| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Unité de Pathogénie Microbienne Moléculaire, Institut National de la Santé et de la Recherche Médicale Unité 389;
Unité de Chimie Organique, Centre National de la Recherche Scientifique Unité de Recherche Associée 2128;
Plate-Forme 5, Production de Protéines Recombinantes et d’Anticorps; and
Unité d’Immunologie Structurale, Centre National de la Recherche Scientifique Unité de Recherche Associée 2185; Institut Pasteur, Paris, France
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Upon natural infection or following vaccine trials, as well as during experimental infection, the Ab-mediated protection has been shown to be serotype specific, pointing to the LPS as the major protective Ag (2). Shigella serotypes are defined by the structure of the oligosaccharide repeating unit (RU)3 that forms the O-Ag, the polysaccharide part of LPS (3). For the predominant S. flexneri serotype 2a, the biological RU is the branched pentasaccharide shown in Fig. 1. It bears a linear tetrasaccharide backbone made of three L-rhamnose residues, A, B, and C, and a N-acetyl-D-glucosamine residue D, that is common to all S. flexneri except serotype 6 and represents the RU of S. flexneri serotype Y. The RU of serotype 2a is characterized by the presence of the 
-D-glucose residue E, branched at position 4 of rhamnose C on the linear tetrasaccharide backbone.
|
75% of the vaccines during a S. sonnei outbreak (4). However, glycoconjugate vaccines based on the use of detoxified LPS present some drawbacks such as the need for accurate control of the detoxification step and also the potential loss of immunogenicity upon coupling to the protein carrier. In addition, to fulfill the requirements of regulatory agencies for always better-defined molecules to be used in humans, an alternative strategy has been proposed. It is based on the conjugation onto appropriate carriers of synthetic oligosaccharides mimicking determinants which are targets for protective Abs, with the hope that they induce an anti-LPS Ab response. Recent data have confirmed the assumption that short fragments of polysaccharides may mimic the native Ag. Indeed, semisynthetic glycoconjugates incorporating fragments of bacterial polysaccharide Ags were shown to be highly immunogenic in mice (5, 6, 7). The "proof of concept" was recently provided in humans with the efficacy of such a semisynthetic glycoconjugate in protecting against Haemophilus influenzae b infection (8). The aim of this study was to identify functional mimics of the S. flexneri serotype 2a O-Ag. We chose to identify the protective serotype 2a-specific determinants by performing an extensive study using synthetic oligosaccharides to define the contribution of each sugar residue in the recognition by protective serotype 2a-specific mouse mAb of the G isotype (mIgG). First of all, five mIgG specific for S. flexneri serotype 2a and representative of each IgG subclass were selected, and their protective efficacy was assessed in a murine model of infection (9, 10). Then, available mono-, di-, tri-, tetra-, and pentasaccharides representative of the RU, as well as longer sequences, were tested for their recognition by these five mIgG using inhibition ELISA to define an IC50. A set of oligosaccharides was next selected based on these antigenicity data. Their potential as accurate mimics of serotype 2a O-Ag, i.e., their ability to induce anti-O-Ag Abs, was assessed in mice upon immunization with the corresponding tetanus toxoid (TT) glycoconjugates. We, finally, succeeded in identifying a potential candidate for further development of a chemically defined glycoconjugate vaccine to S. flexneri 2a.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
M90T, an invasive isolate of S. flexneri serotype 5a, and 454, an invasive isolate of S. flexneri serotype 2a, were the virulent strains of reference. For intranasal (i.n.) infection, bacteria were routinely grown on Luria-Bertani agar plates at 37°C. They were recovered from plates and bacterial dilutions were performed in 0.9% NaCl with the consideration that, for an OD of 1 at 600 nm, the bacterial concentration was 5 x 108 CFU/ml. Killed bacteria for i.p. immunizations were prepared from bacterial cultures at stationary phase, diluted to 5 x 108 CFU/ml in 0.9% NaCl, and then incubated at 100°C for 1 h. They were then kept at –20°C in aliquots.
Production and characterization of mAbs specific for S. flexneri serotype 2a LPS
BALB/c mice were immunized i.p. with 107 CFU of killed S. flexneri 5a or S. flexneri 2a bacteria three times at 3-wk intervals. Mice eliciting the highest anti-LPS Ab response were given an i.v. booster injection 3 days before being sacrificed for splenic B cell fusion according to Kohler and Milstein (11). Hybridoma culture supernatants were screened for Ab production by ELISA using LPS purified from S. flexneri serotype X, Y, 5a, 5b, 2a, 2b, 1a, and 3a, respectively, as previously described (12, 13). Briefly, LPS purified according to Westphal and Jann (14) was used at a concentration of 5 µg/ml in PBS. As secondary Abs, anti-mouse IgG-, IgM-, or IgA-alkaline phosphatase-labeled conjugate (Sigma-Aldrich) were used at a dilution of 1/5,000.
