INFECTION AND IMMUNITY, June 1998, p. 2441–2446 0019-9567/98/$04.0010 Copyright © 1998, American Society for Microbiology
Vol. 66, No. 6
Streptococcus pneumoniae Type 14 Polysaccharide-Conjugate Vaccines: Length Stabilization of Opsonophagocytic Conformational Polysaccharide Epitopes† CRAIG A. LAFERRIERE, RAMESH K. SOOD, JEAN-MARC DE MUYS, FRANCIS MICHON, AND HAROLD J. JENNINGS* Institute for Biological Sciences, National Research Council of Canada, Ottawa, Ontario, Canada K1A 0R6 Received 3 November 1997/Returned for modification 12 January 1998/Accepted 16 March 1998
A simple and convenient method was developed for the preparation of Streptococcus pneumoniae type 14 polysaccharide (Pn14PS)-tetanus toxoid (TT) conjugate vaccines, using terminally linked Pn14PS fragments of different lengths. Native Pn14PS was simultaneously depolymerized and activated for conjugation by partial N-deacetylation followed by nitrous acid deamination which yielded fragments (1.4 to 150.0 kDa) having a free aldehyde at the reducing end. These were then conjugated to TT through their terminal aldehydic groups, using the reductive amination procedure. All of the above conjugates, when injected in rabbits, induced anti-Pn14PS antibodies, whereas the native Pn14PS did not. The amounts of anti-Pn14PS antibodies elicited by these conjugates, as determined by enzyme-linked immunosorbent assay, followed a trend with conjugates containing the highest-molecular-weight Pn14PS eliciting the highest titers. The same trend was also observed in the ability of the antibodies to opsonize and kill live type 14 pneumococci, although the increase in opsonophagocytic activity was more pronounced and did not correlate linearly with increases in antibody titer. Competitive inhibition of the binding of different conjugate antisera to the native Pn14PS, using Pn14PS fragments as inhibitors, established that the conjugates induced antibodies with specificities for different lengths of Pn14PS beginning at 2 repeating units (RU). It was also established, both immunologically and antigenically, that at least 4 RU of Pn14PS were required to form an extended conformational epitope and that approximately 22 RU of Pn14PS were required to duplicate the same epitope on the same saccharide chain. The conformational epitope was found to be essential for the induction of antibodies with high opsonophagocytic activity and that augmentation of opsonophagocytic activity was also dependent on further chain extension. number of terminally linked saccharide fractions of defined length and to carry out opsonophagocytic assays on the induced antisera. We recently reported the results of systematic immunogenicity studies in rabbits, using conjugates which conform to the above criteria and that were made with PS fragments of pneumococcal types 3, 6A, 18C, 19F, and 23F (16). In these studies, we found little variation in the antibody titers and opsonophagocytic titers induced by different conjugates. We now report that in contrast to the above result, there is an increase in the immunogenicity of Streptococcus pneumoniae type 14 PS (Pn14PS)-tetanus toxoid (TT) conjugates and an even more significant increase in the opsonophagocytic activity of the antibodies generated by these conjugates with increasing saccharide chain length. That this result could have implications in the development of pneumococcal vaccines can be established from other studies (10). In these studies, it was found that although a conjugate made with depolymerized Pn14PS produced high concentrations of antibodies to the saccharide component, it was poorly protective in a chinchilla model of otitis media. To explain the unusual length dependency of the saccharide moieties of the conjugates, we carried out competitive inhibition experiments on the binding of the native Pn14PS to the above conjugate antisera, using Pn14PS fragments as inhibitors, to determine which epitopes within the Pn14PS were responsible.
The currently licensed 23-valent capsular polysaccharide (PS) vaccine for the prophylaxis of pneumococcal infections is poorly immunogenic in infants less than 2 years of age (3, 17). To overcome this serious deficiency, efforts have been made to develop conjugate vaccines against the pneumococcus (reviewed in references 12, 14, and 17). The strategy used has been to focus on the few types which are most commonly involved in disease in infants, especially otitis media (1, 10, 25). Their capsular PSs have been conjugated to various carrier proteins, and the immunological properties of the conjugate vaccines were evaluated in various animal models (5, 6, 10, 15, 21, 23, 25) and humans (1) and demonstrated to have T-celldependent characteristics of isotype switching and boosting. The above conjugates are diverse in terms of their different structural parameters, made with either small oligosaccharides (1), intact PSs (1, 6, 21, 23, 25), or saccharides of undefined length (10). Two studies (9, 22) have established that conjugates made with largest pneumococcal capsular PSs are the most immunogenic. However, in both these studies, saccharides of only two different sizes were employed to make the conjugates, and the coupling techniques used resulted in random and probably multiple coupling of the carrier protein to the saccharides. Opsonophagocytic assays on the conjugateinduced antisera were not performed. Ideally for this type of study, it is preferable to use conjugates made with a greater * Corresponding author. Mailing address: Institute for Biological Sciences, National Research Council of Canada, Ottawa, Ontario, Canada K1A 0R6. Phone: (613) 990-0821. Fax: (613) 941-1327. E-mail: [email protected]
† National Research Council of Canada publication no. 39583.
