Mice were immunized with 2.5 µg of either Inaba or Ogawa LPS followed by a booster 5 14 days later of a hexasaccharide conjugate of the contra-serotype to the priming LPS. The individual end-point titers of anti-Ogawa LPS IgM are shown for prebleed (PB, open squares), primary (P, open triangle), and secondary (S, closed inverted triangles) sera. The horizontal bar represents the arithmetic mean of the data. The +cont. is quaternary (Q, closed circles) sera from mice immunized with purified Ogawa LPS, four times every 14 days. (e) The individual end-point titers of anti-Inaba LPS IgM are shown. The symbols are as in d. The +cont. is hyperimmune anti-Inaba LPS sera. (f) The individual end-point titers of anti-Ogawa LPS IgG1 are shown. The symbols and +cont. are as d. The immunization schedule to identify the groups is shown below the graphs. (g) The IgM titers of prebleed (PB, open squares) and primary (P, open triangles) sera to Ogawa disaccharides of mice immunized once with Ogawa or Inaba hexasaccharide are shown. The +cont. are either hyper immune Ogawa LPS antisera (closed circles) or hyper immune Inaba LPS antisera (open circles). h. The prebleed (PB open squares) and primary (P, open triangles) end-point titers of Inaba disaccharides of mice immunized once with Inaba hexasaccharide or Ogawa hexasaccharide are shown. Symbols are as in panel g. (i) The IgM titers of prebleed (PB, open squares) and primary (P, open triangle) sera to Ogawa disaccharides of mice immunized with Ogawa or Inaba LPS are shown. The symbols are as in g. (j) The IgM titers of prebleed (PB, open squares) and primary (P, open triangles) sera to Inaba disaccharides of mice immunized with Ogawa or Inaba LPS are shown. The symbols are as in g. k. The IgM titers of prebleed (PB, open squares) and primary (P, open triangles) sera against Ogawa disaccharides of mice immunized according to the prime-boost strategy. (l) The IgM titers of prebleed (PB, open squares) and primary (P, open triangles) sera against Inaba disaccharides of mice immunized according to the prime-boost strategy. m. The serum vibriocidal response against Ogawa bacteria of mice primed to Ogawa or Inaba LPS before boosting with the contra-serotype hexasaccharide. The open circles represent prebleed (PB) responses while the closed circles represent the vibriocidal titers in primary (P) sera and the closed triangles represent the vibriocidal titers in secondary (S) sera. Horizontal bars are the mean of the individual data sets. The −cont. represents the vibriocidal activity of bacteria and complement only. The +cont. (closed circles) represent the vibriocidal capacity of hyperimmune Ogawa LPS antisera. (n) The vibriocidal activity against Inaba bacteria of PB, P, or S sera of mice immunized according to the prime-boost strategy. The symbols are as in m, except for the +cont. (open circles) which is the vibriocidal activity of anti-Inaba LPS hyperimmune sera. The immunizing protocol is shown below the x-axis.

Having established the control responses of mice immunized with LPS alone, we examined sera from mice that were primed with LPS and boosted with a conjugate of the contra-serotype. Regardless of the serotype of the priming LPS, mice responded with anti-LPS IgM antibodies at days 10 and 35 (Fig. 4d and e). Day 35 IgM titers were not statistically different from those of day 10, although the trend was for the mean end-point titer to be lower. In general, the mice did not respond to the prime-boost immunization protocol with measurable IgG1 titers (Fig. 4f).

Reactivity of antisera to Ogawa or Inaba disaccharide epitope

Mice immunized once with either the Ogawa or Inaba hexasaccharide conjugates produced significant amounts of IgM antibody reactive with Inaba and Ogawa disaccharide conjugates at day 10 (Fig. 4g and h). The average of antidisaccharide IgM end-point titers was 1/10 000, similar to the antidisaccharide response of mice hyperimmune (+cont.) to either Inaba or Ogawa LPS. Mice immunized with 2.5 µg of purified LPS did not have serum antibody that reacts with disaccharide conjugates (Fig. 4i and j). When mice were immunized four times with purified LPS, the response to the homologous (immunogen) disaccharide was always 10-fold higher than the response to the heterologous disaccharide, suggesting that anti-LPS sera has both serotype-specific and cross-reactive antibodies (Fig. 4i and j).