Only the hybridoma cells secreting mIgG reacting specifically with LPS homologous to the strain used for immunization, i.e., recognizing serotype-specific determinants on the LPS O-Ag were selected. Those selected, representative of the four murine IgG subclasses, were then cloned by limiting dilution, and injected i.p. into histocompatible mice for ascites production. mIgG were precipitated with 50% ammonium sulfate from ascites fluid, centrifuged, and dialyzed against PBS before being purified using ion-exchange chromatography as previously described (12, 13).
mIgG sequence analysis
Total RNA was extracted from hybridoma cells by RNAxel kit (Eurobio). mRNA was converted into cDNA with a reverse transcriptase kit (Invitrogen Life Technologies) and used as template for PCR amplification using TaqDNA polymerase (Invitrogen Life Technologies) according to the manufacturer’s protocol. The amplification was performed with the primer of corresponding isotype (IgG1, 5'-GCA AGG CTT ACT AGT TGA AGA TTT GGG CTC AAC TTT CTT GTC GAC-3'; IgG2a, 5'-GTT CTG ACT AGT GGG CAC TCT GGG CTC-3'; and IgG3, 5'-GGG GGT ACT AGT CTT GGG TAT TCT AGG CTC-3'. The following eight H chain variable region (VH) primers were also used: 5'-AG GTG CAG CTC GAG GAG TCA GGA CC-3'; 5'-GAG GTC CAG CTC GAG CAG TCT GGA CC-3'; 5'-CAG GTC CAA CTC GAG CAG CCT GGG GC-3'; 5'-GAG GTT CAG CTC GAG CAG TCT GGG GC-3'; 5'-GAG GTG AAG CTC GAG GAA TCT GGA GG 3'; 5'-GAG GTA AAG CTC GAG GAG TCT GGA GG-3'; 5'-GAA TGT CAG CTC GAG GAG TCT GGG GG-3'; and 5'-GAG GTT CAG CTC GAG CAG TCT GGA GC-3'. For the light chains, the primer sequences were: for the 
-chain, 5'-GCG CCG TCT AGA ATT AAC ACT CAT TCC TGT TGA A-3'; for the variable region (VL), 5'-CCA GTT CCG AGC TCG TTG TGA CTC AGG AAT CT-3'; 5'-CCA GTT CCG AGC TCG TGT TGA CGC AGC CGC CC-3'; 5'- TGG ATG GTG GGA AGA TG-3'; 5'-GAG AGC AGA AAT AAA CTC CC-3'; 5'-CCA GAT GTG AGC TCG TGA TGA CCC AGA CTC CA- 3'; 5'-GAC CCC AGA AAA TCG GTT-3'; 5'-CCA GTT CCG AGC TCG TGA TGA CAC AGT CTC CA-3'; 5'-TTC CCA GGC TGT TGT GA-3'; 5'-GAG CTC GTG ATG ACA CAG TCT CCA-3'; and 5'-AAT TCT AAC TAG CTA GTC GCC-3'. Nucleic acid sequence determination was conducted by Genome Express using PCR products. Sequence analysis was performed with software packages from the Genetics Computer Group (Madison, WI), the GenBank (Los Alamos, NM), and European Molecular Biological Laboratory (Heidelberg, Germany) databases. For the determination of the gene families, analysis of the nucleotide sequences was performed with the international ImMunoGeneTics database (http://imgt.cines.fr) (15). The determined sequences have been submitted to the European Molecular Biological Laboratory nucleotide sequence database (http://www.ebi.ac.uk) with the accession numbers from AJ 784030 to AJ784039.
Protection experiments
Purified mIgG (20 or 2 µg) were administered i.n. using a volume of 20 µl to mice previously anesthetized via the i.m. route with 50 µl of a mixture of 12.5% ketamine (Merial) and 12.5% acepromazine (Vetoquinol). Intranasal challenge was performed using 108 virulent bacteria in a volume of 20 µl (10). Measurement of lung bacterial load was performed at 24 h after infection as follows. Mice were sacrificed by cervical dislocation and lungs were removed en bloc and ground in 10 ml of sterile PBS (Ultra Turrax T25 apparatus, Janke and Kunkel IKA Labortechnik). Dilutions were then plated on trypticase soy broth plates for CFU enumeration. The lung bacterial load in mice receiving the mIgG was compared with that of control mice receiving PBS. Each experiment was performed using 10 mice per group and repeated three times. All of the animal experiments were approved by the Institut Pasteur Animal Use Committee.
Histopathological studies
Mice were anesthetized, their trachea catheterized, and 4% Formalin injected to fill the bronchoalveolar space. Lungs were then removed and fixed in 4% Formalin before being processed for histopathological studies. Ten-micrometer paraffin sections were stained with H&E or labeled with specific Abs and observed with a BX50 Olympus microscope (Olympus Optical, Europa).