MATERIALS AND METHODS Materials. Type 14 S. pneumoniae (ATCC 6314) and native Pn14PS were purchased from the American Type Culture Collection, Rockville, Md. Native Pn14PS had a high molecular weight as it was eluted in the void volume of a Bio-Gel 8.5 column. Dextran T fractions were obtained from Pharmacia Biotech,
LAFERRIERE ET AL.
Baie d’Urfe´, Que´bec, Canada. Goat anti-rabbit immunoglobulin G (heavy plus light chain) [IgG (H1L)] antibodies conjugated to horseradish peroxidase and tetramethylbenzidine substrate were obtained from Kirkegaard & Perry Laboratories Inc., Gaithersburg, Md. TT, obtained from Institute Armand Frappier, Montreal, Que´bec, Canada, was purified by gel filtration on a Bio-Gel A0.5m (Bio-Rad) column equilibrated with phosphate-buffered saline (PBS). The monomeric TT obtained as described above was dialyzed against distilled water, lyophilized, and used for conjugation. Sodium cyanoborohydride (Aldrich Chemical Company, Milwaukee, Wis.) was purified as described by Borch et al. (2). Preparation of Pn14PS fragments. Native Pn14PS (10 mg) was partially Ndeacylated by dissolution in 0.5 M NaOH (1.0 ml). The solution was kept at 70°C for 30 min and cooled, and the pH of the solution was brought to ;4.5 by using glacial acetic acid. The partially N-deacylated product was deaminated by addition of a 5% aqueous solution of sodium nitrite (Anachemia, Montreal, Que´bec, Canada); the solution was stirred at room temperature for 2.5 h and then neutralized with 2 M NaOH. The solution was dialyzed against cold water and lyophilized. The product obtained was fractionated on a Bio-Gel A0.5 (Bio-Rad) column eluted with PBS at pH 7. Appropriate fractions were combined to make four lots, and the average molecular size (Kav) of the individual lots was estimated by comparison of their Kav values calculated from their elution volumes on a Superose 12, HR 10/30 column (Pharmacia) with the Kav values of dextrans of known molecular weights. Smaller fragments (2 to 5/6 repeating units [RU]) of Pn14PS (2RUPn14PS to 5/6RUPn14PS) were prepared by deamination as described above except that the heating time during the N-deacetylation step was extended to 2 h, and the fragments were separated on a Superdex 30 HiLoad 16/60 PrepGrade column eluted with 330 mM phosphate–5 mM NaCl (pH 7.3) buffer (20). Thin-layer chromatography on silica gel (n-butanol–acetic acid–water, 2:1:1), fast atom bombardment-mass spectroscopy (FAB-MS), and 1H nuclear magnetic resonance (NMR) were used to confirm the structures and purity of the fragments. Coupling of Pn14PS fragments to TT. The Pn14PS fragments obtained by deamination contained a terminal aldehyde and were conjugated with purified TT by the reductive amination procedure (11). In brief, a solution of Pn14PS (3 mg), TT (3 mg), and sodium cyanoborohydride (3 mg) in 0.1 M sodium bicarbonate buffer (0.3 ml, pH 8.1) was kept at 37°C for 4 days. The progress of conjugation was monitored by analyzing small aliquots from the reaction mixture by high-performance liquid chromatography using a Superose 12, HR 10/30 column (Pharmacia) with PBS as the eluant. Conjugation was indicated by gradual disappearance of the TT peak with simultaneous appearance of the conjugate peak having a relatively lower Kav value. All conjugates were purified on a Bio-Gel A0.5m (Bio-Rad) column eluted with PBS. Fractions containing conjugate were pooled, dialyzed against water, and lyophilized. The conjugates were analyzed for carbohydrate contents by phenol-sulfuric acid method (7), using purified PS fragment as a standard. Protein contents of the conjugates were determined by the bicinchoninic acid analysis method (24), using bovine serum albumin (BSA) as a standard. Using the same procedure, the 30-kDa Pn14PS (30kDPn14PS) was also conjugated to human serum albumin (HSA), and the resulting conjugate was designated as 30kDPn14PS-HSA. Vaccination of rabbits with conjugate vaccines. Groups of three New Zealand White rabbits (NRC stock) were immunized subcutaneously at two to three sites with 10 mg of carbohydrate either as TT conjugates or as the native Pn14PS, adsorbed on aluminum hydroxide (1 mg/ml) in a total volume of 1 ml. The animals were injected on day 0, 21, and 42. The sera were collected on day 0, 21, 34, and 55, filtered sterile, and stored at 280°C. ELISA. Rabbit antibodies to Pn14PS were quantified by enzyme-linked immunosorbent assay (ELISA). Briefly, each well of microwell plates (Corning no. 25805-96 or Linbro Titertek no. 76-381-04) was coated with 1 mg of native Pn14PS or 30kDPn14PS-HSA in PBS. The plates were then blocked with 0.5% skim milk or 1% BSA in PBS. Serial dilutions of serum were applied, and the PS-specific antibodies were detected with goat anti-rabbit IgG (H1L) conjugated to horseradish peroxidase at 1/200 dilution and tetramethylbenzidine substrate. Tween 20 (0.05%) in PBS was used as washing buffer between each step. Titers were recorded as the reciprocal of the dilution that gave an absorbance of 1 at 450 nm. Competitive inhibition of ELISA using Pn14 fragments. Competitive inhibition of ELISA was performed following the ELISA procedure as reported above, using the native Pn14PS as