Mice primed with 2.5 µg of purified LPS and boosted with the neoglycoconjugates of the contra-serotype to the priming LPS had comparable primary end-point titers to the disaccharide epitope as mice immunized with neoglycoconjugates alone (compare Fig. 4k and l and Fig. 4g and h).

Serum vibriocidal titers of mice primed with LPS and boosted with neoglycoconjugates of the contra-serotype

Serum vibriocidal titers were apparent at days 10 and 35 for mice primed with one LPS serotype and boosted with the contra-serotype neoglycoconjugate (Fig. 4m and n). Immunization of mice with conjugates alone does not induce vibriocidal titers after 5 days (data not shown). Immunization with only purified LPS induced low vibriocidal titers (1/100) in 25% of the mice immunized with Ogawa LPS whereas 60% of the mice immunized with 2.5 µg of Inaba LPS had an average end-point titer of 1/1000. When mice were primed with a low dose of Ogawa LPS, Inaba hexasaccharide conjugates enhanced (>1/1000) the Ogawa-specific vibriocidal response (Fig. 4m). The day 10 sera vibriocidal end-point titers were only about a log lower than the vibriocidal titers of mice hyperimmune (four doses) to Ogawa LPS (+cont.). The ability of the Ogawa hexasaccharide to enhance the Ogawa vibriocidal response of Inaba LPS-primed mice was less apparent, with titers of 1/100 being the average, when Ogawa bacteria were the target (Fig. 4m). These titers were lower by a log than the vibriocidal titer induced by 2.5 µg of Inaba LPS alone (Fig. 4b). The use of Ogawa LPS to prime mice for a booster with Inaba hexasaccharides was ineffective at inducing a vibriocidal response, as measured against Inaba bacteria (Fig. 4n). Priming mice with Inaba LPS and boosting with the Ogawa hexasaccharide resulted in low vibriocidal titers (1/100) against Inaba-expressing bacteria, which contrasts with the 10-fold higher vibriocidal titers (1/1000) in sera for mice inoculated with 2.5 µg of Inaba LPS. Changing the structure or context of the LPS-based B cell epitopes available during the course of the immunization affected the vibriocidal response.

Response of mice preimmune to the rEPA, primed with either Inaba or Ogawa LPS, and boosted with Ogawa hexasaccharide conjugate

This experiment was designed (Fig. 1c) to test the efficacy of the neoglycoconjugates in a circumstance that may represent the immune status of human vaccinees: the presence of memory helper T cells specific for the carrier protein and memory/activated B cells specific for various LPS epitopes. Mice immunized twice with rEPA in an adjuvant before LPS priming and then immunized with rEPA associated with the conjugate at day 5 after the LPS priming, had anti-rEPA titers greater than 1/3.2 × 106 compared to lower serum titers (1/50 000 100 000) of mice that only received rEPA in the form of the conjugate at day 5 (data not shown). Regardless of the LPS serotype used to prime, mice made good anti-LPS IgM responses at day 10 (Fig. 5a and b). On average, the end-point titers in the primary sera were higher in this scenario than for mice immunized according to the schedule in Fig. 1b, which involved priming with LPS but not with the carrier. The anti-LPS IgM end-point titers in the primary sera of mice primed to carrier and LPS before boosting were less than fivefold different from titers of sera from mice hyperimmunized to purified Ogawa or Inaba LPS. As did sera from mice from the other immunization protocols, sera from mice immunized according to schedule C bound both serotypes of LPS.

Figure 5

Humoral response and vibriocidal of mice primed with rEPA and LPS (Inaba or Ogawa) before being boosted with Ogawa hexasaccharides. (a) The individual end-point titers of anti-Ogawa LPS IgM are shown for prebleed (PB, open squares) and primary (P, open triangle). The horizontal bar represents the arithmetic mean of the data. The +cont. is quaternary (Q, closed circles) sera from mice immunized with purified Ogawa LPS, four times every 14 days. (b) The individual end-point titers of anti-Inaba LPS IgM are shown. The symbols are as in a. The +cont. is hyperimmune anti-Inaba LPS sera. ELISA endpoint titers of 100 are considered negative (−cont.). (c) The prebleed (PB, open squares) and primary (P, open triangles) IgM titers against Ogawa disaccharides of mice immunized according to the prime-boost strategy shown in schedule C. (d) The prebleed (PB, open squares) and primary (P, open triangles) sera IgM titers against Inaba disaccharides of mice immunized according to the prime-boost shown in schedule C. (e) The vibriocidal titers of prebleed (PB, open squares) and tertiary (T, closed circles) antisera against Ogawa bacteria. The -cont. is bacteria and complement without sera; the +cont. is hyperimmune Ogawa LPS antisera. Vibriocidal titers of less than 1 : 40 are considered negative. (f) The vibriocidal titers as in panel e. but against Inaba bacteria. The open circles for the +cont. is anti-Inaba LPS hyperimmune sera. The immunizing protocol are shown below the x-axis.