Synthetic oligosaccharides representative of S. flexneri 2a O-Ag
Oligosaccharides representative of fragments of the O-Ag of S. flexneri 2a were synthesized as their methyl glycoside to reproduce the ring and anomeric forms that the reducing residue adopts in the natural polysaccharide (see Table I). The serotype-specific oligosaccharides, thus bearing residue E, were obtained through multistep chemical synthesis as previously reported (16, 17, 18, 19). The 
EC disaccharide (16), which has a non-natural EC glycosidic linkage, was synthesized to probe the influence of such linkage on Ab recognition. Since, based on available building blocks, they were considered to be the easiest chemically accessible longer fragments, the octa- B(E)CDA'B'(E)'C' (20) and decasaccharide D''AB(E)CDA'B'(E)'C' (21) were synthesized to study the length-dependent oligosaccharide-Ab recognition. '' indicates the residue linked at the nonreducing end (left end), whereas ' indicates the residue linked at the reducing end (right end) of the AB(E)CD biological RU. Di- and trisaccharides devoided of the E residue were either previously described or synthesized according to known procedures (22, 23, 24, 25, 26).
|
Characterization of the saccharidic determinant recognized by the available mIgG was performed by measuring the mIgG-oligosaccharide interaction as follows. First, a standard curve was established for each mIgG tested. Different concentrations of the mAb were incubated overnight at 4°C on microtiter plates coated with purified S. flexneri 2a LPS at a concentration of 5 µg/ml in carbonate buffer at pH 9.6 and subsequently incubated with PBS/BSA 1% for 30 min at 4°C. After washing with PBS-Tween 20 (0.05%), alkaline phosphatase-conjugated anti-mouse IgG was added at a dilution of 1/5,000 (Sigma- Aldrich) for 1 h at 37°C. After washing with PBS-Tween 20 (0.05%), the substrate was added (12 mg of p-nitrophenylphosphate in 1.2 ml of 1 M Tris-HCl buffer (pH 8.8) and 10.8 ml of 5 M NaCl). Once the color developed, the plate was read at 405 nm (Dynatech MR 4000 microplate reader). A standard curve OD = f(Ab concentration) was fitted to the quadratic equation Y = aX2 + bX +c, where Y is the OD and X is the Ab concentration. Correlation factor (r2) of 0.99 was routinely obtained.
Then the amount of oligosaccharides giving 50% inhibition of mIgG binding to LPS (IC50) was determined as follows. Each mIgG at a given concentration (chosen as the minimal concentration of Ab which gives the maximal OD on the standard curve) was incubated overnight at 4°C with various concentrations of each of the oligosaccharides to be tested in 1% PBS-BSA. Measurement of unbound mIgG was performed as described above using microtiter plates coated with purified LPS from S. flexneri 2a and the Ab concentration was deduced from the standard curve.
The recognition capacity of anti-LPS mIgG for LPS was determined as described above using various concentrations of LPS that were incubated in solution overnight at 4°C with the predefined concentration of each mIgG. IC50 was defined as the concentration of oligosaccharides required to inhibit 50% of mIgG binding to LPS.
Semisynthetic glycoconjugates
Maleimide-activated: TT (batch FA 045644) was a gift from Sanofi-Pasteur (Marcy l’Etoile, France). It was stored at 4°C as a 39.4 mg · ml–1 stock solution in 0.05 N NaCl buffer. In a typical experiment, stock solution of TT (12 mg, 304 µl, 0.08 µmol) was diluted in 0.1 M PBS, pH 7.3 (296 µl). To this solution was added (
-maleimidobutiryloxy) sulfosuccinimide ester (Pierce) (3 x 1.53 mg, 3 x 50 equivalent, dissolved in 60 µl of CH3CN/0.1 M PBS, pH 7.3, 1:1) in three portions every 40 min. The pH of the reaction mixture was controlled (indicator paper) and maintained at 7–7.5 by addition of 0.5 M aqueous NaOH. Following an additional reaction period of 40 min, the crude reaction mixture was dialyzed against 3 x 2 L of 0.1 M potassium phosphate buffer (pH 6.0) at 4°C using Slide-A-Lyzer dialysis cassettes (Pierce) displaying a membrane cutoff of 10 kDa. General procedure for the conjugation step: following dialysis, maleimide-activated TT in 0.1 M potassium phosphate buffer solution was reacted with each of the known synthetic S-acetylthioacetylated tri-, tetra- penta- (27), deca-, and pentadecasaccharides (28) related to S. flexneri 2a O-SP in a 1:12 molar ratio, respectively. Reaction mixtures were buffered at a 0.5 M concentration by addition of 1 M potassium phosphate buffer (pH 6.0). Then NH2OH/HCl (7.5 µl of a 2 M solution in 1 M potassium phosphate buffer, pH 6) was added to the different mixtures and the couplings were conducted for 2 h at room temperature. The conjugated products were purified and stored as described previously (29). Hexose concentrations were measured by a colorimetric method based on the anthrone reaction using the corresponding oligosaccharides as standards (30). Protein concentrations were measured with the Lowry method using BSA as a standard (31).