the coating antigen and with the other following variations. After the blocking step, 50 ml of 2 to 4 mM Pn14PS fragment inhibitors in 0.5% BSA–0.02% Tween–PBS buffer was added to wells and serially diluted twofold with the same buffer. Then 50 ml of pooled antiserum (diluted in the same buffer to give an optical density at 450 nm [OD450] of approximately 1 in the absence of inhibitors) was added, and the mixture was incubated at room temperature for 3 h. The remainder of the procedure was followed as described above. Native Pn14PS was always included as a control to determine variation. Percent inhibition was calculated as [OD (no inhibitor) 2 OD (with inhibitor)]/ OD (no inhibitor) 3 100. Inhibition-versus-log concentration curves were drawn for each inhibitor with each antiserum, and the curves were extrapolated to determine the concentrations required for 50% inhibition. Opsonophagocytic assay. To test the biological properties of vaccine-induced antibodies, we used an in vitro opsonophagocytic assay using peritoneal cells obtained from glycogen-stimulated rabbits or dimethylformamide (DMF)-stimulated HL60 cells (16). In brief, HL60 cells were cultured in RPMI 1640 (Difco)
INFECT. IMMUN. supplemented with 20% fetal calf serum. Five days before use, the cells were stimulated with 90 mM DMF. On the day of use, the cells were centrifuged and washed three times with cold Hanks buffered salt solution (HBSS) containing 0.3% BSA. For peritoneal cells, rabbits were injected intraperitoneally with 250 ml of 0.1% glycogen solution in saline. After 4 h, the fluid was collected with heparin and filtered on sterile gauze. The peritoneal cells were then washed in cold HBSS containing 0.1% gelatin and 10 Ul of heparin per ml. After lysis of the erythrocytes with ammonium chloride buffer (0.155 mM NH4Cl, 10 mM KHCO3, 1 mM EDTA), the cells were resuspended in HBSS containing 0.3% BSA and counted. For the assay, 0.250 ml of heat-inactivated antiserum (30 min at 56°C) was serially diluted fourfold in HBSS with 0.3% BSA. To this was added 0.05 ml of buffer containing 2 3 105 to 3 3 105 log-phase S. pneumoniae serotype 14 organisms and 0.05 ml of complement (4-week-old rabbit serum [Pel-Freez Clinical Systems, Brown Deer, Wis.], preadsorbed for 1 h at 4°C with type 14 log-phase bacteria). This mixture was incubated for 15 min at 37°C with shaking at 135 rpm; then 2 3 106 to 3 3 106 rabbit peritoneal cells, 6 3 105 to 10 3 105 HL60 cells, or an equivalent volume of buffer was added to the mixture for a final volume of 0.5 ml. Quantitative cultures on Columbia plates (QueLab, Laval, Que´bec, Canada) were done at time zero and after 60 min of incubation at 37°C with end-over-end rotation. Percentage kill was calculated as (CFU preimmune with and without cells 2 CFU immune with cells)/(CFU preimmune with and without cells). Bactericidal activity was also calculated as (CFU preimmune with and without cells 2 CFU immune without cells)/(CFU preimmune with and without cells).
RESULTS Depolymerization and activation of native Pn14PS. Native Pn14PS is a neutral PS having a repeat tetrasaccharide structure (18), and it was depolymerized by using the following procedure (19). Partial N-deacylation of Pn14PS at the interchain N-acetylglucosamine residues reveals a free amine and makes it susceptible to nitrous acid deamination. During the deamination reaction, the equatorial free amino group is converted into a diazo group with nitrous acid, and after rearrangement, the sugar is converted into 2,5-anhydro mannose. Rearrangement also leads to the loss of the aglycon and results in depolymerization of the native Pn14PS and the production of fragments having terminal free aldehyde groups. Although there are alternate products formed during the reaction, 1H NMR spectroscopy did not reveal any significant change (,5%) in the structure of the fragments. The fragments obtained as described above were fractionated on a Bio-Gel A0.5 column, and appropriate fractions were pooled to obtain four lots with average molecular mass of 150, 70, 30, and 15 kDa, as determined from an elution curve constructed by using dextran T fractions of known molecular weight. Lower-molecular-weight fractions were further fractionated on a Superdex 30 column and yielded well separated fragments of 1, 2, 3, 4, and 5/6 RU. The purity and structures of the fragments were demonstrated by thin-layer chromatography and 1H NMR (not shown), and FAB-MS confirmed the structures of the 2- and 3RUPn14PS [1,378.6 (M 1 Na1 1 H2O) and 2,027.7 (M 1 H1), respectively]. Conjugation of Pn14PS fragments to proteins. Conjugation of the fragments to TT by using reductive amination gave conjugates where each saccharide chain was attached by a single covalent bond. Each conjugate is identified in this text according to the size of the PS in the conjugate, and their biochemical analysis is reported in Table 1. A similar procedure was used to conjugate a 30kDPn14PS to HSA. Conjugates containing the higher-molecular-weight fragments had approximately one saccharide chain per molecule of TT, but conjugates prepared from smaller fragments contained greater numbers of chains per molecule of TT. Conversely, the percentage dry weight of carbohydrate in the conjugates was greater in conjugates made with the higher-molecular-weight fragments and decreased in conjugates prepared from smaller fragments. 1 H NMR spectra of all conjugates confirmed that the struc-