The prior exposure to the rEPA resulted in higher end-point titers to the Ogawa and Inaba disaccharides at day 10 compared to mice primed with LPS but not the carrier (Fig. 5c and d). The anti-disaccharide titers were similar to serum titers of mice given four doses of Ogawa or Inaba LPS. The prime boost strategy did not enhance the response to disaccharide of a particular serotype.

Vibriocidal responses of mice primed to carrier and LPS before immunization with Ogawa neoglycoconjugates

Ogawa hexasaccharide neoglycoconjugates are superior at inducing vibriocidal antibodies compared to the Inaba hexasaccharide conjugates (Chernyak et al. 2002; Meeks et al. 2004). We wanted to know if Ogawa hexasaccharides could enhance the vibriocidal responses of mice primed with either serotype type of LPS. Priming mice with rEPA and LPS and later immunized with Ogawa hexasaccharide conjugates induced vibriocidal titers that were cross-reactive with V. cholerae LPS (Fig. 5e and f). The serum vibriocidal response (against Ogawa bacteria) of mice primed with Inaba LPS and boosted with Ogawa hexasaccharide conjugate was on average 1/1000 10 days past LPS priming. The vibriocidal antibodies (against Ogawa bacteria) induced by the homologous system, Ogawa LPS priming and a booster with Ogawa conjugate, were on average 1/2000, but the response was less variable (Fig. 5e). The mean anti-Inaba vibriocidal titer of mice primed to carrier and Inaba LPS and boosted with Ogawa hexasaccharide was 1 : 200 (Fig. 5f). Mice primed with rEPA and Ogawa LPS and boosted with Ogawa hexasaccharide conjugates had a slightly lower (1/100) vibriocidal titer to Inaba-LPS expressing bacteria.

Discussion

Cholera subunit vaccines based on V. cholerae LPS B cell epitopes without the attending Toll receptor 4 agonist component of LPS are being developed (Gupta et al. 1998; Chernyak et al. 2002; Meeks et al. 2004). Ogawa-based neoglycoconjugates (synthetic hexasaccharide variously linked to BSA) induce vibriocidal and protective antibody (Chernyak et al. 2002). Human clinical trials require neoglycoconjugates with a medically acceptable carrier protein such as rEPA which has been used for other O-SP conjugate vaccines (Pavliakova et al. 1999). Synthetic Ogawa or Inaba saccharide fragments conjugated to rEPA were immunogenic and induced vibriocidal antibodies. The rEPA-based conjugates were not as readily soluble nor did they easily induce an IgG response compared to the BSA-based conjugates (Chernyak et al. 2002; Meeks et al. 2004). The utility of rEPA as a carrier for synthetic V. cholerae carbohydrate epitopes is questionable.

Serotypes, immunodominance and epitope

Structurally, Ogawa and Inaba V. cholerae LPS are almost identical, differing only at the upstream terminal sugar where an O-2 methyl group or a hydroxyl at that position defines Ogawa and Inaba, respectively (Hisatsune et al. 1993; Ito et al. 1994; Liao et al. 2002; Villeneuve et al. 2002). Inaba LPS is thought to be a superior immunogen to Ogawa LPS (Longini et al. 2002). The data presented herein supports that notion. The immunologic explanation for this is unknown and at variance with the protective efficacy of the LPS-based Ogawa and Inaba hexasaccharide conjugates we have developed. Ogawa hexasaccharide immunogens are superior to their Inaba counterparts, as the latter only rarely induce vibriocidal antibody (Chernyak et al. 2002; Meeks et al. 2004).