Biotinylated oligosaccharides
Conjugation of the known synthetic S-acetylthioacetylated tri-, tetra- penta- (27), deca-, and pentadecasaccharides (28) related to S. flexneri 2a O-SP to EZ-link PEO-maleimide-activated biotin (Pierce) was run in phosphate buffer at pH 6.0 in the presence of hydroxylamine (32) and monitored by reversed-phase HPLC. Reversed-phase HPLC purification gave the target conjugates as single products, whose identity was assessed based on mass spectrometry analysis.
Immunogenicity studies in mice
Seven-week-old BALB/c mice were immunized three times at 3-wk intervals, followed by a fourth injection 1 mo after the third one, with the equivalent of 10 µg of oligosaccharide per mouse and per injection, in the absence of adjuvant, with the following glycoconjugates: B(E)C-TT, B(E)CD-TT, AB(E)CD-TT, [AB(E)CD]2-TT, or [AB(E)CD]3-TT. Control mice received TT alone using a dose equivalent to the maximum administered to mice receiving the glycoconjugates, i.e., 140 µg/mouse and per injection. For the glycoconjugates incorporating the tri- or the tetrasaccharide and for the control mice, two independent experiments were performed using seven mice per group. For the three remaining glycoconjugates, three independent experiments were performed, two including 7 mice per group and one including 14 mice per group.
The Ab responses induced upon immunization were assessed 1 wk after the third and the fourth injections by ELISA. Purified LPS serotype 2a (14), biotinylated oligosaccharides corresponding to those incorporated into the glycoconjugates (this study), and TT were used as coated Ags to define the anti-LPS 2a, anti-oligosaccharide, and anti-TT Ab titer, respectively. Biotinylated oligosaccharides (0.5 µg/well) were coated on plates previously incubated for 1 h at 37°C with avidin (1 µg/well; Sigma-Aldrich). The amount of TT used for coating was 0.1 µg/well. Anti-mouse IgG alkaline phosphatase-labeled conjugate (Sigma-Aldrich) was used as secondary Ab at a dilution of 1/5,000.
Statistical analysis
Significant differences were established using Student’s test. Values of p < 0.05 were considered to be significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Different subclasses of IgG specific for LPS O-Ag are induced following natural infection with Shigella (33). To test whether all of the subclasses exhibit similar protective capacity, murine mIgG specific for serotype 2a determinants on the O-Ag and representative of each of the four murine IgG subclasses were obtained. Upon screening of hybridomas for their reactivity with LPS purified from S. flexneri serotype X, Y, 5a, 5b, 2a, 2b, 1a, 3a, respectively, five mIgG exclusively recognizing the S. flexneri 2a LPS were selected: F22-4 (IgG1), D15-7 (IgG1), A2-1 (IgG2a), E4-1 (IgG2b), and C1-7 (IgG3). Each LPS-mIgG interaction was characterized by measuring the IC50 that was shown to range from 2 to 20 ng/ml.
The protective capacity of the selected mIgG was analyzed using the murine model of pulmonary infection as previously described (10). In the absence of an intestinal model of experimental shigellosis in adult mice, this model has been shown to mimic the acute intestinal inflammation observed in humans. Indeed, mice infected i.n. with live virulent Shigella develop an acute bronchopneumonia characterized by a massive intra- and peribronchial polymorphonuclear (PMN) infiltrate, in addition to alveolitis (9, 10). Moreover, we recently demonstrated that this model is also relevant for mimicking the Shigella-induced immunomodulation of the host response observed during natural infection (Refs.34 and 35 ; J. Gamelas-Magalhaes and A. Phalipon, unpublished observations). In addition, our previous data emphasize the relevance of this experimental model for testing the protective ability of mAbs that are expected to be effective in the intestinal environment. Indeed, the protective capacity of murine mAbs specific for S. flexneri serotype 5a primarily demonstrated in the mouse pulmonary model (10) has been confirmed using both in vitro and in vivo models of intestinal infection (Ref.36 ; A. Phalipon and B. Corthésy, unpublished data).