VOL. 66, 1998
TYPE 14 PNEUMOCOCCAL POLYSACCHARIDE EPITOPES
TABLE 1. Biochemical analysis of Pn14PS-TT vaccines No. of RU
Mol wt of PnPSa
Mol of carbohydrate/mol
214 100 43 22 5/6 4 3 2
150,000 70,000 30,000 15,500 3,400 2,715 2,026 1,337
55 40 38 33 26 14 10 9.4
45 60 62 67 74 81 86 87
1.22 1.42 3.06 4.76 13.5 8.6 8.3 11.7
a Obtained by comparison of Kav values of the PS fragments with those of dextrans of known average molecular weight on a Superose-12, HR 10/30 (Pharmacia) column or determined by FAB-MS. b Carbohydrate contents were determined by phenol-sulfuric acid assay using purified Pn14PS fragments as the standard. c Protein amounts estimated by bicinchoninic acid analysis using BSA as the standard.
tures of the attached saccharide fragments remain virtually unchanged. Immunogenicity of conjugates in rabbits. The above TT conjugates of Pn14PS fragments of various chain lengths were evaluated for their immunogenicity in New Zealand White female rabbits. Each group of three rabbits was injected subcutaneously with 10 mg of carbohydrate as native Pn14PS or as conjugate. Using 30kDPn14PS-HSA as the coating antigen, preimmune sera taken from each group of rabbits showed trace or nondetectable amounts of antibodies against Pn14PS. The geometric mean ELISA titer of the antisera obtained from individual rabbits injected with native Pn14PS showed no increase with subsequent injections, as shown in Table 2. In contrast, all conjugates elicited statistically significant antibody rises after the first injection; the additional injections on day 21 or 42 also resulted in antibody increases, though these were not statistically significant in the cases of the 30kDPn14PS-TT, 5/6RUPn14PS-TT, and 3RUPn14PS-TT (P . 0.05). When the titers of the individual rabbit antisera were measured, there was no statistically significant difference in any of the titers of the day 55 antisera (Newman-Keuls multiple comparison) (Table 2). All conjugates studied induce antibodies of the IgG isotype in rabbits, and their levels could be boosted with subsequent injections, clearly showing T-cell-dependent enhanced immune response to Pn14PS-TT conjugates. When the titers of the pooled rabbit antisera were measured by ELISA using 30kDPn14PS-HSA as the coating antigen, the differences in titer were small but detectable. However, when the native
FIG. 1. Titration of pooled rabbit antisera induced by Pn14PS-TT conjugates containing Pn14PS fragments of different lengths, using the native Pn14PS (top) and 30kDPn14PS-HSA (bottom) as coating antigens.
Pn14PS was used as the coating antigen, a distinct difference in titers was observed, particularly between the largest and smallest Pn14PS-TT conjugate antisera (Fig. 1). Competitive inhibition of polyclonal antisera with Pn14PS fragments. To characterize the polyclonal antisera elicited by the different-molecular-weight PS conjugates, a competitive ELISA was carried out with the native Pn14PS as the coating antigen and depolymerized Pn14PS fragments as inhibitors to determine the amount of each fragment required to give 50% inhibition. Figure 2 shows the data compiled from the competitive ELISAs of six different antisera inhibited with nine different Pn14PS fragments. Logarithmic axis scales were used in the graph for clarity. Visual inspection of the curves reveals a salient feature. All curves representing antisera induced by 4RUPn14PS-TT through 22RUPn14PS-TT plateaued no matter which fragments from 4 to 22 RU were used as inhibitors. This plateau is indicative of the antisera containing antibodies having a common specificity for an extended groove-type conformational epitope. Thus, the conformational epitope requires a minimum of 4 RU to form and 22 RU to duplicate. The latter can be
TABLE 2. Pn14PS-specific antibody titers of individual sera from rabbits vaccinated with native Pn14PS or its fragments coupled to TT Immunogen
Pn14PS 150kDPn14-TT 70kDPn14-TT 30kDPn14-TT 15kDPn14-TT 5/6RUPn14-TT 4RUPn14-TT 3RUPn14-TT 2RUPn14-TT
Geometric mean titera (95% confidence interval) of rabbit serum obtained on day: 21
,100 1,100 (230–3,300) 2,100 (1,700–3,300) 2,600 (700–5,100) 1,300 (900–1,300) 3,800 (1,200–22,000) 600 (500–800) 2,500 (1,100–6,900) 700 (100–1,800)
,100 11,400 (5,600–40,600) 19,400 (10,200–27,400) 12,600 (5,400–21,400) 9,100 (7,600–12,200) 15,400 (4,900–29,900) 5,900 (3,300–13,500) 10,800 (9,100–12,100) 7,300 (4,500–17,800)
,100 22,600 (10,600–83,000) 22,600 (17,600–26,200) 16,200 (10,000–23,700) 21,900 (20,800–23,200) 17,900 (2,400–138,400) 7,900 (3,800–22,200) 11,300 (6,700–18,100) 18,700 (9,200–27,900)
a Each rabbit was vaccinated subcutaneously at two to three sites with 10 mg of carbohydrate either as a conjugate or pure Pn14PS adsorbed on aluminum hydroxide (1 mg/ml). Geometric mean titer is defined as the dilution giving an ELISA OD450 of 1.0 for rabbit IgG (H1L). Titers of 100 or less are reported as 100. All titers on day 0 were ,100.