The structure of V. cholerae LPS and that of the corresponding neoglycoconjugates differ. The O-SP component of LPS has more monomeric units (12 15 vs. 6) which are linked to the core saccharides. The longer LPS with its additional sugars (perosamine and core residues) could contain an additional protective B cell epitope (Liao et al. 2002; Chatterjee & Chaudhuri, 2003). Another difference between the two immunogens is the ability to induce the innate immune response. Native LPS contains lipid A, a Toll-4 receptor (TLR-4) agonist which can enhance human and murine B cell activation leading to enhanced immunoglobulin responses (Ogata et al. 2000). The neoglycoconjugates do not contain lipid A, and thus attending inflammatory signals required to enhance immunogenicity are derived from RIBI® adjuvant that contains monophosphoryl lipid, a lipid A look alike which was engineered to remove the phosphate group from the reducing end sugar and the ester-linked fatty acidy at the 3-positon to obviate most of the toxic aspects of LPS's lipid A (Baldridge & Crane, 1999).

Do these differences contribute to the different efficacy of LPS and the neoglycoconjugates at inducing vibriocidal antibody? The analysis of the geometric mean end-point titers (anti-LPS IgM and antidisaccharide IgM) and the corresponding vibriocidal capacity of the sera (day 10) from mice immunized according to the different protocols is instructive. Mice immunized with 2.5 µg of Inaba LPS had low titer anti-LPS responses, did not have a measurable anti-disaccharide response, but did have mid-level vibriocidal titers. Low doses of Inaba LPS, and to some extent of Ogawa LPS also, can induce vibriocidal antibody that does not bind the terminal disaccharides of either LPS serotype. This is consistent with a LPS B cell epitope (common epitope, designated as C), other than the two terminal sugars, being more effective at inducing early protective responses than the terminal LPS sugars. The C eptitope was postulated to be the tetronic acid residue of the perosamines (Manning et al. 1994).

Mice immunized with either Inaba or Ogawa hexasaccharide neoglycoconjugates had similar if not higher anti-LPS IgM titers than the LPS-immunized mice (348 606 vs. 200) and significant levels of antidisaccharide antibody (5572 11 143 end-point titers that bound both LPS serotypes. However, these sera were not vibriocidal, suggesting that antibodies to terminal disaccharides of V. cholerae LPS are not strictly correlated with a protective response. Others have reported that a vibriocidal, anti-Ogawa mAb (S-20-4; IgG1) binds the terminal sugars (mono- and disaccharide) of Ogawa LPS (Wang et al. 1998; Villeneuve et al. 2002). S-20-4 was generated from a hyperimmune mouse and is very specific for Ogawa LPS, reacting with Inaba monglycosides with about 800-fold less affinity (Wang et al. 1998; Liao et al. 2002). We speculate that the early IgM (10 day) response to the disaccharides we report, which has not undergone isotype switching and thus is likely not to be somaticly mutated, is the reason these anti-disaccharide antibodies were not protective.

LPS priming before boosting with neoglycoconjugates

If mice are primed with purified LPS of one serotype and boosted with a hexasaccharide neoglycoconjugate of the contra-serotype, there is no effect on the magnitude of the early anti-LPS (IgM) response. Unexpectedly, the antidisaccharide response is lower compared to mice immunized with neoglycoconjugates alone, yet vibriocidal titers are more evident from mice immunization with the prime-boost schedule compared to mice immunized with LPS or neoglycoconjugates alone. These results suggest that priming with Ogawa LPS expands B cells that the Inaba hexasaccharide can also activate. We do not know if the terminal sugars or the tetronic acid epitope is the target for the booster effect. The ability of the humoral response to LPS priming can be manipulated by the epitope composition of the booster immunogen, as is evident for mice primed with Inaba LPS and boosted with Ogawa hexasaccharide. Mice only inoculated with Inaba LPS respond at day 10 with a geometric mean vibriocidal titer of 420. If mice are boosted at day 5 with Ogawa neoglycoconjugates, the titer in day 10 sera is reduced to 80.

The sera of mice primed with rEPA and LPS (Ogawa or Inaba) before being boosted with the neoglycoconjugate had higher anti-LPS IgM and anti-disaccharide geometric mean titers compared to the sera of mice immunized by protocol A or B. It is reasonable to expect that higher anti-LPS and anti-disaccharide titers would correlate with enhanced vibriocidal titers. However, mice immunized according to protocol C had vibriocidal titers within the range of those titers resulting from the other immunization schedules. As in the other prime boost immunization schedule, boosting with Ogawa oligosaccharide-based immunogens was synergistic for the vibriocidal response, whereas boosting Inaba-LPS-primed mice with the Ogawa hexasaccharide resulted in a reduced anti-Inaba LPS response.