Naive mice were administered i.n. with each of the purified mIgG before i.n. challenge with a sublethal dose of S. flexneri. Upon challenge, lung bacterial load in mice passively administered with 20 µg of each of the mIgG specific for S. flexneri 2a LPS was significantly reduced in comparison to control mice receiving PBS (Fig. 2A). Upon passive transfer using 2 µg of mIgG, only mIgG D15-7, A2-1, and E4-1 were shown to significantly reduce the lung bacterial load in comparison to control mice, but with much less efficiency than that observed using 20 µg (Fig. 2A). As shown in Fig. 2B, reduction of lung bacterial load in mice receiving 20 µg of mIgG was accompanied by a reduction of inflammation and therefore of subsequent tissue destruction. In comparison to control mice showing an acute bronchoalveolitis with diffuse and intense PMN cell infiltration (Fig. 2B, a and b) associated with tissular dissemination of bacteria throughout tissues (Fig. 2Bc), only restricted areas of inflammation were observed in Ab-treated mice, essentially in the intra- and peribronchial areas (Fig. 2B, d and e), where bacteria localized (Fig. 2B f). Following passive administration with 2 µg of mIgG, inflammation resembled that of the control mice with a similar pattern of PMN infiltration and tissue destruction, in accordance with the very low, if any, reduction in lung bacterial load (data not shown). Moreover, the protection observed was shown to be serotype specific, as anticipated. Mice passively administered 20 µg of mIgG C1-7 specific for S. flexneri 2a were protected against a homologous challenge, but not upon heterologous challenge with S. flexneri 5a bacteria (Fig. 2C). Similarly, mice receiving 20 µg of mIgG C20, a S. flexneri serotype 5a-specific mAb of the same isotype as mIgG C1-7 (i.e., IgG3), showed significant reduction of lung bacterial load upon i.n. challenge with S. flexneri 5a, but not with S. flexneri 2a (Fig. 2C). In mice protected against homologous challenge, inflammation was dramatically reduced with a slight residual intra- and peribronchial PMN infiltrate (Fig. 2D, b and c). In contrast, in mice not protected upon heterologous challenge (Fig. 2D, a and d), inflammation and tissue destruction were similar to those observed in control mice (Fig. 2B, a and b).
|
A concentration of 1 mM was arbitrarily defined as the maximum ligand concentration to be used in ELISA inhibition assays. The binding of the five protective mIgGs to 23 synthetic mono- and oligosaccharides was evaluated in inhibition ELISA as described in Materials and Methods. None of the mono (A, B, and C are the same; D; E)- or disaccharides (AB; BC; CD; D''A; EC) showed any binding even when used at a concentration of 1 mM. Evaluation of trisaccharide recognition by testing D''AB, BCD, CDA', ABC, ECD, and B(E)C emphasized the unique behavior of mIgG F22-4, which was the only Ab showing measurable affinity for such short oligosaccharides (Table I). ECD was the only trisaccharide recognized by F22-4, pointing out the crucial contribution of both the branched glucosyl residue (E) and the neighboring N-acetyl-glucosaminyl residue (D) to Ab recognition. This was supported by the absence of IgG recognition for AB(E)C or D''AB(E)C at any of the concentrations used. Comparison of the recognition of the branched tetrasaccharide B(E)CD to that of the linear ECD indicated that rhamnose B, accounting for a reduction of the IC50 by a factor of 40, was also a key element in the recognition by F22-4, although of less impact than residue D, since B(E)C is not recognized. Indeed, B(E)CD was recognized by all of the protective mIgG, except A2-1 and C1-7, for which the minimal sequences necessary for recognition at a concentration below 1 mM were pentasaccharides AB(E)CD or B(E)CDA'. Extension of B(E)CD at the reducing end, yielding the branched pentasaccharide B(E)CDA', did not result in any major improvement of Ab binding for the other mIgGs. The minor, if any, contribution of reducing A to binding was also apparent when comparing F22-4 recognition of ECD and ECDA'. Further elongation at the reducing end, yielding ECDA'B' did not improve binding to F22-4.
Introduction of residue A at the nonreducing end of B(E)CD, leading to AB(E)CD, had a somewhat variable impact on Ab recognition with a positive effect in the case of A2-1 and C 7-1 and a negative one by a factor 2 to 5 when considering the other Abs. Thus, for the recognition of short oligosaccharides, two families of mIgGs were identified: the first one, represented by F22-4, recognizing the ECD trisaccharide, and the second one, comprising the remaining four mIgGs, that recognized the same common ECD sequence flanked by the B residue at the nonreducing end, elongated or not with residue A at either end. This observation was confirmed when measuring the recognition of extended oligosaccharides (Table I). Indeed, the decasaccharide D''AB(E)CDA'B'(E)'C' showed the highest affinity for all Abs except F22-4. In the latter case, the octasaccharide B(E)CDA'B'(E)'C' was the best recognized sequence with an IC50 of 0.22 µM. Elongation to the decasaccharide by addition of D''A resulted in a loss of F22-4 recognition by a factor of 20. Interestingly, recognition of the octa- and decasaccharides by the other mIgGs differed from that of F22-4. D15-7, and E4-1 behaved similarly. Extension of B(E)CDA' by B'(E)'C' leading to the octasaccharide, and then by D''A leading to the decasaccharide, both resulted in improving Ab binding by a factor of 4. Contribution of B'(E)'C' to C1-7 binding appeared to be minor, whereas introduction of D''A resulted in an overall gain in binding of 20. In the case of A2-1, addition of B'(E)'C' to B(E)CDA' resulted in a gain in recognition by a factor of 25, and subsequent elongation by D''A at the nonreducing end improved binding further by a factor of 4. Thus, in general, elongating the oligosaccharide sequence increases the Ab affinity.