LAFERRIERE ET AL.
FIG. 2. Fifty percent inhibition curves derived from ELISA inhibition data on the binding of native Pn14PS to pooled rabbit antisera induced by Pn14PS-TT conjugates containing Pn14PS fragments of different lengths and using the same fragments as inhibitors for each individual antiserum.
ascertained from the observation that for inhibitors larger than 22 RU, the descending curves indicate the occurrence of further inhibition, which is consistent with a concomitant increase in the valency of the conformational epitope on the larger fragments. Confirmatory evidence for the multivalency of conformational epitopes within the native Pn14PS and its longer fragments was obtained when they were used as inhibitors of the binding of native Pn14PS to the 150kDPn14PS-TT antiserum (Fig. 3). In these experiments, it was determined that native Pn14PS was the best inhibitor and that the ability to inhibit was dependent on saccharide length. This finding is consistent with increasing valency of conformational epitope. In addition, the importance of multivalency of conformational epitope to binding was demonstrated by including an HSA conjugate containing three to four chains of 24KdPn14PS in the same experiment. The conjugate had increased inhibitory properties over that of the single 24kDPn14PS fragment and was equivalent in inhibitory power to the 90kDPn14PS. Contrary to the results above obtained with the conjugates containing the larger saccharide fragments, the binding of native Pn14PS to antisera induced by 2RUPn14PS-TT and 3RUPn14PS-TT was readily inhibited by 2-RU and 3-RU fragments, indicating that these fragments were not long enough to be able to express the conformational epitope either antigenically or immunogenically. An alternate explanation that these
FIG. 3. ELISA inhibition of the binding of 150kDPn14PS-TT pooled rabbit antisera to native Pn14PS, using as inhibitors native Pn14PS (3), 90kDPn14PS (w), 24kDPn14PS-HSA containing 3.5 chains of 24kDPn14PS (■), and 24kDPn14PS (h).
FIG. 4. Opsonophagocytic killing of type 14 pneumococci by pooled rabbit antisera to Pn14PS-TT conjugates containing Pn14PS fragments of different lengths.
latter conjugates produced mostly pocket-type and not groovetype antibodies can be largely refuted on the evidence that 1RUPn14PS, which had the same nonreducing terminal sugar sequence as all the other Pn14PS fragments, was able to only partially inhibit the binding of Pn14PS to 2RUPn14PS-TT antiserum and failed to inhibit the binding of Pn14PS to the 3RUPn14PS-TT antiserum. It is interesting that 2- and 3RUPn14PS were also able to inhibit, to a small extent, the binding of native Pn14PS to antisera made with the conjugates 4RUPn14PS-TT through to 150kDPn14PS-TT. Thus, it is probable that a minor population of antibodies of specificities other than those attributed to the conformational epitope can be induced by conjugates made with longer fragments of Pn14PS. Opsonophagocytic killing of S. pneumoniae type 14. The opsonization of pneumococci by type-specific antibody is thought to be the major immune mechanism protecting the host against infection with pneumococci (29). To determine the opsonic activity of the antisera, we used an in vitro opsonophagocytic assay. As described previously (16), we found rabbit peritoneal cells to be unreliable as a source of phagocytic cells; hence, we developed the same assay using DMFactivated HL60 cells according to the method described by Whitin and Anderson (28) and demonstrated that the same titer for 50% opsonophagocytic killing of type 14 pneumococci could be obtained with either type of cells (results not shown). The results of in vitro opsonophagocytic killing of S. pneumoniae type 14 bacteria by DMF-activated HL60 cells in the presence of pooled rabbit antisera to Pn14-TT conjugates are given in Fig. 4. Some of the samples of pooled rabbit antisera were able to kill type 14 pneumococci in the absence of HL60 cells; however, this bacterial killing disappeared as the antisera were further diluted, whereas there was still 100% killing of bacteria in the presence of HL60 cells. Further dilutions of the pooled rabbit antisera from each group revealed a trend between different antisera in terms of the ability to kill type 14 pneumococci under the experimental conditions. There was a correlation in the maximum dilution of the antisera able to give 50% opsonophagocytic killing (opsonic titer) with the molecular weight of the PS in the corresponding Pn14PS-TT conjugate, with 150kDPn14PS-TT giving the highest opsonic activity. The opsonic titers for each antiserum were compared with the ELISA titers obtained with native Pn14PS as the coating antigen. This ratio, the relative opsonic activity of the pooled antiserum, is presented in Table 3. It is reasonable to separate these numbers into two groups based on the conformational epitope described above. If we choose the 5/6RUPn14PS-TT
VOL. 66, 1998
TYPE 14 PNEUMOCOCCAL POLYSACCHARIDE EPITOPES
TABLE 3. Relative opsonic titers of antisera against Pn14PS-TT conjugates of different molecular weights Antiserum
Relative opsonic titera Individual
150kDPn14PS-TT 70kDPn14PS-TT 30kDPn14PS-TT 15kDPn14PS-TT
4.00 4.16 2.00 5.38
5/6RUPn14PS-TT 4RUPn14PS-TT 3RUPn14PS-TT 2RUPn14PS-TT
1.33 0.56 0.81 0.51
a Calculated as the pooled antiserum titer giving 50% opsonophagocytic killing divided by the ELISA titer. *, difference of P , 0.02.