Serotype-specific and common structures of LPS: effect on immunogenicity and protective response

What explains the lack of an additive response to prime boost immunization if the immunogens are not identical LPS serotype structures? We do not know the answer to this question, but hypothesize that the immunodominance of the B cell epitopes (B and C) in different LPS serotypes is linked to the effect. The structure of the Inaba epitope may provide a weaker epitope for the anti-Inaba antibody repertoire than that of the anti-Ogawa antibody, and thus the terminal sugars of Inaba are not as immunogenic as those of Ogawa. This idea is supported by the work of Liao and colleagues. (Liao et al. 2002), who suggested that the loss of the methyl group at O-2 influences the polarity of the Inaba epitope, making not only the resulting 2-OH more polar, compared to 2-OMe, but also the oxygen of the neighboring 3-OH less negative due to the lack of the electron-donating effect of the methyl group at the 2 position. Compared to Ogawa LPS, the Inaba terminal sugar may not be immunodominant if presented in the context of LPS, and the B cell response is focused on the C antigen, which is why Inaba LPS can induce cross-reactive protective antibody following immunization (Villeneuve et al. 1999).

In individuals exposed to or immunized with V. cholerae LPS, the initial B cell response is directed against the LPS immunodominant epitope, which differs based on the serotype. Ogawa LPS and a booster with the corresponding serotype conjugate would focus on the terminal sugars. The cross-reaction of anti-Ogawa antibodies to the terminal Inaba LPS sugar would allow for an additive response. Inaba LPS priming followed by a booster with Ogawa conjugates would be the least likely to be effective. The Inaba LPS presents the common epitope in the priming event. The immunodominant Ogawa epitope in an Ogawa conjugate booster would be presented in the context of B cells activated and expanded to the C antigen and, thus, it would represent an initial immunization rather than booster. Immune pressure results in serotype conversion, especially from Ogawa to Inaba. It is hypothesized that a gene product, rfbT a methyl transferase, provides the Ogawa serotype structure. The loss of mutation of this gene results in the Inaba serotype becoming the dominant serotype in the individual or the community. We do not think that immune pressure has selected a different serotype in our studies as there is no active infection. However, our data support the immunodominance of Ogawa over Inaba if the immunogen is in the form of neoglycoconjugates. These data, along with the immune-mediated serotype conversion, highlight the complexity of the anti-LPS response in cholera.

The fact that individuals can have similar levels of anti-LPS antibody (ELISA) and have different levels of protection or not be protected at all, shows there is much to be learned about the epitope specificity to the V. cholerae anti-LPS response. We need to know how to manipulate LPS-based immunogens to achieve maximal protection. Synthetic LPS-based neoglycoconjugates may be useful for this as they can be generated to present restricted B cell epitopes, perhaps the cross-reactive C epitope. In order to manipulate the anti-LPS response, we need to determine the specificities (LPS structures bound by antibody) and the affinities of the antibodies induced by the neoglycoconjugates and by purified LPS. We also need to determine if there is a hierarchy to the LPS-immunogen epitopes. Depending on the LPS serotype, do some epitopes stimulate B cells more effectively and thus optimize early production of protective antibody? The information needs to be evaluated in the context of prior exposure to V. cholerae LPS of individuals living in cholera-endemic areas.

Conclusion

The results we report show that synthetic neoglycoconjugates can enhance the serum and vibriocidal response of mice primed with purified LPS. The novel finding that Ogawa neoglycoconjugates can boost Inaba-primed mice while Inaba neoglycoconjugates can not boost Ogawa-primed mice highlights the issue of immunodominant LPS epitopes that we propose, based on experimental evidence, differ between Ogawa and Inaba LPS. Neoglycoconjugates are being developed for use as boosters for cholera vaccine responses that have waned. The neoglycoconjugates are an attractive alternative because they can be delivered parenterally without the attending inflammation of LPS, and they have a carrier component to enhance B cell memory and antibody-isotype switching. Parenteral immunization obviates the problems of immune interference that have been reported for boosting using the existing oral cholera vaccines.

Acknowledgements

This work was supported by an NIH grant to WFW (AI 47373) and by intramural NIH support to PK.

Footnotes

  • Editor: Artur Ulmer

References

View Abstract