Molecular characterization of the protective S. flexneri serotype 2a-specific mIgG
To analyze whether the differences observed in the recognition of oligosaccharides by the mIgGs reflect differences in the structure of these mAbs, their CDRs were sequenced (Table II) (37). Only two VH and Vk gene families were expressed among the five mIgGs studied. VH J606 (38) and VK24/25 (39) encoded F22-4 VH and Vk, respectively. A2-1, C1-7, D15-7, and E4-1 VH genes were members of the VGAM3-8 family (40) and their Vk genes belonged to the VK4/5 family (39). The joining segment of the F22-4 H chain was encoded by JH4 (41), while A2-1, C1-7, D15-7, and E4-1 H chains shared the same diversity and joining segments, DSP2 (42) and JH3 (41), respectively. The joining segment for the L chain is encoded by JK1 (43) for F22-4, JK4 for A2-1 and E4-1, and JK5 for C1-7 and D15-7. For all mIgGs, the CDRs L2, L3, and H1 were similar to canonical classes (44), 1/7A, 1/9A,and 1/10A, respectively (45). For F22-4, the canonical forms of the loops L1 and H2 were similar to classes 4/16A and 4/12A, while those of the four other Abs belonged to classes 1/10A for L1 and 2/10A for H2. The CDR-H3 of A2-1, C1-7, D15-7, and E4-1 contained seven residues, including several aromatic ones, while the CDR-H3 of F22-4 was very short, made of only four amino acids with a proline residue in the first position. These results suggest that the unique behavior of F22-4 in recognizing the trisaccharide ECD, in comparison to the other mIgGs, could be related to a particular molecular structure.
|
Among the different oligosaccharides tested, we selected 1) ECD since it was the shortest sequence recognized with an IC50 below 1 mM, at least by one of the five mIgGs (Table I); 2) B(E)CD since it was the tetrasaccharide recognized by three of five mIgGs in contrast to ECDA' recognized by F22-4 only and AB(E)C not recognized at all (Table I); and 3) AB(E)CD since it represents the biological O-Ag RU and was almost as well recognized by the five mIgGs as B(E)CDA (Table I). Besides, its synthesis was believed to be less demanding than that of B(E)CDA. Because longer sequences were shown to be better recognized than shorter ones, we tested the decasaccharide [AB(E)CD]2 representing two biological RU and the pentadecasaccharide [AB(E)CD]3 representing three biological RU, although they may not be the easiest targets when considering synthetic strategies. Our choice derived from the following observations: 1) the octa- and decasaccharides used for the antigenicity study were chosen arbitrarily because they were readily available; 2) studies on short fragments have demonstrated the crucial input of reducing D in Ab recognition; and 3) the presence of nonreducing A was thought to be critical since terminal nonreducing residues of carbohydrate haptens may be immunodominant (46). The corresponding TT glycoconjugates were constructed (see Materials and Methods) and the average value for carbohydrate:protein ratio was shown to be 12. Intraperitoneal immunization of mice was performed using an equivalent of 10 µg of oligosaccharide per dose without any adjuvant. The immunogenicity of the different glycoconjugates was assessed 7 days after the third and fourth immunizations, and the last boost was shown to significantly increase the anti-oligosaccharide and anti-LPS 2a IgG Ab titers (data not shown). ECD-TT neither elicited an anti-oligosaccharide IgG response nor an anti-LPS 2a IgG response. Anti-oligosaccharide Abs were induced by B(E)CD-TT, but no anti-LPS 2a IgG response was measured (data not shown). In contrast, the glycoconjugates incorporating 1, 2, or 3 RU raised both anti-oligosaccharide (Fig. 3A) and anti-LPS 2a IgG responses (Fig. 3B). However, we observed that the anti-LPS 2a IgG titer elicited as well as the number of mice responding was highly dependent on the hapten length (Fig. 3B). Whereas only 28.5% of mice responded to AB(E)CD-TT, 85% responded to [AB(E)CD]2-TT and 100% to [AB(E)CD]3-TT. In addition, mice immunized with the pentasaccharide elicited an anti-LPS 2a IgG titer significantly different from that induced by the pentadecasaccharide but not from that induced by the decasaccharide (p = 0.005 and 0.2, respectively). Similarly, mice immunized with the decasaccharide elicited an anti-LPS 2a IgG titer significantly lower than that elicited with the pentadecasaccharide (p = 0.0002). No cross-reactivity against serotype 2a LPS and each of the selected oligosaccharides was detected in sera of control mice immunized with TT alone (data not shown). Taken together, these results demonstrate that the pentadecasaccharide [AB(E)CD]3 is as an accurate functional mimic of the O-Ag and, therefore, is a good candidate for the development of a chemically defined glycoconjugate vaccine against S. flexneri 2a infection.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Five mIgGs specific for the serotype 2a of S. flexneri were characterized for their protective capacity, the amino acid sequences of their CDRs, and the oligosaccharide determinants they recognize on the O-Ag. It is noteworthy that a large number of O-Ag-specific anti-Shigella mAbs were produced for diagnosis purposes (47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57). However, only a few of them were precisely characterized in terms of protective capacity, recognition pattern, or molecular structure (48, 58, 59, 60). To obtain mIgG, hybridoma cells were selected on the basis of their secretion of mAb recognizing determinants specific for the S. flexneri serotype used for immunization, i.e., serotype 2a or 5a, respectively. During screening, most of the hybridoma cells tested (90%) were shown to secrete serotype-specific mAbs. This result differs slightly from previous reports on the isolation of mAbs directed to determinants common to several S. flexneri serotypes including 2a and 5a (53, 54). It may be explained by recent new insights on S. flexneri O-Ag conformation in the case of S. flexneri serotype 5a demonstrating the impact of the E residue specifying the serotype on the overall conformation of the O-Ag (61). Analogously, we may reasonably hypothesize that this residue also protrudes exquisitely from the surface of the O-Ag in the case of serotype 2a, and being repeatedly exposed at the bacterial surface, preferentially triggers B cell receptor-mediated recognition, thus leading to the induction of a predominant anti-serotype specific Ab response. In favor of this hypothesis is the demonstration that during natural infection in endemic areas where several Shigella species and serotypes coexist, the mucosal Ab response is predominantly directed against serotype-specific determinants (2, 62).