antiserum as the division and take the average relative opsonic activity of the groups so formed, the two levels of relative opsonic ability are calculated to be 3.89 and 0.63. This calculation indicates that the opsonic ability of the antisera, normalized for the amount of antibody present, is six times better when the length of Pn14PS used in the immunogen is greater than 5/6 RU (P , 0.02). DISCUSSION In this study, a comparison of the immune performance of Pn14PS-TT conjugates indicated a trend for increasing immunogenicity with increasing size of Pn14PS fragment covalently linked to TT. There was an even more pronounced increase in opsonophagocytic activity which was not linearly related to increases in antibody titer. These results differ from those obtained previously in similar experiments using terminally linked saccharides obtained from the PSs of other pneumococcal serotypes (16). All types (3, 6A, 18C, and 23F) other than type 19F, the conjugate of which exhibited a sequential decrease in immunogenicity with increasing length of saccharide chain, showed little variation in immunogenicity and none in opsonophagocytic activity. As in the previous study (16), the Pn14PS conjugates were made by a procedure which gave sufficient uniformity to their structures to enable a legitimate comparison to be made. All of the Pn14PS saccharide fragments were terminally linked exclusively at the reducing end, and all fragments had the same sequence of sugars at the nonreducing end. However, in experiments using the above conjugates, for practical reasons, we had to accept some variation in structural parameters (Table 1). Improved structural uniformity could have been achieved either by the conjugates having equivalent saccharide loading, a factor which is known to contribute to their immune performance (16), or by them having the same number of saccharide chains. Meeting either of these two criteria is not possible when such a wide range of saccharide lengths are used. From our inhibition data, we have established that the predominant immune response to Pn14PS involves the formation of antibodies with a specificity for an extended conformational epitope. Wessels and Kasper (27) first obtained evidence for a conformational epitope on the Pn14PS. However, this evidence was not conclusive because they did not have access to Pn14PS fragments between 2 and 22 RU and hence were unable to determine the size requirements for this epitope. From our inhibition data, we have established that the conformational epitope cannot be expressed antigentically or immunogenically until at least 4RUPn14PS is used and that
22RUPn14PS is required in order to form two conformational epitopes. We have also provided evidence that increased inhibition is associated with multivalency of conformational epitopes, which was not accounted for in other antigenicity studies on the binding of the native Pn14PS to homologous polyclonal rabbit IgG antibodies and derived Fab fragments (27). Based on NMR and molecular dynamic studies on the type III group B streptococcal PS and the structurally related Pn14PS, it is likely that antibodies to both conformational and shorter epitopes have their origins in the respective extended helical and random coil segments of an intrinsically flexible Pn14PS (4). Extended conformational epitopes have been described for the PSs of group B Neisseria meningitidis (13) and type III group B streptococci (26); in the case of the former, evidence for the epitope being situated on an extended helical segment of the PS was obtained (8). Our data also demonstrate that the conformational epitope is essential for the induction of good opsonophagocytic antibodies and that induction of these antibodies increases with increasing length of Pn14Ps fragment used in making the conjugates. This again could be attributed to increasing multivalency of the conformational epitope with increasing length of Pn14PS as was proposed in the antigenicity studies. However, there is a problem with this interpretation of the data because despite the fact that extensive multivalency of conformational epitope occurs in conjugates containing Pn14PS fragments of 4 to 6 RU, the antisera they produce is poorly opsonophagocytic compared to that produced by conjugates containing much longer fragments. For example, 4RUPn14PS-TT and 5/6RUPn14PS-TT contain about 9 and 14 saccharide chains, respectively (Table 1), and inhibition studies have demonstrated that each chain contains a potential conformational epitope (Fig. 2). If one accepts that 22 RU of Pn14PS are required to form two epitopes, then despite the fact that these conjugates contain more epitopes than the longer Pn14PS, the opsonophagocytic activity of their antisera is inferior. One possible explanation for the above phenomenon is that in terms of increasing opsonophagocytic activity, the number of conformational epitopes in a conjugate is not as important as the presentation of these epitopes to the immune system in extended linear sequences of Pn14PS. An alternative explanation for the above phenomenon could be that due to their intrinsic flexibility, there is inadequate stabilization of the conformational epitope when conjugates containing 4- to 6-RU fragments of Pn14PS are used as immunogens. However, we did not detect this in our antigenic studies, and therefore to support this explanation requires making a distinction between antigenic and immunogenic (opsonophagocytic) epitopes. Interestingly, the above explanation is similar to a certain extent to the hypothesis proposed by Wessels and Kasper (27) in which they extended the same argument to the antigenicity of all lengths of Pn14PS, including the native Pn14PS, and attributed increased affinity of binding of Pn14PS to homologous polyclonal antibodies to continuing stabilization of the conformational epitope with increasing length of Pn14PS. Although on the basis of our data we cannot eliminate the possibility of fine-tuning of individual conformational epitopes occurring in longer saccharide fragments, we have established that stabilization of the epitope does occur in Pn14PS fragments of at least 4 RU and that multivalency of conformational epitope is an important factor in both the antigenicity and immunogenicity of Pn14PS. Our inhibition data indicate that stabilization of the conformational epitope must occur between 4 RU and 22 RU of Pn14PS because at 22RU two contiguous conformational epitopes are formed, and be-
LAFERRIERE ET AL.
yond that, additional multivalency of the epitope occurs. Confirmatory evidence for the multivalency of conformational epitope in intermediate-size Pn14PS fragments was obtained in an inhibition experiment which demonstrated that multiple 24RUPn14PS fragments when linked to an HSA molecule had the same inhibitory power as that of a single chain of Pn14PS approximately equivalent in length to the total length of the individual fragments. ACKNOWLEDGMENTS We thank Michele Lussier for skillful technical assistance. This work was supported by North American Vaccines, Beltsville, Md. REFERENCES 1. Anderson, P., and R. Betts. 1989. Human adult immunogenicity of protein coupled pneumococcal capsular antigens of serotypes prevalent in otitis media. Pediatr. Infect. Dis. J. 8:S50–S53. 2. Borch, R. F., M. D. Bernstein, and H. D. Durst. 1971. The cyanohydridoborate anion as a selective reducing agent. J. Am. Chem. Soc. 93:2897–2904. 3. Borgano, J. M., A. A. McLean, P. P. Vella, A. F. Woodhour, I. Canepa, W. L. Davidson, and M. R. Hilleman. 1978. Vaccination and revaccination with polyvalent pneumococcal polysaccharide vaccines in adults and infants. Proc. Soc. Exp. Biol. 147:148–154. 4. Brisson, J.-R., S. Uhrinova, R. J. Woods, M. van der Zwan, H. C. Jarrell, L. C. Paoletti, D. L. Kasper, and H. J. Jennings. 1997. NMR and molecular dynamics studies of the conformational epitope of the type III group B Streptococcus capsular polysaccharide and derivatives. Biochemistry 36:3278–3292. 5. de Velasco, E. A., H. A. T. Dekker, A. F. M. Verheul, R. G. Feldman, J. Verhoef, and H. Snippe. 1995. Anti-polysaccharide immunoglobulin isotype levels and opsonic activity of antisera: relationships with protection against Streptococcus pneumoniae infection in mice. J. Infect. Dis. 172:562–565. 6. de Velasco, E. A., A. F. M. Verheul, A. M. P. van Steijn, H. A. T. Dekker, R. G. Feldman, I. M. Fernandez, J. P. Kamerling, J. F. G. Vliegenthart, J. Verhoef, and H. Snippe. 1994. Epitope specificity of rabbit immunoglobin G (IgG) elicited by pneumococcal type 23F synthetic oligosaccharide- and native polysaccharide-protein conjugate vaccines: comparison with human antipolysaccharide 23F IgG. Infect. Immun. 62:799–808. 7. Dubois, M., K. A. Gilles, J. K. Hamilton, P. A. Rebers, and F. Smith. 1956. Colorimetric method for the determination of sugars and related substances. Anal. Chem. 28:350–356. 8. Evans, S. V., B. W. Sigurskjold, H. J. Jennings, J. R. Brisson, R. To, W. C. Tse, E. Altman, M. Frosch, C. Weisgerber, H. D. Kratzin, S. Klebert, M. Vaesen, D. Bitter-Suermann, D. R. Rose, N. M. Young, and D. R. Bundle. 1995. Evidence for the extended helical nature of polysaccharide epitopes. The 2.8 A resolution structure and thermodynamics of ligand binding of an antigen binding fragment specific of a(2-8)-polysialic acid. Biochemistry 34: 6737–6744. 9. Fattom, A., W. F. Vann, S. C. Szu, A. Sutton, X. Li, D. Bryla, G. Schiffman, and J. B. Robbins. 1988. Synthesis and physiochemical and immunological characterization of pneumococcus type 12F polysaccharide-diphtheria toxoid conjugates. Infect. Immun. 56:2292–2298. 10. Giebink, G. C., M. Koskela, P. P. Vella, M. Harris, and C. T. Lee. 1993. Pneumococcal capsular polysaccharide-meningococcal outer membrane protein complex conjugate vaccines: immunogenicity and efficacy in experimen-
Editor: V. A. Fischetti