Importance of the E residue and its surrounding sugars is again emphasized in the current study. Indeed, by measuring oligosaccharide-mAb interaction in inhibition ELISA, we showed that the S. flexneri 2a O-Ag exhibits an intrachain immunodominant determinant comprising the trisaccharide ECD as the minimal sequence required for inhibiting mIgG-LPS recognition. Depending on the mAb, additional flanking residues at the nonreducing or reducing end of ECD are required for optimal recognition at ligand concentrations below 1 mM. Since EC is not recognized by any of the protective mAbs, binding of F22-4 to ECD points out to the key contribution of the N-acetyl-glucosamine D residue in the interaction. A F22-4 closely related binding pattern has been reported for mAb SA-3 specific for the cell wall polysaccharide of group A Streptococcus whose RU is a branched trisaccharide (63). Indeed, Abs to branched polysaccharides often recognize the branching point as shown for anti-Salmonella serogroup B mAb Se115-4 (64) and for mIgG C20 specific for S. flexneri 5a O-Ag. The latter, shown in the present article to be protective against a homologous but not a heterologous strain (Fig. 3), recognizes the branched A(E)B trisaccharide epitope (A. Phalipon, unpublished data).
In accordance with the unique recognition pattern of F22-4, we showed that the CDR sequences of this mAb completely differ from those of the four remaining ones. It is noteworthy that the F22-4 CDRs H1, H2, L1, and L2 are quite similar in sequence and/or length to those of SYA/J6, an IgG3 specific for S. flexneri serotype Y, one of the few antibacterial polysaccharide Ab for which x-ray studies of Fab-oligosaccharide interactions have been determined (65). In contrast, the H3 loops, which are the major key of Ab diversity, differ strongly. The CDR-H3 comprises nine and four amino acids for SYA/J6 and F22–4, respectively. SYA/J6 is an example of a groove-like site for binding of an internal ABCDA' oligosaccharide epitope repeatedly exposed on the helical S. flexneri Y O-Ag (60) and its base which includes three Gly residues shows the torso-bulged structure (66). In the case of F22-4, the H3 loop, which can only form a short hairpin, would probably allow a more open binding site that could accommodate the branched E residue.
For the other mIgGs, B(E)CD is the minimal sequence required for recognition. These mAbs bind intrachain epitopes, as indicated by the fact that D15-1 and E4-1 bind to B(E)CD slightly better than to AB(E)CD. This is strongly supported by the increased recognition, by all mIgGs, of the octasaccharide and decasaccharide having a B and a D residue at their reducing end, respectively, in comparison to that of AB(E)CD. In this regard, the carbohydrate-binding specificity of this family of Abs resembles that of a panel of Abs specific for the well-studied Gram-positive group A Streptococcus cell wall polysaccharide, that has a branched trisaccharide RU (67). It is somewhat puzzling to note that although their recognition of the shorter oligosaccharides slightly differs, all of the mIgGs fall into the same pattern of binding when considering the decasaccharide. In accordance with their similarities in oligosaccharide recognition, D15-1, E4-1, A2-1, and C1-7 exhibit similar CDR sequences for both their H and L chains. Further structural analysis of these mIgGs in complex with oligosaccharides will determine which mAb residues are directly involved in Ag binding and will identify the contribution of each sugar to Ab recognition.
As previously reported using the same experimental model for a murine mAb of the A isotype (mIgA) specific for S. flexneri serotype 5a mimicking the secretory IgA-mediated mucosal response (10), we demonstrate here the protective capacity of serotype-specific mIgGs, regardless of their subclasses, in controlling homologous infection with S. flexneri 2a. A similar result is reported for mIgG C20 specific for S. flexneri serotype 5a. In both cases, no cross-protection was shown to occur. With the limits of the experimental model, these data contribute to a better understanding of the role of systemic vs mucosal Ab responses in protection against reinfection by emphasizing the contribution of serotype-specific systemic IgG response in protecting the host against Shigella reinfection as earlier hypothesized (68). This is in accordance with previous results indirectly demonstrating the protective role of the systemic response upon vaccination on the field using a detoxified S. sonnei LPS-based conjugate administered parenterally and eliciting mainly a specific anti-LPS IgG response (4).
By testing the immunogenicity of selected oligosaccharides based on our antigenicity data, we demonstrated that not all of the selected antigenic sequences accurately mimic the natural Ag. In particular, the tetrasaccharide B(E)CD was identified as an immunodominant epitope on the O-Ag, and immunization with the corresponding TT conjugate induced a high titer of anti-B(E)CD Abs, indicating that this oligosaccharide was immunogenic in mice. However, in contrast to other bacterial systems for which protein conjugates incorporating fragments shorter than 1 RU of the surface capsular polysaccharide (CP) were shown to induce anti-CP Abs (69, 70), B(E)CD-TT failed to induce any detectable anti-LPS IgG, demonstrating that this short sequence was not a functional mimic of the O-Ag. An analogous situation was observed earlier for S. pneumoniae type 6B semisynthetic glycoconjugates evaluated in mice, although conjugates incorporating haptens smaller than 1 RU induced anti-CP protective Abs in rabbits (71). Interestingly, AB(E)CD-TT, incorporating the exact biological RU, was able to induce anti-LPS Abs in mice although with a poor efficacy. These results support those obtained with a series of Shigella dysenteriae type 1 glycoconjugates bearing haptens differing by length and reporting that a minimum of 2 RU were needed to induce high titers of anti-LPS Abs (5). In contrast, our data differ from those published on S. pneumoniae type 14 (72) and type 3 (6) and emphasize that bacterial PS should be dealt with on a case-by-case basis. The most striking observation in our study was the strong enhancement of the anti-LPS Ab titer resulting from elongation of the hapten length, when going from one biological RU, to 2 and 3 RU. This is in accordance with previous reports emphasizing that better immunogenicity can be obtained with longer saccharidic haptens (5, 46) and former assumptions that a minimum of 2 RU was necessary for inducing a strong anti-polysaccharide Ab response.
As hypothesized earlier (68), and subsequently confirmed in human studies using classical glycoconjugates based on detoxified S. sonnei LPS (4), the presence of anti-LPS IgG appears as a reliable marker to predict protective immunity induced by parenterally administered glycoconjugate vaccines. According to this assumption, several groups have gone through clinical trials (73, 74, 75), and others are considering doing so (5). We are, therefore, confident in the efficacy of the pentadecasaccharide mimic we identified in inducing protective immunity. The parameters influencing the immunogenicity of the glycoconjugate are currently being studied to envision in the near future a Phase I clinical trial.
Since no data are available, so far, a question that remains open is the feasibility of the semisynthetic strategy for the development of multivalent glycoconjugate vaccines. Actually, the approach we have undertaken may appear more time-consuming than the combination of the relevant classical polysaccharide-protein conjugates, involving polysaccharides purified from the different prevalent strains, detoxified if required, and subsequently coupled to an appropriate carrier. However, we assume that in addition to potential advantages previously mentioned (cf. Introduction), the chemically defined approach is particularly suited for the construction of a multivalent S. flexneri vaccine. Indeed, S. flexneri serotypes only slightly differ in their O-Ag structure and future synthetic strategies could benefit from knowledge gained from our work on serotype 2a. In-house ongoing studies will contribute to the assessment of the accuracy of our assumption.
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work has been supported by the Ministère Français de la Recherche, the Direction Générale des Armées, the Ms Frank Howard Fellowship (to K.W.), the Roux Fellowship (to F.B.), and the Transversal Research Program from the Pasteur Institute (PTR 99 which included a fellowship to C.G). P.J.S. is a Howard Hughes Medical Institute Scholar. 
2 Address correspondence and reprint requests to Dr. Armelle Phalipon, Unité de Pathogénie Microbienne Moléculaire, Institut National de la Santé et de la Recherche Médicale Unité 389, Paris, France. E-mail address: phalipon{at}pasteur.fr
3 Abbreviations used in this paper: RU, repeating unit; mIgG, mouse IgG; TT, tetanus toxoid; i.n., intranasal; PMN, polymorphonuclear; CP, capsular polysaccharide.
Received for publication July 29, 2005. Accepted for publication November 4, 2005.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
-L-rhamnopyranosides with one or two 2-acetamido-2-deoxy--D-glucopyranosyl residues. Bioorg. Khim. 20: 1001-1012. [Medline] immunoglobulin J and C region genes. J. Biol. Chem. 256: 5116-5120. -hemolytic Streptococcus group A. P. J. Garegg, Jr, and A. Lindberg, Jr, eds. Carbohydrate Antigens, ACS Symposium Series 519 111-113. American Society, Washington, DC. 