INFECT. IMMUN. tal pneumococcal otitis media. J. Infect. Dis. 167:347–355. 11. Jennings, H. J., and C. Lugowski. 1981. Immunochemistry of groups A, B, and C meningococcal polysaccharide-tetanus toxoid conjugates. J. Immunol. 127:1012–1018. 12. Jennings, H. J., and R. A. Pon. 1995. Polysaccharides and glycoconjugates as human vaccines, p. 443–479. In S. Dumitriu (ed.), Polysaccharides in medicinal applications. Marcel Dekker, Inc., New York, N.Y. 13. Jennings, H. J., R. Roy, and F. Michon. 1985. Determinant specificities of the groups B and C polysaccharides of Neisseria meningitidis. J. Immunol. 134: 2651–2657. 14. Jennings, H. J., and R. K. Sood. 1994. Synthetic glycoconjugates as human vaccines, p. 325–371. In Y. C. Lee and R. Lee (ed.), Neoglycoconjugates: preparation and applications. Academic Press, Inc., New York, N.Y. 15. Kuo, J., M. Douglas, H. K. Ree, and A. A. Lindberg. 1995. Characterization of a recombinant pneumolysin and its use as a protein carrier for pneumococcal type 18C conjugate vaccines. Infect. Immun. 63:2706–2713. 16. Laferriere, C. A., R. K. Sood, J.-M. de Muys, F. Michon, and H. J. Jennings. 1997. The synthesis of Streptococcus pneumoniae polysaccharide-tetanus toxoid conjugates and the effect of chain length on immunogenicity. Vaccine 15:179–186. 17. Lee, C.-J., and T. R. Wang. 1994. Pneumococcal infection and immunization in children. Crit. Rev. Microbiol. 20:1–12. 18. Lindberg, B., J. Lo¨nngren, and D. A. Powell. 1977. Structural studies on the specific type-14 pneumococcal polysaccharide. Carbohydr. Res. 58:177–186. 19. Lindberg, B., J. Lo ¨nngren, and S. Svensson. 1975. Specific degradation of polysaccharides. Adv. Carbohydr. Chem. Biochem. 31:185–239. 20. Paoletti, L. C., and K. D. Johnson. 1995. Purification of preparative quantities of group B Streptococcus type III oligosaccharides. J. Chromatogr. A 705:363–368. 21. Paton, J. C., R. A. Lock, and C.-J. Lee. 1991. Purification and immunogenicity of genetically obtained pneumolysin toxoids and their conjugation to Streptococcus pneumoniae type 19F polysaccharide. Infect. Immun. 59:2297– 2304. 22. Peeters, C. A., A.-M. Tenbergen-Meekes, D. E. Evenberg, J. T. Poolman, B. J. M. Zegers, and G. T. Rijkers. 1991. A comparative study of the immunogenicity of pneumococcal type 4 polysaccharide and oligosaccharidetetanus toxoid conjugates in adult mice. J. Immunol. 147:4308–4314. 23. Schneerson, R., L. Levi, J. B. Robbins, D. Bryla, G. Schiffman, and T. Lagergard. 1992. Synthesis of conjugate vaccine composed of pneumococcus type 14 capsular polysaccharide bound to pertussis toxin. Infect. Immun. 60:3528–3532. 24. Smith, P. K., R. I. Krohn, G. T. Hermanson, A. K. Mallia, F. H. Gartner, M. D. Provenzano, E. K. Fujimoto, N. M. Goeke, B. J. Olson, and D. C. Klenk. 1985. Measurement of protein using bicinchoninic acid. Anal. Biochem. 150:76–85. 25. Vella, P. P., S. Marburg, J. M. Staub, P. J. Kniskern, W. Miller, A. Hagopian, C. Ip, R. L. Tolman, C. M. Rusk, L. S. Chupak, and R. W. Ellis. 1992. Immunogenicity of conjugate vaccines consisting of pneumococcal capsular polysaccharide types 6B, 14, 19F, and 23F and a meningococcal outer membrane protein complex. Infect. Immun. 60:4977–4983. 26. Wessels, M. R., A. Munoz, and D. L. Kasper. 1987. A model of high affinity antibody binding to type III group B Streptococcus capsular polysaccharide. Proc. Natl. Acad. Sci. USA 84:9170–9174. 27. Wessels, M. R., and D. L. Kasper. 1989. Antibody recognition of the type 14 pneumococcal capsule. J. Exp. Med. 169:2121–2131. 28. Whitin, J. C., and P. Anderson. 1993. Dimethyl formamide (DMF)-activated HL-60 cells in a convenient assay for antibodies opsonic for pneumococcus. Pediatr. Res. 4:187A. 29. Winkelstein, J. A. 1981. The role of complement in the hosts’ defense against Streptococcus pneumoniae. Rev. Infect. Dis. 3:289–298.